This is to certify that the dissertation entitled ...
This is to certify that the
dissertation entitled
METHODOLOGY FOR EVALUATING THE IMPACT OF ANIMAL
TRACTION AT THE FARM LEVEL IN A SMALESCALE
MULTI-CROPPING
SYSTEM
(BASSE CASAMANCE, SENEGAL)
presented by
Alioune FALL
has been accepted towards fulfïllment
of the requirements for
PhD
degree in AE/ATM
Ip,d,MU&
Major professor
o - 1 2 7 7 1
I .._. -_-__--l-.~--...-“--I
---_--
<I--------‘-
i _ _..“- -..-
. .__..” _. “-.
._ _.,-._--.-
.
METHODOLOGY FOR EVALUATING THE IMPACT OF ANIMAL
TRACTIOk AT THE FARM LE-WL IN A SIMALL-SCALE
MULTECROPPING SYSTEM
(BASSE CASAMANCE, SENEGAL)
BY
Alioune FALL
A DISSERTATION
Submitted to
Michigan State University
In( partial fulfillment of the requirements
For the degree of
DOCTOR OF PHILOSOPHY
Agricultural Technology and Systems Management
Department of Agricultwal Engineering
1997 * ‘
/
ABSTRACT
METHODOLOGY FOR EVALUATING THE IMPACT OF ANIMAL.
TRACTION AT THE FARM LEVEL IN A SMALL-SCALE
MULTI-CROPPING SYSTEM
(BASSE CASAMANCE, SENEGAL)
BY
Alioune FALL
A methodology is presented along with an expert system program for evaluating
the impact of animal traction at the farm level in a small-scale multi-cropping system. In
its actual version, the expert system program applies to oxencultivation using Ndama
sipecies in the Basse Casamance region situation. The evaluation is based on a number of
models developed around the draft animals’ daily energy balance. The energy available
for the various field operations is calculated based on (1) the total liveweight of the
oxen(2) the daily amount of feed given to the oxen and (3) the level of draft and power
required to perform the field operation. The liveweight of each ox is modeled by using an
eimpirical equation relating the actual liveweight in kg to the circumference in cm of the
thorax of the ox. The energy is estimated from the characteristics of the field operation
(type of implement, draft, speed, field time) in association with the number of working
days during the rainy season. An optimization module is used to mode1 a cropping system
in relation to the economical environment and endogenous constraints of the farm (land,
la.bor and available energy). The expert system output agreed with current practices and
values found in the literature.
’
------ .
TO my father Macaty, to vieux Adama, to Odette and to Mabeye Sylla, gone too soon
(Peace Uporu Them).
TO my wife: Marne Coumba, my son Ousmane and my mother Fatou Diagne.
. . .
111
ACKNOWLEDGMENTS
1 would like to take this opportunity to gratefùlly acknowledge the assistance and
trust of Dr. R.H. Wilkinson, my advisor and chairperson of my guidance committee.
Also, the completion of this work would have not been possible without the attention and
understanding of the members of my guidance committee, Dr. Burkhardt (AE), Dr. Harsh
(Ag. Econ) and Dr. Harwood (CSS). Their suggestions and valuable ideas helped me
finalize this dissertation and enjoy every moment of the work. Many thanks go also to
Dick Roosenberg and Tillers International.
1 would like to thank a11 the personnel at the ISRA/CRA Djibelor and CRZ Kolda
research stations, for sharing with me ideas and research facilities during my fieldwork.
Special thanks go to Dr. Khouma and Mme Sall, Abdou Fall, Adama Faye, Abibou
Nia.ng, Sabaly, Konte, Bolle, Tuti, Yafaye, Fode, Ba, Andre and Ibrahima Sonko.
1 would like to express my gratitude to ISRVDG, especially to J. Fayes and J.P.
Ndiaye, to the NRBAR/OSU/AID Project, to CADEF, to vieux Diame (Boulom) and also
to EPATA through Mr. Konte for a11 the support provided to make the research possible.
Thanks to 0. Fall, A.M. Mbaye, 1. Diagne, Zeiman, Herman, B. Diagana, C.
B,abou, P. Diop, H. Diakhate, Amadou Fall, Dr. Sall, M. Havard, Sayon K., As Ndiaye,
D. Coulibaly, M.L. Sonko, F. Diame, 1. Thomas, M. Lo, S. Sali, H. Mbengue, S. Kante,
M. Sene, Mme Badiane, D. Sidibe, S. Djiba, 0. Ba, 0. Diop, K. Toure, Ezabele, Brima,
Pape Gueye, Aziz, Lopi’s, and friends, and their respective families for the moral support,
iv
t
~-.--
-.
.- --~---
.~
:
TABLE OF CONTENTS
Pages
LIST OF TABLE% . . . . . . . . . . . . . . . . . . . . . . . . ..<...................... ,.......... . . . . . . . . . . .
xi
LIST OF FIGUREiS.. . _ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
xvi
ABBREVIATION$ . . . . . . . . . . . . . . . . . . . . . . . . . . . .,........................ *.....,.~...*<....
xix
LIST OF VARIAB/LES.. . , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..“.................~........
Xxi
CHARTER 1
INTRODUCTION.. . . . . . . . . . . . . . . . . . . . . . , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . < . . . . . .
1
‘1
1.1. Mechanlzation overview.. _ . . . . . . . . . . . . . . . * . . . . . . . . . . . . . . . . . . . . . . . . . . .
1
l. 1.1, Sub-Saharan Afkica.. . . . . . , . , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2
11.1.2; Senegal. . . . . . . . . . . . . . . . . . . . . ..~.......................................~
6
1.2. Statement of problem.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .,,
8
1.3. Objectivles . . . _. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..<..............<......<~...
9
CHARTER 2
PRESENTATION CbF THE AREA OF STCDY............... ............. <,...........
12
2.1. Presentation of the area of study.............................................
1 2
2.2. Types of production system ...................................................
17
Z!.2.1. Agro-ecosystem 1 (Oussouye). .......... ,, ....... ................
19
Z!.2.2. Agro-ecosystem 2 (RIouf). ....................................
....
2 0
2!.2.3. Agro-ecosystem 3 (Niaguis). ......................................
2 0
2.2.4 .. Agro-ecosystem 4 (Kalounayes). ...................
.............
21
2.2.5. Agro-ecosystem 5 (Diouloulou). ..................................
21
2.3 Croppingj systems................................................................
22
2:.3.1. Groundnut production ..............................................
2.3
2,.3.2. Cereals ................................................................
2 4
2.4. Agricultural productivity. .............................. ,,.................. ....
2’7
2.4.1 .,Land productivity. ...................................................
28
2.4.2. Manpower and labor productivity. ..... ..........................
3 0
2.4. Summary ..................................................................
.......
31
V
.-
.- ._I.. ----~---
-
%.
/
CHAPTER 3
DETERMINANTS OF PERFORMANCE OF DRAFT ANIMALS.. . . . . . . . . . . . . . . . . :
3 4
3.1. Constraints of animal traction utilization......................................
3 4
3.2. Selection of draft animais ......................................................
3 5
3.2.1. Availability............................................................
35
3.2.2. Selection process ....................................................
3 7
3.3. Training of drafl ammals .............................................
-. .........
3 8
3.4. Hamessing systems and team constitution ...................................
41
3.4.1. Types of hamessing system ..........................................
41
3.4.2. Efficiency of hamessing systems. ...................................
4 2
3.5. Working capacities ..............................................................
4 4
3.5.1. Climate effects .......................................................
4 5
3.5.2. Pulling force potential...............................................
4s
3.5.3. Work potential and energy requirements.........................
5 0
3.5.3.1. Instrumentation...........................................
5 0
3.5.3.2. Energy at work.. .........................................
52
3.5.3.3. Energy for animal growth.. .............................
53
3.5.3.4. Energy for maintenance . . . . . . . . . . . . . . . . . . . . . . . . . . ...*...
54
3.6. Feeding systems .................................... ..i ..........................
54
3.6.1. Feed quality and availability........................................
54
3.6.2. Diet for working animals............................................
55
3.6.3. Forage Unit...........................................................
5s
3.7. Implement selèction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5 9
3.8. Soils.. ............................................................................
6 2
3.8.1. Types of soil.. ........................................................
6 2
3.8.2. Mechanical and physical properties ................................
65
3.8.2.1. Physical properties .......................................
65
3.8.2.2. Resistance to traction ....................................
6 7
3.9. Working days ..........................................
..........................
6 8
3.9.1. Soi1 moisture regime.. . . . . ._.............................
,....” . . . . .
68
3.9.2. Weather factors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
71
3.10. Land use and cropping systems .............................................
7 4
3.10.1. Animal traction and cropping pattems ..........................
74
3.10.2. Field performances and trop production ........................
76
3.11. Summary.. .....................................................................
9 4
CHAPTER 4
MATERIALS AND METIIODS ............................................................
8 2
4.1. Research sites ....................................................................
8 2
4.1.1. On-Station ........................... .................................
82
4.1.2. On-Fat-m ..............................................................
82
4.2. On-Farm equipment survey....................................................
8 4
4.2.1. Objectives .............................................................
85
4.2.2. Data collection .......................................................
86
4.2.2.1. Type offarm equipment and draft animal .............
86
4.2.2.2. Management. ..............................................
SS
vi
.-
.-.-_. - .- ~-
4.2.2.3. Utilization and maintenance. ...........................
89
4.2.3. Samplïng techniques. ......................................
........
90
4.2.4. Data analysis ..............................
1':. ........................
9 1
4.3. Estimates of energy requirements ............................................
9 2
4.3.1. Objectives ...............................................
...........
9 2
4.3.2. Experimental design ...............................................
9 2
4.3.3. Draft measurements .................................................
9 6
4.3.4. Power and energy expenditure ........................ ...........
9 7
4.4. Maximum performance ........................................................
9 8
4.4.1. Objectives ............................................................
9 8
4.4.2. Procedures ...........................................................
9 9
4.5. Instrumen/tation .......................................
..........................
102
4.6. Expert System building (see diagram) .......................................
103
4.6.1. Objectives ............................................................
1 0 3
4.62. Rnowledge base design and development ........................
104
4.6.2.1. Presentation of the expert system. ................... .
104
4.6.2.2. Expert system organization ... ........................
106
4.6.2.3. Data processing .................. ,,........ ......... ....
107
4.6.2.4. Program output ..........................................
1 1 1
4.6.3. Program validation .................................................
1 1 1
4.7. Sumn1ary ........................................................................
1 1 1
CHAPTER 5
FARM CHARACTEI@TICS AND SYSTEMS MANAGEME:NT ................
113
5.1.Farmequi’ment ................................................................
1 1 3
5.1.. 1. 1pypes of implements .......................
........ ..............
116
5.1.1.1. Moldboard piow UCF 10”: Tillage ............... ....
116
5.1.1.2. Imported ridger EMCOT ..... ... ................. ., ..
120
5.1 . 1.3. Seeding implement .......................................
120
5.1.1.4. Multipurpose toolbar implements ...................
122
5.1.1.4.1. Multipurpose frame ARARA ...............
1 2 3
5.1.1.4.2. Multipurpose frame SINE 9 ..................
124
5.1.2. Carts ..................................................................
126
5.2. Repair and maintenance .......................................................
126
5.2:. 1. dreakdown fiequencies ................... ........................
126
5.2.1.1. Moldboard plow 10” ........................... .... . .
126
5.2.1.2. Seeder Super ECO ......................................
129
5.2.1.3. Toolbar and EMCOT ridger ...........................
130
5.2.1.4. (lx-cal-t .....................................................
1 3 1
5.2..2. Farm equipment reliability ........................................
1 3 1
5.3. Dali animal management ...........................................
..........
136
5.3.1. Availability of draft animais .......................................
137
5.3.2. Training and working career ............ .........................
137
5.3.3. Sitability ..............................................................
1 3 8
5.4. Land and labor resources ....................................................
142
5.4.1. Land availabihty ....................................................
142
vii
5.4.2. Farm labor ...................................................
144
5.5. Adoption dynamics ....................................................
144
5.6. Surnmary ................................................. ..:: ..............
149
CHAPTER 6
WORKING CAPACITY OF DRAFT AN-MALS ..............................
151
6.1. Environmental conditions ............................................
151
6.2. Maximum performance .................... : ..........................
154
6.2.1. Maximum performance ....................................
155
6.2.1.1. Optimum point ...................................
155
6.2.1.2. Maximum pull force Pf,, .....................
159
6.2.2. Characteristics of power curves ..........................
160
6.2.2.1. Power output mode1 ...........................
160
6.2.2.2. Walking speed (ANSPEED in rn/s). ..........
167
6.3. Draft animals liveweight mode1 ......................................
173
6.4. Management of draft animais .........................................
177
6.4.1. Field work ...................................................
179
6.4.2. Daily working hours .......................................
180
6.4.3. Feeding system .............................................
182
6.4.4. Liveweight changes and body conditions ................
184
6.4.4.1, Body weight losses (LWL in kg/day ..........
184
6.4.4.2. Health and tare .................................
187
6.5. Sumrnary ...............................................................
188
CHAPTER 7
FELD OPERATIONS AND DRAFT REQUIREMENTS .....................
190
7.1. Field operation characteristics ........................................
190
7.1.1. Soi1 characteristics ..........................................
190
7.1.1.1.. Physical properties ..............................
191
7.1.1.2. Water holding capacity ........................
193
7.1.1.3. Penetrometry ....................................
196
7.1.2. Field operations .............................................
196
7.1.2.1. Tillage .............................................
196
7.1.2.2. Seeding ..........................................
199
7.1.2.3. Weeding ..........................................
200
7.1.2.4 .. Harvesting .......................................
200
7.1.3. Soi1 water balance and working days .....................
201
7.2. Pull force and draft requirements ....................................
204
7.2.1. Statistical design ............................................
207
7.2.2. Required pull force ..........................................
208
7.2.2.1. Data collected for UCF 10” plow ..............
209
7.2.2.2. Data for 3-Canadian tines on SINE 9 ...........
212
7.2.2.3. Data for ridger attached to ARARA ............
215
7.2.3. Regression analysis ...........................................
216
7.2.4. Draft requirements ............................................
221
7.3. Implement Field capacities .............................................
225
...
VI11
. .
.,
I
7.3.1.. Fit$d efficiencies ...............................................
225
7.3 .Z!. Fieild capacities .................................................
226
7.4. Summary.. ............................................. :-. ................
227
CHAPTER 8
EXPERT S’YSTEbPVT TO EVALUATE THE IMPACT OF ANIMAL TRACTION
IN THE BASSE C,4SAMKNCE REGION .........................................
230
8.1. Qrganization of the expert system prog:ram ...........................
230
8.2. Database Files structures .................................................
232
8.2. 'L . Faml location and characteristics .............. .,.......... ...
232
8.2.2. Farm implement ............... .............. .................
233
8.23. DM: animais .................................................
234
8.2.4. Feieding system ................ ................................
236
8.2.5. Crops grown at the farm level ................................
237
8.3. Program mqdules ......................................
.......... ........
239
8.3.1” Fi&ld capacities module ......................................
.
239
8.3.12. Energy module ................ ...............................
240
8.3.2.1. Energy doing work .................................
241
8.3.2.2. Energy spent wa1kin.g ..............................
242
8.3.2.3. Energy for maintenance ........ ..................
243
8.3.2.4. Energy fiom LW losses ...........................
244
8.3.3. Feding system module ..................................
...
244
8.3.3.1. Energy tiom the fee,d ............................
245
81.3.3.2. Feed rations ........................................
247
8.3.4. F rm budget module ..........................
....... ......
249
1
8.3.5. 0 timization module ..........................................
251
81.3.5.1. Optimization objectives ...........................
252
d.3..5.2. Constraints and objective functim ................
252
8.3.5.2.1, Decision variables .......................
252
8.3.5.2.2. Objective 1Function ............ . .......
253
8.3.5.2.3. Statements of constraints ........... ...
253
8.4. Optimizati n output ....................................
... ,,..............
257
8.4.1. 1
S rver solution ..................................................
258
8.4.2. Inbelpretation of coefficients .................................
258
8.4.3.
sitivity analysis ....... .................................
..
260
8.5. Program
idation .......................................
.........
...
261
8.5.1.
ample of one session ........................................
261
8.5.2.
cussion on optimization output ...........................
264
8.6. Summary .....................................................................
265
CHAPTER 9
CONCLUSIONS AN-Il RECO~NDATI~SNS.~~ ,........<..,..... * . . . . . . . . *.a... 267
9.1. Conclusion5 . . . . . . . . . , . . . . . . . . . . . . . . . , . . . . . . . . . . ,...,,..,..,,.......I
. . . . . . . . 9 267
9.1.1. Sqrveys , . . . . ,, . < . . . . . . . . . . . . . . . . . . . .,~,.....
. . .
. . . . . . . . . . . . . .
_ 269
9.1.2. FMow-up of field operations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 271
iX
\\,
9.1.3. Maximum working capacity ................................ .,
272
9.1.4. Energy required for field operations ........................
272
9.2. Recoxnmendations ...................... ............:i ....................
273
APPENDIX A ,..._ . . . . . . . . . . . . . . . . . . . . . . ~ .<,.,...” . . . . . . ~ . . . . <.* . . . . . . . . . .
. . . . . I.....
277
APPENDEX B ........ .._ ..................................................
..... ........
285
APPENTIIX C . . . . , . . . . _. .,..........” . . . . *...*. . . . . . . I,.. ._.......... . ,_...........,..
296
APPENDIX a . . . . . . . . , . , . . . . . . , , . . . , . . . . . . . . . . . . . . . . . . . . . . . ..m.........................>.
320
APPENDIX E . . . . . . . . . . . . . . ...<... ~..” . . ..<............................... * . . . . . . . . . . ,,,... 3 2 4
REFERENCES , , . . , . . . . _, . . . , . . . . . , . , . . _ . . . . . . . <
. . . . . . ,
. . . . . . . . ,. . . . . . . . . . . . . . . . . .
346
X
-._. ----
../)UC_.--------I”I_-I
--I
----
--m.-m.--,
LIST OF TABLES
1. Estimated number ‘of draft animals in Sub-Saharan Afiica ............................
3
2. Number of animal drawn impfements in West .Africa .................................
5
3. Land distribution ........................................................................
12
4. Average farm characteristics in the Basse Casamance region ......................
1 8
5. Cereal yields in theiCasamance (kg/ha). ................................................
22
6. Lower Casamance bgricultural production trends in thousand metric tons ........
2 7
7. Reasons for removiing animals from service, .... ............ ........................
3 7
8. Age and weight to initiate training of cattle for work ................
. ............
4 0
9. Environmental factors on draft animals ..................................
..............
4 6
10. Work potential of ban and various animais #.. ......................................
.
4 8
11. Composition and digestibility of feed (DM basis). .......... ............
.........
56
12 Draft requirements! for farm implement (Equatorial Afiica). .....................
6 0
13. Non-working daysi for the West Af?ican regio’n ....................................
73
14. Perceut in yieid indrease with animal traction plowing ............... ...........
78
15. Analysis of variancje.....................................................................
9 4
16. Farm characteristios .......................................................................
115
17. Breakdown frequenci.es on farm equipment ........................................
127
18. Implement repair ahd maintenance. .................................................
135
42. Analysis of variante (LW,, CC). . . . .
. . . r.. , . . . . . . . . . . . . . . . . . .
175
43, Repartition (%) of hraft oxen time., ........................ . ........................
179
44. h/fana,gement and handling of draft animais (@ the fa.rm level.. ....... ...........
181
45. Pasture grass species in zone 4 and 5 ...........................................
183
46. Liveweight (kg) va$ation of pair of oxen.. .
. . . . . . . . . . . . . . . . . . .
185
47. Average liveweightirrariation of Traction Group.. . . . . I . L. , . . . . . ‘
. . . . .
185
48. Liveweight losses abd gained (June to Cctober). . . . . . . . , . , . . . . . . . . . .
185
49. Average soi1 texture of the tria1 site .....................................
. .........
192
50. Annual rainfall (mn$/rank at Ziguinchor (1960- 1996). .............................
202
5 1. Probability (PWD) and number of working da.ys (NIVDA.YS). ...... .........
204
52. Dimension ofyokes used by farmers(survey results). ...............................
207
53. Two+ample t-test for YOKE 1 vs Y0K.E 2., ............................
.........
207
54. Factors’ levels in a hx3x3 Randomized Block design ....... ......................
208
55. Pull Force PF (daNb ,with the UCF 10” piow. ......................................
212
56. Pull Force PF (daNb with the 3-tine SINE 9. ..................
........ ...........
215
57. Pull Force PF (daN) .wi,th the ARARA ridger., ....................................
216
58. Parameter estimates of PF regression ...............................................
219
59. Pull force PF analysisr of varîance ...................................................
220
60. Analys,is of variante (SWC and EQTYPE). .........................................
221
61. Analysis of variante of Specific Soi1 Resistance SSR (daN/cm’). .................
222
62. Field work charactdristics and Field Efftciencies ............ ................. ....
224
63.ANOVAofE(%) .......................................................................
225
64. Average implelment’eflkiency E (%). . . . . . . ” . . . . . . . . . ~
225
. . .
XII1
. ._..
Y
/
65. FW contained in some available feed., ................................................
248
Al. Animal-drawn equipment questionnaire ..............................................
280
BN 1. Maximum draft trial (Gl PI). .....................................................
...
285
B2. Maximum draft tria1 (Gl P2). ..........................................................
286
B3. Maximum draft tria1 (CT1 P3). ..........................................................
287
B4. Maximum draft trial (ci2 PI) ..........................................................
288
B5.Maximumdrafitrial(G21?2) ......................................................
.....
289
B6. Maximum draft trial (G2 P3). .................................................
........
290
B7. Maximum draft tria1 (G3 Pl). ......................................................
...
291
B8. Maximum draft trial (G3 P2). .........................................................
2!?2
B9 Maximum draft tria1 (G3 P3). ...........................................................
293
BIO. Draf’l animals LW and CIRCUMF ..................................................
29 1
Bl 1~ Regression analysis »fthe 18 draft oxen ............................................
293
C 1. UCF 10” Block 1 @SWC-range 6-8% g/g (Extract). ..............................
294
C2. SINE 9 Block 1 @SWC-range 68% g/g (Extract). ............................ ....
296
C3. ARAR.A Block 1 @SWC-range 6-8% ggl (Extract). ............ ..................
298
C4. UCF 10” Block II @@WC-range 8-10% g/g (Extract). ... ........... ........ ..
3 0 0
CS. SINE S) Block II @SWC-range 810% g/‘g (Extract). ...............................
302
C6. ARARA Block II @SWC-range 8- 10% g/g (Extract). .... ........................
304
C.7. UCF 10” Block III @SWC-range lO-13?/0 g/g (Extract). ..........................
306
C8. SINE 9 Block III @SWC-range 10-13% g/g (Extract). ............................
308
C9. ARARA Block III @;SWC-range lO-13% g/g (Extract). ..........................
310
C 10. Output example of the CEEMAT data processing system ............... ........
3 12
.,
;,’
Dl. I;él for maintenanck FM ofworking cattle.. .“, . . .,. . . . ., . . . , . . . . . . . . . I. _..
320
D2. Composition, of so&xr Africain feeds (As-Fed Basis). .......... ................... ..
3 2 1
D3. Composition of sole African feeds (As-:Fed Basis). ..............................
322
D4. Fiekd operation exe/cu.tion tirne., ............ ................ ........................
323
XV
_^. _ _-. .._II_.-. -.--._- -____-
----.-
LIS- OF F JGURE,§
1. Agro-eco:system zones i!n the Basse Casamance region ............................
1 id
2. Is.ainidl distribution fiom k’orth (Bignana) to SOU~~ (oussouye). . , ._, . . . . ~, . .
14
3. Moldboard plow UCF 10” ................................................................
114
4 . EMCOT ridger (Carrbia). ...............................................................
114
5. ARARA toolbar (Ridger). ................................................................
1 1.4
6. Sine 9 toolbar 3-tine .......................................................................
1. 1 4
7. Seeder Super ECO ...... .................................................................
114
8a. Types of implements used by farmers., ................................................
Il?
Ilb. Implements’ manufacturers ...................... ,,.....................................
1 1 7
9. Implem.ents’ working conditions .........................................................
132
10. Acquisition mode. ..... ..................................................................
132
11. DraR animais distribution. ................ .............................................
139
12. Farm ,size (ha) distribution. .............................................................
143
13. Farm labor distribution.........................................................
........
143
14. First year in animal traction......................................... ....................
14s
15. Number of years of experience., ............... .......................................
145
I
16. Temperatures (min and max) at Ziguinchor (1996). ....................... .........
153
17. Relative Humidity (min and max) at Ziguinchor (1996). ............................
153
xvi
-- _ ._-_- -
-<“U”.““l--.“-n-Lq-W”“-
----
m--P
------------
/
.I
1 $a Pc Wet ~CursJe of Tr&tion group 1. .
.
. . . . .. . . .. . . .. . . .. . . .. . . .
.
163
‘_
18b. W,~Iking speed of draction group %
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
‘X63
19a Pclver curve of Traction group 2..............................................
166
190‘. Walking speed of ~Kmion Gpoup 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . .
160
20a. Power curve 0fTr~r:tion group 3..................................................
169
20b. Walking speed of traction Group 3. ...............................
...... ...
163
21a. LW= f(ClRCUM) jregression line (18 oxen). .................................
....
176
21. b. General regressiod line ....................................................
........
176
22a. LW variation of in&vidual ox .....................................................
E84
22b, Liveweight variation of Traction Croups ..........................................
1.86
23. Drained soi1 water dontent (undisturbed cores). ........................
.... . . .
1.94
24. Log-SWC vs time .....................................................................
194
25. Penetrometry Energhr ..............................................................
1137
26. SWC at penetrometby sites ..........................................................
197
2’7a.Monthiyraindistribution(5yearsoutoPIO) ......................................
205
2’7b. PWD at 50% probhbility level.. ..............................................
205
28a. Montly min distribj.rtion (9 years out of 10). .....................................
206
28b. PWD at 90% prob/ability level.. ....................................................
206
29a. PF with UCF 10” at SWC= 6-g% yoke== 90 c:m .....................................
210
29b. PF with UCF 10” at SWC= 8-10% yoke= !30 cm ................................
210
29c. PF with UCF 10” at SWC= lO-13% yoke= 90 cm ..........................
. ....
210
30a. PF with UCF 10” at SWC= 6-X% yoke== 120 cm ...................................
211
30b. PF with UCF 10” $t SWC= 8-10% yoke= 120 cm ................................
21 1
xvii
-...
1,
/
~OC. PF wiith UCF 10” at SWC= lO-13% yoke= 120 cm. . . . _. . . . . . . . . . . . . . . . ~. . . . . . ,. 211
31a.PFwütbSINE9atSWC=6-8%yoke=9Qcm
.,.......... ..-..~ ............LI....I.... 213
3 1 b. PF w:ith SINE 9 at SWC= S- 10% yoke- 90 cm .....................................
213
131~. PF wEth SINE 9 at SWC= 10”13% yoke= 90 cm ...................................
213
32a. PF wltb SINE 9 at SWC= 6-8% yoke== 120 cm.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
213
32b. PF with SINE 9 at SWC= 8-10% yoke= 120 cm ...................................
213
1
3%~. PF wiith SINE 9 at SWC= lO-13% yoke= 120 cm ..................................
213
33a. PF with ARARA at SWC= 6-S% yoke- 90 LL ....................................
217
33b.PFwithARARAatSWC=$-lO%yoke=90cm..
.................................
217
33~. PF with ARARA at SWC= lO-13% yoke= 9LI cm.. ...............................
217
34a. PF with ARARA at SWC= 6-8% yoke=s 120 cm.. , . , . . . . . . . . . , , . . . . . . . . _ . . “. . . 217
34b. PF with ARARA at SWC= S-10% yoke= 120 cm.. . _. . , . . . . . . . . . . . . . . . . . . . “. . . 217
34~. PF with ARARA at SWC= lO-13% yoke== 120 cm... . . I . . . . . . . . . . . . . . . . . . . . .
217
3 5. Animal Traction Eval:uation Mode1 diagram. ............... .... ,,....................
231
C 1. Rairfall distribution at Djibelor Research Station .......................... .........
296
C2. Soi1 water content at Djibeior Research Station.. . . . . , . . , . . . . . . , . . . . . . . . . , . . . I.. . .
296
C3. CEEMAT data recording and processing system., . . . . . . . . , . . . . . . . . . . . . . . . . 319
El-E22 Screen 1 to Screen 22.. , , . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . ..<................. 324-343
/v
. . .
XVIII
lb--
_- ._.- ...---
ABBREWLATIONS
AIAEI.30: .Association des Jeunes Agriculteu~rs et Eleveurs du Departement de Oussouye
(Basse Cas~amance, Senkgal).
AID : Agency for Interriational Development (USAID).
ARC : Agriicultud Research Councii (England)
ASAE: .American Society of agriculturai Engiaeers.
ASCE: American Society of Civil Engineering.
CADEF: Comite d’Acti@n pour le Developpement ,du Fogny (Ba~s~e Casamance, Senegal)
CEEMAT: Centre d’Etude et d’Experimentation en. Machinisme Agricok ‘Tropical
(France).
CIRAD-SAR: Centre de Cooperation 1nterna.tional.e en Recherche: Agronomique pour le
Deveioppement - Departement de:s Systemes Agro-alimentaires et Ruraux.
CPCS: Commission de Pedologie et de Cartographie des Sols (France).
CRA: Centre de Recher/ches Agricoles
CTVM: Center for Tro$ic;al Veterinary Medecine (England).
FAO: Food and Agriculture Organization.
IRAT: Institut de Recherches en Agronomie Tropicale(France)
ISRA: Insti.tut Senegalaks de Recherches Agrkoles (Senegal).
MAC : Mission Agricolk Chinoise (China).
NARS: National A.gricultural Research Systelm.
ORSTOM: Office de Reclherches Scientifiques et Techniques dans les Territoires
d’outremer (France).
X.iX
t
I
l ,
PIIJAC: Projet Integre pour le Developpement de ia Basie
Casamance (Senegal)
SOMIVAC~: Societe pour la Mise en Valeur de la Casamance.
(Senegal).
c
!LX
--
_ _- .-.. -. ..----.
LHI’ OF VARUBLES
AGETRAIN : Training age of draft animais
ANERGY : Energy Dellvered by the Workin,g animals in MJ.
ANIPEX : Draft Animal :Reliability in % or decim:il.
ANNAVAn j. Probability of Draft animal Unavailability in %.
ANPOWER : Power Delivered by the working ianimals in J/sec.
ANSPEED : Draft anima.l Working Speed in m/sec.
ANWKGHR : Numberlof Hours Worked per Day.
AVGLW : Average livbweight in kg.
CI : Cone Index.
CIRCUMF ; Animal th/orax circumference or girth in cm.
CROPRICE : Market price or value of trop per kg.
DE : Digestable Enet-& in MJ.
DM : Dry Matter confent in percent
DN : Soi1 Drainage ini mm.
DEP * depth of penet&&on In the soi1 in cm(penetrometer).
DISTRAV : Distance kraveled in m.
DR : DraR required in d.aN.
DUC, : Drained Upper Limit.
Ed : Penetration Enerbr of the penetrometer a.t depth in J.
xxi
EFC : Effective Field Capacity in ha/day
EFDTM: EtXective field time in sec, min, or hr.
:EPWD: Effective Probability of Working Day in ‘%.
EQTwE: EqUipIIK!nt type
ERE : Energy Required to pull the Implement in Joules (J).
ET : Evapotranspiration in mm per day.
FC : Field Capacity.
FDOP: Field Gperation
FDTM: Field Time (hrs/ha:).
FRFEED : Amount of fie& Feed Required in kg.
Fr : Frequency (decimal number) of precipitation occurence.
FS : Soi1 resistance Force: to the penetrometer bar ,in kgf
FMSIZE : Farm Size in ha.
FU : Forage unit
FUM : Amount of Forage Unit for Maintenance.
NFWKER!3 : Number of Fana Workers.
Gk : Gross Energy in MB.
GSWC: Soi1 water content in g of water per g of soil.
HGT : hei;o;ht of falling weight in cm (penetrometer).
HI : Soi1 Hardening Index.
.
IMPFCGST: Emplement fixed cost.
IMPLEFF : Implement Effoiency in %.
IMPLREL : Equipment reliability in %.
a
XXii
IMFLWDP : implerr@l Working Depth in cm.
IMPLWWD : Implement working width in cm.
LL : SWC Lower L$nit in %
LW : Draft animal Liveweight in kg.
LWL : Liveweight Losses in kg/day
L,WP : Percentage o$LW used for traction in % or decimal.
MAE : Maintenance !Energy in MJ.
MATURELW : Dra& animal mature liveweight in kg.
ME : Metabolizable/Energy in MJ.
NE : Net Energy in h.
PF : Draft animals &lling Force in kgf or daN.
Pr : percentage probability of occurrence.
RF : Amount of raimal in mm.
RO : So~l1 water runoff in mm.
SBD : Soi1 bulk deniity in g/cm3.
SSR : Specific soi1 resistance in N/m2.
SWC : Soi1 Water content in % g/g
SLWGHT : Penetrometer Sliding Weight i.n kg.
TDN : Total Digestible Nutrients in J per kg
VSWC: Volumetric ‘soi1 water content by volume in cm3/cm”.
NWDA’YS : Numbek of Working Days in days.
SWHC : Water holding capacity.
SWP : Wilting Point in %.
a.<
XXIII.
xxiv
l,l. ~~?S~~~~~~i~~ ovt+rview
In eontrast with fa,rrners in industrialized comtries, darxnerrs in de~elc~pkng mmtiies
are gener aily 3rnaBB knld&s who must contend w% increasing ~~~~~J~~~~~~ ~rrxx~re m th:a
land and a !ow levd ofr$eckanizatiotk ta pr,~motz e£âective erop ~intefosification P~~p~~~at~o~
pressure hais skmged m~st of tke traditiorral âamting systems by ~F&%g skort ~C&V!~ RDQ
pennanerd a-opping. Thiese new practices bave led do more soi1 degradation.
in order tu abscirb the impact of yopamla~ion growtk, govemrnent:j irr iJwelo$ng
countries urge farniers to produce more fix>d by improving agrim’ltwaE g;nda;ctivFt;~. Such
improvcments deptxrd irji gcneral,, on tke t-o~nblnâtion of a variety of energy mwces:
mechaniçat brnackinea, $nimals, li:~;r;ans), çfm&xi $ert%zers, pesticides:, ker!Scide~:, ztc; .)
or biolugical (improved~ vanieties, manure, ru:ompmt, etc.). As poinzted OIZ: b:r Esmw and
Hall (19’73), mechanizaiion is one of’the most cri~.ical production inputs 1 o increase cr~p
yields tkrough better cdntrol oftke sai2 watitar I-F: ,;ime (runoff’versus itiltration~, better miF
preparatiori for soi1 aerttion and mtricnts wail;SIity, more &flcient weed wn?rol, and
better timeliness. Tkey dlefined agricultural me: Ihanizatïon as “tke art andi scientific
application ofmec.haniwl aids for increased pmdwtion and presexvation offood and fiber
\\I
?
/’
çrops with less drudgery a:nd increased efftciency’“. Separate studies conducted in most
of the developing countries in Asia, Afirica and South AmeriCa have shown that human and
animal po,wer are probably the most plentifitl source of enere,v (Patrick, 19’93; Campbell,
1990) towards the establishment of a sustainable, mechanized production system. The
switch from hand cultivation to an animal traction based production system is viewed by
Jaeger (1!384) as a Sign&ant step in creating mo:re production opportunities and
increasing retums for farmers through better land. preparation and timeliness. The level of
investment involved in the technobgy is lovv cori..,)G,,. nd to trwtorization. Draft animals are
available and fed locally and thus save foreign exchange (Roosenberg et d, 1987). The
energy provided by the combined two sources (draft animals and tractorization) represent
gO% to 90% of the total eneirgy used in agriculture throughout the world today (Campbell,
1990), The same study repo:rted that there are over 1.5 billion animais used as working
animais in the world compared to 22 million tractors. The species used to generate the
energy needed for agriculture are mainly cattle, water buffalo, donkeys, mules, horses and
camels.
The level of mechanization and intensity of draft animals’ utilization is highly
,variable fi-on1 one region of i;he world to another one. The focus of this research is the
situation in Sub-Saharan Africa and in Sene,gal in particular.
1.1.1. Sub-Saharan Africa
In Sub Saharan Afric.a, animal traction is one of the most important energy SOUK~S
for agriculture (Table 1:). I[n some areas it protides up to 90’5’0 ofthe power required for
trop prodiuction. The level of farmers’ experience varies fronn one country to another. In
-
. -
“**O)UIm--.-
- I I ” - - - - -
--“-
----m-II--
3
Table 1: Estimbted numbers of draft (mimals in Sub-Saharan Africa
RegionKountry
-
-
Cattle
DoRkeys
Horses
- - i
, - -
m--e-
- - - . - _ - - - - - -
West-P&ica
- - - - - - - -r
..-m
Benin
30-40,000
Burkina Faso
7s80,000
Chad
105~130,000
Cote &Ivoire
30-40,000
The Ga.mbia
18”20,000
2S-30,000
s-7,000
Ghana
20,000
1,000
. .
Guinea
100,000
Guinea-Bissau
4,000
Mali
200-3 20,000
Niger
1620,000
Nigeria
1 o’o-200,000
Senegal
130~140,000
200,010O
Sierra Leone
1,000
-
Togo
-
---_.
9- 10,000
-
East Afiica
Central Africa
-
-
-
-
Angola
Cameroon
C. Af. Republique
Zaire - - - -
South Africa
---
Botswana
350-360,000
Lesotho
180,000
Madagascar
260"330,000
Malawi
SO-70,000
Mozarn’bique
100,000
Zambia
180-315,000
Zimbabwe
500~800,000
\\ I
4
/
Botswana, 80% of the fiarmers use animal traction for piowing. Zimbabwe has the largest
draft animal population in th.e region (Starkey and Faye, 1988). Falvey (1986) pointed out
also that one of the most significant impacts of the utilization of working animais is the
increase of cultivated area per active household member: 30 to 40% in Senegal and 40 to
70% in Mali the neighboring country of Senegal.
In ‘n’est Africa, animal traction was introduced before the era of independence in
.i
the early 1960s through the implementation of various agricuItural projects oriented
J
towasds the production and expert of cash crops igioundnuts in SenegaI, cotton in Mali).
The promotion of animal traction for cereal production at that time was almost non-
existent. The introduction of the new technology by the colonial administmtions took
place between 1905 and 1945 in different places stretching between French Guinea,
1
Senegai and Nigeria. The main actions taken were the integration of trop production and
livestock husbandry (mixed farming) in northern Nigeria (192!4), the introduction of plow
farming in northem Ghana (1938) and the implennentation of research on ,animal-drawu
implements in Senegal(1928). In the process, most of the farm implementrs were
originated fiom France or specifically manufactured and adapted for Afiican conditions:
the MANGATM cultivator for Burkina Faso, the A.RIA.NATM, TROPICANATM and SINE?
for Senegal,. the BAJACTM moldboard plow for Mali.
Historical analysiis shows that adopti.on rates by farmers were the greatest in the
first 10 to 20 years after the independence (1960 ,- 1980). This period represents the most
dynamic period in terms of animal traction projects implementation in the regior, (Le
Moigne, l981) (Table 2). This period was also characterized by the creati.on and
development of animal-drawn equipment factories like SISCOMA (1961) in Senegal.
Table 2: Ndmbers of animal drawm impfemen’ts in West Africa
(1957 and 19iW84)
’ MPT: Multi-P’urpoise Toolbars
2 WTC: Wheeied Tooi Carriers
3 Ariana multicuheur
4 Figures from 1974
na: non available
Sourt:e: Extracted fiom Lawrence and IPearson, 1993
.
‘The major research issues addressed during this implementation and adoption
phase to tmprove the technology were as fcbilow:
‘.
- Draft animal :seIection (cattle, horse, donkely)
- Yoke design and fabrication (head, neck-yokes)
- Implement Aection (single or multipurpose)
- Field operation techniques (tillage, weeding,)
1.1.2. Senegal
In the case of Senegal, cattle were the first option for energy supply to agriculture.
The technology was tested in 1925 for the first trime. In the castration and yoking of
ox, farmers were introduced to a new technology that could. generate much greater power
than used previously. The tec;hnology was introduced through the French color-ial research
system RAT. T ‘le main objective was to improve land and labor productivity, and to
ir;.*-case to: 1, cuitivatedl areas and trop yield. The strategy was te produçe enough
groundnuts andIor cotton for expert with regard. to foreign exchange and also to use part
of the production to supply the grain to the emeaging national oil Industry
The major breakthrough in trop production with draft animais came around 1945
when equine and donkey traction were promoted at the farm ieveï to perhorm light works
like cultivatilon and transport. Nong with groundnuts, the cereal production (millet, maize,
and sorghum) reached satisfactory levels satisfying fart fàmily c:onsumption as weli as
supplyin,g local markets with the production SU~~~AS. Progressively, bovine traction was
o..,m~I,m~----..--.,-._ v--m-
------
-----m-œr-
oriented towards the exec@tion of heavier tasks like plotiing in medium and heavy soils,
and carting of bulky loads.
.e
The adoption of dr#& animal technoio&y by farmers was mainly supported by the
rapid increase in areas crogped with cash crops (major source of incorne), the availability
oflocally m’anufactured iqplements and the existence of credit programs to help farmers
purchase their own equipfllent. The animal traction based production system appeared to
be reproduc-ible, as the majior components involved were a11 readily a.vailable to farmers.
Most importantly, drsft ax$mals like oxen were simply borrowed from the household or
village herds.
However, the pro#notion of animal traction by governmental extension agencies
must face regional, cutturti and environmentai diversities throughout the country
Constraints related to thebe aspects have dela;yed for many yea.rs its implementation in the
Casamance area, located Ln the southem part of Senegal.
The latest sutvey çonducted at the national level showed that the draf? animais’
population in Senegal w&; composed of i30-140,000 cattle, MI-1 FLO,OOO donkeys and
200,000 horses (Lhoste, 1988). The Casamance province accounte~ for less than X 5% of
the national totals. Amoig other types of constraints to have delayed the implementation,
were the animal sanitary ic:onditions of the area in relation with the prevalance of the
sleeping sickness. This cbaracteristic has prevented until recently (!in the 1990s) the
introduction of donkeys ia.nd hor,ses for draft purposes.
In the Basse Casmance region, the prevalence of trypfanomiasis disease
transmitted through the tse-tse fly tumed out to be the major barrier as the sleeping
sickness introduces a setious limitation in the working capacity of the draft animals. The
w
/
8
most reliable source of draft energy is represented by the Ndama cattle, &s taurus, (87%
of the total), .while the utilization of horses and donkeys (3% and 10% respectively) is still
marginal. The Ndama breed bas more toleranc;e to ,the disease and is well-adapted to warm
and humid ckmate, The castrated Ndama male oxen are used for traction whiie the cows
are kept in the herd for their milk production. The draft team is always composed oftwo
oxen of simibar weight.
1.2. Statement of problem
TO drlte in the Casamance province, ~gricuhural production based on animai
traction is below expectation. Its promotion .initiated since the early 1960s did not bring
about the expected increase in food production. In fa&, the leveï of adoption of the
technologj by farmers appears to be very low for several reasons mostly sanitary and
economical’. in the process of adoption, fat-mers are very selective in choosing implements
in relation to the level of thel first investment in terms of affordability (Ndiame, 1988). In
general, f&mers justify the use of animal traction for land preparation because it is faster
and presents less drudgeq than using hand tools (FALL, 1985).
Fat-mers in the South did not adopt the technology until some changes were made
to the yoking system. Fa,rmers use different size and multi-purpose yokes for bath
cropping (head yoke) and transportation (neck yotke). Among other factors, the design of
the yoking system was found to affect the working capacity of oxen. In terms of
efEciency, the fitness represents an important criterion when evaluating the work output of
’ The process and level of adoption of improved technologies by farmers in the Basse
Casamanee region is being investigated by a PhD student at Kansas State University.
‘I*.vm%r-,.---“-.- ---a----
---
---I<IuII
I,
9
/
draft anima!s. The yeking system assures cornpatibility In the hnk bettveen the animals
(power source) and the hitched farm equipmcnt (implement irrr cart). A less effective
yoking system is expebted to turn into an appreciable source of potential energy wastage
and tc cause rapid fati!gue if harnessed animais are not working together as a team. A poor
teaming also, afI?ects the quality of the Be1.d operation in terms of timeliness, effective field
capacities: and jpoor ~4ork quality.
Thc real relationohip between the design (width) of the yoke used and the working
conditions (type of im@ement and soil) is not known well enough. to help predict the
amount of pulling forde and energy required to perfarming the diiferent field operations.
In the decision-nnaking process it is important to fit the job to the animnls by choosing the
more suitable team of &en to perform the task How does the size of a head yoke affect
the draft required to pull different types of implement in a sandy textured soif at different
moisture regimes? Thti: determination of this type of relationship is very importa.nt in
evaluating the expectekl role of draft animais in! farmers’ crop,ping system.
1.3. Objectives
The objectives of this study were as follows:
1. TO develop a methodology that Will better assess the Impact of animal
traction at the.farm level.
2. TO develop an expert system computer program to evaluate and prediet
the impact of animal traction at the farm level b;y using a Knowledge
Base; h$anagement System (KBIMS);
1 0
A. significant amount of information on animal trac,tion is available through the
literature but is not systemized or structured into a standard ~&II. The deveiopment of the
expert system a.ppears to be the star-t of such improvement to ease decision-making. In
general, a Kairly Iarge number of technical parameters are involved in the decision making
process. It is a real challenge for decision-makers to integrate local and scientific
knowledge into an accessible structured database. The expert system is expected to
provide the bridge between scientists and the rest of the commumty.
In combining artificial intelligence (AI) and conventional computer programming,
expert systems have been used in a wide range of applications (Raman et al, 1992).
The work presented in this study is organized as follow:
- Chapter 1: Introduc:tion.
The introduction presents an overview of agricultural mechanization in the worid
and specifically in the developing countries and in Senegal in parti,cular, with a focus O~S
sma.Il-stalle farming systems. The statement of problem and the objectives of the study are
afso introduced.
‘- Chapter 2: Presentation of the area of study.
This chapter describes the production and. cropping systerns encountered in the
l
area of study. The notion of agro-ecosystem is introduced and discussed from the point of
view of animal traction utilization.
- Chapter 3: Determinants of performance of draft animals.
The discussion presents different relevant studies conducted on animal traction
a
evaluation in various parts of the world with an ernphasis on Senegal. The main
*a
~**uIII-““,.-
-----.“..---l_l_
I-m--
--
II
11
,’
determinants and parameters of performance are presented.
- C%apter 4: Mat$rials and methods.
The organization of the field data collection aiong with the expert system building
is presented chronological~y with an emphasis on the different statistkal designs.
- C%apter 5: Farnj characteristics and systerm management.
- C%apter 6: Workîng capacities of draft i3Il.imdS.
- Chapter 7: Fielb operations and draft requirements
These t.hree chapters deaf with the presentation and discussi,on of the analysis
performed. In the course if analysis, the resuhs are: compared to existing results in the
literature for first hand validation.
- Chapter 8: Expert System to evaluate the impact of animal traction in the Basse
C a s a m a n c e regqon.
This chapter descrkbes in details the proce ss of building the expert system program.
It shows how the collecteb data and models developed in the analysis part are used. In
some cases, tbe collectedidata are reinforced by relevant secondary data from previous
studies conducted in the &-ea of study by the authcr or by the IS~CRA Djibelor’s
Farming Systems Researdh Team.
‘- Chapter 9: Conclusions and recommerrdations
What are the per(pectives in animal traction studies in the Basse Casamance region
and in Senegsal in generalk What. are the expectations placed on the expert sy!stem
program? Other similar c(uestions are addressed.
Chapter 2
PRESENTATION OF TEIX AREA OF STUDY
2.1. Presentation of the area of study
The Basse Casamance region is located ix1 the southern part of Senegal and covers
7,300 square kilometers (km2), etiending from the Soungrougrou Valley to the Atiantic
Ocean (Figure 1). With population densities ranging from 27 to 60 inhabitants per square
kilometer, there is increasing pressure on land usfe in relation to t:he weather (Table 3).
Table 3: Land distribution
Designation
rea occupied (h
- - -
-
-
-
-
Cultivated areas
109,500
Mangrove (swamp wood)
124,100
Estuary and sedinnent
102,200
Forest
I89,3QO
IResidence area
204,400
- - -
-
‘Tot:al
730,000
-IE-
-m-
-S~u;ce: PIDAC ,Ei by A. FALL (1985)
-
/
'-7I
No a,nimal trac&ion; Wpland
SENk/WL
'.
3: Manding labor organizatlon
b.,,,
Lim.ited animal traction;
-.
Rice. important.
!pAku,
J: Manding ls&or organization
'. -4
Animal traction important;
\\\\
Upland cropa important.
t
k
5: Diola labor organization
\\ "Y =- .----..? *:,.*
ata!drr
b
-+--p,
Anima1 traction important;
- - - J
Uplnnd srops +aad rice.
Figure lr Agro-ecosystem zones in the Basse Casamaniee regiom
i
-I-I.
--CIc-
_..,----
--
‘.
I
- _ .
_ .
. I
_ .
. .
- ”
-
.
-
”
I
. I
-
. _
- ,
.
.
.
.
.
.
.
.
.
.
.
I
.
. ,
.
.
.
1
.
< .
.
.
. ,
.
.
,_
,.
.
.
.
<”
.
.
.
«
.
.
.
.
.
.
.
. ,
.
.
.
.
.
.
.
.
.
.
“ .
.
, .
.
<.
.
.
.
.
. ,
.
.
.
.
.
.
.
. . . ~ . . .
. . . . .
MaY
Jkme
July
m.lg . sept.
act. N o v .
-a-- Normal
* Zigpnfnchor 92
-a- cHxxmly0 92
4Bignona 92
Figure 2: Rainfall distribution frrsm North
(Bignona) to South (Oussouye) and during normal years
The IDiola ethnie group represents 83 percent of the total ipopulation, foilowed by
Bainouck (6%), the Manding (5%) and others (6%).
The climate is sub-Guinean type with a large maritime influence. The occurrence of
one rainy season per year from the month of June to October represents the main
characteristic.
The last two decades bave witnessed a sh’ortage of rainfall throughout the whole
countcy and in the Basse Ca.samance region as well, where the type of landscape has
nJIwr-,.-P..*mm--
_-- -.-._-- -
--m-
--
--Q.
---
induced significant changes in the cultural practices. The amount of rainfall deoreased by
an average of :13X, from 1,500 mm during normal years to 1~~100 mm (39.4 in) (Figure 2).
The lowest amount of railbaIl, 800 mm (3 1.5 in) aras recorded in 1,983 in Ziguinchor,
capital ofthe region.
The topography c$ Senegal is Bat with an average land level elevation of 200 m.
The physical aspect of the Basse Casamance regior: ^ : “aracterized by a dense network of
swamps and srnall stream/s flowing into the Casarnance River. The landscape of the entire
region is d(ominated by a \\Succession of upland and lowland areas. During drought periods,
the smah streams offeredj privilege waterways to :salt intrusion from the Atltintic Ocean
through the Casamance @ver to the iniand valleys following the upstream-downstream
water movement.
With the increase in sait concentration throughout these years, combined with the
low pH of sail and watek, advanced cherni reacrions started takirrg place’in lowland
areas to compromise their agricultural productivity (especially for rice production). Soi1
surveys conducted in the region yie~lded the tremendous figure of I80,OOO ha nosn-
productive, salty soils @art-y, 1986; ORSTOM, 1983). Most ofthese soilis are located
along the CJasamance Dover and its main tributaries (Guidel, Kamoubeul, Soungrougrou,
Baïla).
According to thkir position along the transect, crops are designated as either
upland (groundnuts, maire, millet, sorghum) or lowland crops (mainly rice). AI1 crops are
rainfed.
Several field studlies have been conducted to approach the agricultural and ethnie
diversity of the area. O~IE major finding was the: identification of Gve agro-ecosystem
_P
,’
1.6
zones offering different agricultural production and cropping opportunities. Three criteria
.
were r,tsed to achieve this zoning (Equipe Systemes, 1984):
1. Labor organization at the fkm level between male and female in relation
to the trop to be grown and the agricultural operation to be performed
(tillage, weeding, etc);
2. Ratio of upland and iowland areas in terms of relative importance of
cultivated areas and types of trop;
3. Importance of anime1 traction utilization in the cropping system and the
associated level of rnechanization.
Each of these five agro-ecosystem zones presents different agricultural activities in
reiation to the type of the natural resources avaiiable (type of soil, water quantity and
quality, forest) and the most dominant tthnic group (Equipe Systémes, 1983).
In response to the decrease in Foductivity, farmers developed new strategies to
meet their food security requirements ;Posnler et 4, 1988). The coping strategies to deal
with the, new climatic conditions (drotight, shortening of the rainy season and shortage of
rainfdl) were: primarily based upon a jrogressive transfer of resources from lowland to
upland areas. These major changes ir farmer’s traditional cropping system affected mostly
the use and land tenure at the villag: level. The very first noticeable effect on the
environment was the deforestation, vhich srarted taking place since the early 1970’s to
induce a disequilibrium in the land a!.xation process to different activities. The objective
of the deforestation was to satise tb increasing demand in cropping land. Through this
--
““lll,rl-,.------._..~-..
---
-
-...-..-
3-
--1
--
\\r
/
13
type of management strategy, farmers tried in igemeral to secure production through the
expansion ofupland C$O~S rather than through intensificatio%i.
In an agro-ecobystem characterized by an availability ofupland axeas, the use of
animal traction tumed; out to be the factor contributing most towards the expansion of
fields cx-opped with cabh, trop (groundnuts) to the detriment of cereals such as millet and
maize. The cereals ard now mainly confined in backyards and compound fields. Within a
short period of time (l+s than ten years), th.is type of cropping system afft:cted the
precarious equilibriurn which bas always existed between agricultural and livestock
activities marked by tlie progressive disappearance of long term fallow along lwith village
level grazing areas for! livestock.
In the process,, the meclhanized fieid opemtions on newly cleared 8and brought
about advanced stage4 of soi1 d.egradation to compromise land productivEty. In most castes,
the lack. of good mana/gement practices has led. to alarming soi1 losses.
2.2. Types of Producih system
The agro-ecos@tem zone is defined as a system in which farmers; are carrying out
productive agricultura] activities within their natural environment. The e@vironment is
mainly composed of &ural resources (soil, water, vegetative caver) and: represents an
area homogenous eno$gh in terms of constraints and production opporttinities. The most
Sign&ant observatioi unit to evaluate interacticms and performance of the system is the
fann level (Tabfe 4). $%Lis level of aggregation represents a finalized production system in
itself as it comprises tbe tenter of the decision-making process which identifies agricultural
production objectives ‘and decisions upon resources allocation (Sonko anc! Fall, 19%).
..^_~ . ..- .^ -..
-- .
_--_--_--- -- .
L
d !
-^-_-.
_
_.
18
_-.
--
_...
--_-
_
..--
--
-
--.-
--
2.2J. Agro-ecosyst/em 1 (Oussouye)
Located in the mostjsouthern part of the reg,ion, this agro-ecosystem is composed
of a tigh proportion of lowiand areas. The landscape is domina.ted by swamps, saline and
acid soils (,‘“sol de tannes”)Yiand very few upland fieyds (village sites). The ratio of lowland
to upland areas favors by far lowland areas. Transp~anted rice is the major agricultural
activity (Equipe Systemes, i1984).
Field operations are performed according to the Diola tradition of so&l labor
division by gender: men p&form land preparation while women perform rice nursery
preparation, transplanting bnd harvesting. Women also help in weeding of upland crops
confined iri small fields.
Land preparation i.b manual and starts whenever enoug,h water is stored in the rice
fields. As the cumulative dainfall increases from the beginning of the rainy sehrion, a
network of small dikes difiding the main rice fields into small subplots, 10 m2 to 30 m2 in
area, is used to capture wiater runoff. Farm sizes are generally small averaging 1.3 ha per
household. More than 60% ofthe fields are c:roppc!d with rice.
Ether non-agricuhural acti,vities contribut:: significantly to the total farm revenue:
fishing, prodtrcts from the forest (palm oil ex.trac.:ion for example).
Gvestock activitbs are very important; cattie, poultry, swine are produced. Animal
traction is not used in agricultura~ activities for tiifrerent reasons. The primary constraint is
sanitary with the prevalence of trypanomiasis (G&ssier, 1966; Traverse, 1974). The
alternative to animal traction is to use rentes pcwer tillers fiom non-govemmental
organization (CADEF, &JAEDO). However, \\ery few farm activities are mechanized
through this rental sch&e.
\\,
20
2.2.2. Agro-ecosystem 2 (Blouf)
Located in the northern part of the Casamance River; -agro-ecosystem 2 represents
a continuzrtion of agro-ecosystem 1 but with more available upland areas. More than 30%
of fieldis produce upland crops wbich are mainly groundnuts and millet.
The labor organization (at the household level is the same as in agro-ecosystem 1.
There is no use of animal traction since the big drought event in the late 197Os, which
significantly reduced the avaiiability of animals for draft pur-poses. Farmers are still very
open to the technology (Fall, 1988).
2~2.3. Agro-ecosystem 3 (Niaguis)
Agro-ecosystem 3 is located eastward of agro-ecosystem 1. The population of the
r
area is composed of a multitude of ethnie groups living in different villages, which
r.
continue to practice their original cropping system. Upland areas account for 60% of
cropped fieids.
The introduction of animal traction is still in progress: 30% of cropped areas are
m
mechanicaily tilled with animal-drawn moldboard plows. It is important to mention that
L
faTmers located in this area were the first in the region to experience the use of motorized
equipment for rice land lpreparation with the MAG project (1969- 1973). The
governmental regional extension service SOMTVAC-PIDAC took over the project in 1984
and pursued the rental of mini-tractors and power-tillers to farmers.
I,
. .
-_ --
<~cwY-ll.l~..-<-_
..__- “-
B----,--”
-----
-11--
-----
-.
.,
2 1
/
2,2.4. Agro-ewsystem 4 (Sindian-ECalounayes)
Agro-ecosystem 4 is located in tke northeastern pati: of the region. This area has
been culturally unded Manding ethnie group influences even though the population was’
originally Diola. Theisacial labor division according to gender is organized around the
type of trop to be grbwn. Men are in charge of a11 upland crops which ~cupy more than
84% of cropped areap, .while women are specialize in rice and vegetable production.
The use of animal traction is well implemented and most field operations, from
land preparation to gioundnut lifting, are mechanized. The a.verage fam ,size is around 4
to 5 ha.; rn’ore than S&!! of croplands are mechanically plowed with draft animais,
especially oxen.
0ver the last &y: years, with the emergence of many local farmei-s’ organizations,
mecbanization is beco/ming a major concern ~LYS the population. Farmers iare getting more
involved in the purcba/se of tractors of differernt sizes fused md new) wit&:in a wide pow’er
range (10 I:O 45 hp). The tractorization process is mainly orientecl towards improving the
timeliness of land predaration in the rice fields.
Different econbmical activities are conducted apart from ,the Aeld crops. Livestock
(ca&, sheep, poultry], tree cropping (fruit and palm) and vegeta.ble gar&:ns represent
important sources of *venue.
2.23. Agro-edosystem 5 (Diouioulou)
Agro-ecosysteb 5 covers the northwestern part of thie region. Upland crops are
dominant, and occupy~75% to 80% of a11 crop1:an.d. The labolr organization is Viola type
_ _ ^ .---- -
Y
2 2
/
based on division according to gender in relation to the type of field operation to be
carried out. Men perform land preparation on both the uplatid and lowland areas.
The use of animal traction is also important but to a lesser extent than in the
previous agricultural situation. The level of mechanization barely exceeds land preparation
for 60% of ail cropland.
2.3. Cropping systems
In normal rainfall conditions (1500 mm), the agricultural production of the Sasse
Casamance region is mainly oriented towands riche production. The amount of rice
produced was estimated to be around 75,000 metric tons representing 30% of the national
production. The shortage of rainfkll by 20 to 45%, along with the correlated salinization
process of most rice fieId SC&, has compromised the regional production to the n’oint that
the region has become a net rice importer since 1978 (Equipe Systemes, 1983; Jolly et &
1988). The coping strategies developed by farmers to countleract the decrease in rainfall
were tfo move to more diversification at the regional agricultural production system level.
Ta#ble 5: Cereal yields in Casamance (kgha)
Crops
Millet’Sorghum
.
Maize
Upland rice
Lowland rice
Aauatic rice
S&mp rice
Source: MDR, 1986 “Etude du 5iecteur Agrrcole:
Plan Cérèalier. ”
*‘“sn^la,--.^----.-._r__l--~-.-----
-..- --.-
. ..--.-
-u----.---u-
‘----“N-sll-mwanm,8q~~
.,
/
2:3
A bigger emphasis was put bn the cash trop (groundnuts) while areas cropped in maize
increased hy 19?/0 between 1970 to 1982. Other crops grown ‘&e millet and sor,ghum.’ The
yields measured in farmers bonditions are generally low (Table 5).
2.3.1. Groundnut broduction
The draft animal b+ed cropping system is mainly driven by groundnut production,
the only cash trop support/ed by an organized market system a.t the national l&el. The
cropping ofgroundnuts h& significantly contributed to the extension of croplwnds gained
by rapid deforestation and clearing of new lands: upland crops represent 75O/6 to 80% of
ail cropiands in the regioru/ and fields cropped with groundnuts alone accountis for 45% to
55% of a11 upland crops *ea. The average fa:rm yield is 954 kg/ha (Posner et d, 1988).
In the agro-ecosyhtems loc;ated North of rhe Casamance River (zones 3,4 and 5),
groundnuts are grown eit/her as the first trop in a 2 to 3-year rotation with tiillet!sorghum
production following lan/l clearing or altemating with an annual fallow (2-ybar rotation).
Whereas in the South @@es 1 and 3), the cropping pattern is continuous (I-year rotation:
groundnuts/groundnuts)/as
land availability is a major constraint,
Three varieties o$ groundnuts are grcrwn: 28-206 and 69- 10 1 (110 to 120
days/cycle). These are ijtproved varieties introduced by agricultural researçh institutions
and govemmental extenbon agencies. The third variety is local and is called “Bourkouss”
by the famers. It is a sqort cycle variety (90 days/‘cycle) and Is not suppo&d by the
officiai market system. $‘he local variety is widely used for late secding. The improved
varieties are mostly grobn for their high oil content and are cultivated as cash crops: 69-
101 in the North and 24-206 mainly in the South.
,!
2 4
Groundnuts are the only trop in the cropping system that is highly mechanized
through animal traction. In some agro-ecosystem zones, the level of mechanization
reaches the harvesting operation with the utilization of animal-drawn groundnut diggers or
lifiers. The main difference in the cropping pattem among farmers resides in the type of
land preparation: either flat (with a moldboard plow) or ridged.
if
23.2. Cereals
I !
Cereals (millet, sorghum, maize and rice) account for the remaining cropped area,
45% to 55%, of the total cropped area. Millet and sorghum cari be in either a rotation with
groundnut or altemate with annual fallow. Sorghum is vety seldom in a mono-culture
cropping pattem. It is generally intercropped with groundnuts. Al1 varieties used are local
Millet is important in agro-ecosystem zones with large Manding cuhural influence.
The seeding starts at the very beginning of the rainy season. The varieties used are local
(Sarrio) and have major pest management problems especially with birds and stem borers
(Pentatomidae Coreidae). Millet is generally seeded on ridge plowed land to protect the
seed and Young emerged Iplants against excess moisture. The land preparation is either
mechanically or manually performed in relation to the timeliness of groundnut field
operations.
,, j
Maize production is crucial to farmers as it represents a relay food supply for their
consumption. In general, the harvest takes place during the rainy season and before the
end of the month of September. Maize is mainly grown in agro-ecosystem zones 3, 4 and
5. The cropping pattem is continuous mono-culture (1 -year rotation) in soils rich in
organic matter. Maize fields are mainly located in the backyards of the farmer’s
,-,1-,-1w--
.-<-._ vpn---
--
-w-
IC
-
2 5
compounds on relatlvely small areas (less than 0.5 ha). Lately, there is #ai. trend to cropping
maize out ofthe badkyard on! newly cleared land, but the major concern to farmers’is soi1
fer’tility and the need of field ,protection against pests to secure production.
During the last 10 to 1.5 years, high yielding varieties of 90 day$/growing season
bave been provided do farmers by research institutions and governmental extension
agencies: composite !ZM 10 (yellow grain) and hybrid BDS (white grain). New high
yielding varieties (Synthetic C, Makka, etc.) continue to be distributed.
Maize is the second most commonly rnechanized trop afier groundnuts. The level
of mechanization involves land preparation, seeding and weeding. The intensity of
mechanical cultivatioh is determined by the type of land preparation (flat or tidge).
Ridging ta,kes place gbnerally when farmers are having weed problems ar excess moisture
in relation to the topography. Maize seems to be the trop that offers the largest margin of
progress in terms of ektension of cropped area and increase in level of production
(Ndiame and Fall, 19!!!9).
Most of the ride production is carried out in lowland areas. The cropping system is
rainfed. There are three types of rice fields ranging from strictly raimed (uipland) to aquatic
(lowland). The aquatic rice fields depend greatly on the accumulation of water from runcoff
within the watershed tb low areas inside the vallleys. The third type of rice field is assisted
by the rising water table, which provides enouigh moisture to the rice plant. The types of
soi1 used for rice production vary tiom sandy (upland) to clayey soi1 (lowland). A large
number of rice varieties (local or improved 110 to 120 days/growing season) are used
according to the field conditions in terms of so,il moisture, sahnity, pH, fertiiity, etc.
Tmproved varieties arelwidely used by farmers: DJ 684D, IRAT 133, IR. S., DJ 12-5 19
2 6
Historicahy, the mechanization of rice producti’on in the region started with the
introduction of the first series of power-tillers (DONG FENGTM) fiom The Republic of
China through the MAC project (1969- 1973). One ma:in objective of the project was to
improve productivity throu.gh better timeliness and good quality land preparation (FALL,
1985). The major constraint was related to the size and shape of the plots, which were too
small to facilitate the maneuverability of the power tillers.
The use of animal traction started earlier, in 1967-1968, but was mainly confined
to areas where researchers and extension workers were conducting experimentation
(TUVERSE, 1974). Farmers try to combine animal traction for land preparation and
seeding and tractors for lan preparation whenever possible.
Rice is generally direct seeded by hand broadcasting right after land preparation
except for the aquatic rice grown in lowland areas where transplanting is manually
performed by women when water is still ponding. Rice production requires effective land
preparation before seeding in order to improve weed control and to incorporate organic
matter into the soil. This field operation depends largeiy on the water content of the soil.
”
Other crops are also important at the farm level for their economical and
/
nutritional values. The are,a allocated to these types of crops is generally small: cow-pea,
_j
cassava, sweet potatoes and vegetabies.
The Basse Casamance Agricultural Development Master Plan published the
prediction of production trends for major crops in relation to extreme weather conditions
scemios (drought and normal) (Table 6). The potential increase in productivity to reach
these preciictions is real and depends mainly on how fiirmers Will be able to improve their
capacity for carrying out field activities on time
/ I I
.*UC-,.-.---.-----I
_*_-
- _ - *
__-__
~
..--
--
.-----------
27
Table 6: Lower Casamance agricultural producitiion trends
in thousand metric tqns
-
-
Conditions
-
-
Drought
‘Normal
- -
Year
- - -
1979
1985
1990
2000
1979
1!2%5
1990
2000
1. Senegal
Millet
407
419
430
452
415
497
577
‘779
Rice
6 8
7 0
7 2
7 6
7 7
;93
109
1 5 1
Maize
3 2
3 5
3 8
44
3 4
41
4 7
6 4
Groundnuts
- - -
613
708
798 1015
749
946
1150
1699
-
II. L. Casamance
Millet
2 5
2 6
2 6
2 8
2 6
31
36
4 9
Rice
2 2
23
23
2 4
3 8
4 6
54
74
Maize
6.6
7.2
7.8
9.0
6.8
8.1
9.5
lL!.8
Groundnuts
3 0
35
3 9
5 0
3 6
45
55
82
Source: Master Plan Report, Vol 1 by HARZA (1984).
2.4., Agnicuitural productivity
At the farm level, the increase in productivity to reach a sustainable cropping
systern is a major concern of farmers. A number of decisions are made towards meeting
their production objectives. ‘Ihese decisions are of two kinds (Jouve, 1986):
- gatbering and organizing available resources
- choosing the most appropriate production process and techniques in
relation to the environmental conditions.
According toithe resources available, fat-mers use different strategies to meet their
needs. Hclwever, a number of limiting factors have an impact on their respective
performances and hellp explain why fat-mers do what they are doing.
2 8
2.41. Land productivity
Th.ree major soi1 types, with slopes ranging from 2 to 8%, have been identified
thugh different studies, soi1 surveys and land evaluation in the region (Char-i-eau, 1974).
On upland areas, Oxisol and Ah301 soi1 types dominate whereas in lowland areas, there
are three subclasses: Molisol (rice tïelds without sait), Entisol (Sand deposit in the estuary
of the Casamance river) and peatsol in swampy areas. The Oxisol soi1 type (Ferralsol
according to FAO classification) is sandy-clay, red to reddish in color and characterized by
the presence of kaolinite a.nd a low cation exchange capacity (CEC= 11 rnmol/g). Most of
the upland crops are grown on this type of soil. The Alfisol soi1 type (beige in color) is
mainly located on the transition zones (from upland to lowland) along the transect
between the Oxisol (on top) and the Molisol types. Its organic matter content is low
compared to the Molisol but has a more silty texture than the Oxisol type. It also has a
higher clay content and is recognized by its grey to blackish color.
A number of studies have shown that the increase in land productivity is primarily
bound by the drought situation as farmers do not use minera1 fertilizers due to the
shortage of rainfall (Equipe Systemes, 1984). In general, the fertility level of these soils is
very low because of advanced weathering and lack ofgood management practices
regardless the amount of rnanure provided by the livestock to the cropping system. The
cerelals grown in the farm backyard (maize) and at the village level (sorghum and millet)
usually receive manure and organic matter fi-om different origins, but most of the
groundnut. fields do not receive any type of fertilization. The extensive cropping on newly
cleared lands has generated advanced levels of soi1 degradation. The Sand deposit in the
lowland areas by water runoff from upland fïelds tends to compromise rice productivity.
/ l.a>-/“-l
_-<--.-
.,_._ --.---.
--
_.___.
-.-~
w----
cc
--
2 9
The amount., regularity and distribution of rainfall during the growing season
represent important liimiting factors towards reaching bettef land producstivity. The
decision to start field activities, in terms of working days, for example, is mainly subject to
these twc factors. The date at which the rainy season begins remains the most
unpredictable factor.
The rain disttiibution pattern varies from North to South and fi-om year to year.
The state of drought ihas introduced a significant variation in the amount of rains received
in different localities: 45% deficit in the northern areas and 26% in the southem areas. The
year-to-year variation is the most commonly used indicator by farmers to de’cide about the
cropping pattem (pribrities) and the field operations scheduling towards minimizing risk.
The soi1 rnoisture condition is one important parameter considered by Al-mers to clecide
when ,and how to carry out the field operations. The rainfall pattem is a key factor in the
decision making process to meeting timeliness of farmers’ field activities,
Asronomic tr$als cond.ucted in on-station and on-fat-m conditions have
demonstr.ated the benefits of plowing on land productivity through significant yield
increases (Equipe Systemes, 1984). Usually, land preparation on a11 cropland begins afl:er
.tiater has infïltrated lhe soi1 to a certain depth. The depth of the wetting front at which to
star-t the plowing opdration is mainly dictated by the amount of working power availablle at
the farm level in relation to the soi1 consistency (physical and mechanical properties). On
upland crops, draft atrimals are used when the wetting front is located at around 12 cm or
more aller the useful’rain event (FALL, 1985). In lowland areas for rice production, this
limit is diRerent depending on the type of rice fields described earlier.
/
30
I
2.4.2. Manpower and labor productivity
The availability of manpower throughout the rainy. &ason is the most important
factor for success in carryiing out field activities. The avaiiabiL*y of manpower is mainly
affec.ted by the migration of the Young and active population. Crie way for farmers to
solve this problem is to turn to the mechanization of more and n 3re field operations.
However, the level of access to farm equipment anct gower unitc (animal or tractor) as
factor of production differ fiom one agro-ecosystem to another md fiom one farmer to
another (Fall, 1985; Sonko, 1986; Ndiame, 1987). Xis varia, In in production
opportunit,ies has induced very significant differences in farmers’ performance across the
region. Different studies used the level or degree of mechanization and the availability of
manpower at the farm to divide farmers into three distinct categories at the regional level
(Equipe Systemes, 1985; F’all, 1990; Sonko, 1990):
- Category 1: Non-equipped households (64%)
The manpower (men and women) ranges from 1 to 4. The peak labor demand is
located at the beginning of the rainy season during land preparation and requires 40?/0 of
,total farm seasonal labor demand.
The intensity of the labor constraint is different fiom one cropping system to
anotlner. In agro-ecosystems with important use of animal traction, zones 4 and 5 for
example, some farmers rent a pair of oxen and plowing equipment for a day or two to
meet their fïeld schedule.
For this category of farmers, the area cropped .with groundnuts (cash trop) is small
compared to the cereals: rice fields cari occupy up to 60% of a11 cropped areas.
._,
s _-,.---,-.
-...--m-
_--__--
------.-
--
---
,a--
3 1
- Catbgory 2: Low-levei-equipped households (29%)
Thés categoj of farmers is characterized by the am&int of matipower, ran@,g
between 5 and 8 fati workers. The average is 6 active workers per hobsehold.
Land prepar&.ion is generally the only mechanized field operatibn. The remaining
field operations are rhanually executed which. shifis the peak labor demhnd from land
preparation to weedibg and harvesting. The use of animal traction has a real. impact on the
amount of cropland Iieserved to groundnut production (5 to 8 ha). The’majority of the
!
available Carm workeirs is generally direc,ted to the groundnut fields and penalizes the
cereals. Women invoilved in rice production according to the gender diGsion of labor look
more. and more for oiher alternatives (power tillers or mini-tractors) to’improve timeliness.
- Category 3: High-levei-equipped households (7%) ’
The number cbf farm workers involved in production during the year is large,
rang,ing fiom 9 to 2 1 j persons. These households are well equipped and have 2 to 5 draft
animals.
The area crodped with groundnuts is very important, and avera$es more than 8 ha.
This trop receives thie equipment priority. Farmers are sometimes for&d to use
supplemental extema/l manpower to meet timeliness, as a11 the manpower is not available
during tht: weeding oiperation of a11 cropped land.
2.5. Summary
The identificajtion of the different agro-ecosystem zones at the regional level
represents a major step toward the analysis and understanding of famlers’ specific working
3 2
conditions. The formulation of recommendations to improve their production techniques
has more chance of success ifbased on existing opportunities. The factors involved in
impro,ving productivity in farmers’ conditions are complex.
TO increase productivity and sustainability, it is important to help farmers improve
the soi1 management practices and conservation techniques. Improving tillage practices is
an area to explore in order to reduce soi1 losses from upland areas and to mitigate the
induced and rapid decrease in soi1 fertility. TO meet these objectives, farmers must be
involved at a11 phases of research and development. It is important, for example, that
farmers be able to evaluate the physical properties of soils and Select the right tool and
implernent to perform any given field operation in a timely manner (Henin, 1990). TO
achieve this level of decision-making involving prediction of the power required to carry
out field actïvities on time, the physical and mechanical properties of the soi]: found in the
regïon must be understood in terms of behavior under the influence of various factors.
Farmers in the same agro-ecosystem generally follow similar planting sequences as
a village’s coping strategy against major pests: maize-millet-groundnuts-sorghum with
very little variation in the case of agro-ecosystems 4 and 5 (Equipe Systemes, 1986). The
cropping practices usually dif?er from farmer-to-farmer in terms of type and amount of
agrku:ltural inputs used to carry out the field operations (Ndiame, Coulibaly and Fall,
1988). The degree of combination of these cropping practices also called “itineraire
techniques” explains the diflerences in agricultural production performance among
fkrmers.
The utilization of animal traction by farmers is viewed as an alternative to
improving labor productivity. The potential behind the technology needs to be investigated
.v --r.-s.
-
__--_
_.---
-<----
- - -
__-.__
-
----
w-
more to optimize the output ofthe draft animais. It is important at this point to understand
the difficulty and difference in levels of access to fat-m equipment and draft animals.‘ As a
I
factor of production, the utilization intensity of farm equipment is expect(ed to be different
from one agro-ecosy$em to another and from one farmer to another. The variation in
petiormance among fsrmers belonging to difirent mechanization categories needs to be
better analyzed in ordbr to be able to evaluate the impact of the technology at the fat-m
Ievel and to capture aI1 of the information availabfe in a knowfedge-based system for iater
utilization
Chapter 3
I
DETERMINANTS OF PERFORMANCE~ OF DRAFT ANIMALS
3.1. Constraints to animal traction utiiization
Since the 197Os, development projects involving animal traction have flourished in
.
many developing countries around the world. The utilization of animal traction appears to
- ‘
be very attractive for many small-scale farmers as an alternative way to agricultural
mechanization (Starkey, 1989). A quick overview of the promotion and impiementation of
*
different animal traction-b;ased projects has shown that the utiiization of draft animais is
not a panacea to solve the problem of a11 small-scale fax-mers in the Third World Countries
Eicher (1982) wamed about the “maximum potential benefits” approach used by many
agricultural researchers when carrying out animai traction project evaluation. He pointed
I. I
out the tendency of the approach to inflate the projected retums and long-run economic
I
w.
profitability from the technology when data are mainly collected on experiment stations or
demonstration farms.
The failure of different projects bas demonstrated that the diffusion and
implementation of animal traction have been limited by many factors. The most important
IJ.
are tradition, climatic conditions, diseases, poor infrastructure, lack of appropriate
1.
transportation and communication systems, inefflcient marketing and credit systems, poor
3 4
_-
- _._--
3 5
trainirng, long duration:of the learning process, ineficient extension servides, cropping
systenls types, and inafppropriate veterinary services (Pelissier,.’ 1966; Mum;ciger, 1982;’
Havard, 19:37; Fall, 1900; Ndiame, 1990). Diffi:rent studies have addressed these limiting
factors and their impact on the implementation of animal traction throughout the world.
One exampie reported by Starkey and Faye (1988) in relation to the sanitary constraints
was th.e situation in Caineroon where the use of draft animals was confinad to the
Northern ccltton producing areas of the country with low tse-tse flies infe6tation. Similar
limitations z.re also des@ribed elsewhere: Central African Republic, Conga, Equatorial
Guinea slnd Gabon whdre the use of animal traction is limited to a few mission stations;
less than 1% of the fatiers in Zaire make use of draft animals. In the Basse Casamance
region the diffusion of Animal traction from North to South had to face bath tmditions and
sanitary ‘constraints alofig with the cropping pattem (Pelissier, 1966; Traverse, 1974; Fall,
1985; Sonko, 1990;). Us utilization is mainly confined to the Northem pa& of the region
where 18% (in the Northwest) to 68% (in the Northeast) of the farmers ovjn at least. one
draft animal (Sonko, lg90). These general constraints explain most of the variability in the
geographical distributidn of drait animals in developing countries. A numtier of well-
identified fartors are in+olved in explaining the level of performance of small-scale farms
that ha.ve adopted the tbchnology.
3.2. Selection of draft ianimals
3.2.11. Availabitity
Farmers have a irelatively wide variety of animal species to chose fr,om for energy
supply. Goe and McDawell (1980) in their guid.elines for animal t:raction’s utilization gave
36
the potential of different species ranging from dog to herse. There was a large variation of
dafi potential among species. According to the survey carrüed oit by Schrnitz et a1 (1991)
in 32 countries of Latin America, Africa and Asia, it appeared that cattle were the most
commonly used animal breed fbr draft purposes.
An appreciable amount of quality research is being carried out throughout the
world to increase the work out,put and efficiency of working animals. Crossbreeding
within species is a widely used technique to improve draft potential. Different studies
comparing breeds have shown that it was highly recommended to farmers to use local
indigenous breeds as crossbred animal generally demonstrated inferior working
performance. This was especialky true for tasks of long duration. For extended periods of
time, experience has shown the local breeds always produced the most steady output
(Goe, 1983). Another important advantage of the local species described in .he literature
is theYr availability in numbers and accessibility to farmers, and their adaptability to the
environment. An effective breeding program is always expensive to maintain in order to
produce enough crossbred animais for a continuous supply of draft animais to the farmers.
The supply aspect was found to be very important for the household in relation to the
replacement of injured, dead or sold animais, and for the overall stability of the
technology. In his investigation of the Basse Casamance situation, Sonko (1985) analyzed
the diverse reasons of removing animais fiom service and the impact of the removal on the
household performance (Table.7).
The analysis carried out show& that the technology was unstable for 71% of
animal <traction users who generally possessed one pair of oxen. The loss of one animal
during ,the growing season (2,6%) was likely to compromise the objectives of the farmers’
1 -...rm,--~.--..--~~~
uI,-------
----..
--------.-,
I
37
,
Tdble 7: Reasons for removing animals from service
i
Reasons
Percentages (%)
-
Sales
Death
Trade
Slaughter
ThefI
Rented out
Out for traini/ng
Return to hetd
GifI
U n k n o w n
-
Total
3
/
Sourde: Sonko, 1985
production systems, unless they decided to rent a team of oxen from beh:er off farmers
For .this rcason, he suggested that the optimal. number of oxen at the farrn level is 3 to
secure at ‘east agricuhural production.
3.:!.2. Selectibn process
Specialists bave agreed on most of the criteria used to Select draft animais.
Usudly, draft animalsi for a given breed or species were either selected from the herd of
the household or village, or bought from the rnarket. The conformation (physical form,
neck, legs and feet, hem, etc...), the character, the age, the sex and the weight were the
major criteria for the belection. They also presented the most important fiactors affecting
the work output and working career of a draft: animal. The tractive potential of draft
/
38
,
animals is generally determined by their liveweight LW in kg and their weight distribution.
In comparing crossbred and indigenous cattle, Goe (1983) pointed out that the amount of
‘_
weight distribution over the front legs should enable the crossbred animal to generate
more power. The format and age at which an animal warted its working career were
determinant in its future performance.
The selection must be done from available species in relation to the expected task
to be performed by the animals in the given environmental working conditions. Crosseley
and Kilgour (1983) presented a list of advantages and disadvantages of the most
commonly used draft animals in trop production: horse, ox, donkey and dromedary.
Horses appeared to be very costly to maintain and easy to train while oxen were
characterized as easy to feed and to acquire.
3.3. Training of draft animais
The working capability of animals is determined by in large by the quality and the
level of training. An important number of publications on this topic are readily available to
farmers. These publications are generally used as guidelines by numerous projects during
their ‘early implementation phases. The selection and training of animals are crucial for
their .working career. The rnain objective of the training sessions is to build a working
behavior into the animals in terms of responsiveness and obedience. According to Conroy
(1992!), the trainer and the animais must develop a relationship based on respect and
dominante, as animals deserve feed, rest, shelter and social interaction. During the training
sessions, it is important to Imanage the temperament of the dorninated animals. For this
purpose, animals should not be frightened and negative aspects of the dominante
39
reinforcement must he avoided. One frequent negative aspect seportedl by Keith (199LI)
was to put the animals in discomfort while working which could cause;them to associate
I
the trainer or eventual driver presence with pain and upcoming punishment. The castration
of bu11 is one commdn technique used by fat-mers to calm down the animal and to reduce
their aggressiveness The operation is performed between the age of 1 to 5 years in
relation to the age of! first selection and training.
The age at which anirnals are selected and trained varies amongland within species:
1 to 8 years for buffdoes in Asian countries and 2 to 4 years for cattle ir general (Goe.,
1983:). A good rule oif thumb is to start the trait-ring when the body weight and physical
development of the ahimals have reached 50% to 70% of the matured weight (Goe, 1983,
Lee et al, 1993) (Tadle 8).
Lhoste (1990) reported that farmers with draft cattle in the Groundnut Basin (Sine
Saloum) cf Senegal u!sed different management strategies in relation to the age of train@
and length of working career of the animals on-the-one-hand, and the existing marketing
opportunities at the end of their career on the other. The training ages ranged from 2 to 5
f
years; the ;.ongest carder was 6’ years and the shortest 3 years.
In *:he Basse dasamance region, studies conducted by Sonko (1990) and Ndiame
(1990) shcwed that the training age ranged from (AGETRAIN) 2 to 0 years and the
average working career was 6 years. The training sessions were carried out one month
before the rainy season. Keith (1992) working experience with cattle showed that weaned
calves between the agks of 2 to 6 months were easy to train as they trusted and quickly
accepted the trainer. In such cases, the training would require a firm commitment and
planning fiom the trairher as the use&1 work would really begin at 3 to 4 years of age.
40
Table 8: Age and weight to initiate training of cattle for work
---
Adamawa
Azaouak
-
Maure
2.5-4.0
-
North Soudan
Shorthom
S huwa
N’Dama
Senegal
“,
Fulani
Ankole
Bukedi
Bomro
2.5-4.0
2.0-4.0
-
Ongole Sumba
Khillari
5.0-5.5
2.0-3.0
-
J
Lohani
3.0-3.5
3.0-3.5
-
Valfey
-
Source: Goe and McDowell, 1980.
I
!
4 1
Ah these factbrs are highly interrelated and contribute significantly to building up
the animais’ willingness to work which tumed out to be an important determinant 0fth.e
animals’ working performance. The willingness to work affects the perdentage of the body
weight (PFILW in percent) available for potential power use. The wiilingness affects the
efficiency of the draft! animais teamed through the hamessing system.
3.4. Harnessing systkms and team constitution
3.4.1. Types bf harnessing systems
The main objective of a hamessing system is to allow better transmission of energy
from thle animal to the implement. The ter-m “hamessing” is used in its general sense to
mean yoking as well (Le Thiec, 199 1).
A wide range of hames,ses and yokes are being used around the world. This topic
has been addressed in depth by different resea:rchers (Hopfen, 1969; Starkey, 1989; Le
Thiec:, ‘I 99 1).
Generally, har-nessing systems are divicied into two categories: hom or head yokes
and wither; or shoulders yokes. The tirst type of yoke is tied in fiont or behind the horns.
In both cases ropes or leather straps are used to firmly secure the yoke around the horns.
As mentioned by Le Thiec (1991) and Starkey (1989) a better fit could be guaranteed by
carving the part of the yoke resting on the head of the animal. A head yc&:e must be strong
and light fer better cornfort. Th.e best features of cattle to fit the head yoke turned out to
be a short and strong neck, and strong homs. Impirical experiences have shown that this
type of hamessing syst/em presented advantages of better coordinating the movements of
working animals speci&y when instantaneous maximum draft is required, by allowing a
4 2
rapid adjustment’of the line of draftto reduce the angle of pull. The main disadvantage
found for continuous work was the more rapid state of fatihe induced by the rigidity of
the yoke attachrnent (Le Thiec, 1991). He warned also against the bad quahty of some
ropes used by fanners, which over time tut off the homs.
The shoulder yokes were mainly designed for forward motion. They are adjusted
to pull against the shoulder muscles or the front base of the hump for Bos indicus cattle.
The simplicity of fabrication has caused them to be widely used specially to pull different
types and sizes of carts for transportation. The main advantage of this design allows the
animals to move their heads freely which induces less fatigue than head yokes. Experience
has shown that animals hamessed with withers yokes refused to pull and try to get away
when maximum draft is required.
3.4.2. Eflkiency of harnessing systems
The work effrciency of animais was found to depend primarily on the effectiveness
of the energy transmission. A subjective evaluation technique of this effectiveness by
animal traction users is the classification and codifica.tion of the energy transmission mode
into (1) inefficient, (2) average and (3) efficient hamessing systems. According to Goe and
McDowell(l990), a good and well-design hamessing system would help increase animal
draft capacity by providing at the same time enough comfort to the animal. Starkey (1989)
stressed the quality of a good harnessing system to maximize its positive effects in terms
of mutual encouragement of teamed animais towards achieving a sustained work.
Researchers have been focussing on animal factor ergonomies to improve the energy
_ I
43
transmission for better efficiency. One major step was the improvement of the yoke design
for a better attachment to the working animals (Le Thiec and Havard, 1995; Roosenberg,
1992).
The potential energy and pulling capacity were increased by teaming animals
through a harnessing system which coordinates their movement. The process of tearning
must star-t at the animal selection phase in relation to the task to be pertformed and the
assessment of the expected draft requirements. According to Conroy m.d Rice (1992)
oxen do not need to be identical in order to make up a successfùl working team. The
suggested selection criterian ‘to focus on was animals’ ability to work together. They urged
oxen team trainers to look for simiiar temperament, agility, size, conformation, and speed
This Posit:ion was not far from the recommen,dation made by most animal traction
speciahsts to Select closely identical animals (CEEMAT, 1971; Watson, 198 1). In the case
of anirnals of different size, the strongest should always walk in the fi.rrriow when working
with a moldboard plow in order to balance the draft demand between thc two animals.
The pulling force potential expected fom a team of animals was found to be less
than the cumulative potential of individual animal making up the team. J&tarks (195 l),
cited by Goe (1983) reported the results obtained with different teaming, scenarios. They
showed that the harnessing system induced a iloss of efftciency amounting to 7.5?/0 for a
pair, 15% for triplet, 20% for quadruplet, 3 0% for quintuplet, and 37% for sextuplet. The
researchers of CEEMAT had specificaliy addressed the case of teaming oxea and found
similar results (CEE&T, 1972). Their fmdings showed that the potential of one animal
must be multiplied by 1.9 to estimate the expected potential from the pair of oxen (10% of
efficiency 10s~). Lee et a1 (1993) translated Marks findings into a general formula to
_ I
44
evaluate the efficiency of the team (TE) in relation to the size of the team (TS):
TE = I-(TS-1)*0.07.5
(1)
The average liveweight LW of each animal working in a team cari be determined if
the draft requirement DR for the task to be performed and the percentage of the body
weight (BWP= PF/LW) that an animal could develop into pulling force (PF) were known.
Watson (198 1) proposed another formula by assuming a 15% loss of efficiency for a pair
of oxen:
(2)
LW=
DR
0.85* BWP
The numerical default value of BWP used most oflen for cattle was 0.10 (10%) for
long term traction (CEEMAT, 1975). This value needs to be validated with the local breed
used for traction.
3.5, Working capacities’
In general, fat-mers intend to get the best performance out of their working
animais. The handling and tare of draft animais (health, housing, feed) were found to be
an important consideration in order to keeping them in good working condition. When
conditions were bad or the management poor, animals would become more sensitive to
stresses and diseases.
4 5
.‘3.5.1. Climiate effects
In the hot abd humid climate ofthe tropics, exhaus&d animais will rapidly tend to
lose thenr tolerance ‘to txypanomiasis and mortality Will increase. Acccnrding to Crossley
and %le;our (1983), veterinary services must be made available especially to farmers. They
must look daily for the more! common symptoms related to the appeanance and behavior of
the a.nim,als: urination, mucous membrane, temperature, pulse and respiration rate. Only
healthy animals wer’e found î:o be able to produce the desired power output. Falvey (1986)
defïned health as th$ optimal tinctioning of animal body. This optimu&: could easily be
off:jet by the effects’of stress on the animal?, immune system and by injuries caused by the
harnessing and hitcting system. He warned 1:hat stress could be createql by overworking
animals which are ita a state of malnutrition. This would impair the immunoglobulin
production due to dkficiency of proteins.
Experience dnd tests conducted in different places suggests sttict management
practices to minimize the drastic effects of ciimate on the work output,of draft animais.
Besides the sanitary aaspects, the environmental effects on working anitials’ needs to be
investigated more toi help farmers Select the best animal. Draft animais do react to changes
in working regime ahd envirclnmental conditi.ons by making physiological adjustments to
maintain their body Beat. These adjustments caused thermal stress on thle animais reflected
in their finture behavior (Singhal and Tomar, - ). A few studies have tried to show or
quant@ how much tbe temperature and the humidity really have affected the working
efficiency. These twd factors relate directly to the ability of the animais to dissipate the
heat produced durin$ the task performance. The heat transfer from the animal body to the
environment is a natWa1 phenomenon. The heat exchange is generally conveyed through
/
4 6
conduction, convection, radiation and secretion of sweat. Reported results from Premi
(1979) stipulated that the respiration rate of bullocks doubled ‘with every 10°C rise in
temperature. Goe (1983) reporteci sirnilar results with the Hariana; the respiration rate anc
body temperature increased as the ambient temperature increased from 10 to 28°C. An
increase in relative humidity reported in the same study affected the blood and respiratory
system as it accelerated the puise rate. Zander (1972) gave the range of 30 to 70%1 for air
humidity to be a comfortable climate. The combinat:ion of these environmental factors was
found to play an important role towards the animal’s state of fatigue when at work (Table
9). Animal traction users agree that moming hours represented the best time period to
‘carry out field operations. These important results need also to be validated in different
‘working conditions.
Table 9: Environmental factors on draft animais
Type of
Pulse Rate
- i
Animal
per minute
Respiration
Temperature
under jaw bone
-
Rate per minute
“C
Cattle
40-50
1 O-30
38.5-39.5
Horses
3 6-42
8-16
37.5-38.5
Donkeys
40-60
40-56
36.0-38.0
Source: Crossley and Kilgour, 1983
The most detailed study available which reports the direct environmental factors
affecting physiology deal with the human body (Essay, 1969). Essay developed a linear
,
4 7
mode1 to evaluate a parameter called the Temperature-Humidity Index (TMI) which is
related to the dry bulb temperatu:re and dew point temperature. Gar THI values between
70% and ‘79%, lOo/o of 100 peop1.e were feeling uncomfortable. The THT cari only be used
as an indicator and must *be set for draft animals ,working in a given environment. Other
empirical :methods based on management practices which included rest periods were
recomrnendeci by field researchers to optimize the output of working animal:s in a given
environmental condition. The level of fatigue of working animals was indical:ed through
physiological and behavioral manifestations (Upadhyay, 1989): unwillingness to work,
fiequent stopIj, respiratory problems, etc. In tropi.cal climates, the maximum working ,time
ANWKI-IR in hours must not exceed 8 hrs per da.y. Devnani’s (198 1) investigation showed
that bullocks’ work time during the summer must be kept between 5 to 6 hrs per day and
divided into two sessions (morning and evening). In the Basse Casamance region,
Traverse (1974) suggest$d that the Ndama cattle could be managed in relation to the
intensity 0.f the work:
Light task: 7 hrsVday + rest at mid-day
Medium task: 7 hrs/day + r-est at mid-day f 1 day off/week
EIeav task: 3.5 brs/day t’rom 8:00 to lk:30 am.
This classification was made possible in relation to the draft required to perform
different tasks: transport, harrowing and weeding were considered as light jobs; plowing
sandy soi1 (upland crops):as a medium jobs; and working clay soi1 (lowland rice lïelds) as
heavy jobs. The most recent studies conducted by Fall (1985) and Sonko (1986) have
,
48
shown that in fact, farmers in the Basse Casamance region used draft animals for less
hours per day to perform field work.
3.5.2. Pulling force potentiai
The potential pulling capacity PF in N or daN of a working animal was found to be
mairly determined by the animal species and breed, weight, age, nutritional and health
condition, level of supervi.sion and training, and methods of harnessing. A well
documented research program showed that the pullin,g capacity was directly proportionai
to the animal body weight (CEEMAT, 1974; Goe ami McDowell, 1983; Starkey, 1989;
Campbell, 1990) which in turn was affected by the body condition of the animal and the
harnessing system (Table 10).
The potential was expressed in animal body weight perceniage (BWP in “2). The
. I
average long ter-m pulling force (PF in N or daN) developed by working animals was
1
._
found to be within the ranges of 15% to 20% for donkeys, 12% to 15% for horses and 9?/0
/
_ I
to 12% for cattle (CEEMAT/FAO, 1972). Goe (198’7) mentioned a draft potential of 18YO
to 23% of the body weight from Ethiopian draft oxen
-
t
Table 10: Work potential of man and various animais
I
.a.
I
_.. I
Item
I Man
Horse
0x
Donkey
Weight (kg)
8 0
400-700
300-900
100-300
Pull (N)
100
500
500
400
Speed (m/s)
1.0
1.0
1.0
1.0
Power (kW)
0.1
0.5
0.5
0.4
Daily work hours (h)
6
6
5
4
Daily work output(MJ)
2
10
9
6
Source: Crossley and Kilgour, 1983
Research findings showed that during their growth phase, the closer the draft
animais’ laveweight approached their mature, or genetic weight, MATUREL,W in kg, the
heavier the type of work they could perform. This point has been emphasized by
Upadhya:J (1989) who stipulated that heavy animais possessed more pulling capacity than
light animais but were less suited to speedy work. In evaluating this potential and taking
into account the morphological aspects, Goe (1983) and Lee et d (1993) suggested a
body weight parametier called the condition score CS. The condition score excludes from
tract.ion any animal storing less than 1 (afIer rounding off). This corresponds to less than
5QX of the body weight ratio depicted in the following equation where LW in kgfand
MATURELW in kgf were respectively t he animal live and mature weight:
The liveweight LW could be measured by using proper scaling devices generally
available at the veterihary semices or the governmental extension agencies. In areas
lacking infrastructure and effective technical support services to farmers, researchers bave
developed different models to help estimate the weight of cattle based an the
morpholoçical features of the animal (CEEMAT, 1974; Watson, 198 l):,
(girth $“* L
LW -= -
300
5 0
Where the &th (in inches) represents the circumference (CIRCUMF) of the torso
at the point of heart and L (in inches) represents the length of’the animal from the point of
shoul’der to the rump. However the applicabiIity of this formula was limited to non-
castrated oxen. Cows were also used in areas where they were culturally accepted. The
pulling capacity of cows was found to be approximately 5% lower than for male animais
(Lindsay, 1986). Also cows needed special tare and more attention than maies in order to
maintain their reproductive capacity and milk production. In the Sine Saloum region of
Senegal (known as the Groundnut Basin), survey repo,rts indicated cows were 26% of the
draft cattle and were used as a second or third team member to support the already
existing draft oxen (Lhoste, 1990). In the Basse Casamance area, only male cattle were
used for traction (Sonko, 1985).
3.5.3. Work potential and energy requirements
3.5.3.1. Instrumentation
The evaluation of work output and energy expenditure was found to require from
“!
simple to highly sophisticated instrumentation in relation with the type of parameters to be
included in the energy determination. Research efforts were mostly centered around the
development of different sensors to help measure integrated parameters as heart beat,
pulse rate, respiratory rate, draft, speed, oxygen consumption, body temperature, etc.
l-1
Substantial resources were put into the development of new methods integrating
I
physiological and physical parameters. At the same time, the classic approach using a
simple dynamometer and integrating distance traveled DISTRAV with animal speed
ANS’PEED measurements was still widely used because of its simpiicity compared to the
---
I
51
new methods. The neiv methods were divided into direct and indirect measurements.
These two techniques were well documented by Lawrence arid Pearson (1989).
The direct method involved the measurement of heat output dissripated from the
animal’s body. The he;at generated during the task performance was produced by the
oxidation of carbohydrates, fats and proteins to carbon dioxide, water aad urea. The
method presented some serious limitations. The animals must be kept ericlosed in a
chamber. The theoretical and practical aspects, of the measurement techhque need to be
investigated more in oirder to make the results usable in field operations.
The indirect miethod took the measurements one step further by aiddressing not
only the heat production but also the amount and the quality of gaseous exchange
involveld. The technique was based upon the calculation of the heat dissipated by the body
using a measurable parameter represented by the amount of gaseous exchange. The
underlying principle was that no heat was produced without gaseous exdhange. Oxygen
(OI), ca.rbon dioxide (CO4 and respiratory quotient were the parametersi :mostfy
concerned. Two types of apparatus were used by the Center for Tropical Veterinary
I
Medicine (CTMV) at the University of Edinburg:
- Classic Open Circuit System for use on the animal at wdrk and at rest.
The system was not portable and the use of a treadmill was necessary to Ienable the animal
to work while staying in one place.
- The P!ortable Rreath-by-Breath Analyzer in the ,form of an airtight
facemask fitted with a !low meter was to measure the volume of each breath. Samples of
i
5 2
I
expired air were to be analyzed in the laboratory with the application of correction factors
.
related to temperature, atmospheric pressure and hurnidity. The apparatus could not be
I
‘.
used in the field because of the laboratory intervention, representing a serious
disadvantage.
The AFRC instrumentation package was also developed for continuous
* i
/
measurement of both physiological and mechanical work performance variables (O’Neil et
al, 1989). The data stored in the data logger during the work session were: draft force,
an,gle of pull, speed, heart rate, breathing rate, body temperature and stepping rate. The
only disadvantage resided in the non-availability and high cost of sensors used to collect
the data. The CEEMAT package put more emphasis on the mechanical variables.
Other methods to estimate the different levels of energy required of working
animais are indicated in the hterature.
These levels of energy have been designated as maintenance, work and liveweight
gain by many authors (CEEMAT/FAO, 1972; Goe, 1980; Lawrence, 1985; Falvey, 1986;
Pearson, Lawrence and Ghimire, 1989; Lawrence and .Zerbini, 1993).
3.5.3.2. Energy at work
FAO (1972) reported the energy requirements ,for ruminants at work to be about
2.6 fold the maintenance energy. Pearson and Fall (1993) believed that the FAO method
overestimated the energy required. For the specific case of cattle, the values published by
Lawrence and Zerbini (1989) were the most recent. Thle technique used the factorial
method and was found to be feasible for long measurements in the field. It was based on
4
_
.mmmamqm-.III-IC-----
-.
-
---
.,.
I
5 3
determining the extra dnergy E used by working animals of liveweight LW (in kg) when
walking on a horizontal distance HD (in km), carrying loads LC.(in kg), doing work W (in
W) while pulling loads, and walking uphill a vertical distance VD (in km):
The coefficients given by Lawrence and Zerbini 1989) were mainly fo,r cattle:
A = 2.0 kJ/kg of LW per km of ItD
k = 3.0 kJ/kg of LC per km of I LD
G = 3.3 kJ/kJ of W
1) = 0.033 kJ/kg of LW per km of VD
The values of the coefficients were obtained through laboratory experiments,
which represented a major disadvantage compared to measurements perfomed in the
field. The coefficients need also to be evaluated for local breeds.
3.5.3.3. Energy for growth
IMost draft animals reached maturity during their working career Energy was
required to foster their growth while at work. The level of energy was found to vary in
relation 1:o the management practices and the task to be performed (worktonly versus
work and fattening). Goe (1983) wamed that the energy used by working Ianimals for
growth needed more investigation as the values, published in the literature ipresented
important di,screpanciei. He presented research results obtained in West Africa reporting a
rate of growth of 1 kg/day for cattle trained at 0.5 to 3 years of age. Experiences
conducted in fat-mers0 condition in Senegal showed that the oxen used in the Groundnut
*.1i
5 4
Basin areas gained an average liveweight of 33% per year or 0.16 kg/day, 25% per year or
. . i
0.11 kdday and 17% per year or 0.05 kg/day. This was obse&d for draft animals tramed
!
at age 2-3 years and weighing 170 kg, 3-4 years and weighing 220 kg and 5-6 years and
weighing 300 kg (Lhoste, 1990).
3.5.3.4. Energy for maintenance
The energy required for maintenance has been investigated by numerous scientists
and appears in most animal traction manuals and guidelines. Lawrence and Zerbini (1993)
presented MAFF (1984)‘s contribution in evaluating the energy required for maintenance
MAE (:in MJ/day) for cattle ofliveweight LW (in kg):
M.4E = 0.36 * (L w y
(6)
The working capacity and the source of energy required for draft animais at work
l.i
were described to be largely dependent upon the availability of quality feed and the
efficiency with which this feed was transformed into potential energy .
3.6. Feeding systems
3.6.1. Feed quality and availability
An important amount of information was published in the literature to present
research carried out towards :improving draft animal feeding systems. The rationale behind
providing working animals access to better feed was the maximization of the amount of
work ‘obtained from the limited available resources. Feed is the main source of energy for
II*UIïULU,U-.“..-,-l---
--
---
111-----P
I
55
animcils’. One; important aspect of that research was to increase the nutritional value in
order to provide the nutjrients required for the execution of vari&us physiological fimctions
as work. AS presented aarlier, working animals rely on ,feed to provide energy l‘or
maintenance, growth anti work. ‘The necessary a.mount of feed for a workihig animal could
be calculated if a11 the required energies were known. The only limiting fa&or identified in
the literature was the non-availability of a reliabl:e feed supply (Eicher, 198:!). This was
found to be <.I major technical constraint as feeding prac,tices commonly ustd by
smallholder farmers were simple and dependent on naturally available feed resources:
grazing a.nd utilization of seasonal trop residues and by-products. But moW: trop residues
were regiarded by nutritionists as, poor quality feed (Wanapat, 1989). Ffo&es and
Bamualirn (1989) suggested that the quality of straw feed must be improved by the
addition (of urea. A limited number of natural fo:rages species were found to be high
protein,: Leucaena, Cassava, sweet potato and banana (Table 11). Beside the mine,ral
requirements, draft animals need high enerby feed mainly based on glucose and fat to
satisfi the energy demand of difl?erent physiological functions as sustained muscular
activity. .Aceording to Teleni and Hogan (1989), work was achieved through the
combination of various systems ‘based on adenosine triphosphate (ATP) gtneration,
respiratory and cardiovascular fimctions.
3.6.2. Dlet for working animais
Current research results on nutrition are yet to be transferred to fdrmers. In the
West AfYican region wihh a long dry season, reports show that farmers are still
experiencing a shortage! of energy as land preparation occurs at the end of the long dry
season when animals are in their poorest condition. For most farmers, the diet ofworking
Table 11: Composition and digestibility of feeds (DM basis)
-
Minerals
(db DM)
Types of feed
DM
CP
CF
Fat
DMC
%
%
%
%
%
- - -
-
Ca
- P
- N i
-
S
-
Crop residues
Rice straw
33-95
4-6
33
1
1
38-55
Maize stover
85
6
32-46
0
2
54
Sorghum straw
91
4-8
33
1
.
42-48
Sugarcane tops
2 8
6
35
1
4
6 2
Cassava leaves
2 6
2 0
21
1
2
72
l
Banana leaves
2 4
12
23
1
2
45-66
la
Potato forage
1 1
1 3
14
<l
3
93
Tree legumes
.A
/
Leucaena
3 0
2 9
21
4
1
4
51-60
Glyricidia
2 7
2 4
18
4
1
3
6 7
l
Concentrates
IRice grain
8 6
10
10
2
1
3
1
63-95
Rice bran
8 6
10
2 0
1 l-22
1
18
1
2
43
Maize grain
86
10
2
5
0
3
1
1
65
Soya bean meal
9 0
37-48
5
4
3
6 <l
3
25
” t
Kapok seed meal
9 0
31
3 0
8
5
1 3
50
Cotton seed meal
8 6
4 4
14
8
4
11
28-50
I
Molasses
75
4
0
0
11
1
9
9
100
Tapioca waste
9 0
2
3
4
6
2
6 9
Palm kemel cake
8 9
19
1 3
5
3
7
2
2
= = r =
DM: Dry Matter; CP: Crude Protein; CF: Crude Fiber;
DMD: Dry Matter digestibility.
Source: Extracted from Ffoulkes and Bamualim, 1989
1. IIV.IIIU
___.
--~
. ~ . _
-.-
-I
--
_ I
57
animals is mainly based on roughage and browse. The loss of weight during this dry period
represents a serious limitation to performing heavy fi,eldwork at the beginning of the
cropping season (Starkey, 1989). As pointed out by Wapanat (19893, ver-y few
experiments on feed rations were conducted in on-farm conditions, in order ‘to evaluate
the feasibility and acoeptance of the rations by the farmers.
Th.e literature is documented well in methods and techniques of evaluating the
amount of’ feed necessary for working animal:;. In general, a11 the authors take into account
all forms of energy involved in satisfying the demand of physiological fujnctions. Lawrence
and Zerbirti (1989) included in their estimation the energy for milking CO~S a.t work based
on h44FF (1984)‘s model. Crossley and Kilgour (1983) partitioned the energy content of
feed into gross energy (GE), digestible energy (DE) and metabolizable energy (ME). It
was only at the stage of metabolizable energy that the animal could really use the energy
to perform work. Al1 these levels of energy were interconnected:
A4E =iT8*DE
0
For poor feed (straw for example):
DE =:O.-15*GE
(8)
For good feed (grains for example):
DE =0.85*GE
I
5 8
The energy content of feed was generally determined on a dry matter basis (Db{)
and largeiy depended on th.e percentage of digestible dry matter (DMD). Crossley atid
‘.
Kilgour (1983) indicated in their analysis that the approximate maximum amount of DM
called the Appetite Limit (AL in kg of DM per day), was related to its liveweight:
AL =O.O25*LW
(10)
3.6.3. Forage Unit
Goe and McDowell (1980) reporting different nutritional research activities
conducted around the world used the concept of total digestible nutrient (TDN) and
forage unit (m) to convert the amount of feed into GE and the ME’. For the West
African region, most of the research activities have been performed by the French
institutions CEEMAT and JEMVT. A comprehensive list of nutrient sources for working
I
animais and their energy content using the feed unit (FU) concept has been published
(CEE.MAT, 1975). Average estimates of feed unit for maintenance (FUM in kg)
published by CEEMAT (1975) could be modeled (R2= 0.999) into a relation between the
liveweight (LW in kg) and the number of forage unit FU to provide the required level of
energy2:
FIJM’ = 6.321* IO” * (L w -- 0.7089
’ TDN = carbohydrate -+- protein + 2.25 * fat
1 kg of TDN = 4.409 Mcal or 18.447 MJ of DE
= 3.615 Mcal or 15.125 MJ of ME
1 kg of FU = 0.74 k:g of Barley
” The raw data used to develop the linear mode1 are presented in the appendices.
J
5 9
The mode1 output gave the same TDN results as the values published by Goe and
‘.
McDowell(1980) and Goe (1983) for West Afiica.
’
3.7.. Implements selection and characteristics
Implement selection described by different authors was reveaied to be as important
as the selection of draft animais (Le Moigne, 1981; Starkey, 1989). Foir many farmers, at
the earlier stage of adoption, the selection could be very complicated, a:5 a variety of
equipment designs are available on the market. The introduction of farfn equipment in the
cropping system by development agencies w(as not neutral and was done as a means
towards achieving production objectives. The major challenge in the piocess was to fit
most of the tasks to the draft animals by selecting the right implements, Once the
implement was selected, its utilization would likely predetermine the e4t:cution mode of a11
the following field operations.
Animal-drawn implements were generally designed with ergonomie considerations
for comfort and efficiency. They were mainly composed of three majors parts: working
components (moldboard, ridger, tines, and shares), frame or beam (equipped with or
without uheels) and handles. ‘The shape and sharpness of the working components were
mostly responsible ofthe draft: required (DR in kgf or daN) to move the implement in the
soi1 in the horizontal direction of travel. Each farm impfement had its own requirements in
terms of draft and energy in relation to the type of soil. A number of draft tests conducted
in difI’erent areas of the world have been reported in the literature (CEEMAT, 1975; Goe
and McDowell., 1980) (Table 12).
6 0
Table 12: Draft requirements for farm implements (Equatorial Africa)
-
-
Implement
-
-
I’low
Indigenous
Moldboard
Disc
Disc harrow
single action
double action
Rotary tiller
Harrow
spike or peg
spring tine
I
Roller or puddler
Lfeveler, float
Row-trop planter
Grain drill
Transplanter
=
Source: Goe and McDowell, 198 1
61
For giwn working conditions, the draft required was found to depend on the type
. .
and condition of soil, and the angle of pull (y in degree angle).’ Al1 farm equïpment
involved in trop production were mainly characteriz.ed by
working width (IMPLWWD in cm) and depth (IMPLWDP in cm). If I!F in daN is the
animal pulling force, the draft required DR in daN cari be expressed as:
DR = (Pi?* COS(j)
D!fYerent studies have shown the importance of the angle of pull ‘y and its
!
dependenzy on the type of draft animal and the harnessing system (Dibbits, 1993; Le Thiec
and Haval-d, 1995). The optimum angle of pull was found to give the truc line of draft.
The line of draft or traction was defined as thle straight line connecting the point of draft
located at the hamess level (yoke clevis) and the
tenter of soi1 resistance on the working component of the implement (Watson, 198 1). Le
Thiec and Havard (1995) found the angle of pull for Ndama cattle to be between 15” to
18” and 19” to 23” for Zebu. The variable horizontal and vertical hitch points located on
the implement hitching bar are used to adjust ,the line of draft and the puli.ing angle in
order t,o balance the soi1 forces acting upon the working components and the weight of the
implement .
The utilizatiorj. of animal-drawn implernents requires skilled operators to minimize
eneraT los,ses. Differeht studies found farmers differing in skills and years of experience.
Eicher (19’82) reportitig Barrett et 4 in a study conducted in Burkina FasI; pointed out
i
that there was a slow leaming process.for farmers who were using donkeys or oxen for
_ /
the fir:st time. It took about three to four years before a farmer-knew how to use a
complete package of equipment. Implements were provided with different control systems
I
and adjustment devices. Besides the horizontal and vertical adjustments previously
mentioned to adjust the working width and depth, the wheel was designed to provide
betner stability. Adjustments by triai and error were commonly used to reach the point
where the wheel was not pressing too hard on the soi1 surface and cutting into the ground
to increase the soi1 resistance (rolling resistance and bearing friction) or not lifting off the
soi1 surface (Watson, 198 1). IIandles were also provided to help steer and tilt the
implement to correct the line of draft while in motion. A good adjustment was achieved if
no force was felt at the handles to prevent the operator fiom steering the implement and
contrcrlling the working animals (Crossley and Kilgour, 1983).
3.8. Soils
3.8.1. Soi1 types
The draft DR in daN is mainly defïned in relation to the resistance of the soi1
against advancing implements’ working components (WKGCOMP). In other words, draft
is the force required in the horizontal direction of travel: as stipulated by ASAE Standards
(1984) to initiate and sustain movement. In trop production, the physical characteristics of
the soi1 were found determinant in the evaluation of the amount of force required for
pullïng a given implement.
In the IJS system, soils are generally classified into textural classes or orders.
There are 12 major textural classes according to the size and distribution of the primary
. -.,4~~-*““-I-I~----.-
-
--.
--
-p--e
6 3
partides (percentage of sand, silt and clay). Soi1 structure relates to the arrangement of the
_ .
partkles into peds or aggregates (shape and size) (Foth, 1990). The different soi1 classes
are summarized in the commonly used textural triangle. Within the categories of soi1
taxonoqr, soi1 orders represent the highest category. There are 11 ordkrs divided into
subgroups, family and series.
As pointed out by Charreau (1974), the soi1 survey and classificiatiora of the West
African region had been made possible by a g,roup of more than twenty French researchers
(LRAT and ORSTOM) who helped train Afiican soi1 scientists during a twenty-year
period. They used the French system elaborated in 1967 by the CPCS to develop the Wesr
African scil map. The classification was made of 12 classes subdivided 11nto subclasses,
groups and subgroups, families, series and types. The family was composed of soils
formed fiom the same kind of parent material. The series related to the position on the
toposequence and the types relate to the texture of the surface horizon. They also
presented 5ve vast zones originated from the following parent materials:
” a. In the northern part of Senegal, Mali, Niger and Chad: sands ,lnd sand Idune:s
b. In the Niger Rive?s arc and a great part ofthe Chad basin: alUuvia1 deposits
ranging from pure sand to fine clay
c. Westward, south-westward, and eastward of the Niger Rive?s arc in Burkina
Faso, in two-thirds of Senegal and in Southem Chad are found the Terminal
(Jontînentaliformations. . The materials generally went through a strong
ièrralitic althration.
d. In the southern part of Niger River$ arc, in the central part of Malï, and in the
western part of Burkina Faso is a vast area of Co:mbro-Ordovician sandstones
overlanded by an ancient “lateritic” iron pan . . .
e. In the southem part of Mali, in Burkina Faso, in the central part of Chad:
crystalline shield made up of plutonic rocks, :etr.Torphic rocks and volcanic
rocks. . . . ”
The soi1 map showed that large areas were occupied by grey and yellow-to-beige
I
1
ferrugineous soïls, and red fèrralïtic types of soil. The soils were also characterized by a
I
I
,j I
field capacity (SEC) of 15 to 20% v/v and a wilting point c !VP) of 7 to 9% v/v.
The oxic horizon of the ferrugïneous and ferralitic sons are mainly made of a
mixture of three elements: kaolin, amorphous hydrated oxides (iron and/or alumirum), and
quartz. The amount of kaolinite is highly variable. The soils are desaturated and
characterized by a low cation exchange capacity CEC due to the presence of kaolinite. As
described by many soi1 scientists, they have almost reached the end point of weathering in
their evolution (Charreau, 1974).
These two types of soi1 are dominant in the southem part of Senegal (Ducreux,
1984). The gray ferruginous soils (“Beige soil”) presents a sandy to coarse loamy texture
in the Upper horizon with a subangular blocky structure and a -fine loam to fine clayey
texture in the deeper horizon with angular blocky structure.
The ferralitic soils also called “Red soil” (due to the presence of amorphous iron
oxides) present some Alfic characteristics with subsurface horizons of clay accumulation
and medium base supply. The texture and structure are similar to the grey type.
65
3~.8.2. Mechanical and physical properties
3.8.2.1. Physical properties
Nicou (1975) and Ducreux (1984) preseinted, in their respective studies, the
ferrugineous and ferralitic soils to derive their physical characteristics and mechanical
behavior from the predominance of kaolinite and sandy to sandy-clayey texture of the
surface h,orizon. The dry climate through the wetting and drying cycles had significantly
affected ,the aggregates and structure stability. The stability refers to the ability of the soi1
aggregates to resist the disintegrating effects of water and mechanical manipulation (Jury
et al,. 1?91). The low clay content in the Upper horizons (8 to 12%) and the presence of
kaolinite were described to confer to the soi1 surface a massive to structure:less behavior
when diy. Ir this situation, the a;ggregates tended to harden and a cementation of the
whole soi1 mass takes piace (C%arreau, 1974). This process of increasing cohesion called
“prise en masse” during the drying phase of the cycle is a well-known soi1 characteristic of
the dry tropical regions. The phenomenon has been studied mainly by penetrometry by
differerrt researchers, in relation ‘to the soi1 texture, water content and porosity. The soi1
cane penetrometer was designed and validated as a standard device used to measure the
penetration resistance of soils called the “cane index” (CI in kgf or daN per cm’). The
cane index CI is used to compare mechanical properties of different soils, and to develop
performance and prediction relationships. ASAEI Standard (1984) has recommended two
cane base diameter sizes: 20.27 mm with 15 9 mm shaft diameter for soi? soils, and 12.83
mm with 9.5 mm shafi diameter for hard soils. The hand-operated cane peretrometer use’d
by Charreau (1974) Nicou (197 S) and Ducreux (1984) is described as a simple graduatecl
and Sharp metallic shaft (15 to 20 mm in diameter) with a 60” cane and a shding weight
I
6 6
(SLWGHT in kgf or daN). The sliding weight falling from a given height (HGT in m>
drives the shaft into the soi1 to a certain depth (DEP in m) (Ducreux, 1984). If SR in kgf
or daN is the soi1 resistance to the shafi penetration, then the penetration energy EDEP
corresponding to the depth DEP is given by:
II/. I
1
The penetration forces were found to be 5 to 10 times higher when the soi1 was
dry than wet. At various moisture regimes above the wilting point, field observations and
L !
soi1 manipulation in the laboratory showed that the sandy texture of the surface horizons
I
l
tends to give to these tropical soils a non-sticky, non-plastic and friable state of
consistency (Charreau, 1974; Ducreux, 1984). The mechanism involved in this hardening
process during the dry season and alter intermittent rainfalls followed by dry spells during
the wet season is still under investigation. Nicou (1975) and Ducreux (1984) defined a
textural index called the “hardening index” HI to characterize this physical behavior. The
index was expressed as the ratio of the clay content over the coarse fraction (coarse silt +
coarse Sand). They found tha.t there was a linear relationship between the hardening of the
soi1 Upper horizons and the index HI for soils studied in Senegal and Niger: the higher the
index, the higher the tendency of the soils to harden during the drying cycle. The
phenomenon was found to be important in relation to trop production (tillage and root
penetration). For land preparation, Char-r-eau (1974) wamed that during this hardening
process, the draft required by these ferrugineous and ferralitic soils appeared to be too
high for draft animals found within the area of study (donkeys, horses or cattle). A draft
-w-
,------
-y-------
--
P-m-B.8
6 7
value of 240 kgf or daN from a plow working at a depth of 15 cm in a ferrugineous soi1
‘_
was recorded in the southem part of Senegal.
5
3.8.2.2. Resistance to traction
The most frequent parameter used to characterize the mechanical itlehavior of
different soils is the specific soi1 resistance SSR in kgf or daN/cm2. It represents the sum
of the soi1 and trop resistance and the implement rolling resistance. The draft per unit of
fùrrow cross-section given by the ASAE Standard (1984) relates high-speed (greater than
3 km/h) field operations with implement types and soi1 types. With animal power,
experience bas shown that draft depends mainly upon the soi1 moisture conditions and th.e
type of implements usad.
The specific resistance was found to be highly variable among soils. Smith and
!,
Mullins (1991), cited by Lee et a1 (1993), classified the soi1 specific resistiance in relation
to their textural characteristics. They used the clay content as a criterion to divide the soils
into 5 classes of specific resistance, ranging from 30,000 N/m2 for coarse sandy soils (O?G
to 8% of clay) to 130,000 N/m’! for clayey soils (50% to 100 % of clay). ‘In these
condit.ions, the draft required (DR in daN) to pull an implement with worik:ing width
(IMPLWWD in cm) and depth (IMPLWDP in cm) in a soi1 of known specific resistance
(SSR in daN/cm2) is as follows:
DR,,,,! = (A4PL WfZ’,’ + (IMPL WDP) * (SSR )so,,
w
t*1
6 8
The main limitation of this approach was found to be that the soi1 specific
resistance was not correlated to the soi1 rnoisture regime. Diffcrent studies on monitk-ing
field activities revealed that farmers carried out field operations at non-optimal soi1 water
content in order to meet their production objectives through better timeliness (Charreau,
1974; Fall, 1985; Lee et al, 1993).
The occurrence of the first useful rain event is usually a signai for farmers to
schedule their field activities according to the number of working days available during the
cropping season. The number of working days was considered to be highly determinant in
* !
the process of evaluating how farmers achieved their production strategies in relatio:: with
the ciimate and soi1 type.
3.9. Working days
I
3.9.1. Soi1 moisture regime
I
The soi1 moisture regime described by many authors as the main criterion is used
to evaluate a working day in relation to the type of soi1 and weather. It is generally
described by soi1 physicists to be highly dependent on temperature through
eyapcrtranspiration (ET in mm/day), rainfall (RF in mm), slope of the soi1 surface through
runoff (RO in mm), drainage characteristics (DN in mm) and type of field operation to be
perfolmed (AME Standards, 1984). For every soil, a characteristic curve cari be
developed to relate the soi1 matrix potential yrn in MPa or bar and the soi1 moisture by
volume I& in %v/v. The soi1 water content cari be determined either by in-situ
measurements or estimated b:y using prediction models.
,r . --.- . ..-- -- --._.-__
x*
._I ____“_““__ -_
,
6
9
Thle first techniqwe described in many procedures of determining soi1 water content
in-situ involves the use of either soi1 moisture sensors (tensioniete;, porous blocks,
psychrometer, TDR) for volumetric soi1 water content (VSWC in % cm3icm3), or soi1
sampies with laboratory facilities for gravimetric soi1 water content (GSWC in % g/g>.
These methods are complementary. Once the soi1 particle density (SPD in gkm’), the soi1
bulk density a(SBD in g/çm3) and the water density (WAD in g/cm3) were known it
becomes easy to determine the other related soi1 parameters (Hanks, 1992).
The second technique used by trop modelers was mainly based on equations (soi1
water balance) to predict on a daily basis the soi1 moisture status. The prediiction equations
were generally derived from the hydrologie bala.nce which states that the water stored in a
given volume of soi1 (root zone) is equal to the difference between the amount of water
added (riain and/or irrigation) and the amount ofwater withdrawn (evapotranspiration,
runoff, drainage) (Hillel, 1982). In absence of irrigation, Hunt (1986) proposed the
following equation to determine the soi1 water content (SWC in mm) during day n (the
other variables have been previously defined):
s@C, =:(SWC,.,)-(ET,)+ (R%)-(ROn,h(LK)
(15)
TRO main characteristics of the soi1 water system given as boundxy values, are the
field capacity (SFC in % or mm) and the wilting point (WP in % or mm). The field
capacïty is defined in fthe lïterature as the soi1 water content, sometimes talled drained
Upper limi:: (DUL in % or mm), reached in a few days after wetting of tht: soi.1 layers
located near the soi1 surface (Ritchie, 1972).
70
At field capacity, the soi1 water drainage DN from the lowest soi1 layer of the root
‘.
zone is generahy considered negligible. Jury (199 1) argued that’ a true field capacity. does
net exist, as water Will always continue to drain under gravity reaching an insignifrcant
level after a few days (two days for rnost sandy soil). The water content at field capacity
corresponds to a soi1 matrix potential ranging from Y,-,,:= -0.1 bar (-0.01 h4Pa) to \\Km== -
0.33 bar (-0.03 MPa). During the desiccation process, the wilting point is described as the
Jower Emit (LL in % or mm) of soi1 moisture at which water is not available to plant. Only
adhesion water around the soi1 particles is retained by the soi1 matrix (Foth, 1990). The
soi] matrix potential at this stage of desiccation is Ym= -. 15 bar (- 1.5 MPa).
The soi1 water balance cari be determined with the measurement in-situ of the soi1
volumetric water content at DLL and LL. The soi1 water holding, capacity (SWHC in % or
mm), sometimes called extracrable water is defïned as the difference between the soi1
water content at field capacity (SFC) and the soi1 water content at wilting point (W)
SWHC’ = (SFCT -(HT)
(16)
The average extractable water in the root zone is 13% for most minera1 soils. This
is reiated to the texture, as plant water uptake also takes place during the drainage process
(Ritchie, persona1 communication). The soi1 water content at which the execution of
different animal traction field operations are recomrnended lies within the soi1 water
holding capacity range in the Upper 30 cm of soi1 and 15 cm.
*.MUl-mI-LIIIIIUC
,_-_ --.-
--
--
--v
--UII-*
7’1
3.9.2. Weather factors
.‘- ,
The weather related factors induce a probabilistic aspect into the evaluation ofthe
number of working days. The technique developed was based on determining the
precipitation probabilities (Prah in percent) or frequency (Frai,, in decimal). The probabilities
of occurrence was calculated by using the em.pirical equations adopted by the American
Society of Civil Engineering (ASCE) and reftxred to as the Gumbel’s equation (AID,
1977; Schwab G.O., Fangmeir D.D., Elliot W.J. and Frevet R.K., 1993):
(17)
P ra,n = 100 *(I- F,,,,i
(18)
Where m is the order assigned to the precipitation event ranked in ascendiny order
and n is the total num.ber of data points, The associated return period is the reciprocal of
the fi-e!quency Frai,,.
The probability values for working da:ys (PWJI in decimal) during chosen period
intena.ls (weekly, bi-weekly or monthlv) have been calculated and tabulated according to
difTerent geographical locations. The probabilities are used to help estimate the number of
working days (NWDAYS in days) a specific field operation cari be petibrmed within the
time inter;als at a cefiain level of confidence.
Le Moigne (198 1) had carried out a similar analysis for the West Afi-ican region
He proceeded by det&-mining the number of non-working days. For monthly period
72
inter-vals, he based his analysis on hypotheses built around soil/tool relationships. He
identified four large zones characterized by one rainy season w’ïthin Senegal, Mali, Niger,
Burkina Faso and Chad. For each zone he used the number of rainy days per month
(Table 13) with the following hypotheses:
Scenario 1:
He considered two predominant types of soil: sandy and clay soils. The sandy soils
are located at the highest position of the toposequence and are mainly used for upland
rainfed crops. Their physical characteristics are the same as the ferrugineous and ferralitic
soi1 types described earlier. Generally, these sandy soils have a good drainage systems.
They drain well and fast. In contrast, the clay soils are found in lowland areas located
within the watersheds along the main West African rivers (Niger, Senegal, Volta, Lake
Chad, Casamance, Gambia). They cari stay saturated for longer periods of time compared
to the sandy soils and prevent any tillage operation.
Scenario 2:
He identified the different field operations and cropping practices that are
significantly affected by the amount of rainfall in terms of soil-implement working
component interactions and ground trafficability.
In sandy soils, a11 fields operations involving the utilization of cultivator tines were
classified as quasi-impossible to carry out for rainfall events amounting to more than 50
nun/day. For plowing and ridging, the execution was difficult for rainfall between 30
mmlday to 50 mm/day. Seeding and weeding required 1 day of drainage after rainfall
Table 13: Non-working days for the West African region
:- *
--
!-----II
Zones (rain) /Field operations
Sandy soils
- -
-
-
-
F
-
h?
- 1
A
-
M
<== 350 mm
Plowing
0
0
0
0
Tines land preparation
0
0
0
0
Seedhg
0
0
0
0.5
Weeding - Ridging
0
-
0
-
0
-
0
350r:o600mm
Plowing
0
0
0
0.5
0.5
0
Tines land preparation
0
0
0
0
0.5
0
seeding
0
0
0.5
1
2
0.5
Weeding - Ridging
0
-
0
-
0
-
0.5
- -
1.5
-
0
.-
600t.0800mm
Plouling
0
0
0
0
1
0
Tines land preparation.
0
0
0
0
0.5
0
seeding
0
0
0.5
1
2
0.5
We&ing - RidRing;
0
-
0
-
0
-
1
1.5
-
0
_-
800 to 1200 mm
Plowing
0
0
0
0.5
1
0
Tine.s land preparation
0
0
0
0.5 0.5
0
Seeding
0
0
3.5
1
2.5
0.5
WeedinS - Ridging
0
0
- 0
- 0.5
0
L-
2
-
-
Clay soils
~ -
-
-
-7
- -
<=:35omm
-r
Plowing
0
0
0 0
0.5
1
1.5 3 2
1
0 0
Tïnes land preparation
0
0
0
0.5
1
2 1.5 0.5
0 0
seedirlg
0
0
0.5
1.5
3
5.5 3.5 1.5 0.5 0
Weeding. - Uidging
0
0
- - -
0.5
1.5
3 5
3.5
1.5 0.5 0
-,- - -_-
3.50 to 600 mm
Plowing
0
0
0
0.5
1
1.5
3
4 3.5 1.5
0 0
Tines land prepar.
0
0
0
0
0
0.5
1.5 2 3
1
0 0
seediig
0
0
0.5 0.5
1.5
3
6 7.5 6
3
0.5 0
Weeding - Ridging
0
0
~ - 0.5 0.5
1
2
4
7,5
2
0.5 0
-
,- -
600 to 800 mm
Plowing
0
0
0 0.5
1
2
3
5
4
1.5 0.5 0
‘Tines land preparation
0
0
0 0
0.5
0.5
1 2 1
0.5
0 0
Seed.ing
0
0
0.5 0.5
1.5
4
7.5 9 6.5
3
0.5 0
Weeding - Ridging
0
0
- - ,0 0 . 5
1
- - 3.5
A
5
6.5 5.5
3
1 0
-
800 to 1200 mm
Plowing
P
0
0
0.5
1
1.5
2.5
6 5
2.5 0.5 0
Tines land preparatidn
0
0
0
0.5
1
1.5
:
3 1.5
1
0 0
seeding
0
0
1
2
4
5
8
11 9
5
1 0
Weeding - Ridging
0
0
E Z 0.5 1
.3
3.5
5
8 6
4
0.5 0
z- L
-
I
a
- = -=
-
- - E
Source: Extract froc Le Moigne, 1981
events between 10 to 50 mm. In clay soils, he found that trafficability was greatly limited
by the level of soi1 wetness near saturation. For fïeld operatior$ the amount of ail rainfall
,..
event thresholds found for the sandy soils were shified upwards by 10, 20 and 30 mm.
The major limitation of his work resided in the fact that detailed studies on rainfall
incidence on the execution of field operation did not exist. He did not consider the
limitations induced by the soi1 desiccation and hardening processes near the lower limit of
soi1 water content (LL in ‘Si). The 30 to 50 years rainfall data collected in research stations
throughout the region have never been analyzed in terms of linking amount and
distribution of rainfalls to limitations in using the different farm implements present at tiie
household level.
3.10. Land use and cropping systems
3.10.1. Animal traction and cropping systems
The main objective of the implementation of animal traction projects around Third
World Countries was to increase agricultural productivity through yield increases and
labor savings. The challenge facing many agricultural policy makers was to produce more
than subsistence production levels. The transition from hand hoe to animal-drawn
‘?<l ,
I
implements has led to an increase of farming intensity. New cropping scenarios and
agricultural production opportunities have been created as more energy has been made
available to farmers through the utilization of draft ani,mals. At the same time, timeliness of
field activities has improved. Farmers view the labor savings objective in another way, a
means of reducing drudgery (Campbell, 1990; Ndiame, 1988). Ndiame (1988) reported
75
that the majority of farmers (79%) in the Southern part of Senegal were using animal
traction for plowing because it was casier and faster than hand lools.
Ammal traction has made significant progress and has produced changes in
cropping systems driven at least by one cash trop. The fast spread of animal traction in
different parts of the world was mainly due to the introduction and expansion of cash
crops (Eicher, 1982). In the Sub-Saharan region most of the cash crops bere introduced
and developed by the colonial system for expo:rt: groundnuts in Senegal, cotton in Mali,
Burkina. Faso, and Ivory Coast: etc... These cash crops represented the main source of
revenue to farmers involved in these cash trop oriented cropping systems.
Also, animal traction has brought about signifkant changes in cropping pattems.
Two non-desirable major effects have been the decrease in fallow through the extension of
croppedl areas and the acceleration of land degradation process through erosion. The
extension cf cultivated areas was made possible by the capacity to do more work in the
time available (Crossley and Kilgour, 1983). In. order to use animal-drawn implements on
new fields, not only must the land be cleared, but stumps and big roots also must be
removed which resulted in accelerating water and wind erosion (Pingall et al, 1987). More
detailed studies on the impact of animal traction utilization on the environment need to De
made.
The literature is rich in experiences sho-wing that working animal3 ‘were involved in
different field operations at the farm level: primary tillage, ridging, seeding,, weedimg,
harvesting, transport, etc. Farmers have found the utilization of animal traction for field
activities very attractive. Spencer (1988) pointed out that the degree of adoption by
7 6
farmers was limited by the availability of the different components making up the package
. . . \\
and by the production costs involved in their utilization.
_
TO carry out all field operations required to grow a given trop, farmers must make
strategic decisions in allocating available resources in combination with a variety of
implements present at the farm level to achieve a finalized production system (Equipe
Systemes, 1984). Sebillote (- 1983) addressed this question in terms of cropping practices
and introduced the terminology of “itineraire technique”. The “itineraire technique” was
defined for a given trop cycle as the combination of different cropping practices executed
in sequence, from tillage to harvest. Two “itineraires tecihniques” were considered differenr
if a change was made in the mode of execution of any field operation. An “itineraire
technique” involving the use of a ridge or moldboard plow for land preparation with the
remaining post-tillage operations manually executed would be different from another
“itineraire technique” using a moldboard plow and a seeder, or a moldboard plow with a
seeder and a cultivator (Equipe Systemes, 1984). The level of energy and amount of labor
used were found to be significantly different among “itineraires techniques”.
3.10.2. Field performance and trop yields
Different freld studies have tried to correlate the ievel of energy used for trop
production and yield. The power used per unit of cultivated area was a common indicator
used to demonstrate the benefit of animal traction over hand tools. Morris (1983) argued
that the yield increase induced by more energy per hectare was mainly due to improved
timeliness and better quality or precision of the field operation execution. Most
agricultural policies and decision-makers have always looked at mechanized cultivation as
77
an effective way to significantly increase the income of the household. TO this perspective
many African governments have encouraged the permanent occupation of land by farmers
but prevented at the same time any for-m of private ownership of the land. As pointed out
by Charreau (1974), the Government of Sene:gal, for instance, passed a legislation to
declare the Law of National Domain stating that any land in the country is nationally
owned. The main objective ofthis govemment legislation was to give incentives to
farmers involved in the national agricultural program of intensification tbrough animal
traction utilization. They urged farmers to rernove a11 the stumps in their fields in exchange
of 400 kg of rock phosphate (produced in Senegal) for each hectare cleared.
Agronomie research irrstitutions have carried out field investigations to establish
and quantiti the benefits of the technology. The introduction of the animal-drawn plow for
land preparation had given more opportunities to farmers to perform deeper tillage to a
depth of 10 to 20 cm compared to the super5cial traditional hand cultivation. The yield
increases brought about by this cropping technique were mainly due by the combination of
different factors. These factors are related to the improvement of the soi1 structure for
better water infiltration and root development: better weed control and c’rop residues
incorporation into the soil, better seed bed preparation for good conditions of plant
germination and emergence, and better plant use of water stored in the soil. A fairly large
number of technologies have been developed and, as shown by the percentage of yield
increase over traditionai practices, and most of them make significant contributions to
productivity. Jaeger (1985) presented a summary of station fïndings wid~ely used to
demonstrate these beneficial effects (Table 14).
7 8
Table 14: Percen.t yield increase with animal traction plowing.
Sources
Millet Sorghum Maize
Cotton Groundnuts Rice
Senegal
Charreau and Nicou
3 0
3 0
3 0
2 7
19
50
In Sergent (198 1)
Nicou
19
2 4
5 0
17
2 4
103
In Le Moigne (1979)
Charreau (1971)
2 4
25
35
25
23
Ramond and Tournu (1973
50
130
Tourte (1971)
- Plowing only
- Plowing + OM
25
4 4
43
41
2 0
4 9
4 2
85
3 9
9
Mali
SRCVO (1978)
--T--T--T-l---
In Sergent (1981)
-Donkey
- Oxen
iource: Jaeger, 1985
Iu_I/
---
q-,*m----
-
,,*a-
‘The benefits gained from the utilization of the plow have been demonstrated in
different parts of Africa. Especially in French-speaking coun&es of West Afiica, pl& use
appears to be highly variable depending on the type of soil, the year and the type of trop.
The depressive effects of plowing were mainly explained by poor plowing,.
.Field experiments conducted in farrners’ conditions in the Casamance region for 8
to 10 years showed a rnodest trend of benefits (Equipe Systemes, 1984) when plowing
was cornplemented with different levels of mechanized post-tillage operations like seeding,
weeding, and harvesting. For groundnut production for example, the combination of
mechanized plowing, seeding and weeding (“Itineraire 4”) gave 59% more yield than
plowing (“Itineraire 1 or 2”) alone while mechanized plowing and seeding (“Itineraire 3”)
gave 13% more. These yield improvements were essentially due to better timeliness of the
field operations, the efflcient eradication of weeds and the physical stirring of soi1 for
better aeration and water distribution in the root zone. In this process, labor input per
hectare was also reduced by, as much as, 20% to 25% (Sonko and Fall, 1,992).
3.11. Siimmary
The exhaustive review of literature pre:sented on the determinants of performance
of animal traction’s utilization for trop production has shown a variety of research
activities carried out in different parts of the world Most of the well-published research
activities and results found in the literature were from French (CEEMAT, IFUT), English
(ARC, AFEIC, CTVM] and Gerrnan (GTZ) institutions, American univettsities
(Department of Animal Science at Corne11 University) and NARS in Third World
countries. Investigations were conducted in dilTerent directions without really following
8 0
any predetermined standard or procedure. ,The area of research appears to be new and
‘.
open to more investigations. Lawrence and Pearson (1993) have addressed the issue of
standardization in terms of (a) length of pre-experimental periods to allow draft animais to
reach a ‘steady state’ (training, fitness, diet and climate) before measurements cari be
performed, (b) number of animals needed in an experiment for level of statistical
significance ofthe results, (c) work and performance tests to evaluate continuous and
instantaneous maximum efforts, (d) instrumentation and units to be used, and (e)
physiological tests to understand the difference in draft animal performance when at work.
The lack of standards in the measurement procedures introduced a factor of variability to
explain the discrepancies found in published research results. The nature of the study
involving living species made the task difficult for a11 researchers to agree upon the defauit
values of different indicators used to predict working animal performance.
The main constraints limiting animal traction utilization ar ,i impleme:ltation around
the world has been presented under different aspects, from climatic conditions relrired to
the cultural behavior of target groups. The difficulty in laying out experimental designs
involving weather as a variable made it quasi-impossible to rigorously compare draft
animals working in different environmental settings. Along with weather, local sanitary
conditions were described as a major issue in promoting animal traction. The prevalence
of trypanomiasis in many wet and warm areas of the tropical region for example, has been
considered by many authors to be a significant factor which slowed down the ditision of
the technology. The recommendation gained from the hterature was to help farmers
develop well-organized and feasible work plans to avoid over working their animals. More
important was fitting the job to the draft animals’ capacity in relation to the draft
-
81
requirements and working conditions, the type of impiement, the efficiency of the
hamessing system, the type of soi1 (SWC) and its resistance to’traction, and the
probabilitiems of working days.
The new methods of energy evaluation presented have been maimy impaired. by the
laboratory determination of important parameters. This fact was described to represent a
major disadvantage in the evaluation of some components of the energy balance in
comparison to field measurements. For non-fielld measurable amounts of e:nergy required
from working animais, these parameters cari be used to calculate the default values until
the development of new, more realistic methods.
The quantity and quality of feed given to animal must meet the draft animal
nutritional needs. In general, the right amount and composition of feed stuff was ofien
reported not available to farmers. Most of the feed was mainly obtained through grazing
on natural pasture. As poin.ted out by Ffoulkes and Bamualim (1989), when browsing,
animals naturally tended to Select feeds that were likely to enhance their body condition
Nevertheless. additional feed must be given on working days to avoid weight loss during
peak work demand.
The use of draft animals for field work (plowing, seeding, weeding and harvesting)
has demonstrated opportunities and potentials to generate substantial revenue to farmers
through better productivity (yield increase and labor saving). This is made possible
through better seed bed preparation, weed control, timeliness of field operation,
improvement of soi1 moisture regime, reduction of labor.
9
,”
‘
Chapter 4
. . :
MATERLALS AND METHODS
;
4.1. Research sites
This study was conducted in distinct phases including both on-station an 1 on-farrn
i
sites and focussed on the use of a pair of oxen as draft animals. Par-t of the human
.IB
I
resources (animal scientist and soi1 scientist) needed to conduct the field work were
i
provided by the Senegalese Agricultural Research Institute (MA) through its local
i
Agricultural Research Center (CRA) of Djibelor located 7 kms South of the regional
11.
capital cit! Ziguinchor.
4.1.1. On-station
“1. I
The on-station facilities were mainly used for the evaluation of the draft
requirements of different field operations (plowing, ridging, seeding and weeding).
The Djibelor Agricultural Research Station (ISRAKRA research ‘tenter) is the
major research station in the southern part of Senegal. Until 1979, research activities were
mainly oriented towards the development of rice production techniques in relation to the
amount of rainfall, the type of relief and the importance of hydro-agriculturai production
potential of the region. Since the events of the drought situation (197%82) research
CI
-. . .
83
activities were adjusted to reflect the new production strategies developed by fat-mers to
. .
coumeract the effects of the shortage of rain. Multidisciplinary research ‘leams were
constituted at the station level to approach the diversity of the productioin systems found
in the region from the perspectives of integrated agricultural and livestock activities, an.d
mixed cropping systems. Following the new d,ynamic, research operations on upland crops
(cereals and cash crops) and animal traction were given more emphasis ait both on-station
and on-farm levels.
Djibelor research station has several research facilities: soi1 and w’ater laboratov,
equipped farm shop, experimentation plots, pest management and plant genetics
laboratories. The experimental units of the research station are located aiong the
toposequence. The lowland plots are mainly used for the development of rice cropping
techniques and the upland plots for the improvement of the rain-fed cropping systems
(maize, millet, groundnuts., and sorghum). For this study, the on-station’s research
activit!:es were conducted on the upland areas in the way animal traction is mainly used by
farmers’ circumstances (Posner, Kamuanga and Sali, 1988). The soils are “red soil” of the
ferrallitic type, representative of most upland soils found in the region.
41.2. On-farm
Two research activïties were programmed on-farm: survey on animal-drawn.
implements and test for maximum pulling perfclrmance along with a follow up of the draft
animaXs’ act.ivities during the rainy season.
The on-farm taaget group was represented by farmers located in the two agro-
ecosystem zones 4 and, 5 in the not-them part of the Ba.sse Casamance region where animal
traction is well established. A sample of fat-mers using animal traction for cultivation from
four representative villages (two from each zone) was drawn for’the field research
activities. An important database about the target group is readily available which consists
of information collected by the Farming System Research team (Equipe Systemes) of the
Agricultural Research Center (CRA) of Djibelor during several years since 1982. Data are
mainly related to field operations, production techniques, cropping systems, and calendar
of agricultural production activities (Posner et al, 1985; Ndiame, 1990; Equipe Systemes,
1983). This database Will be used as a reliable secondary data source.
The villages of Boufandor and Bougoutoub in agro-ecosystem zone 4, and
Kagnarou and Sue1 in agro-ecosystem zone 5 were chosen for this on-farm study
consisting of equipment survey and maximum performance trials
4.2. On-farm equipment survey
As part of the methodology, an animal traction survey was conducted at the farm
level (unit of observation). The farm is defined as an autonomous family composed of
I..
members who produce and consume together (Jolly, Kamuanga and Sali, 11988). The fat-m
cari be made of different dependent households. The tenter of decision making process in
terms of management and resources allocation is located at the main farm level.
The focus of this study was put upon different aspects related to the use ofanimal-
drawn equipment at the fat-m level. This type of survey could easily fit into a Rapid Rural
Appraisal @WI) approach (PRAAP, 1992). It was important to complete this survey in
order to identify the range of implements used by farmers before any draft measurements
and other energy estimation was performed.
*.
i
-
--
-
_
-------
85
The enumerators to conduct this survey had to undergo technical training to be
able to identify the implements currently used by fat-mers.
“* .
4.2.1 Objectives
The objectives of the survey were:
1. Identify the types of equipment actually involved in field operations for
which more detailed technical information was needed towards energy
requirement estimates.
2. Evaluate fat-mers’ learning process for animal traction use.
The nul1 hypothesis (Ho) was that there was no difference in fat-mers’ level of
equipm’ent A farmer was considered to be equipped as long as he owed at least one farm
implement (moldboard plow, ridger, cultivator, seeder, cart, etc.). The level of equipment
encompassed two aspects of the mechanization process. The first aspect d.ealt with the
total number of implements possessed by farmers and used to carry out field operations.
The second aspect specified the possible combinations of implements used to perform field
operations from tillage, seeding, weeding to harvest. The higher was the number of
different types of implements, the higher the level of mechanization. Fat-mers with higher
level of equipment were expected to perform better in the production system in terrns of
quality and timeliness of field o:peration.
8 6
m,
!
4.2.2. Data collection
. .
‘_
The questionnaire was designed to help collect thrce categories of information:
type, management, and utilization of animal-drawn implements present at the farm level
(Questionnaire in Appendix Al).
4.2.2.1. Type of farm equipment and draft animals
The main characteristics of animal-drawn implements retained in this survey were
the type and shape of the working components. The configuration and aspect of these
components were generally described to be responsible for the draft required to pulling the
implement in given working conditions (type of soil, moisture regime).
The most popular working components expected to be found at the farm level
..,
were the moldboard plow, ridger with adjustable wings, fùrrow opener, cultivator tines
i
equipped with various size of sweeps, duck foot shovels, or chisel points, and groundnut
I
lifiers.
Multipurpose farm ecluipment is also available in the market for fiumers. These
were designed and used for different field operations f?om tillage to harvest (groundnut
lifting). This was made possible by simply changing the type of working component
attached to the tool-frame. These multipurpose types were also sized to fit the job to the
draft animals through the selection of suitable working components. In general, they
require more ski11 from farmers- compared to an implement with a simple frame design.
A second important c:haracteristic was the type of traction used to pull the
implement (pair ofoxen, horse or donkey). The stability coefficient of the technology
(STAEKOEF in percent) under farmers’ conditions was determined by evaluating the
87
I
difficulties felt by farmers who have the minimum number of draft animais for field
operations and in particular when these animal are oxen. The technology is consideredto
be unstable for farmers with only one pair of oxen, as the loss or unavailability of one
animal Will compromise the production objectives. The stability coefficient cari be
lexpressed as one minus the probability of farmers of having only one pair of oxen
(PBPAIR. in decimal):
STABCOEF,, = [l - @‘BPAIR)]
The readiness of draft animals at work time was considered as a determinant
indicator for better timeliness of field operation.,.~9 Draft animal reliability ANIREL in
percent was introduced at this point to report thle probability that the draft ianimals were
not available to petiot-m field operations on time. It was computed as one minus the
probability of draft animal unavailability, which is the probability of farmers of having one
ox only at the beginning of the rainy season (ANNA.VAIL in percent) when both
probabilities were expressed in decimal form:
Ah’IREL = [I - (ANNA UIL)]
(20)
In general, farm implements are designed an.d built with ergonomie considerations
to fit height and draft requirements for the working animals. Fortunately for the farmers,
most implelnents are interchangeable among draft-animals species through simple
8 8
adjustments of the harnessing system. Switching from a pair of oxen to a single draft
‘”
animal (donkey for example) requires the use of an evener.
’
The types ofhamesses used by farmers to hitch the implements for their fieldwork
were charactenzed in terms of attachment and size to help analyze the efficiency of the
energy transmission mode.
The third characteristic was the type of the design and the pur-pose of the
implement. Animal-drawn implements are generally built to fùlfill specific Ipurposes like
1 .
:
land preparation, secondary Mage, seeding, weeding, cultivation, groundnut lifting,
transport, etc. The utilization of one type of implement usually predeterrmnes the
execution mode of the following field operations. For example, if the land preparation is
done with a ridge plow, it is likely that a11 remaining field operations (seed.ing, weeding,
‘etc.) Will be manually executed. It is important at the farm level to identify the function of
each piece of equipment.
. ”
/
/
4.2.2.2. Management
The status and utilization rate of each animal-drawn implement found at the farm
level was diagnosed to help understand its working conditions. The farm structural
characteristics in terms of family composition (number of farm workers NFMWIERS,
household dependency) was expected to dictate the type of ownership and explain how
priorities were set up in the process of allocating available farm implements to different
fields and crops.
The mode and date of acquisition in relation to the benefits generated were
e:xpected to give indications of the number of years of farmers’experience.
,I
8 9
4.2.2.3, Utilization and maintenance
‘The level of use of farm implements during the grotiihg’season was genera&
reflected by the wear and the number of maintenance performed every yea:.r. From this
infcmnation it was expected to be able to determine the frequency of equipment
breakdown or the probability of equipment failure or the reliability (IMPLREL in percent)
for each type of implement. The farm equipment reliability is defined by ASAE (1984) as
the “statistical probability that a machine Will fimction satisfactory under specified
conditions at any given time. It is computed as one minus the probability of a failure”
(PBRKl~OWN in decimal):
IMPLREL = [i - (PBRKLIO WN ,)mp,ement /
Farmers’ skills in adjusting and using the implements, and the physiical aspect and
nature of the fields (presence of roots and stumps) are described in the literature. These
working conditions shorten the life expectancy of farm equipment in fatiers’ condit.ions
(FAKMLIFE).
13lacksmiths established in the villages were also contacted to evaluate their
capacities for maintaining implements used by farmers in good working conditions.
It was important at this point in the survey to have the farmers’ opinion about the
efficiency of the technical support provided by the extension service and fzlrm equipment
dealers in terrns of traihing, spare parts availability, and access to any type of credit.
9 0
42.3. Samplling techniques
The first step in the process was to determine the sampl’ê sïze n in order to have a
good estimation of the population mean CI. This must be done in relation to the amount of
errer d that cari be tolerated and the level of probability that this error was expected to
hold (Bhattacharyya and Johnson, 1977). Farmers equipped with animal-drawn equipment
represented the target population. The sample was drawn at random in the four
representative villages of agro-ecosystem zone 4 and 5. The sample size n for the survey
was determined by using the concept of standard error (S.E) and length of confidence
intervals (C.1). At this level of planning, the sample size must be calculated in order to
achieve more precision in the est,imation of the different statistics. In this procedure, it was
kept in mind that in general, a survey using a large sample size was more costly to
implement and much more time consuming in terms of data processing. Without loosing
toc) much precision in the results, the sample was sized large enough for the Central Limit
Theorem to apply. The theorem states that, if n is large enough the distribution of the
sample mean is approximately normal. In the case of non-normal distribution:5 because ,f
small n, the Chebyshev’s inequality is still applicable. It stipulates that the probability that
the difference between the population mean p and an observed random variable X mean is
less than a constant d, is greater or equal than 1 minus the variante of the observed
random variable mean divided by the square of the constant d.
_
_
..-
c.
,.,cIUI*-“-II,-,.-,------
.___-
---
‘--WI-*’
9 %
If the levei of confidence is lOO( 1 -a)%, then the upper bound of the sample size is:
,l zz: (-$
(23)
In bhis present study. the standard deviation CT of the population was determined
fiom previous studies condu.ctt:d in the same area (FALL, 1985; Sonko, 1986). The
amount of error d that could be tolerated was set to 10% or 0.1 as suggested by Le
Moigne (1984). The probability that the error of estimation would not exceed 0.1 was put
to 90?6 level for a= 0 10. Since the population variante 0 was equai to 0 2, one
calculated upper bound of the sample size was:
2
,, ZZZ
o.2
= 40
o.r*(O.I)I
TO implement the survey, n (total numiber of observations per variable) was
equally divided into th’e four villages: ni= 10 for i= 1 1.0 4 (i= village numt:ler)
4.2.4. Data analysis
The data collected in this survey were analyzed with the package MIN1T.A.B
Statistical Software of Release 8 PC version. MIXIT.AB is a registered tra~demarkT~l of
Minitab Inc. Basic statistics were performed on the data collected regarding each type ot’
equipment: frequency and distrilbution, mean, standard deviation and staridard error, 90°i
t-confidence Interval for the mean (Bhattacharyya and Johnson, 1977).
The factor leveis b (number of implement types for EQTYPE) and c (number of
different yoke sizes for YOKE) were determined from the où’tput of the smvey. The factor
level a (number of soi1 water content levels for SWC range) was :Function of the rainfall
profile and the soi1 water status. The soi1 water content was determined by the gravimetric
rnethod (GSWC in % g/g) and was converted into the volumetric soi1 water content if
necessary (VSWC in % v/v). Different studies on monitor-mg of field opera.r:ions have
shown that farmers never carried out field activities at optimal soi1 water content. Land
preparation was generally performed at the beginning of the rainy season just afier the first
useful rain. Three levels of factor a were arbitrary chosen within the range jof a working-
day’s soil-water content during the period allocated to land preparation and weed control.
These three leveis of soi1 water content corresponded to three days of fieldwork: one day
was chosen in the month of June and two days in the month of July.
The analysis of variante was performed to determine the significance of each
fiictor and combination of factors on the pull force PF (Table 15).
The overall F-test was performed to detect treatment differences alcng with lOO( l-
a)?/0 simultaneous Confidence lntenal (CI) to compare specifrc pairs of treatments.
1
I’he data analysis was oriented towards approaching the variability among farmers
in terms of the different types of possible implement combi&ons to conduct field
operations from land preparation to harvest
4.3. Estimates of energy requirements
43.1. Objectives
The objective of this research activity was to estimate the average energy required
in carrying out field operations.at the farm level. It was performed through the
measurement of the amount of draft needed to pull an implement in a given work ~g
condition. The draft evaluation was carried out at different soi1 moisture regimes.
The nul1 hypothesis (Ho) was that the amount of draft required to pull farrx
implements in given working conditions wa$; mainly affected by the soi1 moisture regirne
and did r1.x depend +3n the type of energy transmission system used to hitch the impiement
(widt’ of head yoke).
43.2. Experimental desigu
For this purFose, a 2 x 3 x 3 factorial. in randomized blocks desilgn was la::d out to
help determine the cDmbined effects of three factors: .A= soi1 water content SR’C range,
B= type of implement EQTYPE and c‘= size of YOKE on the amount c’f pull force PF in
daN measured. The ;hree factors had respectively a, b and c levels and m obsenations
(number 3f repetitions) tak.en for each combination of the three factor levels, for a total of
abcm ob jervations.
9 4
” I
I
Table 15: Analysis of Variante - .
Source
Sum of
Mian
ii
of Variation
Df
Squares
Square
F
/
I A
(a-1)
SSA
SSAIDf(A)
MWMSE
I
J3
(b-1)
SSB
SSBIDf(B)
MSB/MSE
C
(c-1)
s s c
ssciDqc)
MSUMSE
AB
(a-l)(b- 1)
SSAB
ssAB/DqAB)
MSAB/MSE
AC
(a- l)(c-1)
SSAC
SSAC/-Df(BC)
MSAUMSE
BC
(b- l)(c- 1)
SSBC
SSBC/Df(BC)
MSAUMSE
ABC
(a- l)(b- l)(c- 1)
SSABC
SSABCDf(ABC) MSABUMSE
ERROR
abc(m- 1)
SSE
S SE/Df(ERROR)
Cor-r. Total
abcm- 1
//
Mean
1
/I
II Total
abcm
TSS
II
.1
The general linear statistical mode1 associated with this factorial experimental
design was stated in the form of? (Stapleron, 1995):
I,-
(27)
and Eijti are random errors independently distributed: &ij~~N(O,~).
F o r i=l,...,a
j= 1, . . . . b
k= 1, . . . . c
I= 1, . . . . m.
9 5
It was also assumed that the observations Yijkl (PF in daN) were normally
.
distributed as Yijp N(pijk, 02) and independent.
YijM represented the pulling force PF in daN or kgf and p the expected average
draft in daN or kgf. The parameters a, p, and y were unknown parameters called the
interact.ion effects of A, B and Ç respectively The first and second order interactions
among factors were also included in the formulation of the regression mode1 as well.
The general mean p was calculated.
(29)
(30)
There were a total of 18 different combinations applied each to a 200 mc to .jOO
m2-experimental unit.
Besides this experiment, additional draft measurements were performed to
determine the power requirernents and energy expenditure for other post-tillage field
operations: seeding, weeding.
9 6
43.3. Draft measurements
The measurements performed on each experimental unit were:
Weight of each animal LW in kgf,
Arnount of pulling force PF in daN or kgiT,
. Angle of pull y in degrees (“d),
. Oxen working time or field time FDTIME in s,
. Oxen working speed ANSPEED in ms,
Distance traveled DSTTRAV in m,
Implement working depth IMPLWDP in cm,
Implement working width IMPLWWD in cm,
Soi1 gravimetric water content GSWC in % g/g.
Soi1 penetrometer Cone energy EDEP in J/s.
The pulling force (PF) was measured at the hitch point on the implement and
represented the a:mount of force developed by the pair of oxen to pull the implement on a
continuous working rhythm. For each experimental plot, the range and the average pulling
force were observed. Not a11 the pulling force developed by the draft animais was used to
do usefùi work. Only the component in the direction oftravei was converted into tractice
force to pull the implement in the soil. The amount of tractive force, also called the draft
required (DR) by the implement, was expressed as a fùnction of the angie of pull y in
‘degree:
9’7
DR -= Pr‘ *(Cos y)
.
(31)
This required draft DR was expected to be highly correlated to the effective
working width (IMPLWWD in cm) and depth (IMpL.WDP in cm). The e:lTective working
width (more or less than the measured implement working component width
WWKCrCOMP in cm) was the width over which the implement actually worked.
4.3.4. Power and energy eupenditure
The energy fùrnished by the working animais to perform the field operation was
represented by the amount of energy needed to pull the implement and the distance walked
from he beginning of the operation to the end. The energy delivered to pejform the work
ANERGY in MJ was estimated by multiplying the pulling force PF in daN by the distance
traveled DSTRAV in rn:
ANERGY = PF * (DISTRA V)
(32)
The energy required to pull the implement ANERGYRIQ in MJ was calculated as
the draft: required DR in daN multiplied by the {distance traveled DISTRAV in m:
(33)
9 8
Another approach to the energy evaluation was derived from the power concept
defined as the rate of doing work. The animal power ANP&J%R was calculated by
multiplying the pulling force PF or the draft required (DR) and the animal field speed
,a.
l
ANSPEED in ms. The field speed was defined as the average rate of implement travel in
the field during an uninterrupted period of functional activity (AME, 1984). ‘The rate of
travel was evaluated as the distance traveled (DISTRAV) divided by the fieid time (FDTM
in sec) that the working component of the implement stayed in a non-stop working
position.
ANERGY = k * (ANPOJER) * (ANWKGHR)
PJ)
The energ? spent by the draft animals walking across the field while per%rming the
work must be added to have a more complete formulation of the total energy (ZIquation
5).
The energy required and delivered to perform a given task was used to determine
l^
i
nutrient requirement by converting energy expenditure into amount of availabie feed to be
supplied to the working animals.
IV.
/
4.4. Maximum performance
4.4.1. Objectives
The average available power was expected to increase as the body weight of the
,s.
;
animals increased during the rainy season resulting fiom the availability of more grazing
i
,.,
areas. The objective of this experiment was to measure the maximum effort developed by
I
i
7
*
‘99
a pair of oxen and to determine the efficiency of energy utilization by comparing rhe
energy used and the total energy available.
‘. 3
4.4.2. Procedure
The available animal power was measured at the beginning of the rainy season
during the period allocated to land preparation. The power measurement was used to
determine the maximum draft that a pair of oxen could develop during th;e growing
season. As mentioned by Goe (1983) farmers are facing a problem of animal feed at the
beginning of the cropping season when work labor demand is at its peak.
The research activities were divided into two parts:
- Measurement of the maximum pull force at the beginning of the rainy
season;
- Follow-up with the selected draft animais to monitor the type of
management used by fat-mers during the growing season (see Appendix
A2 for follow-up guidelines).
The measurements were performed in each of the four previousiy mentioned
villages located in the agro-ecosystems zone 4 and 5 and on-station. Two pairs of oxen
were randomly selected per village to make a total of 9 pairs of oxen ( 18 draft oxen) to be
used for the test and the follow up during the cvhole rainy season (June to Oct.ober).
100
- Pulling trials forcontinuous maximum effort
The experiment consisted of pulling triais on soi1 re&iy for land preparation by a
‘_ \\
pair of oxen with a progression of known loads placed on a sledge. The variables observed
were:
Age of each animal AGE in years (yrs),
Weight of each animal LW in kgf,
Pulling force PF in daN or kgf,
Angle of pull y in ‘d,
Distance traveled DISTRAV in m,
Working time PULLTIME in s,
Working speed ANSPEED in m/s.
For thc data analysis, the following graphs were analyzed to determine the
optimum power (AXPOWER,,,t) delivered by the draft. A general 1inea.r modeling
technique was used:
.
Power output ANPOWER:
ANPO WER = f(PF)
(33)
. Specifïc power output ANPO~~IXSPEC:
(36)
The same analysis was conducted with the speed ANSPEED in m/s:
ANSPEED = f(pF
(37)
(38)
Two identical sledges were locally fabricated for the pull trials.
- Follow up of draft animais
Mer the pulling trials., a follow up was agreed upon with the famers to monitor
a11 the activities performed by the 18 draft animals during the rainy season. Information
collected were (,Appendix: AZ):
. Liveweight (LW in kg) at the end of every month.
Type of activities canied out during the given period.
. Feeding system (type of feed and amount given)
Handling and tare.
The foilow-up was performed on a dai1.y ba.sis by two of the survey personnel who
were in continuous contact with concerned farmers.
102
4.5. Iusf rumentation
‘The instrumentation used in this study was developed by CEEMA’T and was
designed for automatic field data collection (CIRAD-SAR, 1993). The anj,mal-drawn
system was composed of two basic elements: a data logger 2 1 X (Campbell Scientific, Inc j
and a series of sensors to be mounted on the implernent ancl connected to the data logger
The data measured by the sensors were made on real time basis through the interna1 clock
of the data logger used for their storage. The main sensors used in Fhis study were
500 daN-dynamometer sensor for measuring the amount of pulling
force (PF),
Dickey John (registered trademark) radar to measure the working
speed (ANSPEED) in m/sec,
Tntcrnal clock of data storage unit to measure the working time
(FDTM in s)
For later data processing, the data logger was downloaded to a computer with the
support software EDLOG via a SC32A RS-232 interface using the software TERM or
SC532 RS-232 interface from an auxiliary storage unit using the ShlCOM The RS-232
interface was to convert the CMOS logic levels of the data logger to the RS-232 levels of
the computer. The SC32A and SC532 provided an effective isola.tion between the
computer’s and data logger’s electrical system, protecting against normal static discharge,
and noise (Campbell Scientifïc, 1992).
103
An additional mechanical dynamometer was also used to supplement the electronic
system in the case it failed to operate properly. The dynaniometer was a spring-loaded
type made by PIAB’M and had a maximum force of 500 daN.
The additional field equipment used to carry out the field measurements were:
I-land held stopwatch,
Measuring tape,
Rain gages,
Soi1 penetrometer,
AMS Soi1 sampiers (Ben Meadows Company).
Electronic weighing scales (Bario Electronic ScaleYM hIode 2 100).
The data collected during the season were used to develop a mode1 to evaluate
draf? animais liveweight.
4.6. Expert System building
46.1. Objectives
The expert system mode1 was built to help achieve an efflcient utilization of animal
traction for trop production through choice of animais and implements to perform specifk
field operations. It was designed for researchers, extension agents, farrners and policy
makers ‘to simulate different scenarios of animal traction utilization at the farm ievel.
13y introducing a confidence threshold level set by the users for “fbzzy reasoning”,
104
the expert system mode1 based on Rnowledge Base Management Systern (KBIvfS)
program was expected to yield more realistic results in farmer’s’ conditions.
._
4.6.2. Knowledge Base Design and Deveiopment
4.6.2.1. Presentation of the expert system
The expert system was developed with the LEVEL 5 Object program (for
Microsoft WindowsTM Standard Edition Release 3.6 Copyright@ 1995 Information
Builders, Inc. NY). The selection of a suitable shell was considered to be the first step in
developing an expert system.
LEVELS OBJECT is a software development tool kit. The program shell is able to
activate external conventional programs called within the expert knowleaclge base and was
able to perform mathematical operations, incorporate uncertaintv and partial id: rmation.
and explain its decision logic. The program was also user-friendly and composed of a
built-in screen editor, a compiler for the inference engine and an interfacing mechanism
with program written in TURBO PASCAL ar:d Microsoft FORTRAN.
The program development was made by using a combination of clecision-makin-
rules, methods and demons with a high-level knowledge engineering laquage called PRL
(Production Rule Language) based on inferences in linguistic form and written in IF-
THEN rules.
Rules are used to express logic, cause-and-effect relationships, between the facts
and conclusions, while a demon proceeds by testing for patterns in data based on
conditions and then executes an action based on the conditions. The method is similar to a
,..k,m.
I
:105
macro with a number of commands ‘associated with an attribute. It defines the attribute’s
behavior and executes a series of actions when the attributë.yAe changes.
Rehak (1983), cited by Raman et a1 (1992), reported that there were three types of
knowledge that could be integrated into an expert system: Heuristic (from experience),
conventional (regarding facts) and inferential knowledge (study of results). They found
that inf’erential knowledge was encountered ün most of the engineering applications.
The representation of knowledge requires an efflcient organization of the
objectives or goals of the expert system and the supporting structure of information.
The inference engine of the shell evaluates the knowledge captured in the rules,
demons and methods by exploring the problem space represented in a form of an IF-
AND/OR-THEN structure.
The program was developed using a backward chaining control structure to
connect the individual rules and forward for the methods. In the backward reasoning, the
reason.ing is performed starting from what the program wanted to prove towards the facts
that are needed rather than beginning from the facts. This system of reasoning described
by Jackson (1990) was more focused than forward chaining, because only potentially
relevant facts are taken into consideration. An example of such reasonimg:
Rule for draft animal selection:
RULE for Draft animal selection
IF the animal is trained
AND ihe animal liveweight > (Mature weight2)
THEN the animal is a draft animal
RULE for trained animal
IF the animal is 2 years old
AND the animal has learned basic instructions
THEN the animal is trained CONFIDENCE SO
106
CONFIDENCE is used to quantify the degree to which one cari be confident in the
‘_ .
accuracy of the conclusion or to give an indication ofthe likelihood of occurrence. The
.. .
lowest confidence level for LEVELS to reach a conclusion or goal is set up through the
THRESHOLD statement.
With the command CHAIN, multiple knowledge bases cari be linked if the domain
of investigation cari be partitioned into separate smaller knowledge bases. This procedure
allows the building of programs of a virtually unlimited size program. It also improves the
readability of the program by enabling the primary knowledge base to activate the
necessary sub-knowledge base depending upon the context.
4.6.2.2. Expert system organization
The mo’del was built around three major components.
- Field capacities evaluation
- Animal energy balance ay,d feed ration
- Annual farm budget and Optimization
The values of the input variables required to t-un the mode1 were either developed
previousiy in the field experiments or taken from a reiiable secondary database
The data input system is composed of disk files and operator entered data through
the user interface system. The mode1 was designed to use different databases related to.
Farm: farm size, number of farm workers,
Farm implements: TYPE, WORKING CQ$@ONENT, PF,
Draft animals: SEX, LW, AGE,
Available feeds: TYPE, NCTRIENTS, FU,
Crops grown: TYPE, VARIETY, YTELD.
Validated default values are used to supplement for missing parameters. The data
entered by the user was made possible by the capability of the expert sys,tem to use prompt
boxes for a11 the information required fiom tbe user. The user could also query the
progra.m to know why a given information was needed and get the explanation,
estabii.shing very interactive communication.
4.6.2.3. Data processing
- Draft and power estimates, and field capacities
The statistical models developed in the second part of the methodology were used
to evaluate the amount of traction delivered, the traction required, the energy delivered.
and the energq required.
The pulling force PF in daN and draft DR in daN required were evaluated for
evety working day in relation with the type of implement used.
In relation to the pull force PF or draft required DR the field capacities EFC in
haiday werr evaluated along wiith the number of working days NWDAYS.
The soi1 water balance was used to estimate the number of working days
NL+DA‘YS affected with a level of probability (5 years out of 10 and 9 years out of 10).
108
The number of working day’s per month was evaluated by subtracting the number
of non-working days from the total number of days of the, &responding month. .The
_ .
execution deadlines of specific field operations were taken into account. For each field
operation, two soi1 moisture limits were considered: drained Upper limit. (DUL) and lower
limit (LL). The difference between DUL and LL, was used to determine: the soi1 field
capacity and trafficability in relation to by the amount of rain. The lower limit was main11
indicated by the level of hardness of the soi1 during the drying process. .4t this level of soi1
water content, the implement worlcing components were barely able to lpenetrate the soi1
surface.
The probability of a working day (PWD in percent) was calculated as the ratio
between the number of working days (NWDAYS) over the total number of daq’ (NDAYS)
of the chosen period (month= 30 or 3 1 days, week= 7 days and SO on). In this study the
probability of a working day was evaluated on a monthly basis:
(39)
A coefficient called effective probability of a working day (EPWD in percent) was
introduced and defined as the PWD weighted by the draft animal reliability (ANIREL in
decimal) and the farm implement reliability (IMPLREL in decimal):
(40)
109
- Energy and nutrient requirements
The nutrients required to carry out the field activities‘were evaluated fi-omthe
energy balance equation or factorial method developed by Lawrence and Stibbards (1990)
(Equation 5). It is quasi-impossible to determine directly in the field the total amount of
energy used by a working animal (Falvey, 1986). The main components used for the
enerjgy balance were the net energy for the work performed by the draft animals and the
net energy used for walking throughout the field.
The calculated energy represented the extra energy on top of the maintenance
energy (h4AE), supplied by the draft animal’s body to perform the work required. In the
process, the digestive system of the working animais converted the feed into digestible
energy (DE). The part of the digestible energy used to perform the work is cailed
metabolizable energy (ME), and is measurable in forage unit (FU). As defined by many
researchers, the feed or forage unit represents the net energy of 1 kg of barley with no
losses in the animal feces, urine, gas, etc. (CEEMAT, 1974; Goe and McDowell. 1980);
Watson, 1981). The number of forage unit per dav for maintenance was calculated by
using Equation 11.
The program was designed to compose the ration from the variety of feed available
to farmers, which could satisfy the daily total amount of forage unit. The amount of feed
requir’ed FRFEEDDREQ in kg of fresh feed per dav cari be calculated by converting the total
FU into fresh feed (forage, grain, bran, straws) on a basi.s of 1000’0 dry matter content
(DM) and intrinsic value (FU) of each type of feed.
110
100
TotalFIl
.
FRFEEDm, = (----
.w
DM) * ’ FiIJFzED )
‘. ,
The upper Emit of the amount of feed an animal could take per day on a dry matter
DM basis was set equal to the Appetite Limit (AL in kg of feed/day) for the corresponding
animal of liveweight LW (Equation 10).
- Production costs and Optimization
Cropping system techniques (“Itineraire t,echnique”) were evaluated in relation
with the level of energy used. The cost associated with animal and implement utilization.
and man-labor used was included in the production cost evaluation with updated prices.
For the benefit calculations, the average yield of the 10 past years’ experience was used for
each cropping technique and level of energy used for four “Itineraire techniques” using
different implement’s combinations.
It was assumed that the cropping system was driven by a cash trop (groundnuts)
and marketed rhrough the officia1 network to help farmers caver a11 the production costs
The optimization was formulated in terms of:
Objective function: What is the trop mixture and hectarage that
would maximize farmers’ profit.
Constraint fùnctions: land availability, labor availability, available
animal energy and required cereals consumptio,n based on FAO
standards,
111
4.6.2.4. Program output
The output of the: expert system was (oriented towards different aspects of farm
equipment selection and management. and decision making process.
- Field Capacities
- Labor balance per ha (available vs required)
- Animal energy balance
- Feed ration
- Complete annuat farm budget
- Optimization solution.
4.6.3. Program validation
The expert system was built around different statistical models and simple
equations to evaluate or calculate variables involved in the decision-making process at the
farm level in terms of resources allocation. The quality of the output is mainly relat’ed to
the validity of the different equations used.
4.5. Summary
The methodology developed in this stuldy was mainly oriented towards the building
of an ex.pert system that would really evaluate the impact of animal traction at the farm
level. The evaluation was centered on trop production activities. The methodology \\Vas
divided into three phases:
112
- ‘survey: the objective of the farm survey through a simple questionnaire
was to identify ail types of farm implements really involved m’crop production at the farm
. v
level. ‘This investigation technique offered an oppot-tunity to Select the implements along
with the working animais on which more detailed technical information vdere needed in
terms of power and energy requirements.
- Energy estimations: the energy estimations were performed on use field
implements and working animals only. The pulling force delivered by the draft: animais and
measured with a dynamometer represented the tore of these estimations. Al1 the
measurements were made in real working conditions in order to include as many variability
factors as possible beside three major ones: type of implement and yoking syst.em, and soi1
moisture regime. It was important to place the animals in their environment to avoid an>
extrapolated data. The mode1 developed would be only applicable to the current soi1 and
weather situation. The expert system program would offer possibiiities to use the methods
in other environments.
- Expert system building: the expert system was designed to integrate a11
aspects dealing with the technology evaluation at the farm level. The pro:gram cari be used
for calculating or predicting the draft requirements, the number of working days during
the growing season, the cost of production and also for estimating the feed rations needed
by the draft animais.
The validation of the expert system was mainly performed through the validation
of the different models used to evaluate various aspects of the technoiogy at the preset
confidence level. The major outputs of the program are expected to compare with actual
/
recommendations made to farmers in terms of resource management and allocation.
‘.
.
Chapter 5
FARM CHARACTERISTICS AND SYSTEMS MANAGEMENT
5.1. Farm equipment
The process to identifl a11 the different farm implements used by farmers took two
fi111 wet:ks to inspect each item surveyed in detail. It required quite an expertise and
patience to evaiuate the quality of the impIement> the effects of utilization and maintenance
in famlers’ condition, and to get the needed information from the users.
A total number of 126 implements belonging to 40 farmers located in agro-
ecosystem 4 and 5 have been surveyed. Al1 the implements involved in trop production
were walking equipment with handles (Figure 3 to 7). At the 9006 confidence level, the
total nulmber of implements consisted of the following (Table 16): 23% moldboard plow
UCF (0.73 $I 0.19 per farm), 22?/0 ARARA and EMCOT ridgers (0.70 L 0.16 per farm),
9% cultivator Sine 9 (0.28 F 0.13 per farm), 19/0O/ SUPER ECO seeders (0.60 2 0.16 per
farm), :and 27% ox-cart (0.85 & 0.16 per farm).
The distribution of each type of implement per agro-ecosystem shows that most of
the UCF msldboard plows were located in zone 4 with an average of 1.25 per farm (CV =
43.2%) against 0.20 per farm in zone 5. On the other hand, most of the ridgers were
confined in agro-ecosystem 5 with an average of 1.10 per farm (CV = 40%) against 0.30
113
IJCF Moldboard Plow
BBG Ridger
Figure 3: Moldboard plow UCF 10”
Figure 4: EMCOT Ridger (Gambia)
Houe She 9 Cultivatx
Figure 5: ARARA toolbar ridger
Figure 6: SINE 9 tooibar 3-tines
Super Eco Seeder
Figure 7: Seeder SUPER ECO
115
in zone 4. The same analysis shows also that the seeders SUPER ECO and cultivators
SINE: 9 were mainly used. by farmers
.
located in agro-ecosystem 4 respectively withi an average of 1.10 (CV = 40%) and 0.50
(CV -= 118%) per fat-m. 0x-carts were mainly
Table 16: Farm characteristics
Farm Fat-m Size Plow
Ridger Weeder Seeder 0x-CartI
Statistics
workers (ha)
UCF
ARARA SINE9 SUPER
i 0x
ECO
Mean.
10.53
4.03
0.73
0.70
0.28
0.60
0.85
2.23
Std .Error
0.71
0 26
0.11
0.10
0.08
0.10
0.10
0.16
Median
9.50
4 00
1 .oo
1.00
0.00
1.00
1.00
2.00
Mode
9.00
4.50
0.00
1.00
0.00
0.00
1 .oo
2.00
Std Dev.
4.51
1.64
0.72
0.61
0.5 1
0.63
0.62
1.03
Variante
20.36
2.70
0.51
0.37
0.26
0.40
0.39
1.05
Kut-tosis
-0.26
0.41
-0.89
-0.54
2.02
-0.54
“0.35
0.96
S kewness
0.50
0.63
0.36
0.25
1.66
0.56
0.10
0.87
Range
19.00
7.00
2.00
2.00
2.00
2.00
2.00
5.00
~finimum
2.00
1.00
0.00
0.00
0.00
0.00
0.00
0.00
-Maximum
21.00
8.00
2.00
3.00
2.00
2.00
2.00
5.00
Sum
421.00
161.35
29.00
28.00
11.00
24.00
34.00
89 00
Count
40.00
40.00
40.00
40 00
40.00 40.00
40.00
40.00
C.l(O.90)
1.17
0.43
0 19
0.16
_~-
0.13
0.16
0.16
0 27
Zone 4
Mean
0.00
0.50
Std dev
0 30
0.59
Zone 5
Mean
1.10
0.05
0.10
Std dev
0.44
0.22
0.30
116
used for transport especially in agrorecosystem 5 with an average of 1.05 (CV == 56?d, per
farm against 0.65 (CV = 88%) in zone 4.
‘_ .
- ,
It appears that farmers in the area of study were more concerned with the
acquisition of tillage tools (5706 of the total) (Figure Sa). This was cons’idet-ed as a
priority by farmers for a number of different reasons that previous studies have tried to
show. Three main reasons need to be mentioned: the improvement of field operation
timeliness, the extension of cropped areas, and the reduction of drudgery (Ndiame, 1980)
The geographical distribution of types of implements is an indica.tor of différent
cropping techniques used in relation to the resources available to the farmers (land,
number of farm workers). The type and number of implements that exist at the farm le) rl
is also related to the agricultural practices. In other words, two farmers l’ocated in the
same area may be expected to use different cropping techniques to produce the same trop
51.1. Types of implements
Most of the implements are Eocally manufactured by SISCOMA ‘(35OG) created in
1963 which became SISMAR in 1984 under a new management staff (FN-L. 198 1). Since
that time SISMAR has manufactured 49% of the implements actually used by farmers
The few implement imported (16%) are mainly found in agro-ecosystem 5 and originatzd
from Ciambia (Figure 8b).
5.1.1.1. Moidboard plow UCF 10”: Tillage
In the Basse Casamance region, a moldboard plow is one of the most important
implements used for land preparation (Figure 3). Besides breaking the soi1 for better
117
.
-.----
30
20
œ
Implenent types
-l_-- --.- --~-- ---- --------I_
Figure 8x Types of implement used by farmers
I.- .-_---
60 ------. _.-_ .______ _-- .._--._ .__-_-_
-- ..-.---.
50
0
s1sMA.R
IMPORTED
Manufactures3
---------.---
- .--.--.-. ------.-__.-__- __--.-.-_-.-. ~--
Figure 8b: Implements’ manufacturer
aeration., its main purpose is to plow. under both weeds for better control, and manure for
rapid decay. The better the bottom scour and turning the soil; the better 1::he moldboard
plow.
Designation: Moldboard plow CFOOOP LJCF type.
Hitching system: chain.
Traction: oxen.
Stability: single wheel for longitudinal stability.
Reversibility: one way,, right-hand plow.
Single bottom moldboard plow.
Nominal cutting width: 254 mm.
Total weight: 38 kg.
Working depth: 17 to 20 cm.
- Beam
The beam is made out of a 50x20 mm flat mild steel. The standardization test has
shown resistance to effort as high as 950 daN. The maximum effort reached in farmers’
conditions never attains this level, even when taking into account the presence of stumpj
scattered in their fïelds. In term of hardness, the value of the Brinell hardness test is 55
kg/mm2.
- Share and landside
The share and landside are made of the same mild steel and tested 95 l@nmL. The
general acceptable working component hardness recommended by research is wirhin the
119
range of65 to 70 kg/mm*. The share is a straight: edge type and its main role is to tut the
soi1 shce or furn~~ before sending it over the curved surface-of the moidboard in order to
,
be twisted and inverted or pulverized. The landside while riding along the Furrow wall
counteracts the throw or side pressure to help keep the plow steady while working.
- Moldboard
The moldboard is a universal type or cylindricol-helicoidal in design. This part of
the bottom of the plow serves to break or invert the fùrrow slice. It plays an important
role t,owards the quality of work to be done. It is also made of mild steel. The rnaterial
used to make the moidboard is of great importance, the better the soi1 scour on its surface,
the better the material.
These three working components along with the beam are bolted on top ofa mild
steel piece of irreguiar shaped metal called a fi-og.
- Hitching sytem:
The hitching system directly affects the maneuverability of the implement, as the
handling and the draft depend to a great extent upon correct hitching adjustments. The
hitching system consists of horizontal and vertical regulator to adjust the working depth
IMPL’WDP and the width MPLWWD It is important that the point of hitch be on a
straight line with the center of draft or resistance and the tenter of power or yoke hitch
point.
120
‘fhe implement is maneuvered by the plowman with the handles weEdeld on the rear
of the beam. The manufacturer SISMAR indicates an average theoretical field capacity of
‘. .
0.45 to 0.50 ha/day in an 8 hr-day.
5.1.1.2. Imported ridger (EMCOT)
This farm equipment is imported from Gambia and is equipped with a bottom
ridger (Figure 4). The implement is also called a middlebreaker. The bottom is built with a
double moldboard plow to invert the turrow slice, half to each side on every pass.
Completing the operation in a-round-trip results into the formation of a ridge f’or planting
crops. The winged designed moldboards are adjustable to control the height RIDGHCT
and the spacing of the ridges RIDGDIST.
The implement is similar in design to the UCF plow described earlier except that
the beam is made with an 1-beam instead of flat steel and -bat the share is in one piece for
the two moldboards.
The implement settings are performed through a clevis equipped with vertical and
horizontal hitch adjustment. The adjustment rules are the same as the UCF 10” plow for
obtaining a correct line of pull.
5.1.1.3. Seeding implement
The success to growing crops depends on the conditions the seed is pla.ced in the
soi1 in relation to different agronomie parameters along wixh the seeding equipment
features: type of seed distribution and dropping mechanism, furrow opener andl seed
coverage system (Figure 7).
121
Designation: Super Eco seeder
One-row and multi-purpose seeder.
‘. ..
Hitching system: chain with or without evener.
Traction: pair of oxen, horse or donkey.
Nominal cutting width: 254 mm.
Total weight: 38 kg.
- Frame
The frame is made ou’t of two 30x7 mm parallel flat mild steel runners, 1062 mm
long and braced in the middle. The frame opens wide progressively towards the front end
to allow the attachment of the hopper (galvanized zinc to protect from rust and
corrosion). The runners a.re joined together at the forefront to form the hitching system
made out of a 30x12 mm flat mild steel.
.4t the rear of the parallel runners two welded vertical standards: /one on each
nmner., are used to fix and adjust a press wheel (two point-to-point conical elements)
which firms the soi1 around the seed. The press wheel action is to complete the combined
action of the seed shoe (furrow opener) and the set of two sweeps (30x12 mm flat. IL!
solid steel) used to caver the seed with soil. The furrow opener and the sweeps are borh
clamped on the runners and are both adjustable in depth. At the very rear end, the runners
are firmly joined by an angle iron brace used also to support the row marker.
1 2 2
- Ilandles
The user guides the seeder with a pair of handles, adjutiable in he.ip,ht fix better
.
comfort. The handles are made with a 25x7 mm flat 1/2 solid steel.
- Metering wheel
The seeder is equipped with a pair of 400-mm diameter metering wheel made from
a 40x4~mm 1/2 solid flat band- steel. When rolling on the soi1 surface, the wheels drive the
distribution mechanism by means of an aile used as driving shaft. The distribution system
is composed of a system of two gears (driving and driven) and a plateau located inside :he
happer. The plateau is equipped with two pins which help hold the seed plates made of
aluminum. Two types of seed plates are more common: Ml-drop round-hoIle and flat-drop
types.
5.1.1.4. Multipurpose toolbar implements
Toolbar implements have been developed by French agricultural research
institutions (RAT, CEEMAT, MOUZON Mfg) working in West Mica to help farmers
increase their field operational capacities (Figure 5 and 6). The strategy behmd the
toolbar is to buy a single multipurpose frame and different attachments corresponding to
specific tasks (tillage. weeding, groundnut harvesting). There are a large variety of teolbar
impiements available to farmers in the whole West African region.
--W
.wœ”“.---
_- _--_
---- - --.
.,XWUU*IU.IIYUIIP-s---‘--
123
5.1.1.4.1, Multipurpose Carne MUR.4
.
Designation: ARF&4 toolbar
1
Utilization: Tillage, weeding, ridging, lifting.
Traction, oxen, horse and donkey for light soil.
Weight: 3 1 to 46 kg with the atta.ched working component.
Tillage attachments:
10” bottom
8” bottom
6’ bottom
Ridging attachments
.250 mm fixed wings
350 mm fixed wings
Adjustable wings
Cultivator attachments:
3 fi-111 and half sweeps (left and right)
5 full and half sweeps (lefi and right)
Optional: double reversible points
Groundnut lifiing attachments
200 mm sweep
350 mm sweep
500 mm sweep
124
- Frame
The main beam consists of a straight squared shaped tube made of mild stee,l. The
. .
stability of the beam in the working position is provided by a wheel. The wheel standard is
clamped at the front of the beam through a welded vertical adjustment device on which a
vertical 5-hole hitching system is also welded.
A pair of 1/2 solid steei braced-handles is bolted at the other end of the beam to
help the user guide the implement.
- Equipment attachment
The standard for either moldboard or ridger bottom and groundnut lifier is
at:ached to the beam by bolts in a special groove located between the handles attachment
and the rear end of the beam.
For the attachment of the other types of working components (spring tooth shanks
or straight-st nmed), two cross bars are used to increase the versatility and working
capacity of tire implement. Up to 4 tine standards cari be damped on : le (cross bars and
one under the main beam.
5.1.1.4.2. Multipurpose frame Sine 9
Designation: Sine 9 Cultivator or hoe
Utilization: secondary tillage, weeding, ridging, groundnut lifting
Traction mode: oxen, herse and donkey on light soi1
Weight: 30 a 45 kg with the attachment.
1
1
125
- Frame
‘.
The main beam of the frame is made out of a 40x20 mm flat milld steel on which
cross bars of the same marerials cari be added to increase its working capacity. The
attachment system is very similar to the ARARA multipurpose toolbar described earlier.
The handles are made out of2Ox27 mm flat mild steel and designed wide enough over the
frame to help its ease in use.
The stabilizing wheel standard is fïxed on the main beam by means of a clamp
equipped with a vise for easy continuous height adjustment.
- Attachments
10” plow equipped with a special shank and standard.
Reversible point mounted on a 25 mm square straight-stemmed shank.
SCanadian spring tines for weeding and stirring the soil.
The shovel and sweep attachment system is composed of one 160 mm centrai
sweep and 2 half sweeps :for weeding. Better action of S;oi1 stirring is performed with a
systern of 3 160 mm-füll sweeps.
Ridger: equipped with two adjustable or fixed wings with 2.50 m.m and 360 mrn
options. The attachment to the main frame is made with a special standard.
Groundnut M?er: The Mer knives or shovels are available in 3 sizes according to
the traction mode and the soi1 tvpe: 300 mm. 350 mm and 500 mm. The
attachment is made with a standard constructed of matrix steel to better
1 2 7
’
Table 17: Breakdown frequencies on farm equipment
Agro- Location
Implement
Repair and maintenance
Freq
Average
eco
type
type
(%)
distance
zones
il,- \\
126
5 . 1 . 2 . C a r t
Designation: 1.500 kg Grand-plateau cart
- .
Traction mode: pair of oxen
Plateau: 2.40x1.50 m
Base: 1.40 m
Pneumatic 145x 14 tires
Me with conical ro.ller bearings
Unladen weight: 220 kg
5.2. Repair and maintenance
The quality of repair and maintenance represents a key factor to t e stability of thc
technoloWq in farmers’ conditions. The survey has revealed that the level of expertise of rhe
local blacksmith was directly related to the farm implement working conditions and
reliability. Most of the blazksmith were not equipped enough to salve a11 the maintenance
and breakdown problems encountered by farmers on their piece of equipment (Table 17)
52.1. Breakdown frequencies
5.2.1.1. Moldboard plow UCF 10”
The straight edge share supplied by the manufacturer (SISCOMA and SIS;L’I.AR!
to equip the plow lasts no more than two years. The locally manufactured share made of
flatten truck Springs by the blacksmith needs to be changed almost every year. Plow shares
r
are d,amaged mainly by the presence of roots and stumps in newly cleared fields as farmers
* l
have the tendency to extend cropped upland areas. In agro-eco system -1 the high rate of
1127
.
’
.
.
Table 17: Breakdown frequencies on farm equipnent
3 Bougou,toub / Cropping Working components: frog
88
8
shares, landside,
sweeps.
Wheel bushings.
Transport Tongue welding
l:!
,-.-‘79
Tube and tires
- -
-
-
-
-
-
-
5
Kagnarou
Cropping Working component: Wings
64
On-site
shares, landsides.
Wheel bushings.
Adaptation and wing
welding on EMCOT ridger
1
Transport Tongue welding
36
On-site
Tube and tires
1 5
-
-
-
- -
5
Sue1
Cropping
Working component : Wings
12
8
reversible shares,
1
and
Wheel bushings
;
On-site
Adaptation and wing
45
welding on ERICIOT ridger
Transport Tongue welding
13
1 7
Tube and tires
-
128
utilization of moldboard plows explains the high maintenance frequencies of 60 and
88%. The landside has a longer life than the share. In gener& 1Aacksmiths buy the raw
‘- .
material used to make the share and landside from junkyard retailers.
The share plays an important role in the amount of draft required to pull the plow,
the penetration, and mostly the stability of the plow in. working position. Different studies
has showed that 75% of the draft is centered along a line passinr across the plowshare and
localized 113 from the landside.
The frog represents *the central piece of the UCF plow. It holds t
(share, lanside, moldboard) and the beam together. It needs to be protected from any type
of wear by replacing worn share and landside on time.
The majority of the implements that are in bad and very bad conditions (Figure 9)
are made of implements with worn frogs. These implements need to be discarded and
replaced as it cari be more costly to fix a plow bottom at this level of wear than to
purchase a new and complete implement.
According to the manufacturer SISMAR, GCF 10” plows are built to last at leasr
for MWLLFE= 5 years against FARMLIFE= 15.30 years in farmers’ condition. The stu&
shows that farmers are still using plows more than 20 years old (Table 18). If bad and
very bad plows are removed from the analysis because of the worn frogs, the life
expectancy of UCF plow, in farmers’ condition becomes FARMLIFE= 8 years. This
parameter is important in the process of evaluating the total number of plows for a given
time period needed by farmers in order to replace existing plows. The implement’s life
expectancy in farmers’ condirion (FAELVLIFE in years) cari also be used to estimate the
129
annual implement utilization fixed cost (IMPFCOS~T in local currency) when performing a
. .
farm budget analysis.
.
5.2.1.2. Seeder Super ECO
The seeder Super ECO requires more expertise from local black:smiths to service
because of the complexity of the seed distribution system. The mechanism housed in cast
iron r,equires little servicing. The manufacturer SISMAR urges farmers to verify the
quality and level of the anti-friction grease in the mechanism and to provide grease to gear
systefnn if necessary. The system is built strong enough to last the seeder’s lifetime.
Most of the other seeder”s working parts is serviceable by local blacksmiths. These
parts include the fùrrow opener, and the two f31 sweeps used to caver the seed with soil.
The weak point of the seeder resides in the plateau which is unservicealble by local
blacksmiths. The plateau made of brass is located inside the hopper and is part of the
distribution system. Tts role is to turn the seed plates held in place with a round brass nut,
a spring and two anti-slide short pins. Two major problems are noticed: the nut and the
spring are ofien lest or the pins are broken. An equivalent nut cari be found in rnarket
places located in nearest biggest town; but the pins, when broken, are difficult to replace.
In terms of breakdown frequencies, the seeder mainly contributed to the overall
breakdown and maintenan.ce frequencies of the shares and sweeps.
In relation to the level of utilization, seeders in working condition were ;found in
the area of study older than FARMLIFE= 11.9 1 years. The manufacturer3 specifications
indicat.e: a lifetime of klANLIFE= 10 years.
!
130
5.2.1.3. Toolbars and EMCOT ridger
There is little breakdown and maintenance on toolbars: The breakdowns identified
in the course of data analysis are classified in the same category of breakdown as the
attached working component. In ail cases, farmers in the area of study considered the
toolbars te, be identified with the working component they came with originally.
words, a multipurpose ARi& or SIXE 9 equipped with plow bottom is no more than a
piow, and the same for those equipped with ridgers. In the long run, the toolbars lose theil
original multipurpose characteristics. In generai, an ARARA toolbar is considered to be a
ridger as most of them are equipped with a ridger bottom and most of the SINE 9 tooibars
are considered to be cultivators or weeders, generaily equipped with 3-Canadian %Il
sweep.
The breakdown rate of ridger bottoms appeared to be closely related to tbc type of
wings they are equipped with. The E,MCOT ridger from Gambia tends to break at the
fiuture level between the wing and the share. Most of these ridgers need welding to repair
the damaged wing fixtures and account for 459’ o of the breakdown in agro-ecosystem 5.
On the average, the lifetime of EMCOT ridger is FARML.IFE= 10 47 years (the
manufacturer given lifetime MXNLIFE ivas not available). If bad and very bad implemenrs
are not çonsidered usable in the analysis, the lifetime decreases to F.AU!LIFE= 8 years.
The ridger bottom attached to the ARARA toolbar faces problems of wear on the
reversible point and breakage or loss of the wing nuts. The wing nuts are important for the
adjustment of their height in the process of monitoring the geometry of the ridges. This
type of failure contributes to 32’io of breakdowns mainly registered on ridgers in aura-
ecosystem 5. The ARARA toolbar with its attachment has a farm lifetime of almost
C
FARMLlFE= 14.60 years and 8 years when the bad and very bad toolbars are not
counted.
The sweeps used on the SINE 9 toolbar are also subject to the physical conditions
of the tields in relation to the presence of roots and stumps. The rate of sweeps wear is
similar to the plow share or ridger reversible point. The lifetime of toolbar SINE 9 in
farmers’ condition is around FARMLIFE= 6 years.
5.2.1.4. 0x-cart
O?c-carts are used year round by farmers. They mainly face problems of tires and
tubes due to the poor quality of the fat-m roads Flat tires on the road cari be changed in
the nearest town. They account for 1 2 ‘?O and 1 j”G of the total breakdowns respectiv,elv in
aoro-ecosvstem 3 and 5.
2
The prim,ary failure of the ox-cart is the break age of the tongue. u;hich generallv
occurs at least once in its Uifetime. Tangue breakage in combination with rhe tire problem
account for between 36 to 4004 of the breakdo\\vns. ‘The liferime in farmers’ condition is
close to FAR!KIFE- 14.67 vears in total and 10 vears when poor quality caris are net
counteci (Figure 9).
5.2.2 Farm equiprnent reliability
The reliability of farm equipment is closely linked to the existing maintenance
network and to the quality of technical services provided to farmers. The studv sho\\+s that
farmers rely on rhree categories of service providers: local blacksmiths. advanced shopr,
located in nearest big towns and farm equipment dealers.
132
,
10
---
---F-
Good
Average
rl
0
Bad
Very bad
State
Figure 9: Implments’ working conditions
60
50
10
Credit
Purchase
Trade
Gift
Acquisition mode
Figure 10: Acquisition mode
133
Until 1980, only govemmental extension services were involved in. making farm
‘_ .
implements available to farmers through the ONCAD and SONADIS network. There
were no private retailers involved in the process. One rnajor advantage of :his :system was
the quality control performed on implements and spare parts, but a11 the cost of
distribution had to be supported by the Govemment. The system did not work well as
fax-mers were net getting the spare parts on time.
The situation became worse alter 1980, when the distribution system administered
by extension agencies carne to a complete stop: no more dealers or spare parts retailers.
The otdy open alternative for the Govemment was to use the newly launched relgional
agricultural projects to support the distribution of implements and spare parts to farmers
located in their geographicai coverage. The PIDAC “Special Program” fmanced by
L?AID in 1979-80, was the only source of credit to farmers in the area of study. This new
iine of credit accounted for at least ZO?/O of the UCF plows, 64% of the seeders and a little
more than IJ”G of the CR-ca-ts (Figure 10). At the same time PIDAC tried to improve
repairing capacities of selected blacksmiths willing to undergo some training and to
purchase a set of tools at the end of the sessions, Currently, the local blacksmiths
represent the central piece of the network of the farm equipment servicing by carry& out
more than SOO’O of the repair and maintenance.
In this context, farmers do not rely completely on their farm implernents. Hand
tools are still read:y to be used to compensate for any implement’s failure.
In these conditions, the most often asked questions by animal traction practitioners
are: How long will it take ta fix a breakdown? How reliable is the technology to fat-mers”
1 3 4
There is no need to put stress on the importance of the reliability parameter when
farmers are in the process of adopting a new technology. The‘parameter is determinant in
- .
evaluating the performance of households using animal traction implements for trop
production. Timeliness is critical in the short period allowed for field operations.
The reliability of each farm implement or equipment (IMPLREL in percent) has
been evaluated and tabulated (Table 18). Two types of reliability coeffkients are
calculated, one on the yearly basis and the other on the implement lifetime basis. On the
implement lifetime basis, two reliability coefftcients are calculated, one on the
manufacturer!3 given life expectancy (IMPLREL man in percent) and the second on the reai
fart-n lifetime (IILPLRJZL rarm in percent). In any case, the reliability is given by folfowing
equations:
(42)
The real implement reliability per year IMPLREZL in percent is expressed by the
combined reliability of the two:
&CPLREL = ( 13fPLREL,,o~) * (LMPLRiq>)
A
I
?
3
-
-7
22;
-=-
-
CI
IC.
--I
N c, 5
X --C
: x
3 -
2
‘I :
136
Among the most used fat-m equipment, the ox-car-t appears to have the lowest
overah reliabihty lMPLREL= 58.2 l?/o. The observed breakdotins are frequent and
.
complex (pale welding, pneumatic problems) and require more time to repair in
comparison with cropping implements (66.07% for UCF plow, 68.75% For AKW,
70.20% for EMCOT ridger, 50.89?/0 for SINE 9 and 69.37% for Super ECO seeder).
According to the level of utihzation, breakdowns on newly purchased farm equipment
happen more ofien afier 2 years of use for cropping impiements and afier 3 to 5 years for
ox-car-t, evaluated from their first date of service.
The longer the idling or immobilization time for breakdowns, the lower the
reliability and therefore the higher the cost of utilization per year or per hectare
5.3. Draft animal management
A total of 89 oxen were counted in the study for the 40 househoids. They
represented more than 80% of the draft animals used in trop production. Al1 the oxen
belong to the Bos taurus breed commonly called Ndama. Small in size and
trypanotolerent, they are physically characterized as humpless and have an optimal or
mature weight MATURELW between 280 and 4 10 kg (Goe and McDowell, 1980) .A
mature weight value MATLRELW= 360 kg is often used for practical applications
The average number of oxen per farm is (2.23 2: 0.27) @ 90% confidence intena
The distribution is slightly different betkveen the two agro-ecosystems with an average ot
1.95 oxen per faml (CV= 47?/0) in zone 4 against 2.50 oxen per farm (CV- 4 106) in the
agro-ecosystem zone 5.
137
5.3.1. Avaiiability of draft animals
Oxen used by farmers were either locally withdrawn,f<om the farmer’s own herd or
purchased (50%) in the village from other farmers. Beside the withdrawal fiom the herd,
the other SO?/0 comprised exchanges against equivalent goods and gifis. I:t appears that the
availability of oxen, for draft purposes is not viewed by farmers as a significant major
constraint to adopting the technology
In the selection process, only male oxen were selected as social and cultural
standards of living prohibit the use of female animais for any physical activity
5.3.2. Training arnd working career
Training of draft animais for teamwork as well as training of farmers to better use
and adjiust farm implements was revealed to be a serious concern to farmers. The only
forma1 training sessions mentioned by very few farmers, occurred in the late 1970s. The
sessions took place at the Guerina Training Center located South of Bignona. The tenter
*as created in 1964 as a pilot unit in the Basse Casamance region, fmanced by the I,XDP
Ils ma.in objective was 10 train 32 farmers in the span of 9 months each year to acquire
skills in using draft animals and the variety of implements. The training \\V~S stopped in
1980 when the administrat.ion lacked financing to support the activities. Xowadays,
farmers are lefi to themselves to learn by doing Common farming practices were to train
animais at a much younger age. Most of the draft animals in the area ot‘study ivere trained
between the ages (AGETRQX in years) of 3 t.o 3 years. On some rare occasions. older
animais (5 to 6 years) were trained. The older the animal, the more dificult to be train&.
Staning one month prier to the beginning of the rainy season, the draft animals ivere gi~en
!
i
138
names, ad then taught basic moves through pulling and working as a team. During the
cropping season, animals are literally forced to work togetherqas a pair. It is only.during
their long working career ,that the pair of oxen begins to demonstrate more willingness to
work. A minimum number of 3 years (YRSEXP) is generally accepted in the area as
needed for a pair of oxen to respond correctly to commands. The average working
experience of’individual draft animal was found to be YRSEXP= 3.86 :? 0.79 years @
90% confidence inter-val. The longest working career was 9 years.
Farmers are facing more technical learning problems. An early study conducted 0‘
the Farming System Research team of ISRA’s Agricultural Research Center of Djibelor
(Basse Casamance) reported in comparing farmers with and without animal traction chat it
took 5 to 6 years of experience for farmers to master the technology. The mastering is
necessary before steady yield improvements becomes apparent. This includes implement
hitching and adjustments, adjustment of line of draft, keeping plowshare Sharp and
moldboard in good scouring conditions, ox-team driving, resources allocation and choice
of appropriate cropping techniques.
5.3.3. Stability
The key factor to a successtùl growing season is to maintain draA: animais in good
working condition. Stability of the technology from the draft animal perspective is what
reliability is to farm implements.
A close look at the data shows that 13?/0 of the farmers had only one draft animal
I
(Figure 11). They needed to borrow another animal to constitute a pair in order to car-
out their tïeld activities. This situation is risky, especially if the transaction takes place at
--- --
139
,
---- ----. ----_--.-.-___---.-..---~-..-._
-.------_-__--.
70 -------_--
- .____- -_---- -.-_^ -.--~-. _----. --.
.
60 -r
250 -
-
-
$40 -
d?
(r0E 30 -
u
-
b
L 20 -
10 -
---m
2
3
1
O -
--.,.-v-1.-
4
a
Number of draft animais
Figure 11: Draft animals’ distribution (zones 4 qnd 5)
140
the beginning of the rainy season. Areliability factor on the draft animais’ availabiiity
(ANIKEL in decimal) is therefore attached to this situation. Using Equation 20, the value
‘. .
of AIVIREL is equal to:
ANIREL= 0.87
The next group of farmers (66%) disposed of one pair of oxen. This repsesents
also a risky situation. If one animal happens to be unavailable because of in-jury, disease.
death, thefi or some other reason, then the production objectives Will be seriously
compromised.
For these two groups of farmers (79%) the technology is not stable, associating a
large amount of risk in the res’ource allocation process at the farm levei. The coefficient ot
stability using Equation 19 is equal to:
STABCOEF= 0.21
The third group of fat-mers (3?,$) showed more stability with 3 draft animais The
third animal was used to replace any unavailable one. The animal must be trained to work
on both the left and right side. This aspect was the most difficult for farmers to achieT+.e, as
working draft animais need to be familiar to one another for better and sustained
performance.
The fourth group (15%) was more concerned with continuity than anything else.
Replacement of the actual working pair of draft animals was a concem and planned well
ahead of time. With two pairs of oxen (4 draft animais), the second pair was trained to
111
replace the first aging one. The average ages of rhe first and second pair of oxen were
given. as follow (Table 19):
.; ’
Agro-ecosystem 4
.lst pair of oxen
---A---
Right side: 6.79 years with a CV= 3 1%
Left side : 5.80 years with a CV- 33%
2nd pair of oxen
-
-
Right side. 4.33 years with a CV= 44?/0
Left side : 3 .OO years with a CV== O?/o
Agro-ecosystem 5:
1st pair of o‘ten
Right side: 8.15 years with a CV== 309’0
Left side : 8.16 Yeats with a CV= 3796
2nd pair of oxen
Right side: 4.20 years with a CV= 51s’;
Left side 4.40 years with a CV= 4796
In any case, the second pair was not expected to yield the same productiv.ity as the
first pair. Switching to the second pair would s,low down the working speed. cau.sing both
fieid capacity and quality of work to decrease.
142
Table 19: Average age of 1st and 2nd pair of oxen
The fifth group of farmers (3%) had 5 draft animais to make up two pairs and one
ready for replacement. They planned for both the future, by training younger animais and
were ready for the replacement of any animal unable to work in the first pa.ir.
The stability analysis shows that for the majority of the farmers the technclogy is
unstable. In many cases they depend on their neighbors to borrow missing components or
on community work organized at the village level to achieve their production objectives. lt
was not the focus of this study to analyze the alternatives but previous studies conducted
in the same area have showed that in most cases, they still rely heavily on 4:raditional hand
toois to complete the unfinished work of the draft animals (Fall, 1985; Ndiame, 1986).
This fact demonstrates the importance of the presence workers at the farm level for animal
traction users until stability or reliability is improved.
5.4. Land and labor resources
5.41. Land availability
The average farm size in the area of study in terms of land area plowed by draft
animais was FMSIZE= 4.03 5 0.43 ha @ 9006 confidence inter-val (CI) (Figure 12). The
143
.
30 .--_--.-_
----_..~-----
25 -
20 _-
15
10
5
0
1
2
3
4
5
6
7
8
Fann size (ha)
Figure 12: Farm size (ha) distribution
--------_
--. --.-._____- _~-~ - ___.__ - ...-___.^ --.- --..--..- --.. ~- -- . . . .._ -..-
50 -I---~-.--_I--_----.--__-_-~-----.-_ ~---~- _-
40
1 to 4
5 to 9
-110to-----141
0
15 to 19
20 to 24
Nmnber of farm workars
Figure 13: Farm labor distribution
144
average farm size is slightly different from agro-ecosystem 4 with FMSIZE= 4.26 ha
(CV= 36%) to agro-ecosystem S with FMSIZE= 3.8 1 ha (CV= 44%). The potential for
- .
land extension is limited by the presence of valleys and lowland areas generafly decoted to
rice cultivation whenever possible.
5.4.2. Farm labor
Most fat-mers were equipped with tillage implements (57041 of total farm equipmenr
surveyed). The lack of post-tillage implements has relegated draft animals to a low level of
utilization during the growing season.
The average number of fat-m workers NFWKERS present at the fa.rm level during
this study was NFWK.ERS= 10.53 + 1.17 @ 90% probability level (Figure 13) The
difference between the two agro-ecosystem zones were not significant with an average of
NFWKERS= 11.60 fat-m workers (CV= 360/0) for zone 4 and NFWKERS- 9.45 farm
workers (CV= 4104) for zone 5, The availability of the farm workers throughout the
growiny season is crucial to meet the production objectives. The temporary absence of
active family members was pointed out to be a constraint at weeding time
5.5. Adoption dynamics
‘ .
As expected during the implementation stage of many agricultural development
projects around the world, three phases have been identifïed in the process of animal
traction adoption in the Basse Casamance region (Figure 14):
/* :
*
145
m
*
?
0 ! m
? ? ? ?
? ? ? ? ? ?
? ? ? ? ? ? ????? ? - __.-- ----_-.~.-
7??
75
80
85
90
95
Year
-~--~
_-
.~---.--
Figure 14: First year in animal traction
-. ----~.
-.-- --_- .-..
100
.
. .
T
L
rn 80
E
f
. .
c
kW
= 60 -
5
,
. .
0
h
/
. .
0 40
i
j-
.-
I "
z
+
I
3
. .
E 20
1
E
.
6
. . . .
.
0 -..-a222--- __ - --_- . . .._. ._. _.-. -._--- . ..--. _
0 5 10
15
20
25
Number of ‘eut-s of experience (years)
-- -. __--.
Figure 15: Number of years of experience
Early adopters:
This phase extended from the end of the 1960s to 1974.. Animal traction adopters
were made of farmers who acçepted the challenge of securing agricultural production
through diversification and extension of upland crops, mainly on groundnuts (cash trop).
They represented no more than 1 5 ?/o of the farmers in the area of study. F’armers located in
agro-ecosystem zone 4 were the first to adopt the new technology.
Adopters and innovators
The better performance of households using the new technology through the
increase of gross ‘cash trop production generated a lot of interest from manual farmers.
The adoption rate was the highest between 1976 and 1980 when no less chan 509.0 of the
farmers became involved. This cari be explained by two decisive factors. First, the
incentives were put on groundnut production at the governmental level through the
application of the National Agricultural Policy.
The policy was characterized by an easy access to credit and organized markets for this
trop. The quasi-total groundnut production was used to supply the national oil indust-,
Second, the progressive encroachment of drought in the area forced farmers to look for
other alternatives to better secure production at the household level. Rice production \\~as
reduced by the decrease in the amount of rainfall (30% to ~0% in eariy 1980s) to rhe point
that rice was no longer driving the farmers’ production system in the area of stud?
147
a d o
Late p t e r s
Tbe decrease in the rate of adoption starred in the ear& 1980s and went tiom .~-~“o
to an insigificant lesel of 2.0 in the early 1990s The National Agricultural Poiicy
administered through the Lo\\:ernmental agexies was ended in 1980 at the national 1~x4
for many reasons The main reason was rhe inability of farmers to psy back the medium
rerm (5 years) credit loans on fartn implements afier the Ion3 lasting droug:ht of 197S-
1982. The farmers’ primnrv source to acquire farm equipment was put to an end ar.d as
was pointed out by ?Jdiame i 19%) it is unlikely that a similar program w41 be initiated in
the filture. Following this,, che credit systems fosrered by the Government and polie‘
makers to finance agriculture Lvere merged into the collaboration with private banks anci
newiy implementing rural development projects Along with other reasons. the conditions
for eligibiiity ofthe Kxmers to be par-~ ofthe newiy implemented credit systems iLere net
eupect~rd :o increase t.he nunber ofnew animal tract::on adopters.
The dynamics ofadoption appeared to be i.ery sompiex in regards to thr: \\xriabi?s
in\\,ol\\.ed As the study sho\\\\;ed. almost 70°,ù ofrhe farm implements ac:uaII~: used lx
farme:-s lvere financed through a credit system (Fipu.re 110). From 1930. mcst oìrhe
farmerr, who benetïted. fro:m the neivlv formed credit lines. tvere already animal rraction
users The rate of new adopters slo\\ved doivn for différent reasons that included :he C~G
of transactions invoived:
- Awareness ofrhe credit systern: farmers with a past credit systcm’s
experience L\\-ere able to reduce the cost of transaction by getting correct information El+
access tu information helped thym Lvork thelr l.vav through the svstem. Besides having
148
easy access to information, they had scrong connections for rapid administrative paper
work and were also more aware of ways to avoid sanctions .or,penalties for no annuity
.
payment
- Members of farmers’ production group or fax-mers’ organization: There is
more and more tendency for farmers’ organization to guarantee the credibility of their
members. They carry out preliminary negotiations on behalf @fat-mers to speed up the
process of acquisition. Farmers’ organizations administer have their own sanctions to
default repayment in accordance with the organization byIaws. In the area of study and
since 1988, the CADEF farmers’ organization guaranteed 5 1 % of the implements acquired
by fat-mers.
- Conditions of eligibility: the involvement of private banks increased the
cost of transaction. Interest went up and farmers were asked to constitute a down
paument in order to be eligible for the credit. The National Agricultural Credit Bank of
Senegal (CNCAS), the main actor in the credit system, was asking for a 20% down
payment and 14% of interest on the capital. With these conditions, farrners generally
believed the cost of money was just too high.
- Non existence of second hand implement markets: Onl:y 26% of the
implements used by farmers were purchased in cash (Figure 10). These farm implements
were mainly bought from out of state markets, across the Gambian border (16%). These
implements imported by Gambian dealers were mainly ridgers fabricated by the English
8.
i
-
. .---
*RINI-,-’
.
149
manufacturer EhKO’T Local markets, especially those located in North and South
Central regions of Senegal accounted for nearly 10% of thé $urchases. Xt appears that
there is a general tendency for farmers to keep their implements on the farm (19%j until
they were completely lworn out, out of service and finally discarded (Figure 9). The ver-y
few occasions whereby farmers got rid of their implements were only when they traded
them off against other equivalent goods (504) or gave them away as gifts to relatives and
fi-iends ( 1%).
- Learning period: It required a long time for fat-mers to master the new
technology (5 to 6 years). In agro-ecosystem 4 and 5, 10 to 15% of farmers were
classified inexperienced (Figure 15). In these conditions, a new adopter experienced
mixed results 0fusin.g animal traction. Animal cultivation cari be ver-y fi-ustrating and using
farm implements cari appear to be a surn of tough skills to master. Lack of working skills
cari prevent fat-mers from performing proper adjustments. and identifving the signs of
inefficiency.
Farmers located in the area of study do not have a long history of animai traction
The average number of years of experience was (16.43 :: 1.5 1) years @ 90?/0 level of
probability with a maximum of 25 years.
5.6. Summary
This broad overview of animal traction in the Rasse Casamance region shows that
many factors were involved in the implementation of the technology.
150
The quality of the implements made available to farmers is of major importance in
relation to its life expectancy. Quality control is a must. The design also is ,a.s important.
The simpler the implement, the better as farmers are not generally skillful in the handling
and management of complex toolbars. The analysis shows that toolbars were generally
,.
identified with the working component that they were originally equipped with. The
introduction of toolbars must go along with specific training to show their flexibility and
to demonstrate how they cari improve the overall farm working capacity.
The quality of maintenance and servicing is important to the viability of the
‘” ,
technology. The skills and expertise of local blacksmiths must be improved through
!
i
training and better tool supply. Increasing reliability of farm implements is hkely to
decrease production cost in the short r-un.
There is a need to improve performance by facilitating the adoption of
complementary implements. This Will help to avoid the side effects of the shifting of
bottleneck to fil1 the labor deficits. TO meet this need it is crucial to reduce tlne learning
period by better training in order to achieve rapid returns on investments that cari be
reinvested.
In order to tie these diRerent aspects together, it is important to carry out an
analysis of actual and potential use of the technology. A correct evaluation of the impact
of the technology Will help farmers to have a better understanding of the effects of animal
traction on the achievement of their production objectives. It Will also help differentiate
between yield increases and land extension effects. A good starting point is the evaluation
of the working capacity of draft animals in the farmer’s condition.
h
Chapter 6
WORKING CAPACITY OF DR4FT ANIMALS
Two pairs of oxen per village and one pair on-station for control were used. A
total number of 9 pai,rs were classified into 3 traction groups according t:o the pair? total
liveweight LW:
Group 1 : 6 0 0 t o 7 0 0 k g
Group 2: 5 0 0 t o 6 0 0 k g
Group 3: JO0 t o 5 0 0 k g
The 18 draft animals were monitored separately throughout the growi.ng season:
workllng schedule (type and time spent), liveweight follow up, disease, and feeding system
6.1. Ebironmental conditions
In the Basse Casamance region, farmers are urged to carry out their fïeld
operations as early in the morning as possible to avoid heat stress and to rest rhe animais
according to the level of intensity of the fieldwork. The heavier the work., the shorter rhe
time alliocated to the task and the longer the rest.
151
152
Accorchg to the weather data collected during the study, it shows that drafi oxen
in the Basse Casamance region faced the weather related fatigue between midJuly and the
end of September when the average minimum relative humidity RHmin approached 70%
(Figure 16 and 17):
- From onset of rain to July 14
. Minimum Temperature Tmin:
24.21 “C
Maximum Temperature T,,,:
33.73 “c
: Minimum Relative Humidity RHmin: 52.56
. Maximum Relative Humidity FUI,,,: 94.77
- From July 15 to end of September
. Minimum Temperature Tmin:
23.88 “C
Maximum Temperature Tmar:
31.04 “c
: Minimum Relative Humidity FELin: 67.82
Maximum Relative Humidity REL: 99.49
The difference in RE%,, shows that the oxen have a narrow windcsw of working
days before the draft animals entered the uncomfortable working period of the year. This
situation dictates a working strategy based on the number of animal working hours
ANWKGHR per day (in hrs/day) allocated to field operations: more working hours before
mid-July and less working heurs for the remaining of the growing season.
It must be noted that both minimum temperature TA~ in “C and maximum
temperature TmX in “C were on the decreasing phase from the onset of the rainy season
and reach the minima towards the end of the year. Minimum temperature ‘I’ti was
decreasing from 27.00 “C and maximum temperature T,, from 43.80 “C. Records show
that the highest temperature has always been recorded during the month o-fbfay and the
lowest in December-January. The effects of temperature alone need more imvestigation.
i
/
i
1
15 ~~~li:i~“.‘:
M
J
J
A
s
0
N
Months
,.-~_- _--,
/
1
I_ Tmin -.Tmax
---~
[-__- - - - - - - - -~-----
__-_..- -- -.-.-..
Figure 16: Temperatures (min and m.ax) at Ziguinchor (1996)
-~---~-------.__
__- .--.-.-.--_--._------ ~_~--_-_-- -.- -.-
-- Rh min -- Rh max
Figure 17: Relative humidity (min and max) at Ziguinchor (1996)
1 5 4
The main objective of the recornmendations on working strategies is to optimize
the energy available from the draft animais to yield a better’&ciency. At this stage of
investigation it is important to evaluate the level of animal energy output ANERGY in
h43’day that fat-mers cari Count on, especially at the beginning of the rainy season when it is
needed the most.
6.2. Maximum performance
The maximum performance test was conducted on-farm during morning hours.
The test took place on already cleaned fields and ready to be plowed for the cropping
season. The time constraint was such that only two villages out of four could be visited
per day.
The pair of oxen from each traction group used the same loaded sledge for each
puiling triai. The pulling distance or distance traveled DISTRAV in m (100 to 200 m) was
sometimes adjusted to fit the geometry of the fieids. Each tria1 lasted approximately 2
hours starting with a warming up of the pair of oxen yoked to walk together without load
for 10 to 15 min. The head yoke type was used for the trials. The characteristics of the
draft animal’s performance are given in Table 20 and 21.
Different adjustments by triai and error were performed at the sledge hitching point
to reach an acceptable line of draft. Two variables were recorded during each repetition:
the pulling force (PF in N) and the time (PTIME in s) to travel the given distance
(DISTRAV in m). A regression analysis was t-un to mode1 the power output of each pair
(ANPOWER in J/s) in relation to the required pulling force PF in N to move the loaded
sledge. The same analysis was carried out on the average performances.
-- Y’--’
1
I
155
6.2.1 Maximum performance
i
.
6.2.1.1. Optimum point
‘_
-- Optimum power (ANPOWEK,t’)
’
!
The optimum poser ANPOWER,pt in J/s is used to compare the performance of
each pair of oxen. The power is calculated by multiplying the PF (N) and ANSPEED
(~/S:I. A one-way analysis of variante is conducted with each of the factors contributing to
Table 20: Performance of team of oxen
/LIV Speed Optimum Max PF
(mis) p o w e r
PJ)
(JN
1267.49 353 1.60
-
1796.70 34, 0.00
-
-
-
0.90
1241.27 2256.30
1282,.29 3256.00
1.09
1603.94 2950.00
-
-
933.64 1962.00
817 68 2158.20
1377.81 2530.00
-
-
( 1) P stands for pair of oxen
(2) The pair of oxen showed unwillingness to pull
Table 21: Average performance of team of oxen
l
II
3 ( 4 4 8 . 3 3 <I 1058.34/--J 0.98--j 1 0 4 3 . 0 4 ( 2 2 1 6 . 7 3 11
I
( )The pair of oxen P2 of Group I was not included in the average
1 5 6
the puwer: the pull force PF in N and the draft animais qketl ANSPEED in m/s (Table 2f.
and 21).
The analysis of variante on the optimum power ANPOWERopt Hn J/s shows that
the difference among the 3 traction groups is not ver-y significant (Table 22). The
arithmetic difference arnong the groups cari help evaluate the energy balance involved in
order to determine the exact amount of feed required for a given group.
p,
Table 22: ANOVA of the Optimum Power (3s)
l
SOURCE
DE:
S S
M S
F
-
w
P
-
Group
2
322285 161142 2.05
0.224
E R R O R
5
393748 78750
TOTAL
7
716033
1
,.* l
INDIVIDUAL 95% CI’S FOR MEAN
BASED ON POOLED STDEV
I
l
LEVEL N
MEAN
STDEV ____- + ________- + ---_---__ + --___-___ +-
Group 1 2
1532.1
374.2
( -m-----m--- * --“v----“m1- 1
Group 2 3
1375.8
198.6 -..e----em * -- -w--mve-
(
1
Group 3 3
1043.0
295.7 (-------- * ---- --..-)
----* + ..--* -----+ ---- ‘. --_- + -------__ f-
lr
POOLED STDEV = 280.6 800 1200 1600 2000
I
- Optimum Pulling Force (PFQ,~)
The arithmetic difference in optimum power ANPOWER,,t is mainly due to the
significant difference found between the pulling force PF,,t of the different traction groups
at the optimum point. The 95% confidence level analysis shows that difference in PF,,,t is
!J
157
signifcant between traction group I/group 2 and traction group 3. The diflerence is not
significant between traction gsoup 1 and 2 (Table 2;3).
‘_
.
,_ ,
The ratio of pulling force to liveweight (PF/LW)o,t at the optimum point is the
same for the 3 traction groups. The difference is not significant (Fisher-F= 0.92 and p=
0.45) and the average for the 3 groups is (PF/LWJ,rc= 0.24 with a CV= 15% with pulled
Table 23: ANOVA of PF in N at Optimum point
- - -
-
-
-
SOURCE
D F
S S
MC;
F
P
Group
2
32 1668
160834
8.19
0.019
E R R O R
6
117780
1.9630
TOTAL
8
439449
----
-
-
-
INDIVIDUAL 95% CI’S FORMEAN
BASEDONPOOLEDSTDEV
LEVEL
N
M[Em
STDEV w___.._ + _____..___ -/- __-______ + _________ +
Group 1
3
1489.7
113.0
( -----.- * ------- 1
Group 2
3
1419.9
47.1
------- * ------ 1
Group 3
3
11058.3
209.6
POOLED STDEV := 140.1
1000
1250
500
1750
standard deviation Stdev= 0.036. Monnier (1965) and Nourrissat (1965) repot-ted in
different field triais that the Ndama cattle could pull continuously 10% to 14% of their
body weight for a long period of time. Field operations requiring pulling 20% or more of
the body weight were considered to be from medium to heavy work and could not be
sustained for a long time. According to Goe and McDoweil(1980), citing Swamy Rao
I[ 1964), an improved harness yielded a pull to weight ratio of 24%.
158
- Draft animais’ speed ANSPEEDopt
The speed ANSPEEDopt in m/sec is the most important factor in the power
.. .
estimation. The significant difference in the pulling force PF is rapidly buffered by the level
of low speed of the pairs of oxen across the groups. The analysis of variante shows that
the speed of travel ANSPEEI& is almost the same for the traction groups (Table 24)
Table 24: ANOVA of the walking speed (misec) (Optimum point)
SOURCE
D F
SS
MS
F
P
Croup
2
0.0018
0.0009
0.07
0.935
ERROR
5
0.0666
0.0133
TOTAL
7
0.0684
INDIVIDUAL 95% CI’S FOR MEAN
BASED ON POOLED STDEV
L E V E L N
ME)Jj
STDEV
mm__ + ______.._m + _____m___ + ____ _ me_.. -+-es
Group 1
2 1.0050 0.1385
(----_ ----* --*..--------)
Group 2
3
0.9667
0.1069
(__-______-- * ___--I_____-)
Group 3
3 0.9767 0.1041
( c~~~~~~~~-*-- *-------- )
---- + -------mw f.. -------- + --------- j--a,
POOLED STDEV = 0.1154
0.84 0.96 1.08
1..20
Considering the level of pulling force PF required for fteld activities, it would be a
slight improvement to accept the level of significance of the optimum output. This cari
help to better allocate the length of time to perform field operations and also, to evaluate
the amount of feed required for draft oxen to carry out specific fïeld activities
-cm
-. ~-
v--
-,,yY,u-II~--’
..***.-YY(I*YUYI
159
6,.2.1.2. Maximum Pull force RF,,,
The maximum pulling force PF,, in N represerits the va%ue for whi.ch the
. .
pair of oxen had difficulties in walking while pulling. It corresponds to the animais’ speed
ANSE’EED almost equal to 0 m/sec. In other words, PIF,, represents a continuous
maximum pulling effort in comparison with the instantaneous maximum pull
The analysis of variante shows that the difference is significant between the groups
at p= 0.038 (Table 25). The 95% confidence interval separates the means and shows that
the maximum pull PF,, of group 1 is significantly different from group 3.
Table 25: ANOVA of maximum pull force (N)
- - -
-
-
---~
SOURCE
DF
S S
A?s
F
P
v-------
---.-
Group
2
1892816
9416308 6.76 0.0
<
ERROR
7 0 0 2 2 3 14004,5
TOTAL
;
259304 1
---...---p_
- - - ~--
INDIVIDUAL 95% CI’S FOR MEAN
BASED ON POOLED STDE%
.-----
_
-
- - - - -
LEV -_ N
MËAN S T D E V
-------+----e-.--e + --e---w-- + e---e--<.-
Group 1 2 3465.8
93 1 (
--L---m--
*
-------.-)
Group 2 3 2820.7
512.4 m------ * o------ )
Group 3 3 2216.1
288 fi (__--___ 5; ____.___)
-.. ----- -i- -....--4-.-- + --------- + ---------
POOLED STDEV := 374.2
2 100 2800 3500
The maximum pull force PF,, cari be translated as the capacity of draft oxen to
overcome the peak draft errcountered in the fields while continuousiy working. The
average ratio of pull force to liveweight (PF/LW),, at this point is 0.5 1 with CV= 14?$
and Fisher-F= 0.08 at p= 0.923. At this stage of the triai, the pairs of oxen were asked to
give the maximum effort which yielded this high level ‘of pull ratio PFILW. The ratio
I
1
1
II
/
160
corresponds to pulling force PF,, tif almost half of the body liveweight LW. In any case.
it is not expected in the reai working situations that draft oxen be submitted to this Ievel of
. .
pull on a regular basis.
Values found in the literature shows that the tria1 results for the Ndama breed are
in the expected range. Goe and McDowell(I980) reported a ratio of 0.5 1 or 5 196 during
tests conducted in India with different pairs of oxen.
6.2.2. Characteristics of power curves
A regression analysis was carried out to mode1 the power output ANPOWER in
J/s of each pair ofoxen. The results are presented in the Appendices. A regression analysis
I
WI
l
was also performed on the average performance of each traction group. The different
models are used in the expert system computer program to help evaluate and refine the
energy balance at the farm level.
6.2.2.1. Power output mode1
The analysis was carried out on two determinant parameters directly involved in
the performance of the cropping system: the power output ANPOWER in. J/s of the draft
animals (pair of oxen) and the level of speed ANSPEED in rn/s at which the task was
performed. Firstly, the parameters were estimated for the general linear regression,
secondly an analysis of variante was performed, and thirdly, a look at the cor-relation
coefficient R among variables to help explain the relationshïp. The RZ is also important as
it helps to determine the amount of variation explained by the regression lime. The best tit
was obtained with the quadratic form of PF using the variables (PF/lOOO) and (PF/lOOO)’
161
A general comment: most of variation accounted in the linear regression was
.
mai:nly due to the border values at:
PF = 0 and PF = PF,,,
At PF == 0, the pair of oxen was walked without load to caver the path. Farmers
had different feeling of walking with the draft animals in terms of how fast ta perform the
task. .At the other end ofthe trial, pulling the maximum load depended on the willingness
of the pair of oxen to do the job. It did not matter h.ow well the oxen were trained. the
greater the amount of draft required, the less the willingness of the oxen. to pull. As a
consequence, the pulling of the maximum load was put to an end to manage the
temperament and excitement of the animals. In such. cases, the value recorded for the
maximum Pull force PF,,,, would be underestimated. The analysis is presented by tracrion
group’
- Traction Group 1: 600 to 700 kg of :LW
The average performance of the Traction Croup 1 is summarized in Table 26 The
regression equation is:
1 6 2
“/
Table 26:‘Average performance of Traction Group 1
PF/LW,
0.00
0.04
0.08
0.14
0.16
0.19
0.21
0.23
0.24
0.28
0.31
0.35
0.37
042
0.52
. 1
.
. 1
Table 27: Parameter estimates (Group 1)
Predictor
Coef
Stdev
t-ratio
P
-
-
No constant
(PF/I 000)
44.350
2.120
20.92
0.000
(PF/ 1000)’
-12.6875
0.8319
-15.25
0.000
-~
s = 4.454
R-squared = 0.88
Table 28: Analysis of Variante (Group 1)
- -
SOURCE
DF
SS
MS
F
P - -
Regression
2
13687.5
4843.8
344.96
0.000
Errer
13
257.9
19.8
Total
15
13945.4
--
163
i
‘2. ,,
?
m œ
?
N
.
m
?
?
.,.m... .,.
T
I
l
0 &--.-------------~-,-,,k----+-
1
0
5 0 0
1 0 0 0
1 5 0 0
2000
2500
3000
3500
Puli Force (N)
_._--~-_-_l_-.--.-----
---
Figure 18a: Power curve of Traction Group 1
1.4 T------
r
~\\
1.2 .i
’
. . .
?
?
m
? ? ? ??
1 .-
‘\\ . a
j.
z
,-L
2 0.8 - -
/
\\
?
?
??
?
3 0.6 --
OF
cnt
*
?
?? ? ?
? ? ? ? ?
?
m
Y
?
/
0.2 -
c
1
\\
I
-\\
0 2 --7--------l-‘
-.-- _...*_ -.*--_
0 0.1 0.2
0.3 0.4
0.5
0.6
Pull Force iliveweight
-- Fitted line
Figure 18h: Walking speed of Traction Group 1
164
,* \\
The fitted linear regression equation R2 explains 88% of the total variation, which
I
,
is ver-y significant (Table 27 and 28). Most of variation was mainly introduced by the,
border values at PF = 0 and PF = PF,, (Figure 18a). Accord& to this model, the
unusual observation at PF,, suggests that the pairs of oxen in the Group 1 çould have
pulled more that was recorded. This influence is not negligibl’e in the amount of variation
translated into the value of the R2.
In the region of the optimum power ANPOWER+ the amount of pull applied
PFCJpt = 1750 N started to become unsustainable and to induce the unwillingness to pull
mentioned earlier. This fact indicates the needs for more studies to improve the harnessing
systems.
- Traction Group 2: 500 to 600 kg of LW
The average performance of the Traction Group 2 is given in Table 29:
-1
The regression equation is:
I..,
1
I
(46)
The fitted regression equation shows a good coefficient of detetmination R’= 0.93
which is very significant (Table 30). However, the analysis of the data shows two
singularities. Firstly, according to the model, the amount of maximum pull PFmas= 20 13.9
N developed by the pair of oxen was also underestimated (Figure 19a). Once again, the
unwillingness to walk: at this level of pulling induced part of the variation. Secondly, the
value of the maximum pull force PFmx, I‘nfluenced the trend of the curve.
165
Table 29: Average performance of Traction Group 2
-
-
-
-
I?F (IN)
Speed
PF/lOOO
(PFA 000)2 P o w e r :. (Power)‘” PFLW
- - -
(mkec)
-
-
(N) CN)
(J/s:)
(Jhec)
-
-
-
0 . 0 0
lu32 0.00
0 . 0 0
0 . 0 0
0 . 0 0
0 . 0 0
8 6 8 . 9 0
1.25 0.87
0 . 7 5
1086.13
3 2 . 9 6
0.16
9 2 9 . 7 0
1.04 0.93
0 . 8 6
-966.89
3 1 . 0 9
0 . 1 7
1209.90
1.11 1.21
1.46
13442.99
36.45
0 . 2 2
1285.l.O
0.89 1.29
1.65
1143.74
3 3 . 8 2
0.23
1409.1.0
0.91 1.41
1.99
1282.128
35.81
0.26
1471 .io
0.99 1.47
2 . 1 7
1456.79
3 8 . 1 7
0.:27
1917.410
0.76 1.92
3 . 6 8
1457.22
3 8 . 1 7
0.35
1970.10
0.75
1.
.97
3 . 8 8
1477.57
3 8 . 4 4
0 . 3 6
2013.90
0.35 2.01
4.06
7 0 4 . 8 6
2 6 . 5 5
0 . 3 7
2412.10
0.29 2.41
5 . 8 2
699.51
2 6 . 4 5
0.44
2452.50
0.42 2.45
6.01
1030.05
32.09
0 . 4 5
3103.00
0.00 3.10
9.63
0 . 0 0
0 . 0 0
0 . 5 6
CI==
-
-
-
Table 30: Parameter estimates (Group 2)
l’redictor
Coef
Stdev
t-ratio
P
Noconstant
(I)F/ 1000)
48.167
2 . 0 0 9
2 3 . 9 7
0 . 0 0 0
(PF/1~OOO)2
- 1 5 . 3 3 7
0 . 8 7 2
-17.58
0 . 0 0 0
s = 3.604
R-squared := 0.93
Talble 31: Analysis of Variante (Group 2)
---y-
-
-
-
SOURCE
D F
S S
MS- F
P
-
--
-
-
-
-
Regression
.2
- 12505.1
6252.6
481.32
0 . 0 0 0
Error
11
142.9
13.0
Total
1 3
12648.0
- - - -
-
-
-
-
166
e----p -. -_.
-~- --_--
50 -j
--.
I_-
/
.- .
/
t!
I
?
?
i
?? ?20 4I
&
i
10 1...
l
l
I
/
oh-
:
/-+--+-m--i
0
500
1000
1500
2000
2500
3000
3500
Pull Force
--
--.
Figure 19a: Power curve of Traction Group 2
1.6 -
-
- -
-
,I
----\\.
1.4 fy-,...
.._
.__..__
1.2 7..
:r-,.,
!!
:
I.I
t
.
I
oc
/--
--,
-_
0
0.1
0.2
0.3
0.4
0.5
0.6
Pull Forceaiveweight
Figure 19b: Walking speed of Traction Group 2
.--
.‘mwa-----
167
- Traction Group 3: 400 to 500 kg of LW
The average performance ofthe Traction Group 3 is summarized in Table 32.
The regression equation is:
P F
P F
(;4NPOR’ER /+== 56.1 I * ( -/ - 75.63 * I- Y
1000
1000
(47)
The parameter estimates and the analysis of variante show a very goodl fit with an
R2 =: 0.91 (Table 33 and 34:). The draft animals had a tendency to walk fast in the area of
lower required PF a.t the beginning of the trial. The unwillingness to pull appeared in the
same region of the graph and. mostly beyond the optimum power point (Figure 20a).
Most of the variation was introduced once again by the maximum pull force PFmaX,
underestimated according to the model.
6.2.2.2. Walking speed (ANSPEED in m/s;)
The walking speed ANSPEED in m/s while performing a task is the second
important parameter towards the evaluation of the power output ANPOWER. :Regression
analysis has also been performed to give more insight in order to explain the accounted
variation in the power output models. The analysis has been conducted for each; Tra.ction
Group. The best fit was obtained by plotting the speed ANSPEED as the dependent
variable and the ratio of pull force PF over the average liveweight of the group PF/L,W
The speed plotted against the PFLW yielded a better cor-relation than the speed plotted
against pu.ill force PF.
168
Table 32: Average jkformauce for Traction Group 3
PFfLW
“--
0.00
0.04
0 . 0 8
23.92 0.12
0 . 1 6
0.21
0 . 2 8
0.31
0.33
0 . 3 8
0 . 5 0
Table 33: Parameter estimates (C;roup 3)
f
-
_--
Predictor
Coef
Stdev
t-ratio
P
-
Noconstant
(PF/ 1000)
56.111
3 . 1 0 6
18.06
0 . 0 0 0
(PF/ 1 ooo)*
- 2 5 . 6 3 2
1.816
-14.11
0.000
-s= 3.417
R-squared = 0.91
Table 34: Analysis of Variante (Group 3)
-
_--
SOURCE
DF
S S
M S
F
P
----
-
Regression
2 -
5 3 6 2 . 4
268 1.2
229.66
0 . 0 0 0
Errer
9
105.1
11.7
Total
11
5 4 6 7 . 5
169
50
7
‘_
‘.
, ’
oa----------
------ --..--_m
i
0
500
1000
1500
2000
2500
Pull Force (N)
------p_
- - - .~.-
--v.---
Figure 20a: P’ower curve of Traction Group 3
-~--
-.
1.6 --p
---.--_--~---_~.--
i
,
1.2 f. -..hw,
û
j
‘-m .._.
1
‘.. m
!
,$
7’ .’ ... ..
. . . .
+
-+
E0.g ;.
‘.\\
,,
FI
I
8
00.6
_-.
“. .*.
rn
i
0.4
1.
m -.
4
1,
/
‘.._
Os2
7
.’
+. . .-.
i
‘..
i
t-J
GIC---
--.4-.-
-1
0
0.1
0.2
0.3
0.4
0.5
Pull ForceILiveweight
- Fitted line
/
- - - - I
- - _ l _
- -
Figure 20b: Wialking speed of Traction Group 3
170
Generahy, the fitted straight lines with their respective negative slopes show that
the speed ANSREED of the pair of oxen was decreasing with higher required pull force.
The higher the pull force PF, the lower the speed level. This sceéd reached the value of
zero at the maximum pull FF,,.
- Traction Group 1
The regression equation is:
PF
ANSPEED =1.40-2.53*(Lw)
(48)
The fitted straight line to mode1 the speed is highly significant as the correlation
coefficient R = 0.94 (with R2= 89.2%) is very close to 1 (Table 35 and 36). The
relationship between the pull force PF and the speed ANSPEED
is ver-y strong and in opposite direction (Figure lgb). The straight line accounts for 89%
of the variation. One observation at PF = 1618.7 seems to be high and introduces part of
the total variation. The high value of the speed at this point gives the impression that the
pair of oxen could have develop more power than actually delivered. The unwillingness
facxor induced by the amount of pulling seems to be the only reason they did not.
,
171
Table 35: Parametet; estimates of walking speed (Group 1)
.
--<
Predictor
Coef
Stdev
t-r&
P
Constant
1.39930
0.06719
20.82
0.000
(PF/Lw)
-2.5326
0.2450
-10.34
0.000
--
s= 0.1326
- R-sq == 89.2%
R-sq(adj) = 88.3%
--
--
Table 36: Analysis of Variante of walking speed (Group 1)
--
-
-
-
SOURCE
DF
SS
MS
F
P
Regression
1
1.8795
1.8795
106.86
0.000
Errer
1 3
0.2287
0.0176
Total
14
2.1082
- Traction Group 2
The regression equation is:
PF
ANSPEELI =1..53-2.56*(~w)
(49)
The parameter estimates and the analysis of variante shows that the iit is good
with a negative slope and a high correlation coefficient of R = 0.94 (with R2= 88.5%)
(Table 37 and 38). The mode1 explains more than 88% of the variation.
172
Tqble 37: Parameter’estimates of walking speed (Gronp 2)
-
Predictor
Coef
Stdev
t-ratio
P
. .
-
Constant
1.52727
0.09084
16.81
0.000
-2.5580
-
_--
s=O.1421
R-sq = 88.5%
R-sq(adj) = 87.4%
Table 38: Analysis of Variante of walking speed (Group 2)
-
---
SOURCE
DF
SS
MS
F
P
-
Regression
1
1.7040
1.7040
84.40 -
0.000
Error
11
0.222 1
0.0202
12
The low speed ANSPEED = 0.29 m/sec recorded at PF = 2412.1 N was :not
singled out by the analysis even though it introduces part of the variation. This low speed
is also due to the unwillingness to pull under heavy load regardless of the quality and level
of training (Figure 19b).
- Traction Group 3
The regression equation is:
P F
AAWEED = 1.15 - 2.95 * (---$
(50)
I
l
173
1
Table 39: Pariameter .estimates of walking speed (Group 3)
Predictor
Coef
Stdev
t-ratio
P
:* ..’
. .
Constant
1.44785
0.04920
29.43
0.000
(PF5W)
2.9540
0.1862
15.86
0.000
s -0.09128
R-sq = 96.5%
R-sq(adj) = 96.2%
_----
-
Table 40: Analysis of Variante of walking speed (Group 3)
---
SOURCE
DE
SS
MS
F
P -
Rekession
1
2.0969
2.0969
251.69
0.000 -
En-or
9
0.0750
0.0083
Total
10
2.1719
The best fit line for the Traction Group 3 is shown in Figure 24. The parameter
estimates and the analysis of variante (Table 39 and 40) shows a very good cor-relation R
= 0.98 (with R2= 96.5%) to translate the close relationship between the amount of pull
required a.nd the speed. There is still a certain amount of unexplained variation related to
the unwillingness factor under heavy load, which leads to more energy losses.
6..3. Draft animal liveweight mode1
The liveweight LW in kg of draft animais is a required input in a11 models
developed to this point. It is important to be able to evaluate this parameter before using
any of these models. The simpler tbe way to do this the better. The liveweight of draft
1 7 4
animals is generaily u’nknown ta the majority of farmers. During this 6eld study none of
,
them was able to give the liveweight of their draft animals. Also, weighing scales are not
-* ,
generally available to farmers to help them determine the weight of their animals.
In order to develop the mode1 at this stage of the research, more draft animais
were included in the sample to bave enough point to generate a better fit. The extra draft
animais used belonged to the farmers in the 4 villages visited during the maximum pull
trials. The simplicity and speed of the measurement technique allowed this increase in
number of oxen, from 18 (original sample) to 73 cases. The technique consiisted of
measuring the circurnference at the point of heart (in cm) of individual animal and
determining the liveweight LW with an electronic weighing scale (Bar10 Electronic ScaleL:
Mode1 2 100).
The regression analysis performed with the first 18 draft oxen gave ai coe&ient of
determination of R2= 0.87 which is an acceptable level of significance. The cor-relation
coefficient between liveweight LW in kg and circumference CIRCLJMF in cm was R =
0.93 to demonstrate a close relationship between the two variables (Figure 21a)
The regression equation is:
L W = -367.80 + 4.78 * (CIRCUMF)
(51)
The parameter estimates and analysis of variante are presented in th,e Appendices.
Only, the complete statistical analysis of the broadened base sample (73 oxen) is given.
-
1 7 5
The corresponding regression equation is:
.:-
.
.
L W == -515 + 5. II * (C’IRCUMF)
(52)
T:he broadened base sarnple gave a better cor-relation R= 0.97 and expiained more
of the variation observed. The coeffkient of determination R’ = 0.94 shows a better fit
(Table 41). Most of the variation, is due to the errors
Table 41: Parameter estimates for liveweight vs circumferenee
Predictor
Chef
Stdev
t-ratio
P
-
Co&tant
515.13
23.88
-21.57
0.000
CC (cm)
5.10
0.15
33.41
0.000
-
--
--
-
-
s = 14.97
R-sq = 94.0%
R-sq(adj) = 93.9%
Table 42: Analysis of Variante (LW, CC)
-
SOLkCE
-
-
-
DF
SS
MS
F
P
-
Regression
- - 1
250037
25003 7
1116.48
0.000
Error
71
15901
224
Total
7 2
265938
1 7 6
400
350
2
i300
5
g 250
z
.C
4
200
150
135
140
145
150
155
160
165
170
li’75
Circumference (cm)
- Fitted Line
l
i
I
1
Figure 21a: LW= f(CIRCUM) regression Iine (18 oxen)
/
450
I
I
400
!j- 250
200
150
130
140
150
160
170
180
1190
Circumference (cm)
- Fitted Line
Figure 21b: General regression line (73 oxe.n)
177
committed when performing the measurements of the thorax circumference (Table 42).
Draft animais do net appreciate being touched by outsi.ders. The measurem.ent had to be
made with the collaboration of the farmers. The best solution was to put them under a
yoke before making the measurement. The coefficient of correlation R= 0.97
Sdemonstrates the positive close relationship between the two variables.
Fiigure 21 b shows that a couple of points are unusual observations. They deviate
fiom the line and present a fairly large standard residuals. Those points are represented at
location (CC LW) = (140, 238) (157, 255), (159, 264), (155, 306), (166, 364). Two
other observations need to be pointed out at (CC, LW) = (188, 442), (189,445) as they
are located at the extreme end of fïtted regression line. They have a real influence on the
line even with their small standard residuals.
The fitted regression line works well in predictirrg liveweight of humpless Ndama
breeds. Caution must be taken when performing the measurements. It must be donc as
early as possible, in morning hours. It is better to do the measurement before the draft
animais glo for grazing, to avoid introducing an extra source of variation. The more feed
they take before the measurements, the more the variation is introduced in the model.
The liveweight LW mode1 developed in this study needs to be validated for :specific
applications. The mode1 Will be very useful in future calculations of draft animal energy
balance ancl feeding system evaluation.
6.4. Management of draft animals
According to their production objectives, the farmers’ systems management of
draft animais was mainiy oriented towards minimizing risk and securing cash trop
,” I
178
production. Farmempointed out that thé strategy of using draft oxen on different crops
depended greatly on the timeliness of the rainy season onset. Two main allocation patterns
.
./.
. . ,
are used:
- If the rainy season is on time: The. cereals (maize and millet) will receive
a ,
:
the draR animais first. Afterwards, they are allocated to the cash trop (groundnuts) for the
time remaining. Their utilization on sorghum and rice fields is occasional depending on the
pattem of rain distribution and on the number of working days.
- If the rainy season is late: The cash crops (groundnuts) are given priority
in using the draft animals. The remaining crops (cereals) are usually manually cultivated.
,*
,
In general, the amount of area for cereal production is fixed compared to the cash
i
,* 1
crops. The total cereal production is mainly used for the farmers’ own consumption. On
,..
the other hand, the amount of land allocated to the cash crops depends on tlhle size of the
i
farm in hectares. In agro-system 4, for instance, the groundnut field cari occupy 65% to
70% of the farm (Equipe Systemes, 1983; 1986; Fall, 1986)
I.
The management of the draft animals during the rainy season deals mainly with the
following questions: What type of field work cari be performed?; How many days per
I . .
t
I
week cari the draft animais be used?; What type of feeding system is used?; How are the
problems that affect the animal’s health resolved?
l
The type of management has an impact on the liveweight and the working capacity
,.
ofthe oxen during the rainy season. The working schedule of the 18 draft oxen was
179
,
monitoredi throughout the rainy season on a daily basis t.o better understand the type of
management practices.
‘- .
6.4.1. Field work
During the rainy season (June to October), the draft animals were mainly used for
three maijor field operations (tillage, seeding and weeding) bksides transport. For the rest
of the season (November), the draft animals were idling or used occasionally for transport
They were not used for harvesting as groundnut diggers were almost non existent at the
fiirm level.
Table 43: Repartition (%) of draft oxen time
Field Operations
June- July
August
September
October
-
-
-
Tillage
83.33
66.72
22.22
0.00
Seeding
38.95
lP.13
0.00
0.00
Weeding
5.56
33.33
0.00
0.00
Harvesting
0.00
0.00
0.00
0.00
Not usled
0.00
22.22
72.22
77.78
-
-
-
Transport
27.78
22.22
5.56
22.22
It appeared that most of draft animals working time was used for tillage activities,
especially in the months of June amd July where 83.3% of the oxen were used for seedbed
preparation purposes (Table 43). Alter plowing, seeding (38.9%) was the second most
important field operation carried out by farmers. The concomitante of these two first field
1 8 0
II
activities was such. that fat-mers with more than one pair of oxen were better off. In
contrast, weeding was not as mechanized (5.5%) in favor of transport (27.8%).
‘. .
During the month of August, more weeding took place with 22.2% but plowing
remained the most important activity. Alter the cash trop, sorghum and rice fields were
the crops to receive the oxen. Starting the month of September, most of the draft animals
stopped being used for trop production as 72.2% were just idling, 22.2% used for tillage
.,*
on other secondary crops and 5.6% for transport. When the growing season reached the
month of October, more oxen were idling (77.8%) and the rest were used for transport
6.4.2. Daily working hours
The l&el of intensity of draft oxen utilization was very high at the beginning of the
ramy season, during the months of June and July. Farmers were working the draft animais
for an average of 7 hours per day with a minimum of 5 hours and a maximum of 9 hours.
This number of hours included the time spent to travel to and from the field location and
the time spent yoking the animais and hitching the farm implements. This working
intensity was repeated 6.47 days per week with little rest, 0.53 day per week. (1 day
maximum) (Table 44).
An average field time of 6 hours per day were used by farmers to perform field
operations. Field time was adjusted according to corresponding field operation efficiency
E in percent (plowing, seeding, weeding). A little time was left for other activities like
transport.
V:i-.--
--
--W
--
-.-‘--
--
-.
__-_
__
-_,_
.
.-
-...
-
--
-
Table 44: Management and Handling of Draft Animalr @ the farm leve!
Days restedIweek_June-July
Days rested/week-August
Grazing hrs/day-June-July
Graz@ hrs/day-August
Grazing hrs/day-September
F. Days with Diseases
7.00
7.08
5.63
1.45
0.00
2.00
1 8 2
6.4.3. Feeding system
TWO types of feeding systems were identified: one corresponding to the rainy .
._ ,
season and the other, to the dry season. The main focus was on the growing season’s
feeding system.
According ta farmers, the feeding system of the oxen during the dry season was
mainly extensive grazing, along with the whole farm or village herds. A complement of
feed based on trop residues, especially groundnut hays was given late in the afternoon.
During the rainy season, draft oxen were kept separate from the herd at tlhe farm
level. TO complement the natural grazing, during the breaks oxen were given a daily ration
of cereal grain (maize, millet) or cereal by-products (stover). Farmers across agro-
ecosystems 4 and 5, used the same strategies as the resources available for fized were the
same: 2.5 kg of corn per animal in June-July and part of August, and 2 kg of millet for the
remaining of the month of August. The amount of complementary feed decreased along
I
with the decrease in amount of work to be performed towards the end of the rainy season,
1
I
as natural grasses were more abundant. The natural grazing played a key rolle in the
.r.
I
feeding system of working animals. The grazing time passed from 5.58 heurs per day at
th’e beginning of the growing season to more than 9 hours per day towards the end. The
most important grass species found for grazing are given in Table 45.
In addition to the grass species, a number of tree leaves were also used to
complement the grass:
” I
- Vene (pterocarpus eraniceus)
- Bu pumbapumb (calotropis procera Ait)
-
1 8 3
fable 45: Pasture’grass species in zone 4 and’5
:=
Gass local names
Corresponding Scientificlnames
.-
Elimpey djikel
cyperus rotundus
Edjilangaye
panicum maximum
Eboussaye
digitaria cilia.ris
Enoname
bracharia plantaginea
Elbongaye
chloris pilosa
Kohing kohing
desmodium hirturn
Essabay
urena lobota
Ebucay
bracharia pallide-lùsca
Essoulangay
pennissetum pedicellatum
Baarafita
Oryza species
Essiteye
Acanthospermum hispidum
33
=-a
In general, the quality of feed given to draft animals is ranked according to the
energy content expressed in forage unit (FU). The total forage unit FU is converted into
energy b:y the draft animals digestive system, based on the t.otal digestible nutrient (TDN)
c,ontent of*the feed. Pasture grasses are generally considered as poor feed. Based on the
work carried out in Senegal by CEEMAT (fiench research institution), grass tut has only
0.14 FU whereas feed used for complement have a higher FU: groundnut hay = 0.35,
maize grain = 1.10, for example.
In ,the energy evaluation, the number of FTJ used to provide the energy required for
the maintenance of the draft a.nimal were used to evaluate the energy balance in relation to
the intensity of the task to be perfbrmed (Equation 11). The lack of the energy required
to petiorm the task Will impact growth, health and liveweight LW of the wolrjking animais.
T:he health problem is translated into the loss of resistance to sleeping sickness.
1 8 4
6.4,4. Liveweight changes and’ body condition
In order to evaluate the impact of the intensity of work on the draft animal body
‘- ,
and health, the liveweight of each draft animal of the 18-base sarnple was monitored by
”
:
means of weighing with an electronic scale (Bar10 Electronic ScaleZ Mode1 2100). This
measurement was made at the end of every month throughout the growing season.
The measurements performed on the pairs of draft oxen are shown in Tables 46
and 47.
6.4.4.1. Body weight losses (LWL in kg/day)
The analysis of the variation of the liveweight shows that both individual ox and
pairs of oxen across the traction groups lost weight during the period of intense work in
June-July. The recovery period started at the end of August and correspondled to the
decrease in field work demand. The coefficient of variation CV ranged from 2% to 9% for
C?oup 1,2% to 6% for C+roup 2 and 1% to 3% for Group 3 (Figure 22a amd 22b).
The average liveweight variation LWV in kg and in percent per Traction Ciroup
was calculated and summarized in Table 48.
From June to July, each pair of oxen went through high level of utilization. The
amount of body weight lost during the process was converted into energy (15 MJ/kg of
LWL given in the literature) towards the accomplishment of the job. This means that the
execution of most field operations required more energy than available from the feeding
system. Thés effect of work on body weight has been reported in many publications but
needs to be evaluated for different species.
.
,-: II
_-
1 8 5
,-
Table 46: Live&igbt (kg) variation of pair of oxen
Month (
Gl
G 2
._._, <
G3
"
Il
l Pl P2 P3
Pl P2
P 3
Pl P2 P3ll
June ' 683
690
648
535
600
517
469
435
441
July
699
673
583
522
544
499
438
427
419
Aug.
713
755
605
567
571
523
423
441
461
Sept.
717
742
635
580
600
537
472
456
487
oct.
718
753
668
598
605
538
486
466
505
Table 47: Avcerage Liveweight variation of Traction Group
Month 1
431
G 2
G 3
Il
AvIs
std
Avg
std
Avg
st:d
(kg)
(kg)
0%)
-
(kg)
(kg)
---
Ju ne
673.67
18.37
550.66
37.96
448.33
14.82
July
651.67
49.70
521.67
18.37
428.00
7.79
Aug
691.00
63.18
553.67
21.75
441.67
15.52
Sept.
698.00
45.70
572.33
26.28
471.67
12.66
oct.
713.33
34.88
580.33
30.07
485.67
15.92
Table 48: Liveweight losses and gained (June to 0ctober)l
Period
I
Gl
G 2
G 3
I Lw-v
%
LWV
Oh
LWV
%
0%)
--
(k%)
(kg)
i
-
--
June-July
-22.00
3.27
-28.99
5.56
-20.33
4.53
i
July-Aug.
39.33
6.03
32.00
6.13
13.67
3.19
Aug.-Sept.
7.00
1.01
18.66
3.37
30.00
6.79
Sept.-Oct.
15.00
2.15
8.00
1.40
14.00
2.97
186
500
400
300
200
100
0
Figure 22a: Liveweight Variation (LWV) of individual ox
l
1
I
4
/
j
!
!
j
4 0 0
- -
J
i
A
S
0
Months
-A-- Group 1 -C Group 2 -+ Group 3
L__~~~
--
--_-
--.----
.
Figure 22b: Liveweight Variation (LWV) of Traction iGroups
“. B
l
I
. ,
1
:
1 8 7
The average number of days worked during this period was’19.47 with a CV=
44% (Table 44). Based on this number of worked days, the liveweight Iosses LWL in
. . .- ,. .’
kg/day per traction group were:
- Group 1
LWL= 1.13 kg/day
- Group 2
LWL= 1.47 kg/day-
- Group 3
LWL= 1.04 kg/day
In accord with a nurnber of cases observed throughout the world, animal traction
researchers agree on the fact that even a good feeding :system would not prevent draft
animals from losing weight during this peak working time of the season. The body weight
losses converted into energy
:must be taken into account when calculating the daily energy balance of draft animais
6.4.4.2. HeaI,th and tare
I:n the majority of cases, farmers benefited from ,the veterinary services from the
local governmental extension agency. Vaccines were given twice a year to protect oxen
from losing tolerance to sleeping sickness and treatments were used against other different
parasites Cases of diarrhea have been mentioned by farmers in relation to thle changes of
tlhe feeding system from the dIry season to newly growing grasses after the first rains. This
never stopped fat-mers fiam using the animals as cases were not very serious and were
treated quickly.
188
,
6.5. S u m m a r y j,
This study showed with detailed calculations that the d@rence in ,optimal
performance an-mg the weight classes of draft Ndama cattle L&e not ver-y significant. A
further analysis would have showed that the coefficients or parameters estimated for the
diff+erent mode1 developed were aiso not very significant. It means that one equation cari
be used to mode1 the working output of Ndama cattle in the Basse Casamance region
environment. The level of significance cari be used, if needed, to differentiate specific
behavior between traction groups in order to make marginal improvements Ion the energy
balance calculations. This cari be done with consideration given to the cost and amount of
human resources available to conduct the study. The nature of the problems may play a
role in keeping the three groups separate. The difference in farmers’ circumstances
required to satisijr both animals needs (feeding and tare) and production objectives may
justify the treatment of each pair of oxen as belonging to a traction group that bas specifïc
needs in relation to their body weight.
The Ndama is a well-adapted species to the Basse Casamance region and needs to
be properly managed in order to satisfy the level of energy required. It is important to
monitor the feeding system and the amount of work demand. Draft animals need at least 1
day’off per week for better efficiency. Body weight losses LWL per working day are
expected from working draft animals but the amount in kg/day cari be kept to a minimum
through good management. Farmers need to develop expertise in fitting the job to the
animais by being more aware of their maximum working capacity. This objec.tive c.an only
be achieved if they know their animals well. One major improvement would be for a
farmer to follow up on the liveweight of his draft animals throughout their working career.
“_ ._ _ _ ___ - ._ . . .
7
I
189
The means of doing this follow up using weighing scales is not always possible in the local
areas, however other means cari be used. The mode1 developed.iq this study is significant
. .
.
and cari be used by any farmer equipped with a simple measuting tape.
In addition to the liveweight, it is important for farmers to develop strategies and
set up priorities in the execution of the field operations in relation to the specific ener+q
requirements. This cari be achieved by planning activities during the working days and in
selecting the appropriate implements to carry out the field operation.
-
‘.
.
‘.
.
Chapter 7
FIELD OPERATIONS AND DRAFT REQUIREMENTS
The evaluation of the pull force PF in daN and draft DR in daN was conducted on
station to better control the variables involved. The technique of land prepar,ation used
was the same as fat-mers’ and the size of experimental plot was in the range of field plot
unit delimited by farmers for plowing.
7,,1. Field operations characteristics
7.1.1. Soi1 characteristics
The soi1 type is the primary factor in the process of evaluating the draft required to
move different farm implements through the soil. The physical and mechanical properties
of the soi1 are the most important parameters involved in its behavior at different water
content levels (SWC). These two characteristics were used in selecting the sites for the
pull force tria1 and draft measurements. The physical properties were: soi1 texture, bulk
density, and soi1 moisture. The soi1 resistance to cane penetrometer was also used to
I
rninimize variability among plots.
A field of 2 ha was cleaned of natural vegetation composed of grasses (Pennisetum
Polvstachvon, Eragrotris tremula) and small trees. The area was characterized for
l
;
*.i I
1 9 0
-1
192
” Table 49: Average soi1 texture of the tria1 site
1 Soi1 depth (cm)
16 - 58 cm
% Clay < 2 um
7.6
116.7
?? ? ?
% Silt 2 - 20 prn
3.5
8.0
i
% Very fine sand 20 - 50 um
6.2
6.5
% Fine sand 50 ?? 200 urn
47.9
38.3
% Sand 200 - 2000 um
34.8
30.5
% Organic matter
0.9
0.8
Source: Extract fiom Niane A., 1984
In the American soi1 classification, the very fine sand is called coarse silt, combi.ned with
the silt to represent 9.7% and 7.6% for the clay. In the USDA textural tria.ngle, the type of
soi1 corresponds to Sand. The high soi1 groundmass which is the ratio between coarse to
,,.
fine particles shows that the soi1 matrix tends toward low soi1 water holding capacity and
hîgh permeability (7 crnihr).
The clay minera1 (1: 1) is kaolinite type and is mainly responsible of the s’oil red
color. The presence of kaolinite and Fe sesquioxides act as bridges between adjacent
particles and play the role of cernent in the soi1 aggregation processes. The level of organic
/
matter is too low (0.9%) to provide enough cernent to the particles. These linkages are
weak in general and confer a massive character to the soi1 structure that becomes hard
when dry and very malleable when wet. The wetting-drying cycles in relation with the
/
‘1 ,
tiequency of dry spells during the rainy season play an important role in the dete:rmination
<”
of the number of working days for different fïeld operations (FDOP).
I
193
IBeside the s&l texture, the soi1 bulk density SBD in g/cm3 of the tria1 site was also
determined with the aid of a tore sampler. Several sets of three undisturbed cores were
-:
‘. .
sampled fiom the field in the first 10 cm-horizon, weighed, oven dried and weighed again.
On a d:qr weight basis the a,verage soi1 bulk density SBD in g/cm’ was calculated
SBD= 1.40 gkm” std dev. = 0.01, CV= 1%
The average value of SBD found by Montorio (199 1) in the Djguinoum watershed
(agro-ecosystem zone 4) was 1.45 g/cm3. He used in his investigation a gamma-ray
attenuation densitometry (CAMPBELL type CPN 507-l. 5)
7.1.1.2. Water hoiding capacity
The soi1 water holding capacity SWHC in % g/g was determined at the
ISRA/Djibelor station soi1 laboratory. The values obtained from this experiment were used
only for indication purposes. The real soi1 water regime must be evaluated in-situ for
better results. For the laboratory determination, three weighed undisturbed soi1 tore
samples were taken from the tria1 sites in the first 10 cm-horizon and broug;ht to saturation
by capillarity. Alter a11 the initial weights were measure:d, the soi1 samples were lefi to
drain for a number of days. ‘The ,weight of each tore sample was determined on a regular
basis (every hour, every two-heurs, every day) until the sa.mples were completely air-
dried. The recorded data we:re used to calculate the soi1 gravimetric water content SWC in
% g/g. Ti’ne SWC was plotted against number of days (NDAYS) to have a drainage curve
(Figure 23 and 24).
194
I
-- ..-~
-.--.
20
.a.
.
5
/
/
4
t
1
I
!
I
I
/ +..++-~
/
I
8
I
1
1
0
1
2
3
4
5
6
7
8
9 10 11 12 13 14 15 16 17
Days
’ -I- Core 1
+ Core 2
-c core3
- Average :1
L-
-
_-.---~..-
Figure 23: Drained soi1 water content (undisturbed cores)
100 i
- -
-~-
t
I
<*
3r
-si
-3
= 1
E 10 I ..-.-.
_.
,... .
..I
w m
!
?
1
ii
i
?
?
m
I
?
i
i
?
1
/
?
/
L-
-+- - ..+A
--.--.- /I
0
5
10
15
20
Time (day)
Figure 24: Log-SWC vs time
1 9 5
The analysiç ofthe graphs shows that the 1/2 ha plot was uniform as regards to the
drainage profile and soi1 water content SWC, One soi1 tore showed a preferential flow at
.
the beginning of the experiment but quickly followed the two other cores drainage pattem.
The SWC in % g/g at characteristic points were calculated:
@ Saturation (MT)
: SWCsal= 23.69% (CV= 5.28%)
@ Upper Limit (DUL) : SWCDCL= 20.00% (CV= 4.08%)
@ Wilting point (L-L) : SWCLL= 6.00% (Charreau, 1974)
@ Air-dry (AD)
: SW&-,= 3.00?4 (CV= 13.60%)
T’he DUL in % g/g corresponding to a point close to the field capacity SFC in %
g/g was reached about 1. day after saturation while the air-dry moisture level under
ambient air temperature was’ reached afier 14 days. As defined by Ratliff et a1 (1983) the
DUL is the highest field-measured soi1 water content after being saturated and allowed to
drain until drainage becames, practically negligible. The soi1 water holding capacity SWJX
is the diiyerence between SWCDUL and air-dry SWC)m.
The SWC at which, fieid operations EDOP are performed is an important factor
towards the evaiuation of the amount of energgy required for the task. Fat-mers are
expected to schedule field operation SO soi1 is worked ait a moisture close to the optimum
soi1 water content SWC,,,.
196
7.1.1.2. Penetrometry
The penetrometry described earlier was the investigation- tool used to evaluate the
mechanical properties of the tria1 site. The dynamic method was’ukd whereby a known
weight falling from a predetermined height hammered the 50 cm long cane penetrometer
rod into the soil. The depth of cane penetration was measured and the associated energy
calculated using Equation 13.
Three locations were chosen at random along one diagonal of the 112 ha plot
Along with the soi1 resistance to penetration measurement, the SWC profrle of the spot
was also determined at the same time.
The graphs of the Cone energy in J against the depth in cm and the SWC profile
” f
helped describe the soi1 resistance to penetration at the given SWC levels (Figure 25 and
I
26). It appeared that the 1/2 ha plot was uniform enough to minimize the amount of
variation expected in the pull force PF and draft requirement DR trials.
I,.
7.1.2. Field operations
In relation to the monitoring of the utilization of draft animals (previous ch,apter), a
number of field operations performed by farmers to meet their production objectives were
characterized. Among those, plowing was the most important followed by seeding.
7.1.2,1. Tillage
Two types of implements were used for seedbed preparation: moldboard and
ridger plows. In the Basse Casamance region, plowing was recommended as a met.hod to
control and slow down weed pressure at the beginning of the rainy season.
197
--
--
1 0 0
2 0 0
300
400
5 0 0
600
Cone Energy (J)
e-Site1 +Site2 + S i t e 3
-
-
Figure 25: Penetrometry energy
0
--m
,------ - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
-5
- \\
Ê-10 't -------
_"----------_--_--_-----------------
0
c
k-15
- - - - - - -
---_-_-__-__---_-----------------
8
a3-20 ------,-- ----_-_-__------.- - - - - - - - - - - - - - - .---
!
:I
-25
-- -.---.--
t
ila--_-___ I -__-.--_-- *.--------------. I--
-30 jv
+
-
-
t
-
v
-
t
-
-
-
t
-
-
-
t
-
0
5
20
25
-
I
?? Location 1 V Location 2 + Location 3
.
Figure 26: SWC at penetrometry sites
l
1 9 8
“‘s<
- Flat land preparation
The most common method of flat land tillage performed !y farmers witih the
*.
‘_ .
moldboard plow (IJCF 10”) is to divide the field into small plot units of 0.6 ha in average
size and to plow each plot until the total field is done. The technique also known as the
Fallenberg method consists of starting at one corner of the delimited plot and of plowing
around the perimeter with the pair of oxen. In the process, the fùrrows are tumed outward
1
i
on one side ofthe delimited plot and adjoined side by side until the tenter of the plot is
reached. The majority of the farmers prefer this technique because it is faster than other
techniques like plowing in lands or by starting from the middIe of the plot.
The average plowing depth in farmers’ field was 1 O- 12 cm and the width 23 cm.
These two parameters depended greatly on the pulling capacity of the pair of oxen and on
the level of expertise of the farmers to properly adjust the farm implements.
- Ridging techniques
The second most common technique of land preparation is ridging. The data
analysis shows that farmers used two techniques of ridging. The first technique was
performed with the ridger equipment attached to the ARAR4 toolbar or with the EMCOT
ridger imported from Gambia (Figure 4 and 5). The technique of ridging is similar to the
plowing in lands technique with the difference that equi-distant ridges parallel to ‘one side
of the plot are made.
The second ridging technique was executed by means of UCF 10” moldboard
plow. In this case, the making of a good standing ridge required two furrows placed back
to back to create a crest of soi1 narrow enough to be planted with a single row trop. This
199
,,
/
method seems to be‘too slow to farmers who are concerned with timeliness. The common
practice Twith a moldboard plow is to make one-way ridges instead. of a back to back .’
:
ridges. This technique is faster as only half of the total field is plowed. With this practice,
the ridge quality is very poor as ,the height subsides faster with rainfall and water erosion.
The characteristics of the ridges made were highly variable depending on the piece
of implernent used and also on the ability of the farmers to properly use the pair of oxen.
The average depth and ridge height, and between ridge distance were 12 cm, 15 to 18 cm
and 50 to 60 cm respectively.
7.1.2.2. Seeding
The seeding equipment that farmers used was the Super ECO one row seeder
!
((Figure 7). The seeder was designed for flat land preparation. This represents a major
hmitation in its utilization. Al1 the fields that have been ridged were manually seeded.
It appears that seeding was not felt as a major constraint as the surplus of labor
generated by the mechanization of the land preparation was mainly used to hand seed
plowed areas. During the rush pe.riod in the months of June and July, only 38.7% of the
pairs of oxen were used for mechanical seeding a,nd it shows the importance of hand
seeding. hJost of the seeding was performed on the cash trop fields (groundnuts) which
occupied more than 50% of a11 cropped land. Al1 the Super ECO seeders were equipped
with at least one groundnut seed distribution plate and 59.9% were equipped with
groundnut seed plates only. Afler groundnuts, maize was the second most mechanized
trop with 3 1.82% ofthe seeders equipped with maize seed plates, followed by sorghum
I
( 18.18% of the seeders) and millet (13.64%).
f
IIl
200
7.1.2.3. Weeding
Mechanical weeding was not found to be as important.as.plowing and seeding. The
‘_
.
cultivator SINE 9 was the only implement used by farmers for weeding purposes. Al1 the
SINE 9 toolbars were equipped with 3-Canadian fYi.tll sweep tines. For better efficiency, the
type of sweep used to equip the cultivator must be able to stir the soi1 surface at a certain
depth to destroy Young weeds in favor of trop growth. The average depth in farmers
conditions was around 8 cm. The weeding period generally statts about 20 days alter
seeding (DAS) and constitutes a peak activity for the available farm labor. The extension
of cropped areas through mechanized plowing and seeding has created a labor bottleneck’s
shifiing situation that needs to be properly managed. The analysis shows that the weeders
owned by farmers were mainly used on part of the cash trop fields (groundnuts) while the
rest of the crops were manually weeded. The level of utilization of the weeders was not
enough to complete the field operation on a11 crops. The low level of utilization was
mainly due to a lack of expertise and to a fear of damaging the crops at advanced growing
stage when the weeding operation was not scheduled on time.
7.1.2.4. Harvesting
Farmers were not equipped with animal-drawn harvesting equipment. Harvesting
of’all crops is still manually performed. The only available animal-drawn harvester is the
groundnut digger built by SISMAR. The timeliness of the field operation in ter-m-; of
satisfling the labor demand is not critical according to the farmers’ point of ,view. Farm
labor is generally available in numbers suf?icient to carry out the field operations by hand
in time for upland crops. Only rice harvesting is felt to be a problem.
201
7.13. Soi1 water balarrce and Working days
Since 1960, the Sahel is globally experiencing an overa’l $ying of the climat?.
.
Following the two major droughts in 1970s and 1980s farmers’ behavior in resources
allocation have shown that the year-to-year variation in total amount of precipitation is an
important and determinant factor in planning field activities. Under these conditions over
the years, fat-mers have developed a risk aversion attitude towards investment.
The weather in the area ‘of study is mainly characterized by the seasonality of the
precipitation. During the wet season (June to October), rainfall does not occur every day
making daily rainfall highly variable,
The intra-annual variation of precipitation dictates in general the farmer’s strategy
to decide on the types of crops to grow and the corresponding field operation to perform
in relation to the types of implement available at the farm level. The analysis of daily
precipitation amount is important to soi1 water balance calculation to heip predict runoff
and drainage and to help eva.luate the field trafficability and determines the suitabihty for
performing daily field operations.
The most important factor in determining the number of working d?ys (NWDAYS
in days) is the soi1 moisture regime. This factor significantly influences the timeliness of
performing field operations. The soi1 water content is generally influenced by highl:y
stochastieal weather variables such as rainfall. relative humidity, temperature, wind speed.
solar radjia.tion, etc. The main effect of the weather variables is to associate a certain level
of uncertainty in the process of scheduling farm operations
The method applied in this analysis is based on the approach using probability
levels The probability level represents a degree of certainty. Two probability levels are
b-
202
considered 90% (9 years out of 10) and’ 50% (5 years out of ten). A database of 37 years
(1960 to 1996) of daily rainfall for the region of Basse Casamanqe. was used (Table 50).
The data was pulled from the Ziguinchor Weather Station databàse which is part of the
National Weather Stations Network. The town of Ziguinchor is 50 km away from the on-
farm research sites. The Gumbel(l941) approach is used (Equation 17 and 18)
Table 50: Annual rainfall (mm)/rank for Ziguinchor (1960-1996)
--_
Y ear
Amount
Rank
Year Amount Rank
Year
Amount:
Rank
(nm
(mn-9
w-w
1960
1274.60
1 6
1972
9 5 1 . 8 0
3 1
1985
1379.90
10
1961
1549.30
5
1973
1289.40
1 4
1986
975.20
2 9
1962
1567.50
4
1974
1240.40
1 7
1987
1042.6~0
28
1963
1429.40
9
1975
921.90
32
1988
13 10.60
12
X964
1222.80
2 0
1976
1297,lO
1 3
1989
1175.40
2 4
J. 965
1756.60
2
1977
7 9 0 . 3 0
3 6
1990
l’I14.30
2 5
1966
1603.80
3
1978
1512.10
6
1991
1223.00
1 9
1967
2006.50
1
1979
1187.30
23
1992
967.10
3 0
1968
8 8 2 . 5 0
3 4
1980
6 9 8 . 5 0
3 7
1993
1481.70
7
1969
1460.70
8
1981
1221.50
2 1
1994
1204.20
2 2
1970
1282.00
1 5
1982
8 9 7 . 8 0
33
1995
10’95.40
2 7
1971
1098.60
2 6
1983
8 1 7 . 2 0
35
1996
1310.90
11
1984
1236.00
1 8
TO determine the number of working days, the days suitable for field operations
are counted for every month of the rainy season for each of the 37 years. The counting is
manual and complex in relation to the main characteristics of a working day (see AME
203
StandOard D230.4,~Agricultural Management Data, Section 8-Working Days, Timeliness).
In addition to the scenarios introduced by Le Moigne (198 l), the following empirical
._ ,
‘. .
assurnptions for the Bass#e Casamance region were used:
Data collected are norrnally distributed for the long period of 3’7 years.
Onset of the rainy season is “the date afier 1 May when rainfall
accumulated over 3 consecutive days is, at least 20 mm and when no dry
spell within the next 30 days exceeds 7 days” (Sivakumar, 1988).
Plowing and Ridging are difficult to perform afIer a lO-day dry spell.
Seeding and Weeding are difhcult to perform afIer a 2-week dry spell.
The pair of oxen is at rest 1 day per week.
The confidence limits were used to estimate the range for the 90%, and 509’0
probability levels. Assuming a normal distribution (mean annual rainfall = 1229.08 mm and
standard deviation= 274.77 mm), the 9 years out of 10 for Ziguinchor lies withinf the limits
given by the mean plus and minus 1,6440*standard deviation (777.36, 1680.88) mm. The
5 years out of 10 limits are within the mean plus and minus 0,6745*standard deviation
(1043.88, 1414.27) mm.
The number of working days NWDAYS and their associated probabilitier; (PWD
in percent) (TILLWDY for tillage and SWWDY for seeding and weeding) are summarized
in Table 51
i
204
Table 51 Probability (PWD) and number of working days (NWDAY’S)
%Rain
Tilldays
TILLWDY
S@ays
SWWDY
fYr
(days)
e4
. - @vs>
pi)
- - - -
SO% Confidence level
8 . 2 0
12.18
0.41
9 . 1 8
0.3 1
2 6 . 4 4
25.53
0 . 8 2
16.88
0 . 5 4
3 2 . 1 8
2 6 . 1 2
0 . 8 4
15.59
0 . 5 0
2 5 . 1 6
2 6 . 8 2
0 . 8 9
17.76
0 . 5 9
7 . 0 7
18.88
0.61
15.88
0.5 1
90% Confidence levei
June
7.85
11.59
0 . 3 8
8 . 8 2
0 . 2 9
July
25.03
25.85
0.83
17.09
0 . 5 4
A.ugust
3 1 . 0 2
25.41
0 . 8 2
14.88
0 . 4 8
September
2 6 . 7 4
2 6 . 7 6
0 . 8 9
17.03
0.513
Bctober
8.38
19.94
0 . 6 4
16.26
0 . 5 2
The probabilities are monthly averages. TO adjust for the 1-day rest per week. the
PWD must be multiplied by an average coefficient of 0.86 (6 days per week). In the Basse
Casamance region, the rest day is usually on Fridays.
The difference is not very significant between the SO?& and 90% confidence levels
in relation to the rainfall distribution and intensities during the rainy season (Figure 27
and 28).
7.2. Pull force and draft requirernents
The total draft animals’ liveweight used for the tria1 was 683 kg with an average
age of 7 years and 5 years of working experience. The pair of oxen belonged to the
Traction Group 1.
-.
205
.- .
35 _-~--.---_l_ll__l---..
---.-
.
- - -
:
z
30:
.-
\\ '\\
$
:
m
025 -
.
hi. ,_..
R
*
izo*
II
‘\\,...
t:
.g
1
,,,’
‘~,,:
!Xl5 -
.\\
&'
\\
10:,I m
5
---_--_-_l___-.-_l_l--
---
J
J
A
S
0
Months
Figure 27a: Monthly rain distribution (5 years out of 10)
+
-
J
J
A
s
0
Months
,-- Seeding - Weeding
- - -
~.--_-.-..-
Figure 27b: PWD at 50% probability level
c
2 25
8
u
g20
c:
+--
J
À
s
0
Months
~-
--l_.-. -
Figure 28a: Monthly rain distribution (9 years out of 10)
.,
13
-
/+
I
i-‘.
30.6
P
3 0 . 4
!...
/
/
l
!
1.
:
; 0 . 2
!-_
,.I
/
;
!
l
0
i
J
J
A
s
Months
l-e Seeding - Weeding
~-~
---.---
-~ .-
Figure 28b: PWD at 90% probabilit! level
207
‘7..2.1. Statistical design
The 3-way factorial statistical design (2x3~3 Factorial in Randomized Blocks) .was
.A .
.- .
laid out within the 1/2 ha plot already tested for homogeneity in soi1 moisture regime and
mechani.c.al properties.
‘4 certain number of studies bave indicated that choosing the right type of
harnessing system could improve the pulling capacity and .the energy output of draft
animals. The head yokes used by farmers for their pair (of oxen cari be classifïed into two
<groups according to the lenggh: 90 cm and 120 cm (Table 52).
Table 52: Dimension of yokes used by farmers (survey resullts)
.-----
--
- -
-
-
-
Yoke type N
MEAIN
MEDIAN
STDEV
s E ME A\\
-
-
-
-
-
-
- -
-
-
-
-
-
YOKE 1
16
0.9694
0.9500
0.1080
0.0270
YO.KE 2
2 9
1.2069
1.2000
0.1043
0.0199
- - - - - -
- -
- -
-
-
-
Table 53: Two-sample t-test for YOKE 1 vs YOKE 2
- - - -
-
-
-
Yclk:e type
N-
MEAN
STEV
SE MEAN
- - - - - - -
- -
--~--
---.-
YCKE 1 16
0.969
0.108
0.027
YOKE 2 29
1.207
0.104
0.019
--I_--~-
-
-
-
_--
- - - -
95% C.1 for ~YO~I - ~‘You~!: ( -0.305, -0.170)
T-test ~YOKEI = ~YOI.E (VS N-E):
T= -7.15
P=O.OO
DF= 30
--~- .-...---_
-
-
-
-
-
~
I
208
The average dimension of YOKE 1 (96.9 cm with CV= 11.14%) was fcwnd to be
slightly different from the 90 cm-head yoke recommended by Research over the years.
‘_ .
The comparison between the number of the two sizes of yokes (90 cm and 120
cm), using the t-test distribution has showed that the difference is highly significant (Table
53).
The factors and their levels used in the 3-way factorial experimental design are
summarized in Table 54.
Table 54: Factosrs’ levels in a 2x3x3 Randomized Block design
swc
Implement type
w Eh)
-
-
.
6 to 8%
UCF 10” plow
8 to 10%
3-Canadian tines on SINE 9 Tooibar
! 10 to 13%
Ridger equipment on ARARA Toolbar
A total number of 18 treatments were blocked by SWC in % g/g. Each treatment
was performed on a 200 m* size plot. The whole experiment was conducted on 4590 tn’
with alleys between elementary plots included.
7.2.2. Required Pull force
The pull force PF in daN was the observed or dependent variable measured by
means of 500 daN dynamometer (PIABTM).
209
_,
7.2.i.l. Data collected for IJCF 10” plow
The distribution of the data with the 90 cm-yoke at SWC.= 6-8% (avg SWC=
.
7.01% gjg) is slightly skewed to the lefi (Figure 29a) co:mpared to the distribution at
SWC= 8-10% (avg SWC= 9 34% g/g) and SWC= 10..13’% (avg SWC= 11..97?/0 g/g). The
SWC= El- 10% distribution ;presents two peaks, the first at PF= 100 daN and rhe second at
PF= 160 daN (Figure 29b). The first peak (less than 5% of the data) is not as signifïcanr.
The distribution at SWC= 1 O-l 3% seems to be more normal than the first twio (Figure
29c).
With the 120 cm-yoke, the tendency in the data. distribution looks the same. A
slight skxwness to the right at SWC= 6-8% (Figure 3Oa) must be noted. The most normal
distribution was achieved at SWC= 8-lO?/o (Figure 301)) and the most skewed at SWC=
1 O- 13% (Figure ~OC).
The data collected fclr the moldboard plow UCF 10” are summarized in Table 55.
The avera.ge PF values shows that the lowest required PF for the UCF 10” plow was
obtained within the 8- 10% SWC range. The 90 cm-yoke gave slightly higher values, about
5.29 % more than the 120 cm-yoke type. The 90% Confidence Interval for the average PF
required by the UCF 10” moldboard plow is:
PFITFIW* = (146.98, 168.03) daN
The corresponding implem,ent average working width and depth were respectively
IMPLWWD= 23.17 cm and IMPL,WDP= 10.62 cm.
Pull Force (daN)
Figure 29a: PF with UCF 10” at SWC=G?% Yoke= 90 cm
I
4o i
/
90 100 110 120 130 140 150 160 170 180 190 200
P u l l F o r c e (daN)
-_
Figure 29b: PF with UCF 10” at SWC=S-10% Yoke= 90 cm
-<
. /
j
/
/
I
o 30 ;
!
1
/
I
k
1
I
oh
1
5 20 -
s f
E=3y():./
!
,km
1.11,
100 110 120 130 140 150 160 170 180 190 200
PuIl Force (daN)
‘Figure 29c: PF with UCF 10” at SWC=lO-13% Yoke= 90 cm
211
”
130
140
150
160
1’70
MO
190
200
Pull Force (da.N)
‘Figure 30x PF with UCF 10” at SWC=6-8% Yoke= 120 cm
40, - -
- - . - - - - -
/
I
30
;.
.s
à--
0
/
52(-J-
.-.
5
4
e
;
10
+
____
(
/
1
0 '
UJ-LL'
110 120 130 140 150 160 170 180 -
Pull Force (daN)
Figure 3Ob: PF with UCF 10” at SWC=S-10% Yoke= 120 cm
50 -----.-.-.__l--~-..-/
!
i
410
_
/
/
g
I
/
g* 30 I
-'.
5
5 20 --
II
k .
10
+
OL-IZ
100 110 120 130 140 150 160 170 180
Pull Force (daN)
-
-
-
-----.-_-.I_-
.-.
‘Figure ~OC: PF with UCF 10” at SWC=lO-13% Yoke= 120 cm
iI
2 1 2
Table 55: @Il Force PF (daN) with the UCF 10” plow
7.2.2.2. Data for 3-Canadian tines on SINE 9
The distribution of the data collected with the 120 cm-yoke is more nornnal than
the 90 cm-yoke. However, the variation between the graphs is not significant. CMy two of
the six are negatively skewed (Figure 31a and 32~) while the other three are slightly
positively skewed (Figure 31 b, 31c and 32a). The distribution at SWC= 8,- lO?G with the
120 cm-yoke is normal (Figure 32b).
The statistics on the data collected with the SINE 9 cultivator are summarized in
Table 56. It appears from the data table that the lowest PF required correspond to the 8-
10%~SWC range for the 90 cm-yoke type. For the 120 cm-yoke, the difference betw-een
the 8- 1004 and 1 O- 13% SWC range is not significant. Higher SWC has the tendency to
generate more pulling effort as the wet soi1 adhered to the sweeps. The 90% Confidence
Inter-val of the average pulling force, required by the 3-tine SINE 9 across the SM’C (6
13%) is equal to:
PFsIs~~ = (102.84, 135.20) daN
213
._.
30 J
_
i
,
ô
I
/
&
1
-
.
!
8
520;
3
8 .
r= 10 -
n .-ILI--
”
110
120
130
140
150
160
170
180
Pull Force (dz&)
Figure 3111: PF with SINE 9 at SWC=6-8% Yoke= 90 cm
---~--~___----~_-
--_I
40
----.~--------
-
30 - . .
h
<
I
e
3 20 -
z-
/
y”
80 90 100 110 120 130 140 150
Pull Force (daN)
Figure 31b: PF with SINE 9 at SWC=&lO% Yoke= 90 cm
40 __---_--_--------_ .-.- ---^._I_--
3 0 -
ô
o\\
Q
= 20 -
0
ii
-b ’ 1
k 10 I
,IL!
o.mm.
-
100
110
120
130
140
150
Pull Force (daN)
-.------- .---. -.-.-- _.._- ~---.--- -.-- ---~
Figure 3 Ic: PF with SINE 9 at SWC=lO-13% Yoke= 90 cm
214
-~
_-
----__-
30 ---
1
. . .
/
[
T 20 t
0,
7’
+-
l
0
110
120
130
140
150
160
170
Pull Force (dam
Figure 32a: PF with SINE 9 at SWC=&S% Yoke= 120 cm
--
/
70 -
/
J
!
60:
1-m m-
80
90
100
110
120
Pull Force (daN)
Figure 32b: PF with SINE 9 at SWC=8-10%
Yoke= 120 cm
35
30
0
80
90
100
110
120
130
Pull Force (daN)
Figure 32~: PF with SINE 9 at SWC=lO-13% Yoke= 120 cm
_.- i-
215
Table 56: Pull Force PF (daN) with tbe 3-tine SINE 9
-=
-
--
Yoke (cm)
9 0
‘.‘_ .*
120
-
-
-
-
-
SWC-range
16-S%
8-10%
10-13%
6-8%
8-10%
IO-13%
SWC(%gig)
7.01
9.34
11.97
7.01
9.34
11.97
- - -
-
-
-
-
Mean
149.09
111.91
130.45
124.91
100.00
97.76
S.E
1.93
1.97
1.58
2.63
1.31
1.51
Median
150.00
105.00
130.00
120.00
100.00
100.00
Mode
150.00
100.00
140.00
120.00
100.00
HOO.
Std dev.
14.3 1
19.08
14.92
19.52
9.38
13.12
Min
110.00
70.00
100.00
100.00
80.00
60.00
Max
180.00
150.00
200.00
170.00
120.00
130.00
m
-
-
The average implement working width for this PF range was IMPLWWD=- 52.08
cm and the depth. IMPLWDP= 9.60 cm.
7.2.2.3. Data for ridger attached to the ARARA
A similar analysis of the graphs shows that most of observed data dis,tribution are
slightly positively skewed (Figure 33a, 34a and 34b). The skewness coefficients af the
distribution corresponding to the IO- 13% SWC range are positive and greater than 1 (1.18
for 90 cm-yoke and 1.57 for 120 cm-yoke) to witness higher frequency values towards the
low required pull force PF (Figure 33c and 34~). Only one set of data is slightily
negatively skewed with the 90 cm-yoke at 8-10?/0 SWC :range (Figure 33b).
TIhe lowest required PF with the ridger working component attachment was
recorded within the lO-13% SWC range different from the IJCF 10” plow and the 3,tine
SINE 9 (‘Table 57). The 90% Confidence Interval for the average required PF in daN
within the ó-13% SWC range is equal to:
PFARA.RA = (109.57, 135.29) daN
Table 57: Pull Force PF (daN) with the ARARA ridger
The average corresponding implement working width and depth were
IMPLWWD= 22.59 cm and lMPLWDP= 10.44 cm respectively.
?.2.3. Regression analysis
The general linear regression mode1 for this analysis uses different levels of
qualitative independent variables: equipment types, soi1 water content ranges and yokes
type. The following dummy variables were used to describe these levels: SWC2 (1 if SWC
range is 8- 10% and 0 otherwise), SWC3 (1 if SWC range is 10- 13% and 0 otherwise), HS
(1 if the equipment type is SINE 9 and 0 otherwise), ARA (1 if the equipment type is
ARARA with ridger and 0 otherwise) and YK120 (1 if the YOKE is 120 cm and 0
otherwise).
- . ..-_
7
0
100
110
120
130
140
150
160 170
180
Pull Force (daN)
Figure 33a: PF with ARARA at SWC=4-8% Yoke= 90 cm
80
90
100
110
120
130
140
150
,-.-.---.---
.__-
Pull Force (daN)
Figure 33b: PF with ARARA at SWC=E)-10% Yoke= 90 cm
~ ----.---. --_--~.-----_. -.- . ..- - .-.---_--..~
.--,
l
(j(J _----_--_---
,
/
i
100
110
120
130
140
150
Pull Force (daN)
Figure 33~: PF with AURA at SWC=lO-13% Yoke= 90 cm
218
-
-
-- - - - - - -
100
110
120
130
140
150
160
170
180
Pull Force (daN)
-~-
--------_.
Figure 34~1: PF with ARARA at SWC+8% Yoke= 120 cm
100 110 120 130 140 150 160 170 180
-!'ull
Forfia)-. ..__ ._. ~ .._. ~_.
Figure 34b: PF with ARARA at SWGS-10% Yoke= 120 cm
/
” *O go 100 110 120 130 140 150
Pull Force (daN)
I-.--_---_-._.
Figure 34~: PF with ARARA at SWC=lO-13% Yoke= 120 cm
.-.-
‘
.-,rllUY-~---
219
- Full mode1 with interactions
PF = 177 - 23.1* (SWC2) - 17.5 * (SWC3j - 35.5 * (KY) - 38’.2 * (MA) - 9.99 * LX1 20)
- 8 0 * (SWC2KY) -L 16.‘9 * (SWcJI2ARA) - 2.0 * (SU’C3HS) - 11.0 * (SW~:3ARI)
(53)
The regression equation explains a significant part of the total variation (:$=
88.5%) (Table 58 and 59). Most of the non-explained variation cornes fi-om the lack 0.f
normality in the distribution of data collected within some elementary plots. One part of
the variation is also induceId by the variation of the different implements’ working width
(LMPL,WWD) and depth (IMPLWDP). The low values of the skewness coefficients, less
than or equal to 1 in absolute value for 88.89 % of the observed distribution, did not
strongly indicate a. need for data transformation. The assulmption of normal distribution
Table 58: Parameter estimates of PF regression
----~
- - - -
---.-.
Predictor
Coef
Stdec
t-ratio
P
-1-
-
-
-
-
-
Co:nstant
1’77.45
- 8.77
2o.r
0.000
SWC2
,-23.08
11.77
-1.96
0.086
swc3
-17.54
11.77
-1.49
0.174
HS
-3 5.46
11.77
-3.01
0.017
ARA
-38.19
11.77
-3.24
0.012
M(l20
- 9.99
5.55
-1.80
0 109
SWC2HS
- 7.96
16.65
-0 48
0.645
swc2ARA
1.6.89
16.65
1 .Ol
0.340
SWC3HS
- 2.00
16.45
-0.12
0.907
SWC3ARA
-10.99
16.65
-0.66
0.528
s= 11.77
R-sq = 88.5OX,
R-sq(adj) = 75.6?,0
Table 59: Pull force PF Analysis of Variante
--
----<
SCWRCE
D F
S S
M
S
““F
P
-
--~
Regression
9
8561.0
951.2
6 . 8 6
0 006
Errer
8
1108.6
138.6
Total
1 7
9 6 6 9 . 7
-.
- - - -
Variables
D F
SEQ SS
-
_~
s w c 2
1
336.5
swc
1435.8
HS
1708.0
3 9 3 6 . 0
YK120
449.3
SWC2HS
2 9 2 . 9
swc2ARA
3 3 4 . 0
SWC3HS
8.2
swc3ARA
6 0 . 4
holds as the central limit theorem allows when the number of observation n from the
random variable PF in each elementary was large (n= 5 1 to 149).
The probability values in Table 58 of testing the hypotheses that the interaction
coefficients are equal to zero are quite large. This implies to remove the interaction terms
as they are not ver-y important. The same analysis for the yoke suggests its removal from
the model.
- No-interaction mode1
I’F y 173 - 20.1* (iwc2) - 21.9 * (SiK’3) - 38.8 * (HS) - 36.2 * (AM)
(54)
2
2
1
This new Xinear equation simple to use explains 76.7% of the variation and has
slightly higher probability l’evels (Table 60). However., it also has a slightly larger standard
.
errer.
Table 60: .Analysis of Variante (SWC and EQTYPE)
-----
----
-----
SOURCE
1%
SS
MS
F
P
Regression
4
7416.3
1.854.1
10.70
0.000
Ekror
13
2253.4
173.3
Total
17
9669.7
Variables
DF
SEQ SS
WC2
1
336.5
swcri
1
1435.8
HS
1
1708.0
ARA
1
3936.0
The PF averages corresponding to the SWC-range and EQTYPE factors are of
particular interest i:n the evaluation process of the PF required to perform a given field
operation.
‘7.2.4. Draft requirelments
Another approach to the evaluation of the working capacity of draft animals is to
star-t from the draft requkements DR in daN of a given field work (Equation 12 or 31).
The draft is defined as the force needed to move the farm irnplement in the direction of
1:ravel. The conversion factor between the pull force PF :in daN and the required draft DR
222
in daN is the angle of pull y at the implement hitching point, During the trials, the angle of
pull, which gave the best line of traction, was maintained throughout the experiment. The
average y was 22.22 degrees with a CV= 1.15% and a 95% CI off22.09, 22.35) with a
coefficient of traction COS y = 0.9257 (Table 62).
For the type of sandy soi1 in the Basse Casamance region, a more interesting
parameter in relation to the working width (IMPLWWD in cm) and depth (IMPLWDP in
cm) is the draft required per unit cross-section of furrow slice at low speed This
parameter is called the specific soi1 resistance SSR in daN/cm2 (Table 61).
Table 61: Analysis of variante of Specific Soi1 Resistance SSR (daN/cm’)
_ - I _
-
SOURCE
DF
SS
MS
--
---_---
-- swc
2
0.10629
0.05315
EQTYPE
2
1.62410
0.812OC
ERROR
13
0.03464
0.00266
TOTAL
17
1.76503
---
-+
Individual 95% CI
-_--
s w c
Mean
-+ m-d-c ----f---- __-__ f _ec- *--“-i--------
6-8%
0.704
(mm--..-* _-e..m-)
8-10%
0.502
(
*--....--)
m---w
lO-13%
0.526
(”
___- * -wC-er)
0.490
0.560
0.630
0.700
EQTYPE
Mean
-+ _________ + ____--_-- + ------- --+-------
UCF 10”
0.613
( 3 - )
SINE9
0.222
t-*-j
l4RARA
0.957
t-*-J
-+ _________ + -_------ “+ ---- -----+-------
0.200
0.400
0.600
0.800
-..
-
7
223 .
The difference in SSR in daN/cm2 at the higher SWC ranges is not significant, 0.56
daN/cm* for the 8- 10% S’WC versus 0.53 daN/cm2 for the 1 O,lJ% SWC range. The
difference is highly significant between the lower and higher SWC ranges.
The implements show also a highly significant difyerence between their respective
draft requirements per unit area within the whole 6-13% SWC range. In relation to the
characteristics of the field operation performed, the ridger attached to the ARARA toolbar
has the highest draft requirements, followed by the moldboard plow. The sweeps attached
to the SINE 9 toolbar required less draft. A good estimate of the SSR daN/cm2 fi)r
commo:n use when evaluating the draft required DR in dtiN or the required pull fcbrce PF in
daN during an average working day (8-13% SWC range) is:
Moldboard plow: 0.55 daN/cm’
Ridger plow
: 0.89 daN/cm2
Full Sweep
: 0.20 daN/cm2
From this perspective, it Will be important to know ahead of time the implement
working width IMPLWWD in cm and depth IMPLWDP in cm in order to use Equation
14. The value of the pulling angle y Will also be needed. mer calculating the
corresponding traction coefficient COS Y, it Will be possible to determine the required PF in
#daN to be developed by the pair of oxen (Equation 31). Other important field
performance characteristics must also be taken into account in order to achieve a good
evaluation of the draft animals working capacity.
- --. - -- -_ --
_ -
l.l
-
Table 62: Field work cbaracteristics and Field eff~ciencies
ARARA with ridger
120.00
22.50
10.10
480.00
11.33
9 10.00
690.00
75.82
22.00
3-tine SINE 9
ARARA with ridger *
120.00
23.00
10.00
1238.01)
92O.W~
7
225 .
7.3. Implement Field capacities
7.3.1. Field effkiencies
‘_ .
The field efficiency E (in percent) of a farm implement is defined a.s the ratio of the
implement’s effective field capacity (EFC in ha/day) to its theoretical field capacity (TFC in
ha/day). An equivalent approach is to determine the ratio in percent of effective feld time
(EFDTM in hrs) to total field time (TFDTM in hrs). The effective field time of an
implement is the time spent performing a fimctional activity.
(55)
Table 63: ANOVA of E(%)
-----
--~--
---~-
Source
DF
Seq SS
Adj SS- Adj MS
F
P
-PV--
----
WC
2
378.08
378.08
189.04
--~-
4.74 0.030
EQTYPE
2
571.56
571.56
285.78
7.16 0.009
YOKE
1
480.09
480.09
480.09
12.03
0.00s
Errer
12
478.86
478.86
39.90
Total
17
1908.58
Table 64: Average implement efficiency E (%)
.-----
---
EQ’TYPE
N
MEAN
MEDIAN ST?&
SEMEAN- C’G(O4,) -
t-6 10”
6 ---
61.0’7
60.93--- 6.87
2.81
--~
11.35
74.54
72.63
11.43
4.67
15.33
65 19
64.69
9.46
3.86
14.51
4..
t
226
The summary in Table 62 sugge’sts that the field effkiencies E depend no’t only on
the type of implement but also on the type of yoke used. The analysis of variante shows
‘_ .
that the difference in E is highly significant (Table 63).
Field efficiencies are best given in terms of field operation. It makes more sense to
use the eficiency attached to the type of implement to perform the field activities (Table
64).
The effective field time (EFDTM) to perform a field operation is affected by the
losses of time when turning at the end of the furrow and when the implement is grounded
because of continuous high draft requirements caused by the rooting system of tree stumps
scattered across the fields.
Another source of variation was introduced by the reluctance to pull on a
continuous basis of the pair of oxen and their. They ofien made frequent stops to tope
with the draft requirements close to their optimum draft performance, with an average
PF/LW in the following ranges:
UCF 10” plow
: PF/LW = (0.21, 0.24)
3-tine SINE 9
: PF/LW = (0.15, 0.20)
ARARA with ridger : PF/LW = (0.16, 0.20)
7.3.2. Field capacities
Based on the field eficiency E in percent, the effective field capacity EFC in
ha’day is equal to the theoretical field capacity TFC in ha/day times the efficiency E!,:
2 2 7
ANWKGHR
EFC = (WWKGCOMT * ANISPEED) * (E) * (
utn,
j*30 *m4
(56)
‘_
.
Where the WWKGCOMP in cm is the width of the implement’s working
component; ANISPEED in m/s is the animal walking speed and ANWKGHR in hrs/day is
draft animal daily field working time. The coefficient .36*‘10-” is used to convert the units
into halday.
The field capacity depends significantly on the number of hours ANWKGHR the
draft animais are effectively used in the field to perform a task. The time spent traveling to
the field and yoking the pair of oxen are not included in the number of animal-work-heurs.
The total area cropped in ha during the season is calculated using the EFC’ (haiday)
and the number of days worked in relation to the probabihty of working day (PWD in
percent).
7.4. Summary
14 number of decision variables are involved in farmers’ daily activities during the
rainy season. Some of the variables are weather t-elated, therefore out of the fàrmer’s
control. The soi1 moisture rlegime SWC (Oh g/g) remains the most determinant faclor when
it cornes to performing farm activities. From the onset to the end of the rainy season.
farmers rely on their good judgement to decide when to perform a field operation in
relation to the available resources at the farm level.
The number of days suitable for tillage and other tïeld operations is dictated by the
type of sd along with the configuration of the rainy season. One of the best approaches 10
predict the number of working days NWDAYS is to use the level of probability or level of
1
,I
228 .
certainty. Two levels are generally used the 50% or 5 years out of 10 and the 90% or 9
years out of 10.
.- .
It is unfortunate that precipitation data are not always normally distributed. The
duration of the period to analyze is important to draw acceptable statistical inferences: the
shorter the duration, the less the likelihood of a normal distribution. Long duration periods
are preferred as the World Meteorological Organization WMO recommends a minimum of
30-years data. Even with this range of data, annual rainfall data are still skewed
(positively), with fewer totals significantly greater than the mean. Sometimes the use of
the median instead of the mean cari provide a more realistic and usable results especially
for the 50% level of occurrences (Sumner, 1988).
During a suitable day for fieldwork, the level of intensity of the draft animals
utilization represents a determinant factor in the evaluation of their field working capacity.
At the beginning of the rainy season, the work demand is always high and farmers
consequently try to maximize their utilization for greater output. The number of working
hours per day ANWKGHR along with the speed ANISPEED to perform the work will
determine the farmer’s capacity to meet his cropping objectives. It is important to choose
the correct combinations of implements, which Will manage the available working capacity
ofthe draft animals. As demonstrated in previous chapters, farmers do not have a wide
range of implements to Select from to carry out the tield operations. Only the expected
draft required DR in daN from the working component-soi1 relationships would help make
a decision. The amount of draft found in this study shows that farmers normally use the
draft oxen within the area of their optimum energy output (PF/LW= 0.20). This level of
utilization corresponds to medium-to-heavy work intensity. Working an average of6
A
1
229 .
hours/day in these conditions reinforces the management strategy of 1 day-rest per week.
It is important to fit the job t’o the draft animals in ternis of work,load to improve their
eficiency and productivity.
The use ofdifferent size head yokes does not seem to have a Sign&ant effect on
the pull force PF developed by the pair of oxen. It is a good management p:ractice to look
aRer dra.A animal’s comfort by fitting the yokes to the animal as energy is converted to
usefiA work through the yoking system. A poorly fitted yoke Will reduce power output.
On the other hand., the type of yoke has a significant effect on the field efficiency 13. The
use of a larger yoke has the tendency to increase the working width without really
performing a better quality job. In order to use realistic parameters, the use of the
implement field efficiency (highly significant) is recommended in the process of field
working capacity evaluation.
‘.
.
Chapter 8
EXPERT SYSTEM TO EVALUATE THE IMPACT OF ANIMAL
TRACTION IN THE BASSE CASAMANCE REGION
8.1. Organization of the expert system program
The expert system developed in this study is aimed at evaluating the working
capacity of a pair of draft oxen in the Basse Casamance region conditions. The program is
built with Rexibility in order to be expanded in the future to include other draft species and
other geographical regions.
The program is organized around different modules (Figure 35).
- Field capacities module
- Energy module
- Feeding system module
- Farm budget module
- Optimization module
The main module (Energy module) is used to calculate the traction potentiai and
the traction delivered in terms of energy.
230
231
r-------‘
/
/
Default Values I---------b’
CALCULATIONS
Traction Potential
~--_
ANIMALS i+ -_.-_ --e-d
i-.------
I Traction Delivered
I_-----.
..~
Traction Required
/
OUTPUTS
------l--- .--. -_ __. 9 Nutrients Requirements k-b
rSOIL FILE /--7
Field CapacQ
Power Requirements
_-----_
Feed Rations
WEATHER 1
Farm budgets
WORKING DAYS
LYE-__d i-
ry--------
/
l
Recommendations
/ PRODUCTION -+
FIELD
/
COSTS
/ ()pERjjTI()NS k----------- .------- +!
i
-17-r
i
- costs
l OPTIMIZATIQN ~
Figure 35: Animal Traction Evaluation Mode1 Diagram
7
232 .
The program was built with the LEVELS Object Oriented 3.6 for Microsoft
Windows Copyright 1995, Information Builders. The language is a high-level knowledge
engineering language called Production Rule Language (PRL). The syntaxes are organized
within Methods (When changed and When needed), Rules, and Demons which contain
eomrnands that tel1 the inference engines to initiate one or more actions at run time.
The program at run time needs a certain number of external files ma
DBASE III program and graphie files made with the Paintbrush software of the Microsoft
Windows 3.1 package.
Most of the information and default values supplied to the program were data
gathered from the field or models developed in the previous chapters.
For the Optimization module the program is set to use the Optimizer tool of an
external Windows spreadsheet server program like Microsoft Excel, Lotus 123 version 3
or later versions, and Quattro Pro.
8.2. Database Files’ structures
8.2.1. Farm location and characteristics
Barne of the file: FARM.DB3
Fields Name:
COUNTRY : Senegal.
COUNTPIC
: Country picture (Senegal.bmp).
REGION
: Basse Casamance.
REGPIC
: Region picture (Basse. bmp).
SITE
: Southwest (Geographical Location).
NZONE
: 5 (Number of Agro-ecosystem zones).
ZONE
: Agro-ecosystem zone
VILLAGE
: Village name (optional).
FMSIZE
: Farm Size in ha.
NFWKERS
: Number of Farm Workers (active workers).
MCASCROP : Groundnuts (Main Cash Crop).
WPESOIL
: Sandy clay (Type of dominant upland soi]).
2
3
3
It is assumed that small-scale farms’ production systems are in general driven by at
least one main cash trop (MCASCROP). The information is needed for the program to
execute the energy evaluation module and the optimization
Two types of information need to be input in the p:rogram at this point by the user:
- The types of crops grown at the farm level which make up the cropping
system (Groundnuts, Maize, Millet, Sorghum, Cotton and Rice).
- The First year the farmer started using animal traction. The name of the
fa.rmer is optional (I-irst screen of the program).
.8.2.2. Farm implement
Name of the file: EQUIP.DIB3
Fields’ Niame:
NAME
:UC.F 10” (Current Name of the implement)
WKGCOMP
: Moldboard (Working component)
NWKGCOMP : 1 (Number of working component)
WWKGCQMP : 25 cm (Working component Width)
WKGDEP
: 12 cm (Avg Working component depth)
,QVGPF
: 145 daN (Average Pull force required)
RETPRICE
: Retail market Price
IMPLPIC
: Implement picture (* .BMP)
MANLIFE
: Years Life according to Manufacturer
F.ARMLIFE
: Years Life in Farmer’s conditions
IMPLEFF
: Field Efficiency (in percent)
IMPLREL
: Implement Reliability (in percent)
The program Will use the given average pull force (AVGPF in daN). The mode1
developed in the study (Equation 54) is used to supplement a default value for AVGPF
The efficiency IMPLEFF in percent and reliability IMPLREL in percent were
Idetermine:d previously. New data entered by the user Will overwrite the current ones.
234 .
Information on the price of the implement (RETPRICE in local currency) cari be
supplemented at run time.
.
A bitmap filename with extension .BMP for the implement picture IMPLPlC is
needed to be loaded for display to the user.
8.2.3. Draft animals
Name of the file: DRAFTANLDB3
Fields’ Name:
TYPE
: 0x.
BREED
: Ndama.
SEX
: Male.
AGE
: Age of the younger animal if in pair.
CIRCUMF
: Circumference at point of heart in cm.
AVGLW
: Average Liveweight in kg.
MATURELW : Mature Liveweight in kg.
NDRFTANI
: Number of draft animals.
MKTPRICE : Market price.
MATPRICE
: Mature draft animal’s market price.
AGETRAIN : Age trained.
YRSEXP
: Number of years of experience.
ANWKGHR
: Animal number of field work hours per day
ANIPIC
: Animal Picture (*.BMP).
- Readiness for traction
The liveweight LW of draft animals are not generally known or readily available
from the majority of farmers. The mode1 developed previously (Equation 52) is used by
the program to estimate the value of the liveweight LW in kg. The information neelded is
the average value of the circumference CIRCUM in cm, measured at the point of hean of
each draft animal The average liveweight (AVGLW in kg) is used by the program to
evaluate the pulling capacity of the draft animals. For the specific case of cattle, different
studies have shown that oxen with a liveweight less than half ofthe value of the species’
235 .
optimal or mature liveweight (MATURELW in kg) must be excluded from traction (Lee
et al, 1993). When a pair o:f oxen is rejected on the basis of that rule, the program Will
. .
display a message and Will mot proceed fùrther:
Not ready for traction
- Age versus Years of experience
The program Will us’e the AGE in years of the younger animal (if a pair of draft
animais is used) to test the readiness for traction in terms of capacity to perform field
activities. The average training age (AGETRAIN in years)‘s threshold of draft animais of
current qpecies are usually well known to animal traction users. In the Basse Casamance
region, oxen are trained when they are two years of age. To be ready for field work an ox
must be a.t least more than 2 years. The threshold cari be overwritten.
The number of years of experience (YRSEXP in years) is used to rate the field
Yworking capacity of the draf? animais. In the Basse Casa.mance region 3 years of
experience is an accepted number ofyears to carry out good quality fieldwork. The
number of years of experience and the age for training alre compared to the age o:f the
animal to verify the consistency of the data entered in the session. The program. willl reject
inconsistencies and ask for new data. If a draft animal with correct liveweight has less
experience than required, the program Will inform the user through a displayed message
that:
Need more traiuing
236 .
The program Will proceed from’that point but Will treat the draft animals as
inexperienced.
.
- Working potential
Draft animals with correct liveweight and adequate working experience, are
characterized by the program as:
Good for Field work
The average number of working hours (ANWKGHR in hrsfday) expected from the
draft animals in relation to their pulling capacity is needed to evaluate the effective field
capacity EFC (Equation 56).
Information on draft animals’ market prices MKTPRICE and MATPRICE in local
currency are used in the cash flow analysis in the farm budget module. This information
cari also be supplemented at r-un time.
The draft animal’s picture ANIPIC is a bitmap file type with .BMP extension.
8.2.4. Feeding System
Name of the file: FEED.DB3
mds’ name:
MAINFD
: Main feed (Pasture grazing, tut grass, .)
MFDQUANT : Quantity of main feed in kg
DMCMFD
: Dry Matter content of the main feed in %
ERGCMFD
: Main Feed Energy Content (in MJ NE/kg DM)
SUPLFD
: Supplemented Feed in kg (Grain, stover, ,..)
SFDQUANT : Quantity of Supplemented feed in kg
DMCSFD
: Supplemented Feed Dry Matter Content in %
.
237 .
ERGCSFD
: Supplemented Feed Energy Content in MJ/kg DM
S F D P R I C E Price of Feed per kg
In farmers’ conditions, the main feeding type (MAIlWD) during the rainy season is
the naturall grazing system. T’he quantity and quality of feed (MFDQUANT in k:g) fiom the
natural grazing is never enough at the beginning of the rainy season to provide enough
energy to the draft animals. Farmers use different sources for feed supplement (SUPLFD
in kg) like maize grain, sorghum grain or stover. The groundnut hay is generally used
during the dry season along with the other trop residues. Fe:ed information during working
days is required.
,
The price of the supplement feed (SFDPRKE in local currency) is used to
evaluate the cost of the daily maintenance of the draft animais
8.2.5. Crops grown at the farm level
Name of the file: CROPS.DR3
celds’ naEg:
TYPE
: Type of Crop (Groundnuts, maize, millet, . .)
VARIETY
: Variety of trop grown (69-101, ZM 10, . ..)
CYCLE
: Number of days from seeding to hawest
SDRATE
: Quantity of seed in kg/ha
SD LINE
: Distance in between rows in cm.
RIDGDIST : Distance in between ridges in cm.
FTRATE
: Fertilizer rate in kg/ha.
SDILIMIT
: Date limit of Seeding for potential yield
SDPRICE
: Seed Market price per kg.
FTIPRICE
: Fertilizer Market Price per kg.
CROPRICE : Crop h4arket Price per kg.
YIELD
: Avg Yield without animal traction in kg/ha
LAIBCDAY : Farm labor cost per day
II
238
,.. &‘.. .
CMy information related to thc types of cr:;?
:rown at the fam level are entcrcd
in the current session. The pro,. r-am Will not recognize $3 trop that has not been selected at
the beginning of the session.
1st of the variables are used for the elaboration of the farm
budget.
Somc variabIcs listed in the file are net activated I* ,* ’ current
of ,the
version
program but wili bc in the r
:turc. Variable such as I-
:
‘I’ (date limit of swdirig
fc)r a givcn trop) are used t
.valuate t he timeliness cost,
-,ariable FTRATE
(fer-tilizcr rate) to include farmers that use fertilizers. The 11.
.f farmers in the Basse
CJasamancc. rcgion do ,t use ? l-tilizers. Different studies ha\\
,,,,
. ed that the actual
aseragc rate of fuliliz(.
under 10 hg/ha of NPK (8: 18:27)
Thc YIELD in L requircd by the program correspol
JC Icvel ofyield
observed before the uti!
dion of animal traction. In thc pmct :s of evaluating the yie?d,
tlic program assumes th
.u-mers with less than 6 years of expcrience or using non-
experienced dra.ft animai.\\ il not benefit from the yield increase induccd by the
mechanical plo\\ving. The yield inçrcments givcn to experienc.c.d fxmers come fro!n the
average rcsults of past experienccs 05 aitir,, .: : raction induc.ed yield increases over m;:.nual
farming. A percentage increase on : id i:.
.! for diffcrent level of mechanization for
cxh trop grown at t.he farm level. ‘The yi~,~~ ~;icreases generated by the mecharr::-ation of
land preparation were found to be most significant and consistent according to ‘i’ourte
(‘197 l), Nicou (1379) Clcarreau and Nicou (1981) cited by Jaeger (1985) (Table 14).
Results fi.on field work carri<,: out by thc FSR tcam (Equipe Systcmcs, 1982-90) in the
Basse Casamance region werf ,-‘w used. Rice is the trop with the highest yield increase
(S@.OO%), follom,ed by coti~ j_, i .:, .lOOh), maize (46.71?/0), Sorghum (34.60%), groundnuts
adjusfed by a coeficient (0.86) for the 1 day of rest per week.
trop and for each field operation at given probability Icvel: 5 years out of E 0 md 9 ycars
out of 10. The LABOIIBALh, in mn-days/ha is the diifercnce between thc labor Sait&Eable
(AVI..A.BOR.t,, in mn-days/ha) r$nus the labor requit-cd (T,ABORE@, in mm-da:~~%a) to
The output on the fart labor shows the yotential trade- offs and thc Mm enjxts
uf‘each levd of mechanization for each trop. The farmer must choose among thc t rades-
of’f, tnwnrds meeting his production objectives. Qne major strategy oAen perfomc-d by
fxrrwrs is to use the iabor surplus generated by the FmxhaFizatiorl of land preparation to
perfCm~ seeding oyeration instead ofusiug seeders.
8.32. Energy module
fhe energy balance uses the rnost recent scientiGc work on animal emrgy
evahtation. The factorial method (Equation 5) is used to perforrn the difYerent energy
culc~lations. The coeficients used in the equation for this study are mainly pulled fi-or-n the
dissertation work carried out on cattle in the West Africm region by Abdou FALL (1995)
(59)
i
244
8.3,2,4. Energy from L,W bsses (LWLERG)
The beginning of the rainy season represents a peak energy dcmand period because
of intense iield activities. Thcse activities especially tillage, take place within the numbcr of
warkings days in the months of Jrrne and July. The level of draft oxen utiliz.ation is orten
SO intense that draft oxen uses their body reserves to mcet energy needs, resulting in a loss
ofbody weight (LWL in ke/day). ,AnimaI traction researchers in general, bave reported
nc:gative LW change,s during this time ofthe year. Even well fed draft animafs were found
to Jose weight (Goe, I983).
The analysis of field activities in far’mers!.I conditions have shown in yrevious
chapters that draft oxen in the Basse Casamance region lest an average of LWL - 1 2 1
kdday. Lawrence and Becker (1994) suggested the use of a coeficient a’f 15 MS NE/kg
to convert every kg of LWL into usable energy by the draft animais. The en~gy
(I..WI.,ERG in h4J hFlday) is calculated as followed:
The LWIXRG is expressed in MJ NE/day and t.,WL~fiotip in kg’day.
8.3.3. Feeding system nodule
The calculation of the feeding system is based on rhe ratio k of the energy used
during every working day (daily maintenance + work donc) over the daily :majntcnarice
k, 0.019 * {xj -: 0.503
where x in MJ/kg IX.4 is the equivalent cnergy content of the total diei.
246
- NE available for worir. frcm the feed ~~~~~~K~~)
Only part of the energy made available from the metaboliz&le enetgy ME is tised
to perform work and tlie rernainirrg part used for maintenance pur-poses. The net energy
available for work from the feeding system (FDWRKNE in M.J NEiday) is calculated by
subtracting the requir-ed energy fGr maintenance per day MAE from the total net available
eFlCX&$
(h$NWRKNE
in NJ NE):
According to the fa]-mers’ systems management of draft animais, the pair of” oxen
maly be used 6 days a week. The amount of energy available for work M’RKNE in MJ
NE/day is evaluated for the working days only and adjusted to rake into accourt tfie
energy available fiom LW losses LWLERG in kg/day:
This level of energy is available and expected ta he expended from tl-te pair of
oxen. It is important tu mention that the quantity of feed used to carry nut a11 these
c,a!culations corresponds to the total amount given to bath animais workiog in the team
k
248
A fist of available feed stuffs along with their FU cC)ntent per kg of DM (Table 65)
is presented for ~he user to select and make up his own ration FDRATION exyrcssed in
FI-J.
=:-.-..I -_._-
-1------
----.-
I_-_-~
--
.--vv.
---
~kemn&----.
Airailable Feed
FU/kg
___--
~~~
_---_- -_--.,
Sorghum grain
0.90
Millet grain
0.70
Ma& grain
1.10
Cow peas grain
1 .oo
Cotton grain
1.10
X.
Groundnut cake
1 .oo
Millet/Sorghum stover
0.85
Cut grass
0.14
Pasture grazing
0.06 - 0.19
Good quality hay
0.29
Groundnut hay
0.35
R.ice straws
0.26
------
m7=l=:lP-----
Sowce: CEEMAT, 1975.
Most of these fecds are available on a yearly basLc’ It is important to select the feed
avaifable during the intense working period, mainly June, July and August.
The program w-il1 present to the user a calculator ;i)~pe of display to sum the N-J of
the selected feed.
0‘
(7’7)
~~gj,_l variable cost wo labor (TVCWOLAB in CFA.4ta)
TVCWQLAB = SDCQST + FERCOST
Total fixed cost
--_---..-_
(TFCOST in CFA/ha)
TFCOST =i JMPFCOST + ANJFCOST
~~~OZ&CI~~QSJ (TCWLAB; TCWOLAB with and without tabor in CFNha):
‘TCWLAB = TVCWLAB -t TFCOST
TCWOLAB = TVCWOLAB + TFCOST
GIMWLABha = PRQDVAL - TVCWL,AB
(82)
CMWOJ,ABha = PRODVAL - TVCWOLAB
C8.3)
l~&tkjaygin NM per ha (NMWJ,ABha; Wh’fWOLABha w and wo labor in CFA/ha)
MiWLABha = PRQDVAL - TCWLAB
w
K.MWOJ~ABha = PRODVAL - TCWQJ,AB
(85)
.I.
P--1-....---
..~-,-^--..~--.--L=--
----n-
.msœweewIm--
(88)
252
‘E-E objectives of the optimiza~ion problem GUI be expressed as followcd:
8.3.5.1. Optimhtion objectives
Goal 1: Maximize revenue from the cash trop
Gai 2: Produce enaugh cereals for farmer’s own consumption.
Instead of using a goal progr-amming approach, the second objective of the
problem is considered as a constraint in the search of solutions to producing cereals
corresponding to no less than the farmer’s needs.
In small-scale farming system, farmers’ objectives are not alwaps monetary.
Trades-off are sometimes possible in relation to the current economfcal production
environment. A major adjustment performed in the Basse Casamance region was to
sacrifice the cereals’ production in general and rice in particular, in favor of the groundnut
cash trop. Rice was kept to a minimum and was allocated the least resources compared to
other upland crops. The main reason was the risky wealher situation characterized by a
persistent shortage of rainfail. Among the cereals produced by farmers to meet their oyVvn
consumption, rice production always plays a role of regulator and oc,cupies a fixed
percentage of land in the low land areas acc,ording to the spatial distributbon of crops.
8.3.5.2. Constraints and objective fonction
8.3,5.2.1. Decision variables
The decision variables used in the program, in the form of area allocated to a trop
(CRCIPAIEA in ha) are defined as follow:
Land constraints
&,,, (LARUREQ,, * CROPAREA) < AVLABOR
,e.
C-W
The available farm labor (AVLABOR in man-days) is evaluated in relation to the
number of active farm workers NFWKERS at the farm le\\el multiplied by the number of
working days during the growing çeason.
,,
The labor required per L~O~Qi-,, in mari-days’ha corresponds to the total
amount of man-days per ha needed to conduct the trop from tillzge to harvest.
Animal energy constrairrts
The animal energy must be evaluated for the time the fïeld operation FDOP Iasts
and for as many field operations FDOP as performed (tillage, seeding, weeding). If the
pair of oxen is used for plowing only then t.he competition betueen different crops
bccomes a constraint in relation to the total energy available for the number of’working
days (NWDAY S in days) for tillage field operation.
The available animal energy (AVANERG in MJ NE) is equal to:
AVANERG = (k,,,)*(MAE)*(~,,,,, NWDAYS)
(92)
f
3’1~ consomption ccrnstraints are evaluated usin@, the FAO standards for food
dried, thrcrshcd, de-hulled and milled. AfIer proccssing, only approximatcly 65% of the
procc:ssi;~g from harvest to the final stage for cC)ï:Iî1IrrlPti(311 The cotton is net edible (094,)
team bave repo1-te.d that most farmers provide more than 60% of their groundnut seed,
arxxnd 20 to SO kg per ha, ‘The total household ~SL’ for groundnuts is estimated to be 5%
of the harvest per year.
l’he constraint on consumption cari lx formulattx! as follow: The sum of
ho~xxho’lcl total consumption needs (CONSNEED in kg:).
256
C’ONSNI:GD = 200 * (NFW)
The consumption constraints cari be written as follow:
Y
~~r.R,zp (CONSVAL * CROPAREA) > CONSNEED
(94)
Cropping system constraints
The cropping system constraints are expressed in terms of minimutn area of the
farm land to be reserved to specific c.rops. Rke production in the Basse Casamance region
car-ries a heavy load of cultural behavior a.s the traditional cropping systems in the past was
mainly centered around this trop. A number of social scisnce researchers refer to the
Basse Casamance population as the “Rice Civilization” people (Pelissier, 1966).
For the Basse Casamance region, the percentage of land allocated to rice
production (RCPCENT in percent) varies on the basis of type of agro-ecosystem. In agro-
ecosystem zone 4, for example an average of 15% of the farm land FMSIZE are used for
rice produc.tion against 20% in the agro-ecosystem zone 5.
The value of RCPCENT in percent is expected to be different for individual
farrners but the present version of the program is designed to use a default value pcr agro-
ecosystem z.one.
_ .
._,._.
“...
-
____
__-__._
-
-.-
-
-.
-_., <,
_~
”
.--
-“.- ‘.~,.l...‘: . . .-*-! -
. ..“..Tll / 2, .j ..W./” Xe“ ^A .-. . . .: I _,
:
-.*a
.__.<
.
.,_
The cropping systcm csnstraint is net includcd in the table of constraints prrxnted
‘Il~e nn~ount of land resouxes akcatcd to a trop (CRQPAREA in ha) ~nust be
positive or equal to zero. W’hen using the CROPRREA amiable in the eyuation, C‘ROP
has IO be replaced by GN fo,r groundnuts, MZ for ntaize, ML for millet, SG for sorgh~~rn,
CT for cotton and RC for rice. This general condition on the decision variable also called
the normegativity restrictions will guarantee a rcal value to the solution
3.4, Opltrn&ation outyut
The optimkation oul:put is bound by the amount and quality of information the
server is ahle to process. The Sofw- tool froc Microso~R Excel cx Optimizer from Quattro
i
258
Pro for Windows have their own technique to handle such problems. The current version
ofthe expert system program is designed to fit the Quattro Pro for Windows’ forrnat
8.4.1. Server solution
It is important to mention at this point that the solution to the optimization
problem refers to the combination of the different decision variables and not to the
maximum value of the net profit (NPRO’FlT in CFA). The main concern for the user is to
determine the trop mix to be grown and the area of land that will maximize the profit,
usjng the available resources.
The solution is given in ha of cropped land for each trop. A zero value means that
the trop must not be included in the cropping system. As the decision-maker, the user
must weigh ah the information given by the optimisation output before fc~rmulating, a
decision. The cropping system chosen should reflect the specificity of his farrn. Qualitative
variables not used in this analysis cari help reach a better viable solution.
8.4.2. Interpretation of Coeflkients
A number of information items are given along with the solution to the problem.
Thc information items are presented to the user on the same display. The main çoeffkients
given are:
- SlacWSurplus
The coeflkients represent the difference between the level of utilization of a given
resource and its availabihty.
considered a “free good”. A free good means that an increase in its availability ,wvill not
increase or change the }>tofits. The slack must aha~ays be zero or positive. Slack. is better
equations.
In surplus situations, the difference bctween the Ievel of utihzation of a resource
and a minimum requirement i s positive. This difference is used to measur-e how much the
requirements are exceeded.
Dlfferent names are used to represent the same coef!kient, like opportunity cost,
margina! or dual values in relation to the \\~;ty the optimkation problem is stated. Shadow
prices are applied to Jesources to estimate thcir opportunity cost when not one unit of tbe
conespc>ndding r-esource is included ~I!I the. solution. Jn other words, it expresses the change
in the value of the objective func.tion when Ione additional unit of the resource is added to
261
8.5. Pragraïkl vnfidwtion
In the proccss ofc.ontinuous validation, the expert system program is built to \\+Bork
with val:‘c!afed input data. Ruilt-in :ncssagcs guide the execution of thc program. ‘fhc ert-or
routines execute and vcrify at the same titr\\e tbe quafity of t he da.ta entered. In case of
/nconsistencies, the program will prompt the user to enter correct data. The amoi~m. of
TWO modules need to be testtd and validated against other methods of c\\~alllatiom:
260
any othcr \\wiable enfers the basis, it cm only cif her decrcases thc profit or ieave it
- Weduced cost
The reduced cost for the decision variables reprewtts the amount by which the
objecl-ive tinction will be reduced if the variable is introdtIced into the optimal solution. In
other words, it expresses the penalty imposed on the objeclive function if one unit of the
variable net currently accounted is entered into the optimal solution
A fully used rcsource is “hinding”. It constitutes a bottlcneck for any improvernent
or reorganization of the environrnent. The Kuhn-Tucker conditions stipulate that ifthe
shadow price is nonzero then thc resource is fully utitized mcaning there is no slack OI
wplus (Turban and Meridith, 1994).
8.4.3 Sernsitivity aaalpsis
An explicit module to run a sensitivity analysis is mt included in this version. ‘I’he
progmm musl. be re-run in order to change the parme.ter and observe its efiect on the
optin;al solution. The purpose of the sensitivity analysis is tu give to the user additional
decision making information.
261
: 4 years
262
- Prqpm? sutput
Effective fieid cqacities CMdayj
G N
M Z
ML
S G
R C
l’illage
0.18
0 . 1 8
0.18
0 . 1 8
0 . 1 8
Seeding
0.04
0.07
0 . 0 5
0.07
0.05
Weeding
0.03
0.03
0 . 0 5
0 . 0 6
0.05
</.
Harxsting
0.05
0.08
’
0 . 0 8
0 . 1 0
0.02
R C
.
T’illage
39.69
WcedFng
<.
.,.
-
*.
Maintenance (h4J NE/day) = 37.02
~1s-g evaluatjon rnod~~.&~-W!?&
G N
MT2
ML
S G
R C
‘TII age
223.78
223.78
2 2 3 . 7 8
223.78
223.788
W’eeding
+.
Seeding
I-farvesting
Maintenance (MJ NEYha) = 208.76
..-..___
._ --.-
.m”e..-...-..-“.-
I<oI---
-11111
.*-.-.1-m
Q~ti!~~~i~3-~!~t_lll--
Land (!m)
Reduced cost (CFA/ha)
4 . 7 5
0 . 0 0
a.35
0 00
0 . 0 0
-34264 3 1
0.00
-43607 64
a.90
0.00
Slack
Sh2ldOW prices
0.00
108154 90
0 . 0 0
509.5 1
1 OI 4.40
a 30
44 24
0.00
0 . 0 0
-,54468.06
consiraints (ha, mm-day, MJ NE, kg equi\\xletzt-cereal and ha of rice in the ordel- written).
dL---
-- -
Field
canacities
In this farmer’s situation, only land preparation is mechanized with the UCF 10”
moldboard plow. AH the post-tillage &Id operations are manually performed.
According to the tillage effective field capacity EFC in halday, it will require from
the pair of oxen an average of 5 days (6 hours/day) to plow one hectare of land. It is
a ~r.umed that flat land preparation is executed with the moldboard plow.
In comparison with the values pubiished in the literature, the program output for
ca& oper ation is within acceptable range. For example, L)e Moigne (198 1) gave the I ange
of6 to 8 days for a pair of oxen, \\vorking 4 to 5 hrs yer day to complete: the task.
Draft animais Energy balance
The validation of the animal energy balance is more dif’frcult in terms of fInding
similar utilization environment. For testirtg the consistency of the information given by the
program, the output is compared to the experiment repor-ted by Falvey (1986) in terms of
procedure. Al1 the steps were followed to reach the em gy balance for ea.ch field
operation. The value given in the output is mainly refated to the level of confidence given
to the raw data and to the validity of the diflerent equations used.
Qptimization output
The solution of the optimization rccommends that tbe user grow ‘79% of the farm
with groundnuts which is consistem with the practices in agro-ecosystem zone 4. In terms
of cash slow, it was expected from the. program to allocate more land resources to the
cash crl~p. In a small~ scale farming system, the x,alue attached to the cash trop :qresents
the drivin~ force for trades-of’fin relation to thr: pricc levels of goods in the officia1
mar-ket Su& decisians in turn cari reshapc the type ofcrapping system to the point that
thcy become the determinant factor in the evcrlut icsn of the global production sl’stem.
getting îhe rwded information to run the program. As this sttrdy has show, the
ewluation of the dif%wnt variables a.rBd yarameters rquired conriderable ti:.ne atxf human
266
resources. The good news is that once the data are validated and tested for stnbility, they
remain consistent and ready to be used.
Expert systems in agriculture represent a valuable rool to conduct farming system
diagnostics, field operations scheduling and management. The existence of such programs
gives opportunity to farmers, govcrnment servants, non-gsvemmental organizations, to
access technical information and to interpret simulation in relation to the socio-economiçal
The organization of the program in modules makes information flow and outputs
easy 10 interpret. The energy evaiuation module uses in its important routines, animal
x
energy equations developed elsewhere by other researchers. This approach is aimed to
stimulate the continuity of the research on that topic. The factorial method developed by
Lawrence (1985) to evaluate the extra energy used by draft animals to perform work is
actually widely used around the world. An effort has been niade in this study to use
coefiçients vaiidated in the West African region with cattle.
The annual farm budgeting and the linear optimization are classic tools used in
decision-making problems. The program output was designed 1.0 fürnish as much
decisional infomlation as possible to the user.
As a general comment, this version ofthe program needs to be used for more
validation and improvement. Some actions are going to be taken in the near future to
expand the versatility and the geographical coverage to other regions of Senegal and,
possibly to other countries.
--
._--
-.__-.
-.
.
__.
.-
_.--
..-.-
.-
‘~l-~llr-.F.------
#,,-..-.“.
e.--u
-r.IIIIIu---
?
9.1. corscllusions
Aninul traction is providing more !han 80% ofthe energy needed in the
agricu1tura.l activitie s, in %ne&d in piifliCtJli4r. l’he animal species used for traction ai-e
rnainly catl:le, horse and donkey. In the Southern part of Senegal, in the Basse Cassmance
rlegicn, cattle are the llia;n source of power. The Ndama brecd is \\,vell adapted in the
environment characterized by harsh sanitary cunditiorrs fix working animais. The level of
~tilizik~~ is highly variable across the region, and yet it needs to be evaluated at the farm
level Tht: purpose of the evaluation is EO estimate its potcntial and cwrent impact. Tt
267
268
represents an important factor of evolution of the cropping systems and towards farmcrs’
achievements in building up a sustainable production system.
In this study, the evaluation technique used was built around an expert system to
simulate different levels of utilization of animal traction in differcnt agro-ecosystern zones
in the Basse Casamance region. The expert system was structured according to modules
linked by routines and subroutines using a high-level language cafled PRJ, (Program Rule
Janguage). The PIU language is similar to natural langu.age (like natural English) with
. “ <
interactive editors for more flexibility and accuracy in building applications. The LEVELS
OHJIXT~~ (Release 3.6 for Microsoft Windows, Copyright@ 1995 by Information
J3uilders, Inc.) shell was u.sed as the main programming environment and development rool
kit. As any expert system derived fiom the artificial intelligence (AI) domain, it needs a
considerable amount of quality information in order to be used for decision making. ‘The
i.nformation supplied to the expert system cari be performed through a well-organized
knowledge base management system (KBMS).
The knowledge gathered from animal traction utilization in the area ofstudy was
processed into 5 modules:
- Field capacities c.alc,ulations
- Animai Energy evaluation
- Feeding system
- Annual Farm budget
- Optimization of farm resources’ utilization.
-.v
.-,x--.*-s
“,.s.“ma
----
Y
269
Information kd in the program has betn collertecl directly from the fields in
farmers conditions and from reliable sources in the. existir!g scientitic literature. A
trials. TO carry out the fkld activitjes, the farm was wnsitlered to be the unit of
obser\\;ation ar!d was viewed as the basic ele:dnent of the prod~~ction system
ofthc: diffcrewt types of farm implements cwrcntly used by farmer-s for trop prodwtion.
The second N~S a follow up ofdifkrent kld operations ]xrf?nrmed with the draft ~~X~II.
The field triais were desi,gned 1.0 evahatc two things: (1) The amount of poser
a..nd ewrgy required to pelform difkent ficld operations; oncl (2) TO meawre the
rcaximum and optimum poser that the draft oscn ~wuld ~~evclop at the beginning ofihe
rainy season.
9.1.1, Sumeys
- Farrn hplenrents
It V:~S possible in a relatively short pcriod of time (two weeks for 40 farmers) to
idcntify all the implements actually used by farmers. The C’ata collected ranged from the
type and number of implement s, the acquisition mode, the system management, the repail
and mainterxtrxe problems, the type and number of dl aft cnimals to the local b!acksn-ith
activities. I%~~X data were ana.lyzed to better understand ihe agricultural cropping system
alternatives available to fumers in relation to the opportw~ities ofEred by the type of
agro-ec.osystem %OIles.
270
It appears that the range of implements is not very large. Only one mode1 per type
of implement was used: UCF 10” moldboard plow was the only type available; the same
for the cultivators including the SINE toolbar; also one type of one-row seeder (Super
I-ICX)). Only the ridging implements were available in two versions. One was attacbed to
the ARARA toolbar and the second, imported from Gambia.
AN the farm implements experienced several breakdowns. This was associated with
the quality of the materials used to make the working c.omponents and the conditions of
use. Most of the soi1 working components were victims of the natural wear and ofthe
presence of roots and stumps scattered in farmers’ fie&. The problems were aggravated
by the absence of technical support. The local blacksmiths were not able to handle some
types of breakdown. The unavailability of required tools and good quality raw materials
prevented timeliness of the services. In these conditions the reliability, evaluated for each
piece of farm equipment was generally found to be low.
- Draft animals
Even though oxen were available to farmers, the distribution was found to be
highly variable. The main source of draft animais remained the herds in the village. The
s:~alysis shows how unstable the technology was as the majority of farmers refied ou only
one pair of oxen. They represented more than 70% of the farms. One pair of oxen wifl net
protect farmers from production Eosses in case of animal injuries, diseases or deaths
tn
such cases the only alternative is to borrow another animal, if available, or to rent for a
day or two, or to simply use hand tools. The majority of the farmers in that situation use
hand tools. For some fax-mers, this situation cari last a Iongtime before they find a
271
replacement. The frequency of these types of event placed a certain Jcvel of unre~iability in
the utilization of dl-ait animais for a given year.
How the technology was helping fa1 mers to fi.&3 their production needs WC+ the
main conccr~~ ofthe follow up. The aspects ineluded in the daily follow up were: type of
field operations pcrforrned, hoxw it MJ~S performcd in relation to the type and ar~~~rrt of
resourcx:s Ernrolved, the health and tare of,the animais and, the fceding system uscd to
maintain r:heir working cxpacilies. The draft animals ~~ere \\veighcd every n~onth to
evaluate the liveweight L.W variation during the rainy season.
TO gcrt the best perfol mance from the animal, fa]-,iners had to manage the oxcn
working time in relation to the climatic conditions. They mainly used their animals during
morning heurs. An average o,f 6 to 7 heurs of field\\i~nrk per day put a lot of physical stress
on the animaks, to the peint Ihal. they lest weight to covcr the le\\el of energy demand. The
main feeding system through rlatural graxing aAer work or during breaks was net Ienough
to s~;pply 1:he enerby necessary i-or the ta&:;. Feed x5’as supp~emcnted early in the nrc~rrzing
on a daily basis. The supplement las mainly made of rnaizc, millet, sorghos grain or
stover.
The bveweight of draft animais is a :necded paranreter in the process ofevalu;~ting
tke impact of animal traction. ‘The nrodel dcvelopcd herei:r cari he nsed to estinsate thc
liveweight through the measurcrnernt of the circumference or girth ‘The circumference and
the liveweight are highiy correlated.
212
9.1.3. Maximum working capacity
The maximum working capacity n’as evaluated through the amount of power
developed to pull different loads. The pull force PF deLeloped was measured along with
the speed used to pull the load. These two variables were combined to evaluate the pouer
output of the draft arnmals at different level of utilization intensities. Drafl animais were
classified by liveweight into traction groups and the power output of each group modeled
using a linear regression. The results showed that there was no significant dif3erence in the
power output among the t.raction groups, therefore one general mode1 could be used to
mode1 the power output.
Two interesting points on the power curve were anaIyzed to compare the groups’
performance: the optimum and maximum power. The amount of pull force PF
corresponding to the optimum power revealed the existence of a range of forc.es too
dif%cult to sustain. NI pairs of oxen used in the pulling tria1 showed a certain level of
unwilhngness to pull in that range and induced an underestimation of the corresponding
power output. At the maximum point, it was found that a pair of oxen could develop
around 50% of its liveweight into force. This level of pull was an extreme and was not
expec,ted on a tegular basis in farmers’ working conditions.
9.1.4. Energy required for fie!d oyerations
The difTerent implements in the repertoties at the far-m level were tested to evaluate
the level of energy reyuired for their daily utiiization. The amount of pull force generated
for moldboard and ridger plows corresponded closely to the optimum area of the power
.
_ __ -_._ .- -- -
-
273
curve. The bel ofworking intensity could be c~assifkd in the range of mediurn to heavy
job
‘The effective field capacity 1342 ofeach field operation was calculated a!ong v,zith
the number of working days during the rainy. Tsvo lcvels of probability were used: 5 years
out of 10 and 9 years out of 10.
‘The mount of energy delivered by the pair of oxen was evaluated using the
factxial mefhod with rh use of local cctef?icit:~~~fs, The atiimal energy balance was
detsr-mined by taking into account the encrgy used for working, for t+xlking, for
liveweight losses, and for the daily maintenance.
A main output of the energy balance was the calcu1ation of the feed ration fiorn
fized availabfe to fanners.
This methodology will nee:d more refinernents in the friture to better evah~atc the
infomation needed to build up the KBMS.
‘The first reconmendation is tIo improve the ~mthodc~logy of coflecting information
fi-or-n farnters, in par~iwlar, da.ta. on anjmal traction. The technology is wjdcly used in
developinq~ comfries and need more attention. ‘Thc other aspects to investigate furrher are
as follow.
The technology has brought about significant changes ta famers cropping
systems. l’he levei of utjlkation is still low as many famers me draft anirnals fx land
274
preparation only. There is room for improvement by designing better equipment to carry
out post-tillage field operations. Weeding is a serious problem which farmers are faced
every year due to fluctuation of available family labor during the season.
The use of draft animais for fieldwork is very attractive to farmers as their
utilization cari generate substantial revenue. By raising the level of utilization of the draft
animais, it Will be possible to increase the rate of adoption of the technology. One aspect
that will increase adoption of animal traction is the reduction in production costs resulting
fi-0111 mechanization.
- Draft animal selection
The environmental effects on working animais need to be hetter addressed to help
filrmers Select the right format of draft: animais and to fit the job to their pulling capacity.
The stress generated by the local climate has an effecf on their effrciency. The only existing
recommendation is to work morning heurs but farmers’ working needs go be)lond morning
hours.
The format and the age to star-t working drafl anima& are determinant for future
performance through better training. Farmers are not generally consistent on these two
factors because of lack of technical information.
. ”
- Feeding system
Meeting the feeding requirements of their working animais at the begInning of the
rx.iny season is a serious concern for fiirmers. The state of the art shows that agro-
_-. ._-..
ec.osystem zones offer w-y limited possibilities. It is important to focus on what is feasible
in farnws’ conditions. More research is nccded.
Different triais dto\\~~ed tllat draft animais are wtsting considerable amount of
energy in relation 1:o the high level of pulling intensity. Regular field working c.onditions
\\vill require pull force around the optimum power out@ Ievel. One aspect to comider in
the futcre is the improvemeut of the yoking systcm to haw animais pull in a more
cc-mforMAe attd &Sent way The loss ofthe willingness to pu11 by the ox-team does help
the farmer or improve the quality of the ficvldwot-k.
The major challenge is to fit the job to the tizorking animais. This cari be done. by
desi:gning and adapting imptements to bettcr fit the wor&.ing conditions. This must be done
in consitieration with the ergonomie factors intolved in 1 he implement desigII fo a.cc;:ount
for -!he animal format, the soi1 characteristics, and the u!:er.
- Learning pwiod
It is crucial to find ways to reduce the Iearning period. Fal-mers do not see the
bcnc:f3s ;If using animal traction because they do net master the technology. The lack of
skiils shows that most of Ihe advan2ages of using animal traction are offset. One
alterrrative is to reinforce the capacity of the farmers’ organization to train farmers on a
APPENDICES
i. .,
i
Ekh---
..__
_-.
276
regultir hasis. The classic approach of regional training ceriters depcnds to a large exilent
ou exm-nal fhding smmes
The evaIuation of work output and energy espenditure stilf requires highly
and labora~~:ory mcasurements is almost the only icchnique used. Tt should be possible to
develop pc,rtable heupensive semors to meatwre integrateti paranieters used in the tzrrergy
evaluation: draft, spced, heat, temperature, alid healt rates. ‘The instrumentation developed
by the French research institution CEEMA’T PWds to be completed with a computer
software capable of analyz.ing data recorded in the data lol;!,ger anti c.arrying out basic
power atld cnergy calculations.
- Ccbmputer progrnms
More computer programs dealing with animal traction management need to be
de\\ eloped. This Will errable information to be shared by difTèrent users through a common
database and to better simulate czxnditLxx of utilir,ation. The present expert system is
expected to be improved, to be ,tra!idated, and to be cstended to other regions in Sencyal,
and to 1 he M’est Afiican region.
AWENDJX A
i,k
c%J*ctives
-
-sx.-m-I-
A s p a r t o f t h e met:hod01.~3gy, t h e a n i m a l t r a c t i o n
qucstj onnaire i s conducted at t h e household l e v e l .in order
to evaluate t h e t y p e and number of farm e q u i p m e n t avai! able
to farmers. T h i s sulvey m u s t be comp.?eted b e f o r e any draft
a n d other energy evaluat.ion .is perfo:r-med. %-JC? maj n focL;s o f
t h e q,lest.ionnaire cleals with different tecIlni.cal aspects
re.lated ‘to t h e utiliza.tion c f an.imal drawn-equipment at t h e
housel-lold l e v e l . T h e o b j e c t i v e s of this survey are:
1. TO iC!entify t h e t y p e s o f equiprnent actually
i n v o l v e d i n ficld. operations for which m o r e d e t a i Ied
investigations are needed to determine power
.reqi2irements.
2 . TO evaluate the leaxning process for animal
t r a c t i o n u s e b y correlating perfor:mance a t t h e frirm
‘1 cvel against years of exper i.ence.
Mel:hQ(.1l31c>qy
--- .. . . .
-.-~.I~-I.LI~c
‘Cne t a r g e t grou;p j ncl.ude faxmers equj.pped w i t h 3nd
usj ng an:-mal tractio’n implemcnts.
T h e enumerator for thi s
survey must u n d e r g o technical traj ning i n i à e n t i fyj ng t.he
differ:ent: pieces o f Iequipment 2nd their u s e b y farrnerz.
The questionnai.re addresses a 1. imited number of
v a r i a b l e s :
1
-. . Agro-ecosystem zone :
There C?xiS’:S f ive agro-
eCOsÿsten1 z o n e s i n t h e area Iof s t u d y and formers locnted i n
two of them a r e considcred e f f e c t i v e a n i m a l t r a c t i o n users.
;!. T,ocation: B y l o c a t i o n i t i s understood v i l laqes
wi t h i n the i d e n t i fied ,agrrc--ecosystem :<one.
-1
- ,P kmseho3d:
I d e n t i f i c a t i o n o f the chief o f the
hoc sehol d.
4 <I fhmber of EamiLy workera: r-iunher of persans
belonging t o ‘the househo.3d and carrying out f i e l d
activi t:ies (labor f o r c e ) .
_
_
-.
_
_
--..ee-
_
-
-. -._.
_
_
_
-..-_.
__.-
5. Eq-uipment type: equipment must be identified
properly. A variety of equipment is being used by farmers.
T'he most common items are:
. Moldboard plow UCF 12,"
. Multipurpose beam .?W4RA
. MUltipurpOSe beam SINE
. Ridger EMCOT
. Seeder SUPER ECO
. Transport cart
6. Field aperation: this variable relates to the piece
of: equipment and its use in the field for land preparation,
weeding, etc...
7. Working tools: this variable is typical of the
multipurpose tool bars which cari be equipped with
moldboards, ridgers or cultivator tines.
8. Traction mode and age of the animal(s): Three
species are used: donkey, horse and cattle as source of
energy.
9. Yoke type: in relation to the field operations, the
activity and the species, yokcs are derigned to provide
efficient transfer of energy from the animal to the
implement . The type Will be described by its name and
length.
10. Acquisition of the equipment: information is
collected as regard to the Date of acquisition, Mode
(credit,
cash payment, gift, etc...), State (new, used),
Price if bought.
11. Origin: farmers use equipment of different trade
names (locally made or imported).
12. Management: what is the status of the equipment at
the household level: does the farmer own the equipment? 1s
he co-owner with a neighbor? 1s he borrowing, renting, long
term leasing, etc . . .
13. Actual condition af the equipment: 1s the
equipment in good condition (working parts must be
checked), does it need repair and maintenance ? Should it
be discarded ?, . . .
14. Repair and maj.nl:enancs: what repair has heen
performed and at what cost? Who took carre of the
maintenance?
15. Obse:rva t-i QX1S : important infoxnatj.on corning up
during t h e s u r v e y
-
.__.
..-_-l-_-_.-___-__------“-
--..--.
---_-
..-
-II
..
_ -
.”
_,-..
-._
_
---,m
,“*
I,.
.3$iaMmm
1
1
i
:
---I”m---
N 0 c
Qf
.
_--
0”
._
_.
-.-
-
f
282
APPENDIX A2: Draft Animal8 Follow-Wp Guidelines
1. Draft animals utilization
-
1. Utilization Timeframe (heurs per day)
1.1" Field work
- Tillage
- Seeding
- Weeding
- Harvesting
1.2. Transport
- Destination
- Distance
- Rate (paid or non-paid)
1.3. Custom work
- Number of days
- Type of field work
- Site
- Rental rate
2. Draft animals feeding system
2.1. Dry season
- Types of feed
- Daily Quantity (average)
- Origin
c
2 8 3
2.2. Rainy season
- T;ypes o f
feed
- - Jxily Querit,ity (average)
-- ‘Supplement (type, quant.ity)
-. IOrigi:n (Fl2r natural g r a z i n g , sample t h e
type of grasses a n d hays f o r i d e n t i f i c a t i o n
a n d eva.l.uate i ts prcr;3ortion i n t h e feeding
SySt:ep) ,
I I . Health
.e-.,--
1 . Health Proble~s antl diseases
- Dry season
2. Manifestations / Xdentification / Freqvency
3. Length of irmmobi3..izI-tt.i.on
4 . Gare
.5. Housing
--l.-l .- .---_ .-_-_.< --
8
III. ADDITIONAL OBSERVATIONS 284
Table Bl: Maximum draft tria1 (GI PI)
/June 25,1996
LOCATION :
Ijibelor Station
1
RAFT ANIMAL NAME :
Etoile Double Les
Right
Left
LENGTH (cm) :
85.00
82.00
166.00
162.00
357.00
326.00
8.00
7.50
XPERIENCE (yrs) :
5.00
5.00
200.00
IlAvg. Draft STD S.E
cv
Min
Max
Spe@d
Powcr
fY%-
Draft/LW
Powcr/LW
N
4r
N
N
ds
J/a
N/N
J/s-N
I/
0.00
0.00
1.28
0.00
0.00
0.00
46.89
11.37
18.06
196.20
294.30
1.17
303.81
0.04
0.05
259.67 517.58
50.82
9.44
9.03
392.40
588.60
1.14
590.04
O.ûR
n na
“SU,
9îû.37
62.90
;ü.ôû
6.a5
76‘4.ti‘u
98i.00
1.09
1003.21
0.14
0.15
/!
1086.95
91.53
16 . iô
8.44
951.00
1275.30
1.02
1108.69
0.16
0.17
/e 1409.32 1270.49
83.78
13.08
6.59
1079.10
1471.50
0.92
1168.65
0.13
0.17
65.92 i 3.83
4.68
1275.30
1563.60
0.90
1267.49
0.21
0.19
1531.44
66.51 I 9.05
4.34
1373.40
1667.70
0.75
1148.58
0.23
0.17
1866.84
155.59
27.08
8.33
1667.70
2452.50
0.53
989.43
0.28
0.15
2138 -09
145.97
23.37
6.83
1962 n n
.Y”
2452 -50
0.49
1047.66
0.32
0.16
2344.59
68.67 I 21.72
2.93
2256.30
2452.50
O,?O
79?.16
0.35
O.iî
2777.01
351.39 i 97.46
12.65
2452;50
3433.50
0.34
944.19
0.41
0.14
3531.60
I1
3531.60
3531.60
0.00
0.00
0.53
0.00
286
-
??
?
I
I
3
0
0 0 -
r-!
r-7
3 _-.l
m 1
??
3 3 ---
30000
3
2dmc\\10
^-=-ï-=1-===7=~---i-;====r:=~:---.-
?
.__,
?
?
.__-_--.I-._-^_-.”
?
???
?
_---_.__,
.
?
?
0
w
?
, . ri
?
?
r-c\\iI-l
???
r-mw
InLcr
07
??
??????
??
???
cn 0
c-i r-l r‘ r-
--..
OI 0
ri
WI--W ??
??
w w
??
??
.
??
?
?
?
w w
*
?
??
mm ??
????
??
m 0 -----.-.-
???
r-l
Dl
w 0
??
?????????????????????????????????
u-1
TJ’
IT‘I
w
l-4
??
r-i
Cl
??
?
II
?
0
, m ?
?
?
?
.
??
“l-----ll.-.ll-
0
-l._----.--.---
II
.
?
?
?
??
Ln 0
r-i
?
??
???
cn
PJ
0 0
CO
w CO
r-i
?
m
??
?
??
?
??
??
0
0 ??
0
0
??
??
r-4 0
C\\I
w
8-i 0
0
0 0
I
I
0
.
.
?
--
.
..--.
-------1
-....m---aE--^--.-.--
~_I
.
--
..-.
.
-.,.1.-1-.-1
<_-.“--m.-_-
-
---.-.--
_-.-,
-_--.-.sl.
--.----.--
--
__II,.
--_
i
Table B3: Maximum Draft tria: (Gi P3)
>ATRE :
;OCATION :
!ONE :
1RAFTANIMALNAME:
Amerique Gaston
SIDE :
XNGTH (cm) :
:IRCUMF'ERENCE (cm) :
7EIGHT (kg) :
iGE (yrs) :
EXPERIENCE (yrs) :
'ath Length (m) :
1.16 11386.26
Lt c;In
l-cl
ti
4.J
0 0
0 0
0 0
c c,
0 0
0 0
..“_.._
-----_”
.<“--
--“-,,--
v; hl
ci*
----
p
._-_
01 Q
to
0
-.1..-----.
0
00000
--“_.-I-~-.--<---l.l_
r’
0
0
-“--^------.
0
0
or-r-r0
-.-ll”.-~-~“-.l-~-.-^-~“----
<
.
0
0
0
CO
03
ru3
r-lwr-l
P
Cd
.
.
0
Lt-l
0
cn
0
r-lr-l
N
t-4
h)
.
*
0
0
0
u-l
w
0
P-l
u-l
.
.
_.--~
0
CT
0
0
0
0
0
.
.
---.--“<..-<
-.--__
._.
-_.-.
---
-----
l-
--
----
--
4 Y
QZ k ti
m
w
5
c 5
B
.
x
&È (u d
*
2
2
289
-L
;
:
I
;
:
: 3 t0
; )
; :
; n r:
-
, , 43cnl-icooo0U-Jl-i&030 I
, 34 300000,
(
, 34 0, --
r
OI VI
m
61 OI
‘9
a w
N
u-l 0
0
I
0 0
-ll--__-_
---.~_
.
.
__i u 2 r; Ef f2 z
r-i s)
0 0
r-4 Fi?
0 0
.
.
CJ :
0 0
4.
C\\I cs 0 0
.
.
7-I 0
0 0
0-r
0 0
u-l
0 0
cs
0 0
.
.
?
-.~..<-__._-I~-
??
1P.l u
?
?
.._
?
--_-
?
???
UJ 0-l
????
LIT
?
?
?
?
---.----
?
????
???
i‘-
???
r
r-i v-i
?
?
?
??
??
????
CO
???
r-l
0 m
??
?
??
?
??
?
0 N
?
r In
?
?
?
?
??
?
?
?
w as
=cr N
?
?
.--l---.l._l-_.-
?
0 0
?
?
-.I,---.--_-_
--
Table B7: Maximum draft tria1
(GB PI)
JATE I
L'3CATION :
ZONE :
JRAE'TANIMALNAME:
SIDE :
;ENGTH (cm) :
~IRCUMEZRENCE (cm) :
GIGHT (kg) :
iGE (yrs) :
3XPERIENCE (yrs) :
== v: a . .
-.A- z E jj r*
-xIi- Hfd u x- E-4 ?? ??? . .
????
---. ----
--
2
rb ti
ca 2:
-
:
292
3 3 ----_--._---.
?
?
c-r) r-t
?
?
w r
?
?
0
CU
?
?
r r-
?
u-l
?
0
?
??
w 0 ---
-__-___
j
.
1
ui w
293
i
:
r)
;
n
t
3
-
0
0 0
.
0
(v 0
*
0
M 0
.
(3
P-l m
*
0
m 03
.
----l_
Table B10: CiraEt animais LW and CIRCUMF
.---
----. --..-..
, -*_,- ..__~-
_---.--
lec.
-.--II
LW
ieng th
:1KIc1s
bx! .
IN
,ength
~IRCWMF
#
.~__
km)
(cm)
#
- - -
(kg)
-._- -__.
km)
( cm)
_-_--
(kg')
--<.--- .----_- -_---
1
166.00
96.00
135.00
38
273.00 loo.00'
152.00
2
173.00
90.00
135.00
39
277.00
95.00
159.00
3
186.00
89.00
137.00
40
278.00
101.00
15'4.00
4
186.00
94.00
138.00
41
279.00
99.00
158.00
5
190.00
90.00
1.39.00
42
282.00
95.00
159.00
6
.207.00
88.00
145.00
43
286.00
103.00
155.00
'7
208.00
99.00
141.00
44
290.00
99.00
16:3.00
3
212.00
95.00
144.00
45
294.00
98.00
15r3.00
3
213.00
93.00
142.00
46
297.00
103.00
160.00
1 0
1213.00
91.00
140.00
47
300.00
102.00
1 5 "7 . 0 0
11
215.00
89.00
140.00
48
300.00
99.00
158.00
1 2
222.00
91.00
146.00
49
301.00
92.00
162.00
1 :3
227.00
88.00
148.00
50
304.00
96.00
164.00
1 4
228.00
91.00
148.00
51
305.00
101.00
158.00
1!3
233.00
89.00
145.00
52
306.00
103.00
155.00
1 6
233.00
95.00
148.00
53
3 0'7 . 0 0
99.00
1 5 9 . 0 0
1'7
234.00
96.00
148.00
54
31?.00
97.00
1 6 0 . 0 i3
1 8
234.00
84.00
148.00
55
311.00
106.00
1 5 6 . 0 0
1 9
235.00
94.00
1.50.00
56
322.00
104.00
166.00
2 0
238.00
90.00
140.00
57
325.00
103.00
1 60 . 0 0
2 1
2 3 8 m 0 0
96.00
150.00
58
.32 6 . 0 0
82.00
1 162 * 0 0
2 2
2 4 0 ,, 0 0
9 3 . 0 0
149.00
59
.3 3 15 . 0 0
100.00
1 6 -? . 0 0
2 3
2,4 3 " 0 0
95.00
152.00
60
:3 3'7 . 0 0
100.00
1 6 9 . 0 0
2 4
2 ,4 4 " 0 0
98.00
150.00
61
:3 4 4 . 0 0
99.00
1'7 2 . 0 0
2 5
2.51 1 0 0
91.00
150.00
62
:3 4 6 . 0 0
97.00
1 6 9 . 0 0
2 6
2 .54 <I 0 0
92.00
l!jO.OO
63
:3 4 8 . 0 0
107.00
168.00
2 7
2 .5 5 a. 0 0
100.00
157..00
64
:3 5-7 . 0 0
85.00
166.00
2 8
2 5 6 m 0 0
95-00
1 5 3 m 0 0
65
:3 5 8 . 0 0
108.00
172.00
2 9
257.00
Ill.00
ltj2 ,, 00
66
3 6 4 . 0 0
105.00
1 6 6 . 0 0
3 0
2 61 . 0 0
93.00
154,.00
6 -7
3 7 :1 . 0 0
107.00
1-71.00
3 1
2 6 4 < 0 0
95.00
159,, 00
68
3 7 6 . 0 0 105.00
1 '7 Cl , 0 0
3;'
2 65 . 0 0
100.00
152.00
69
3 7 6 . 0 0
105.00
1-7 4 . 0 0
3 3
L! '7 1 . 0 0
l 0 0 * 0 0
156,,00
70
3 7 9 . 0 0
108.00
17 0 . 0 0
3 4
2'71 . 00
:1 0 2 * 0 0
156.,00
71
4 0 2 . 0 0 118.00
181.00
3 5
2' 7 2 . 0 0
97 < 00
153..00
72
442.00
114.00
188.00
3E
272.00
94.00
151.00
73
4 4 5 . 0 0 119.00
189.00
3-Y
- -
273.00
----,m
87.C!O
---.
152.00
- -
-_.l--ll_
--. -.,-
295
_.
Table B11: Regression analysis of the 18 draft oxen
--
.--.-
LW
JLENGTH
--m CIRCUMF
LW ks CIRCUMF
(cm)
139.00
20'7.00
88.00
145.00 Constant
-407.8(
213.00
91.00
140.00 Std Err of Y Est
19.4i
222.00
91.00
146.00 R Squared
0.8'
234.00
84.00
148.00 No. of Observations
18,01
238.00
90.00
140.00 Degrees of Freedom
16.0(
256.00
95.00
153.00 X Coeff
4.78
261.00
93.00
154.00 Std Err
0.45
264.00
95.00
159.00
272.00
97.00
153.00
279.00
99.00
158.00
290.00
99.00
163.00
301.00
92.00
162.00
304.00
96.00
164.00
326.00
82.00
162.00
344.00
99.00
172.00
357.00
85.00
166.00
376.00
- - -
105.00
170.00
APPENDIX C
SUMMARY OF DATA COLLECTED DtJR.iNG ON-STATION TRIALS
100
80
8’
20
!j/IiiI!l/ !/ii
0
Ii<,
4-n
M
J
J
A
S
0
N
Month
/
_.~ ._... --- ____ -.---_._-..-.--.. .----_-.-. ..- .-... .~. ---.-.-..
_-
-...
Figure C.1: Rainfall distribution at Djibelor Research Station (1996)
14
u 8
& 6
150
200
250
300
350 I
Date (Julien)
= O-2 cm
??
2-7cm
+ 7-15cm
3 15-26 cm
_~__-.-.--- _________ --~-- _-_._ _.- .._.____ - _____ -_-_. ._
Figure C.2: Soi1 water content at Djibelor Rescarch Station (tria1 site)
296
297
1m:plerrent
P1.o~ UCF
Wiclth (cm)
23.00
23.50
Depth (cm)
9.150
9.00
Total time (s)
li366.67
840.00
Effect. time (s)
'740.00
460.00
Efficiency (%)
54.15
54.76
Angle of pul:L
22.30
22.00
Yoke (cm)
90.00
120.00
Draft (ciaN)
PU 1. 1
ReCJ,
E)u11
Req.
130
120.53
160
148.35
180
166.89
160 148.35
150
139.08
160 148.35
120
111.26
150
139.08
190
176.1.6
170
157.62
170
1. 5 -7 . 62
150
139.08
210 194.71
140
129.8:L
190
176.3.6
170
157.62
180 166.89
180 166.89
190
176.1.6
150
139.08
180
166.89
160 148.35
200 185.44
180
166.89
170
l. 5 -1 * 62
180
166.89
190
176.16
150
139.08
180 166.89
150
139.08
160 148.35
160
148.3'5
150 139.08
170 157.62
160 148.35
160
148.35
180
166.89
150 139.Of3
190 176.16
160
148.35
190 176.16
170
157.62
190
176.16
190
176.16
210 194.-71
170 157.62
210
194.'71
160
148.313
210
194.71
170
157 .62
200 185.44
150
139.013
160 148.:35
180 166.8!3
190 176.16
170
157.62
220 203.98
170 157.62
170 157.62
160 1413.35
298
(Table Cl continued)
160
148.35
170
157.62
21 0
194.71
190
176.16
160
148.3.5
170
157.62
170
157.62
160
1.48.35
160
148.35
170
157.62
180
166.89
170
157.62
200
185.44
180
166.89
190
176.16
270
157.62
170
157.62
180
166.89
170
157.62
160
148.35
150
139.08
170
157.62
190
176.16
170
157.62
180
166.89
160
148.35
140
129.81
150
139.08
180
166.89
190
176.16
170
157.62
160
148.35
150
139.08
180
166.89
160
148.35
160
148.35
1'10
157.62
160
148.35
170
157.62
170
157,62
190
175.16
170
1137.62
200
185.44
170
157.62
190
176.16
170
157.62
210
194.71
170
157.62
200
185.44
160
148.35
Avg (daN)
178.91
162.92
166.00
153.91
Std dev.
21.38
29.47
11.05
10.25
CV(dec.)
0.12
0.18
0.07
0.07
Max (daN)
220.00
203.98
190.00
176.16
Min (daN)
120.00
0 . 0 0 1 4 0 . 0 0 1 2 9 . 8 1
299
Table Ci2 : 'SINE 9 Bioek 1 @ SWC-range 6 - 8 % g/g (Extract)
I:mpl.ement
:3-.tine SINE 9
Width [cm)
s 0 . 0 0
Cl1 . 5 0
Depth (cm)
9 . 3 3
9.00
Total time (s)
7 7 4 . 7 5
520.00
Effect. time (s)
4 9 0 . 0 0
350.00
Efficiency (%)
6 3 . 2 5
67 , 3 1
Angle of pull
9 c
L. L .‘50
22,50
Yzke (cm)
9 0 . 0 0
120.00
Draft '(kg)
Pu1 1
Req.
Pull
Req.
160
147.82
1. 0 0
92.39
1 1. 0
101.63
17 Cl
157.06
150
138.58
1.2 0
110.87
1 4 0
129.34
1. 5 0
138.58
160 147.82
1. 4 0
129.34
1 4 0
129.34
1 3 0
120.10
140
129.34
:1. 1 Cl
101.63
1 3 0
120.10
140
129.34
150
138.58
1. 5 0
138.58
160
147.82
110
101.63
160 147.82
1OC
92.39
140
129.34
12Cl
110.87
17 0
157.06
120
110.87
150 138.58
100
92.39
180
166.30
140
129.34
130 120.10
130
120.10
150 138.58
100
92.39
140
129.34
130
120.10
1 3: 0
120.10
100
92.39
140 129.34
130
120.10
160
147.82
120
110.87
130
120.10
:1 3 0
120.10
170
157.06
:1 1 0
101.63
150 138.58
:1 3 0
120.10
150
138.58
:1 2 0
110.87
140
129.34
:1 00
92.39
130
12c.10
:L 00
92.39
150
138.58
140
129.34
150 138.58
120
110.87
180 166.30
120 1 1 0 . 8 7
300
(Table C2 continued)
150 138.58
130
120.10
150 138.58
110
101.63
150 138.58
160
147.82
140 129.34
170
157.06
150 138.58
120
110.87
140 129.34
100
92.39
160 '14'7.82
120
110.87
160 147.82
160
147.82
130
120.10
100
92.39
130 120.10
110
101.63
160 147.82
150
138.58
150 138.58
110
101.63
150 138.58
110
101.63
170 157.06
140
129.34
170 157.06
120
110.87
150 138.58
100
92.39
130 120.10
140
129.34
150
138.58
150
138.58
140
129.34
140
129.34
170 157.06
110
181.63
160 147.82
130
120.10
130 120.10
160
147.82
150 138.58
110
101.63
160
147.82
120
110.87
160
147.82
120
110.87
Avg (daN)
149.09
137.74
124.91
115.40
Std dev.
14.18
13.10
19.34
17.87
CV(ciec.)
0.10
0.10
0.15
0.15
Max (daN)
180.00
166.30
170.00
157.06
Min (daN)
110.00 101.63
100.00
92.39
301
Table C.3 :
'ARARA Block 1 @ SWC-range 6 - 8 % g/g (Extrac:t)
Impl.ement
Ridger ARARA
Width (cm)
23.00
22.50
Depth [cm)
10.00
10.10
Total time (s)
'7 1 0 . 0 0
480.00
Effect. time (s)
3 9 0 . 0 0
340.00
Efficiency (%)
.54 .,!33
70.813
A-rgle C>:E pull
2 2 . 0 0
22.00
Yoke (cm)
13 0 . 0 0
120.00
D:raft r:
kg)
Pull
Req.
PU11
Req.
12 0
lli.26
13 0
120.53
120 ll.l.26
:1 7 0
157.62
130 120.53
:i40
129.81
100
92.72
l 2 0
111.26
100
92.72
1. 1 0
10 1 . 9 9
120 111.26
1. 1 0
101.99
140 129.81
1. 2 0
111.26
100
92.72
1. 4 0
129.81
130 120.53
1. 3 0
120.53
11.0 101.99
1. 2 0
111.26
140
129.81
100
92.72
130 120.53
1 2 0
111.26
120
111.26
130
120.53
150 139.08
14c
129.81
120 111.26
180
166.89
100
92.72
160
148.35
110 101.99
:170
157.62
130
120,53
:t 2 0
111.26
160
148.35
:170
157.62
110
101.99
:1. 3 0
120.53
150
139.08
:i. 6 0
148.35
130 120.53
260
148.35
140 129.81
120
111.26
150
139.08
130
120.53
150
139.08
1.30
120.53
130 120.53
1. 50
139.08
150
139.08
3. 7 0
157.62
160 148.35
1.30
120.53
150
139.08
1.50
139.08
130
120.53
1 4 0
129.81
302
(Table C3 continued)
100
92.72
150
139.08
120 111.26
170
157.62
100
92.72
140
129.81
150 139.08
130
120.53
130
120.53
110
101.99
110 101.99
140
129.81
130 120.53
110
101.99
130 120.53
140
129.81
130 120.53
150
139.08
120
111.26
140
129.81
140
129.81
130
120.53
170 157.62
140
129.81
110 101.99
160
148.35
130
120.53
110
101.99
160 148.35
140
129.81
160
148.35
160
148.35
110
101..99
130
120.53
150
139.08
160
148.35
120 111.26
15 0
139.08
130 120.53
120
111.26
120 111.26
180
166.89
180 166.89
160
148.35
120 111.26
130
120.53
150 139.08
110
101.99
110
101.99
100
92.72
Avy (daN)
130.18
118.55
138.36
128.29
Std dev.
19.49
24.01
20.61
19.11
CV(dec.)
0.15
0.20
0.15
0.15
Max (daN)
180.00
166.89
180.00
166.89
Min (daN)
100.00
0.00
100.00
92.72
303
Table CU: 'U"S' 10” Hock II @ SWC-range 8- 10 % g,'g I(Extract)
Impl~?rnerlt
I?l.ow 1JCF
Width (cm)
23.43
24.44
Depth (cm)
12.11
9.33
Tcltal. time (s)
2050.. 00
1033.00
Ef feet:. time (s)
1374.00
710.00
Efficiency (3)
67.02
68.73
Angle of pull
2.2 II 50
22.50
Yoke (cm)
9 0 ., 0 0
1.20.00
Draft (daN)
Pu :Il
Req.
Pull
Req.
160.00
147.82
120.00 110.87
160.00
147.82
150.00 138.58
150. 00
138.58
160.00 147.82
170.00
157.06
170.00 157.06
1 s 0 . 0 0
138.58
160.00 147.82
160.00
147.82
150.00 138.58
150.00
138.58
160.00 147.82
150.00
138.58
160.00 147.82
150*00
138.58
140.00 129.34
150.00
138.58
140.00 129.34
160.00
147.82
150.00 138.58
200.00
184.78
140.00 129.34
150.00
138.58
160.00 147.82
150.00
138.58
150.00 138.58
140.00
129.34
150.00 138.58
l.50.00
138.58
150.00 138.58
140.00
129.34
130.00 120.10
:150.00
138.58
150.00 138.58
160.00
147.82
150.00 138.58
l70.00
157.06
150.00 138.58
150.00
138.58
150.00 138.58
150.00
138.58
160.00 147.82
140.00
129.34
160.00 147.82
160.00
147.82
140.00 129.34
l.50.00
138.58
1~40.00
129.34
l80.00
166.30
140.00 129.34
;160.00
147.82
160.00 147.82
1180.00
166.30
140.00 129.34
170.00
157.06
1150.00
138.58
3. 5 0 . 0 0
138.58
1:30.00
120.10
304
. .
(Table C4 continued)
160.00
147.82
150.00 138.58
160.00
147.82
150.00 138.58
180.00
166.30
140.00 129.34
160.00
147.82
150.00 138.58
160.00
147.82
130.00 120.10
160.00
147.82
160.00 147.82
170.00
157.06
150.00 138.58
180.00
166.30
160.00 147.82
160.00
147.82
160.00 147.82
160.00
147.82
130.00 120.10
160.00
147.82
120.00 110.87
160.00
147.92
120.00 110.87
150.00
138.58
140.00 129.34
160.00
147.82
150.00 138.58
180.00
166.30
130.00 120.10
170.00
157.06
130.00 120.10
160.00
147.82
140.00 129.34
170.00
157.06
150.00 138.58
160.00
147.82
150.00 138.51:
170.00
157.06
140.00 129.34
160.00
147.82
140.00 129.34
160.00
147.82
150.00 138.58
170.00
157.06
130.00 120.10
150.00
138.58
150.00 138.58
160.00
147.82
150.00 138.58
180.00
166.30
140.00 129.34
160.00
147.82
160.00 147.82
Avg (daN)
153.22
141.56
145.52 134.44
Std dev.
19.60
18.11
11.82
10.132
cv (%)
12.79
12.79
8.12
8.12
Max (daN)
200
184.78
170.00 157.06
Min (daN)
90
83.15
120.00 110.87
305
Table C!i:
"SINE 9 Block II @ SWC-range 8- 10 % g/g (Extract)
Implemerh
C u l t i v a t o r H S
W i d t h ( c m )
54 . 50
5 2 . 0 0
Depth (cm)
10 . ii! 5
9 . 0 0
T o t a l . tiine (s)
1 3 4 2 . 0 0
9 5 4 . 0 0
E f f e c : t i v e t i m e (s)
9 5 6 * 0 0
9 1 6 . 0 0
Effic:iency (8)
7 1 e S! 4
9 6 . 0 2
A n g l e o f p u l l
2 2! m 5 0
2 2 . 5 0
Y o k e ( c m )
90 . 0 0
1 2 0 . 0 0
D r a f t (CiaN)
P u l l
R e q .
P u l l
R e q .
1 4 0 * 0 0
1 2 9 . 3 4
100.00
92.39
1 3
1 2 0 . 1 0
8 0 . 0 0
7 3 . 9 1
? ? ? ? ? ? ?
1 5 0 . 0 0
1 3 8 . 5 8
1 0 0 . 0 0
9 2 . 3 9
1 4 0 . 0 0
3 2 9 . 3 4
1 1 0 . 0 0
7 3 . 9 1
3 4 0 . 0 0
1 2 9 . 3 4
1 0 0 . 0 0
9 2 . 3 9
1 4 0 . 0 0
129.34
9c. 00
8 3 . 1 5
1 5 0 . 0 0
1.38.58
1 0 0 . 0 0
8 3 . 1 5
1 4 0 . 0 0
1 2 9 . 3 4
1 0 0 . 0 0
9 2 . 3 9
1 5 0 . 0 0
1 3 8 . 5 8
1 1 0 . 0 0
7 3 . 9 1
14 0 . 00
1 2 9 . 3 4
1.0 0 . 0 0
9 2 . 3 9
14 0 . 0 0
l. 2 9 . 3 4
9 0 . 0 0
8 3 . 1 5
1. 4 0 . 0 0
l 2 9 . 3 4
3. 0 0 . 0 0
8 3 . 1 5
3. 3 0 . 0 0
l 2 0 . 1 0
3. 0 0 . 0 0
9 2 . 3 9
15 0 . 0 0
1 3 8 . 5 8
1 0 Cl . 0 0
9 2 . 3 9
3. 0 0 . 0 0
9 2 . 3 9
1. 0 0 . 0 0
9 2 . 3 9
14 0 . 0 0
l29.34
1. 0 0 . 0 0
9 2 . 3 9
3.3 0 . 0 0
1 2 0 . 1 0
1 0 0 . 0 0
9 2 . 3 9
1.0 0 . 0 0
9 2 . 3 9
3.2 C) . 0 0
8 3 . 1 5
1. 5 0 . 0 0
1 3 8 . 5 8
^ 0 0 . 0 0
9 2 . 3 9
l-2 cl . 0 0
1 1 0 . 8 7
:_ 10 . 0 0
7 3 . 9 1
:1 4 0 . 0 0
:129.34
:. 0 0 . 0 0
9 2 . 3 9
l4 0 . ‘0 0
1 2 9 . 3 4
9 0 . 0 0
8 3 . 1 5
:100.00
9 2 . 3 9
l 0 0 . 0 0
9 2 . 3 9
l2 10 . 0 0
1 1 0 . 8 7
l 1 0 . 0 0
7 3 . 9 1
1 2 0 . 0 0
1 1 0 . 8 7
l 0 0 . 0 0
9 2 . 3 9
: t o o . o o
9 2 . 3 9
:t 1 0 . 0 0
7 3 . 9 1
1 1 0 . 0 0
1 0 1 . 6 3
:t 0 0 . 0 0
9 2 . 3 9
: t o o . o o
9 2 . 3 9
8 0 . 0 0
7 3 . 9 1
: t o o . o o
9 2 . 3 9
:t 0 0 . 0 0
8 3 . 1 5
1 3 0 . 0 0
1 2 0 . 1 0
8 0 . 0 0
7 3 . 9 1
386
(Table C5 continued)
120.00
110.87
100.00
92.39
120.00
110.87
80.00
73.91
100.00
92.39
100.00
92.39
140.00
129.34
100.00
83.15
130.00
120.10
100.00
92.39
100,00
92.39
120.00
110.87
120.00
110.87
120.00
110.87
100.00
92.39
100.00
92.39
110.00
101.63
90.00
83.15
100.00
92.39
100.00
92.39
90.00
83.15
100.00
92.39
110.00
101.63
110.00
73.91
120.00
110.87
120.00
83.15
90.00
83.15
90.00
83.15
100.00
92.39
100.00
92.39
120.00
110.87
100.00
92.39
100.00
92.39
90.00
83.15
120.00
110.87
100.00
92.39
110.00
101.63
100.00
92 * 39
110.00
101.63
100.00
92.39
110.00
101.63
100.00
92.39
Avg (daN)
111.91
103.40
100.00
87.32
Std dev.
18.98
17.53
9.29
8.64
cv (SI
16.96
16.96
9.29
9.90
Max (daN)
150.00
138.58
120.00
110.87
Min (daN)
70.00
64.67
80.00
73.91
307
Table CE; : 'ARARA Black II @ SWC-range t3- 10 % g/g (Extra&)
1mplemen.t
Ridger -MaRA
Width (cm)
2 1 * 57
22.50
Depth (cm)
11.00
11.33
Total time (s)
1120 0, 00
91.0.00
Effective time (s)
624 (8 00
690.00
Efficiency (%)
55,.71
75.82
Angle of pull
'L! 2 . 0 0
2'2.00
Ycike (cm)
90, 00
120.00
Draft (daN)
Pull
Req.
Pull
Req.
100.00
92.72
160.00
148.35
80.00
74.17
150.00
139.08
90.00
83.45
170.00
157.62
100.00
92.72
150.00 139.08
100.00
92.72
180.00 166.89
90.00
83.45
200.00 185.44
120.00
111.26
200.00 185.44
90.00
83.45
130.00
120.53
lC~O.00
92.72
200.00 185.44
110.00 101.99
140.00 129.81
120.00
111.26
150.33
139.08
100.00
92.72
160.00
148.35
120.00 111.26
150.00
139.08
140.00 129.81
160.00 148.35
100.00
92.72
150.00 139.08
90.00
83.45
130.00
120.53
100.00
92.72
140.00
129.81
90.00
83.45
140.00
129.81
90.00
83.45
150.00
139.08
120.00 111.26
130.00
120.53
80.00
74.17
120.00 111.26
140.00 129.81
110.00 101.99
150.00 139.08
130.00
120.53
150.00 139.08
130.00 120.53
100.00
92.72
100.00
92.72
130.00 120.53
140.00
129.81
150.00 139.08
140.00 129.81
120.00 111.26
100.00
92.72
110.00 101.99
110.00 101.99
(Table CG continued)
130.00
120.53
120.00
111.26
100.00
92.72
130.00
120.53
140.00
129.81
100.00
92.72
130.00
120.53
110.00
101.99
120.00
111.26
130.00
120.53
1'20.00
111.26
140.00
129.81
120.00
111.26
130.00
120.53
110.00
101.99
140.00
129.81
120.00
111.26
140.00
129.81
120.00
111.26
150.00
139.08
140.00
129.81
110.00
101.99
130.00
120.53
120.00
111.26
130.00
120.53
120.00
111.26
100.00
92.72
150.00
1.39.08
110.00
101.99
120.00
111.26
130.00
120.53
140.00
129.81
140.00
129.81
140.00
129.81
120.00
111.26
130.00
120.53
130.00
120.53
150.00
139.08
130.00
120.53
120.00
111.26
140.00
129.81
120.00
111.26
140.00
129.81
120.00
111.26
130.00
120.53
160.00
148.35
100.00
92.72
140.00
129.81
120.00
111.26
120‘00
111.26
110.00
101.99
100.00
92.72
Avg [YdaN)
118.50
109.87
137.64
127.61
Std dev.
17.54
16.27
23.35
21.65
cv (81)
14.80
14.80
16.97
16.97
Max (daN)
150.00
139.08
200.00
185.44
Min (daN)
80.00
74.17
100.00
92.72
309
Table C7: 'UCF 10" I3loc:k 111 @ SWC-range 10 - 13 % g/g (Extra&)
i'mplement
Plow UCF
Width (cm)
22.75
22.50
Depth (cm)
Il.80
12.00
Total time (s)
2308.00
2472.00
Effect. time (s)
1.262.00
1634.00
Efficiency (8)
54.68
66.10
AngILe of pull
22.00
2.2.00
Yoke (cm)
90.00
120.00
Draft (ciaN)
Pull
Req.
Pull
Req.
:14 0 ,, 0 13
129.81
140.00 129.81
150 * 00
139.08
150.00 139.08
:t 5 0 0 0 0
139.08
150.00 139.08
l 6 0 0 0 0
148.35
160.00 148.35
:1 6 0 . . 0 0
148.35
150.00 139.08
1 8 0 ., 0 0
166.89
150.00 139.08
190.00
176.16
140.00 129.81
3. 6' 0 . 0 0
148.35
150.00 139.08
2 0 0 . 0 0
185.44
150.00 139.08
1. 5 0 . 00
Z-39.08
140.00 129.81
1 6 0 . 0 0
3.48.35
:i50.00 139.08
1. 7 0 . 0 0
157.62
I.40.00 129.81
160.00
1.48.35
l.50.00 139.08
160.00
l.48.35
3.50.00 139.08
15o.oc1
139.08
1.40.00 129.81
150.00
139.08
140.00 129.81
160.00
148.35
140.00 129.81
160.00
148.35
140.00 129.81
150.00
139.08
140.00 129.81
150.00
139.08
150.00 139.08
1.50.00
139.08
150.00 139.08
160.00
148.35
140.00 129.81
150.00
139.08
170.00 157.62
17~0.00
157.62
150.00 139.08
180.00
166.89
150.00 139.08
160.00
148.35
l50.00 139.08
200.00
185.44
150.00 139.08
150.00
139.08
170.00 157.62
160.00
148.35
160.00 148.35
170.00
157.62
140.00 129.81
310
(Table C7 continued)
160.00
148.35
140.00 129.81
170.00
157.62
150.00 139.08
160.00
148.35
140.00 129.81
160.00
148.35
150.00 139.08
120.00
111.26
150.00 139.08
130.00
120.53
140.00 129.81
150.00
139.08
150.00 139.08
160.00
148.35
130.00 120.53
170.00
157.62
140.00 129.81
150.00
139.08
150.00 139.08
170.00
157.62
150.00 139.08
150.00
139.08
140.00 129.81
170.00
157.62
150.00 139.08
150.00
139.08
150.00 139.08
200.00
185.44
140.00 129.81
160.00
148.35
150.00 139.08
150.00
139.08
160.00 148.35
150.00
139.08
160.00 148.35
140.00
129.81
150.00 139.08
200.00
185.44
150.00 139.08
180.00
166.89
160.00 148.35
160.00
148.35
160.00 148.35
160.00
148.35
160.00 148.35
160.00
148.35
150.00 139.08
160.00
148.35
170.00 157.62
160.00
148.35
160.00 148.35
180.00
166.89
150.00 139.08
Avg (daN)
154.72
143.45
146.67 135.99
Std dev.
15.99
14.82
8.46
7.84
CV(dec.)
9.93
9.93
5.68
5.68
Max (daN)
200.00
185.44
170.00 157.62
Min (daN)
120.00
111.26
130.00 120.53
311
Table C8 :: 'SINE 9 Block III @ SWC-range 10 - 13 % g/g {Extract)
Implement:
3-tine SINE 9
Width (cm)
52.50
52.00
Depth (cm)
10.00
10.00
Total time (s)
1480.00
1:332.00
Effect. time (s)
1116.00
986.00
Efficienc:y (8)
7 5 . ,4 1
74.02
Angle of pull
2 2 . 5 0
22 . 50
Coke (cm)
9
:L 2 0 0 0
.
???????
Draft :daN)
Pull
Req.
Pull
Req.
1. 3 0 . 0 0
120.10
80.00
73.91
1.2 0 . Cl 0
110.87
60.00
55.43
14 0 . 0 0
:129.34
70.00
64.67
1.2 0 . 0 0
110.87
80.00
73.91
1. 3 0 . 0 0
120.10
90.00
83.15
1. 4 0 . 00
129.34
80.00
73.91
14 0 . 0 0
129.34
90.00
83.15
1 5 0 . 0 0
138.58
9c.00
83.15
1 5 0 . 0 0
138.58
80.00
73.91
1 50 . 0 0
138.58
90.00
83.15
1 3 Q . 0 0
120.10
90.00
83.15
1 4 0 . 0 0
1L29.34
9Q.ûO
83.15
13 C) . 0 0
720.10
110.00
101.63
1 30 . 00
120.10
80.00
73.91
12 0 . 0 0
110.87
:100.00
92.39
14 0 . 0 0
129.34
120.00
110.87
1 50 . 0 0
138.58
:120.00 110.87
15 0 . 0 0
138.58
:100.00
92.39
14 0 . 0 0
129.34
l.10.00
101.63
14 0 . 0 0
1.29.34
120.00 110.87
14 0 . CIO
1.29.34
110.00
101.63
130.00
120.10
120.00 110.87
1 5 CI . CI 0
1.38.58
110.00 101.63
130.00
120.10
:too.oo
92.39
130.00
120.10
100.00
92.39
12 01 . 0 0
110.87
80.00
73.91
120.00
110.87
:1 0 0 . 0 0
92.39
120.00
110.87
:100.00
92.39
140.00
129.34
100.00
92.39
140.00
129.34
80.00
73.91
312
(Table C8 continued)
150.00
138.58
120.00
110.87
130.00
120.10
90.00
83.15
140.00
129.34
80.00
73.91
140.00
129.34
110.00
101.63
140.00
129.34
110.00
101.63
150.00
138.58
100.00
92.39
14'0.00
129.34
110.00
101.63
140.00
129.34
110.00
101.63
110.00
101.63
100.00
92.39
140.00
129.34
100.00
92.39
200.00
184.78
80.00
73.91
150.00
138.58
100.00
92.39
140.00
129.34
110.00
101.63
130.00
120.10
110.00
101.63
140.00
129.34
110.00
101.63
140.00
129.34
90.00
83.15
120.00
110.87
90.00
83.15
130.00
120.10
110.00
101.63
120.00
110.87
80.00
73.91
140.00
129.34
110.00
101.63
140.00
129.34
110.00
101.63
120.00
110.87
90.00
83.15
130.00
120.10
100.00
92.39
140.00
129.34
100.00
92.39
140.00
129.34
130.00
120.10
110.00
101.63
100.00
92.39
110.00
101.63
90.00
83.15
Avg (daN)
130.45
120.52
97.76
90.32
Std dev.
13.22
12.21
14.53
13.43
CV(%)
9.67
9.67
14.80
14.80
Max (daN)
200.00
184.78
130.00
120.10
Min (daN)
110.00
101.63
60.00
55.43
,313
Table C9: *AR.A,RA B~OCIC 11x @ SWC-range 10 - 13 % g/g W:xtract)
Implement
Ridger ?!iRARJi
Wi.dth (C:m)
23.00
23.00
Depth (cm)
10.33
1.0.00
Total time (s)
1520.00
1238.00
Effect. time (s)
fi 90 . 0 0
920.00
Efficiency (%)
El8 . 5 5
74.31
Angle of pull
22.00
22.00
Yoke (cm)
90.00
120.00
Draft (daN)
PL11 1
Req.
Pull
Req.
83.45
100.00
92.72
92.72
120..00
111.26
101.99
10 0 * 0 0
92.72
11.1.26
10 0 * 0 0
92.72
92.72
1 0 0 * 0 0
92.72
92.72
1 5 0 S 0 0
139.08
111.26
1 2 0 (I 0 0
111.26
l-1.26
10 0 * 0 0
92.72
139.08
9 0 ., 0 0
83.45
139.08
10 0 I) 0 0
92.72
139.08
10 0 <) 0 0
92.72
139.08
1 0 0 ., 0 0
92.72
92.72
1 0 0 I) 0 0
92.72
129.81
1 1 0 <) 0 0
101.99
120.53
100.00
92.72
111.26
!3 0 ., 0 0
83.45
101.99
1 0 0 <) 0 0
92.72
92.72
1 0 0 <) 0 0
92.72
1X1.26
1 0 0 <, 0 0
92.72
92.72
12 0 * 0 0
111.26
101.99
9 0 ,. 0 0
83.45
129.81
1 0 Q * 0 0
92.72
111.26
!3 0 , 0 0
83.45
111.26
140.00
129.81
i32.72
100 .oo
92.72
120.53
1:30.00
120.53
132.72
100.00
92.72
92.72
80.00
74.17
111.26
100.00
92.72
192.72
100.00
92.72
3 1 4
3.
.>w
(Table C9 continued)
130.00
120.53
100.00
92.72
.<,
100.00
92.72
90.00
83.45
100.00
92.72
90.00
83.45
100.00
92.72
90.00
83.45
,,
10.0.00
92.72
90.00
83.45
100,00
92.72
90.00
83.45
100.00
92.72
100.00
.1
92.72
110.00
101.99
100.00
92.72
100.00
92.72
100.00
92.72
100.00
," *
92.72
110.00
101.99
100.00
92.72
100.00
92.72
100.00
92.72
80.00
74.17
100.00
92.72
100.00
92.72
100.00
92.72
90.00
83.45
90.00
83.45
100.00
92.72
100.00
#~
92.72
90.00
83.45
90.00
83.45
100.00
92.72
120.00
111.26
90.00
83.45
*
100.00
92.72
130.00
120.53
100.00
92.72
100.00
92.72
100.00
92.72
90.00
83.45
110.00
101.99
90.00
83.45
120.00
111.26
100.00
92.72
100.00
92.72
100.00
92.72
100.00
92.72
90.00
83.45
100.00
92.72
90.00
83.45
110.00
101.99
100.00
92.72
Avg (daN)
109.88
101.98
100.00
9 2 . 7 2
Std dev.
16.16
14.99
13.19
12.23
C'V ( 8 )
14.60
14.60
13.10
13.10
Max (daN)
150.00
139.08
150.00
139.08
Min (daN)
90.00
83.45
80.00
74*17
Table. ClO: Output taxample of the CEEMAT data processing
system (WCF' plowr -Bugutub rice field)
Tl
T2
T3
T4
T5
#
hour:min Second Draft:
Radar
hr ::mn
S
daN
pulse
0103.
12:36.
0 0 0 0 .
108.5
53.00
0103.
12:36.
002.0
131.9
138.0
0103.
12:36.
004.0
102.7
162.0
0103.
12:36.
006.0
135.8
126.0
0103.
12:36.
008.0
132.0
122.0
0103.
12:36.
010.0
131.8
099.0
0103.
12:36.
012.0
158.3.
130.0
0103.
12:36.
014.0
122.3
115.0
0103.
12:36.
016.0
086.0
146.0
0103.
12:36.
018.0
152.6
147.0
0103.
12:36.
020.0
108.5
148.0
0103.
12:36.
022.0
094.4
140.0
0103.
12:36.
024.0
165.8
149.0
0103.
12:36.
026.0
117.2!
134.0
0103.
12:36.
028.0
26.84
28.00
0103.
12:36.
030.0
23.49
0.000
0103.
12:36.
032.0
1.195
0.000
0103.
12:36.
034.0
2.690
0.000
0103.
12:36.
036.0
17.38
18.00
13103.
12:36.
038.0
23.62
48.00
'3103.
12:36.
040.0
079.2
08.00
J3103.
12:36.
042.0
099.4
48.00
13103.
12:36.
044.0
148.Î
125.0
3103.
12:36.
046.0
1 0 1 . 0
096.0
0103.
12:36.
048.0
122.3
117.0
0103.
12: 36.
050.0
097.4
143.0
0103.
12:36.
052.0
099.8
196.0
0103.
12:36.
054.0
092.0
104.0
0103.
12:36.
056.0
088.5
082.0
0103.
12:36.
058.0
085.2
102.0
0103.
12:37.
0000.
104.1
127.0
0103.
12: 37.
002.0
131.8
134.0
0103.
12:3-l.
004.0
105.5
123.0
0103.
12:37.
006.0
136.'7
181.0
3103.
12:37.
008.0
103.,4
164.0
316
(Table Cl0 continued)
0103.
12:37,
010.0
132.7
139.0
0103.
12:37.
012.0
090.8
144.0
0103.
12:37.
014.0
117.8
143.0
0103.
12:37.
016.0
121.8
166.0
0103.
12:37.
018.0
096.8
138.0
0103.
12:37.
020.0
116.1
088.0
0103.
12:37.
022.0
158.8
113.0
0103.
12:37.
024.0
166.1
100.0
0103.
12:37.
026.0
169.7
40.00
0103.
12:37.
028.0
138.9
09.00
0103.
12:37.
030.0
108.3
094.0
0103.
12:37.
032.0
133.5
59.00
0103.
12:37.
034.0
146.6
144.0
0103.
12:37.
036.0
127.2
094.0
0103.
12:37.
038.0
180.8
110.0
0103.
12:37.
040.0
187.5
115.0
0103.
12:37.
042.0
142.8
092.0
0103.
12:37.
044.0
171.0
0.000
0103.
12:37.
046.0
165.7
0.000
0103.
12:37.
048.0
120.0
0.000
0103.
12:37.
050.0
130.1
0.000
0103.
12:37.
052.0
076.0
0.000
0103.
12:37.
054.0
103.1
08.00
0103.
12:37.
056.0
123.1
082.0
0103.
12:37.
058,o
115.4
073.0
0103.
12:38.
0000.
116.0
65.00
0103.
12:3a.
002.0
079.0
137.0
0103.
12:38.
004.0
127.4
072.0
0103.
12:38.
006.0
130.5
090.0
0103.
12:38.
008.0
136.4
089.0
0103.
12:38.
010.0
162.0
071.0
0103.
12:38.
012.0
170.8
086.0
0103.
12:38.
014.0
134.1
085.0
0103.
12:38.
016.0
136.9
63.00
0103.
12:38.
018.0
147.9
39.00
0103.
i2:38.
020.0
07.63
0.000
0103.
i2:38.
022.0
3.016
0.000
0103.
12:38.
024.0
0.872
0.000
0103.
12:38.
026.0
43.77
0.000
317
(Table Cl0 continued)
0103.
12:38.
028.0
091.7
53.00
0103.
12:38.
030.0
099.9
114.0
0103.
12:38.
032.0
69.18
120.0
0103.
12:38.
034.0
090.4
114.0
0103.
12:38.
036.0
121.9
100.0
0103.
12:38.
038.0
096.5
100.0
0103.
12:38.
040.0
093.8
081.0
0103.
12:38.
042.0
106.9
45.00
0103.
12:38.
044.0
089.3
092.0
0103.
12:3#8.
046.0
104.8
101.0
0103.
12:38.
048.0
2.494
5.000
0103.
12:38.
050.0
10.81
0.000
0103.
12:38.
052.0
0.934
33.00
0103.
12:38.
054.0
0.414
63.00
0103.
12:38.
056.0
32.81
3.000
0103.
12:38.
058.0
078.6
132.0
C103.
12:39.
0000.
113.5
176.0
0103.
12:39.
002.0
107.1.
182.0
0103.
12:39.
004.0
091.0
189.0
0103.
12:39.
006.0
075.9
183.0
0103.
12:39.
008.3
080.5
198.0
0103.
12:39.
010.0
088.51
184.0
0103.
12:39.
012.0
0 92 . 6
174.0
O103.
12:39.
014.0
110.8
167.0
0103.
12*3Q
*.. 2.
016.0
105.3
201.0
0103.
12.39
. . 2.
018.0
086.7
154.0
0103.
12:39.
020.0
114.5
174.0
0103.
12:39.
022.0
0 9 2 . 6
191.0
0103.
12.39
. . 4.
024.0
077.6
141.0
0103.
12:39.
026.0
080.8
144.0
0103.
12:39.
028.0
103.4
170.0
0103.
12::39.
030.0
109.3,
169.0
0103.
12::39.
032..0
092.3
108.0
0103.
12:39.
034..0
078.6
108.0
0103.
12::39.
036.0
096.2
154.0
0103.
12 : :3 9 .
038-O
089.9
121.0
0103.
12::39.
040,.0
090.2
127.0
0103.
12 : :3 9 .
042,.0
084.:3
176.0
3 1 8
(Table Cl0 continued)
0103.
12:39.
044.0
107.4
111.0
0103.
12:39.
046.0
67.62
130.0
0103.
12:39.
048.0
083.2
147.0
0103.
12:39.
050.0
071.3
140.0
0103.
12:39.
052.0
070.7
127.0
0103,
12:39.
054.0
60.34
120.0
0103.
12:39.
056.0
51.88
088.0
0103.
12:39.
058.0
65.32
084.0
0103.
12:40.
0000.
56.15
076.0
0103.
12:40.
002.0
071.2
101.0
0103.
12:40.
004.0
073.5
57.00
0103.
12:40.
006.0
30.47
31.00
0103.
12:40.
008.0
45.81
0.000
Storage
.L *..
c
--. s . . . . . ~
unit
P-
i
i
1
1P/i 1 :...:. : . . . . .- . ..-. : . . . . . . . . .7
R P M
Fuel cmmmption
Pull (3 pQiow
-T
--r-- 7-
A
Fgm EDLOG
Figure C3: CEEMAT data recording and processing _
APPENDIX D
APPENDJX D
FEEDLNG SYSTEMS .AVAILABLE TN M’EST-AFRICA
Table Dl: FU for msiintenance FUM of working cattle
-----
--=zzz.~----.-
11-----1---
Weight
Feeding Units (FU)
&e:)
y - y - - - - - - -
1 1 _ - - - - .-._----
- -
200
1.9’5
250
2.30
300
2.60
35’0
2.9:5
4010
3 . 2 5
450
3.55
500
3 . 8 5
--I
-=e-
..I-~-~~-
Source: CEEMA?‘, 1975
320
321
Table D2: Composition of some African feeds (As-Fed Basis)
-
Feeds
Dh
100 % DM bas,is
T-FU
-
-
-
-
n,
P
C a
Green grass pasture
Sudan (Andrononon gavanus)
18.0
6.2
0.02
0.35
0.66
Guinea (Panicum ma.ximum)
16.6
6.7
0.27
0.45
0.57
Elephant (Pennisetum purnureum)
18.5
6.9
0.26
0.60
0.58
Digitaria
15.3
9.4
0.23
0.49
0.60
Mixed fi-esh shoots
17.7
18.2
0.45
0.29
0.77
Mixed Young pasture
24.8
5.8
0.24
0.31
0.61
Mixed Young pasture + legumes
25.6
7.5
0.10
0.87
0.72
- _ .
hier pasture (standing hay)
suclan #1
29.7
4.2
0.12
0.34
0.57
Sudan#2
34.6
3.5
0.15
0.39
0.63
GUiJSfX
34.7
3.6
0.22
0.54
0.57
Digitaria
32.7
2.5
0.33
0.64
0.54
Mixed - average quality
30.8
3.2
0.18
0.30
0.38
Mixed - pour quality
39.6
2.0
0.10
0.48
0.35
Para gr&s (Br&hiaria)
28.6
0.08
0.22
0.58
-”
3rred hay - average quality
- -
64.0
2.6
0.09
0.69
0.46
-.-
;traw
SUdan
93.2
0.06
0.39
0.49
Guinea
95.7
0.08
0.48
0.49
Elephant
93.2
0.04
0.17
0.15
“Ivory Coast hay”
87.0
0.24
0.42
0.44
Brachiaria
94.1
0.06
0.62
0.37
Mixed &andina or tut)
96.4
0.10
0.40
.-
'0.30
?reen legumes
Stvlosanthes gracilis
21.0
12.4
0.23
1.65
0.73
Stylosanthes gracilis - older
44.0
5.9
0.18
1.22
0.72
Centrosema uubescens
21.8
!5.5
0 . 2 1
1.04
0.70
Centrosema pubescens - older
29.4
14.7
0 . 1 8
1.15
0.68
Egume hay
Stylosanthes
92.8
6.1
0 . 1 7
0.54
Centrosema
87.0
.2.5
0.27
0.54
Peanut hay (as-f4 basis = afb)
92.1
6.3
0.13
0.40
Bean hay (f)olichos 1ablab)afb
93.9
14.2
1.2
0.56
Cowpea hay (Vinnas sinensis)afb
89.0
7.6
0.26
0.61
Mai~ - whole Young plant
21.9
4.8
0.20
0.34
0.89
Maize - sarne + cob forming
48.2
0.25
0.69
0.75
Maize - dry Ieaves+stalk+husk
93.6
ii
w
m
0.53
Millet (Sorghum almum) - whole
27.4
Ii.9
0.54
0.62
0.70
Millet - whole plant - younger
15.0
4.7
0.35
0.30
0.41
Millet - same plant - older
30.9
2.7
0.28
0.23
0.42
Millet - ~IV. leaves + stalks
85.0
-
1.9
0.14
0.55
0.36
L_
-
-
ource: Watson, P.R. 198 1,
,322
Table D3: Composition of some African ffeeds (As-Fed Bah)
-SE--
-
WZ%J-=L
-6
Feetls
D M
100 % DM basis
1
-
-s-z
DP
Ca
----.
--_---
-
- -----
P
----- -. --,-
ksirls
Maize (Coastal West Africa)
87.0
6.61
0.34
0.03
1.05
Maize (Savanna West Af+ica)
92.6
7.73
0.33
0.02
1.08
Millet (Burkina Faso)
91.8
'1.3
0.37
0.40
0.97
Paddy rice
87.3
4.7
0.26
0.06
0.82
Sorghum
89.9
.5.9
0.29
0.02
0.92
Forlio
3.6
0.06
0.07
0.86
-~-
--~--------
88.4
-----.-
-._-_ -
-
--.-
Brans
Maize (traditional milling)
86.0
6.23
0.72
0.06
0.92
Millet
92.3
9.0
0.61
0.09
0.86
Rice
88.6
4,.16
0.41
0.09
0.42
Sorghum
- - - - - - - - - - - - - -
90.7
- - -
6.8
0.64
----- - - -
0.09
0.78
- - -
Cakes md me&
Peanut cake
92.7
47.3
0.65
0.11
1.93
Cottonseecl meal (industria])
-
- - - - -
- - - - - -
--.-
212
1.2
-.--_ -
-
0.15
0.56
-I_
seeds
Cottonseetl (whole)
94.4
!2.6
0.49
0.11
II .05
Peanuts
90.8
14.0
0.29
0.12
II. Il
Cowpeas (Vigna)
90.4
19.1
0.42
0.17
Il .06
Beans 1(Dolichos lablab)
- - - -
--_-_--II
89.6
----
19.9
0.29
-._-.-.
-
-
0.26
- - -
L.-caves
Banana (Tanganyika)
16.2
1.26
0.03
0.17
0.14
Manioc (Cassava)
27.3
9.4
0.51
0.92
0.64
Sahel acacia
60.8
10.1
1.90
3.05
0.46
Acacia Albida - dry leaves
92.8
7.2
0.15
0.23
0.69
Baobab - dry, West Africa
- - - -
---s--v
91.0
.----
5.7
0.40
-.---.
- - -
1.08
0.41
-_I
Miscellaneous
Brewcrs wet meal (local)
30.7
17.6
0.41
0.26
0.80
Brewers meal dried (local)
92.3
21.6
0.33
0.03
0.87
BreadfGt (Artocarpus m.)
30.0
0 0
0.04
0.02
0.3 1
Mango (greenlpulp) (munis)
14.5
0.16
0.01
0.02
0.18
Yam (fresh)
36.8
0 0
0.05
0.11
0.39
Yam (dlnecl)
89.6
5.04
0.12
0.19
OI.59
Cassava (manioc) - fresh
34.2
0.0
0.04
0.04
OI.34
Cassava - dried
88.0
0.0
0.08
0.09
Ot.98
Rice husks (&a.@
92.0
0.12
0.08
0.08
0.29
Ek3.n bulls (Vim sinensis)
39.3
5.60
0.44
0.85
Cocoa :puds
92.1
4.13
0.15
0.20
0.46
Maize cobs
88.3
0.0
0.03
0.01
0.40
Maize .. mature cob + grain chop
93.3
5'.4
0.01
0.22
0.84
Whole ‘banana
24.1
0..68
0.02
0 . 0 1
0.26
Banana pulp
24.1
0.55
0.02
0.02
0.25
Sugar cane moiasses
83.3
0.9
0.03
1.49
1.04
Nere powder (Mali)
-
1 .oo
Rice straw
0.35
-
-
-
-
c--x=
92.5
0.19
0.19
0.29
- - - _I=D - - =Es
Source: Watson, P.R. 198 1.
323
Table D4: Field operation execution time
1. Tillage
Manual: 28 a 35 man-dayska (8 hrs/day)
Animal traction: 6 a 8 ox-team-day/ha (4 a 5 hrs/day)
2. Manual seeding
Millet
20 man-day/ha
Maize
15 man-day/ha
Sorghum
3 5 man-day/ha
Groundnut
24 man-day/ha
3. Manual Weeding
Millet
20 man-day/ha
Maize
32 man-day/ha
Sorghum
17 man-day/ha
Groundnut
3 8 man-day/ha
4. Manual Harvesting
Millet
13 man-day/ha
Maize
13 man-day/ha
Sorghum
10 man-day/ha
Groundnut
18 man-dayi’ha
(Source: Le Moigne, 198 1)
“2
-.-x”..“‘-“mm~l.~--
_ - - _ . _ . - “ . , . - -
APPENDUIE
. -l
APPENDIX E
.
ONE COMPLETE SESSION OF THE EXPERT SYSTEM PROGRAM
EXPERT SYSTEM
MSU: Midigan Statt University
Figure El: Screen 1
324
3 2 5
Figure E2: Screen 2
3 2 6
Figure E3: Screen 3
--
I
~~.u--~u”NIIIIIIR(~
.--.
“ . _. .I_. _ -m..,-.-
-mm”.
327
-=SiBZZE-
mnj.m,-
II
iÏÏ’
..v
3
.:..
EXPERT SYSTEM FOR ANIMAL TRACTION EVALUATION
-,-v.
._L------p.---
The farmcr’s name is optional
Figure E4: Screen 4
328
Gewraphical Location and Fart Chractuistics
DB-1 9-t 997
I. Agro-eco-systcm Zone:
r-7
4
- Boulandor 1
1. Farm Sizc(ha]:
1 6.00 1
. Numbcr of Farm Workers:
r--iq
i. Main Cash trop:
Figure E5: Screen 5
329
Geographhal Location and Farm Charactuistics
08-l 3-l 997
.-
Types of Crops Grown
Type wf Soi1 @ the
@ the Fat-mi Level
Agro-ecosystem Level
Figure E6: Screen 6
330
Figure E7: Screen 7
331
- - , - -
I I - ~
r
UCF Moldbcaard Rw
AI
I component 1
Ratlng for field work:
Figure ES: Screen 8
3 3 2
.-
Figure E9: Screen 9
3 3 3
Figure ElO: Screen 10
c
_
._ .I_.--._.
..._.” ---.
.-...
“.*~..“,-“--.m”..-
334
Figure Ell: Screen 11
335
Farm Ekpipment Inwdment
-
-
II --- ~
Total Imglomcnt Cort
0
\\ 881df
Figure E12: Screen 1.2
.r
336
Figure E13: Screen 13
1.
-Il-.---I--P
“..
- - -
-
.s.‘-IIII,LIIP--
_- __.-_.____ “.1--.II__.-.- ..-- -““___. ..“.” -, l”,.- -1-1.-.--
r
338
P
F E
I R
EA
L 1
D 10
- _..--_ -. _
22.6
.---_._-_
17.6
-".-.-_ .-.
34.191
.---- -_ -i
339
_____ __..._,..-_.._._.. _ . . ..__ _ .._.. . . . . ._.
..__” _...-..,...
_ _ _ _ . . . . . _
. .._. __.__“.” . . ..-...<....-.....
_ . _ . . . -
. . _ . _ . _ _ . .
_ _ . . . . .
.
t
f
Figure E17: Screen i 7
342
Resuhd
FA- Fwd
- -I~F&
-
@‘U-j:
-_-
AvüikMe
.--.y
feed:
+.“....,.-.--
Figure E19: Screen 19
343
3 4 4
Figure E21: Screen 21
.
- .u-. .‘““.*I-“--lc
f
>.
REFERENCES
. “ .
R E F E R. E N C.: E S
AID, 1997. World water for agriculture (Climatc, Prccipitation Probabilities and
adequaçies for rainfed agriculture). Agency CI~ Intcr~mlional De\\~elopmcnt, \\j’ashingl:on
DC. L%A.
ATRIC: 1986 f’ield performance of maresha. Test report ‘Nos 4, 16 and 18. AIRIC:
Ethbpia.
Andersen, F.M. 1985. Draughf animal pou;er systems in %-lb-Saharan Afr-ica: their
production l:mpact and rescarch needs. In Copland, J W. (ed.) Draught Animal Pc~er li71
predrtction. Procee.dings of an internatiioral WCH k shop held at James Cook University,
Tow-îsville, Qld, Australia, 1 O- 16 July 1985. Canbei,ra AclAR Proceedings No. 10, 26-
31.
ARC. 1980. The nutrient lequirements of rurnbant li\\,estcKk Slough: Commonwealtla
Agi~uIfural Ri.ireaux.
Barwell and Ayre, 1982. The harnessing of dr-aught animais. Intermediate Technology
Publications, Ardington.
Retker J. and Klut&ac:.h H.D 1991. The role of donkeys in Agricul?ural mechanization in
Niger: Potsntial and limitations In Fielding ad Pcarson (eds), Donkeys, hiules, H~tses in
Tropic:al Aqicultural Dcvelopnxnt, CTV?vf Ed~nburgh: 1991.
Campbell, J 1930. Dipple s?ic.ks, dorkeys and diesels. IRRl, Los Banos, Philippines.
CFXMAT 1995. klanuel de cult~m avec tractjon arrimale: techniques rurales en .Urique.
Cadre d’Etudc et d’Experimentation du hfachinisme Agricole Tropical (CXEMAT) et
h%nktere de la Cooperation, Parc de Tout-voie, Antony, France. 336~.
CEEMAT 1974. Aide mcmeoire du moniteur de la culture atîelee. Centre d’Etude et
$‘Experimerttation du Machinisme Agricole TropicaE (CEEMAT) et Ministere de la
Cooye.ration, Parc de Tourvoie, Antor~y, France.
CEEWIAT/‘P’AC~.
1972. Tlsc ctnployment of draught animals in1 agriculture. Rome, Italy
346
347
C’IMMYT. 1984. Report of a networkshop on draught power and animal feeding. 4-6
October 1983, Ezulwini, Swaziland. Networking workshop Report 2. CIMMYT Eastern
and Southern Afnca Economies Programme: Mbabane, Swaziland.
Charreau, C. and Nicou, R. 1971. L’amelioration du profil cultural dans des sols sableux et
sabla-argileux de la zone Tropicale seche Ouest-Africaine et ses incidences Agronomiques
(d’apres des travaux des chercheurs de I’IRAT en Afrique de l’Ouest). Agronomie
Tropicale 26:903-978.
Conroy D. 1992. Advanced training techniques for oxen Tiller’s TechGuide. Tillers
International. Kalamazoo Michigan, USA.
Conroy D and Rice P. 1992. Selecting and training axer-. Tiller’s TechGuide. Tillers
International. Kalamazoo, Mjchigan, USA.
Crossley, P. and Kilgour, J. 1983. Small farm mechanization for developing countt’ies
John Wliley and Sons, Chichester, UK 2S3p.
RERBAC. 1992. Rapport d’activites. Campagne agricole 1991-l 992. Ziguinchor, Senegal.
Devnani, R.S. 1981, Design considerations for hamesses and yokes for draught animais
Technical bulletin no C%%E/8 1/30.
Dibbits, H.J. 1993. Human and draught anima.l power in trop production. Past experiznces
and outstanding problems. In Human and Draught Plnimal Power in Crop produc.tion.
Workshop proc.eedmgs, Harare, Zimbabwe, 18-22 January, 1993. Silsoe Research
Jnstitute-Agricultural Research Food Council. CEE-FAO, Rome 1993, Italy.
Ducreux, A. 1984. Caracterisation mecanique des sols sableux et sablo-argileux de la zone
tropicale seche de l’Afrique de l’Ouest. Etude d’un prototype d’outils permettant de les
travailler en periode seche. These de Docteur Ingenieur en Sciences Agronomiques-
Option Pedologie. Academie de Montpellier, Universite des Sciences et Techniques du
Languedoc. Montpellier, France.
Eicher, K.C. and Baker, D.C. 1982. Research on agricultural development in Sub-Saharan
Africa: A critical survey. MSU International Development Paper no 1.
Fall A. 1995. Factors affecting feed intake, energy expenditure and work output of oxen
and bulls used for draught purposes in Semi-Arid West Africa. Submitted to the University
of Edinburg in partial fulfiilment of the requirements for PhD degree.
Fall, A. 1985. Situation actuelle de l’environnement et de l’utilisation du parc de mater-iels
de culture attelee en Basse Casamance. Memoire de titularisation. Departement de
Recherches sur les Systemes de Production et Transfert de Technologie en Milieu Rural.
JSRA. (document interne).
--
.----s
-1
348
Fall, A. 1’390. Adoption et Principales cxmtraintes a la dif%rsictn des equipmcnts de traction
animale en Basse Casamance, Senegal. In. P. Starkey, M. Goe and A. l’aye (eds), Animal
traction for Agriculture Development. Wagcningen: Technical Center for Agricultural and
thF@JatiOIl (k%@-EEC hJJli2 ~OM3ltiOll).
FALL, A. 1990. Rolcs ac&.rels et potentiels des forgerons traditionnels dans la
mainfenance du materiel agricole en Basse Casamance 1 Senegal. In: P. Starkey, M. Goe
and A. Faye (eds), Animal traction for Agriculture Dçve!opment. Wageningen: Tcçhnical
Cent~- for Agricuhural and Cooperatiun (ACP-EEC Lorne Convention).
Fail, 13.N. 198 1. Etude sur la situation et les perspectives de l’industrie du machinisme
agricole au Senegal. Premiere consultation regionale pour le developperncnt de l’industrie
du machinisme agricole en Afrique. ON1[_!Dl-O‘CTA-CI:A-FIZO.
F.40. 1972. Manrral on the Emplo:yment ofdraught Animais in Agriculture. CEEMAT
FAO. 198Z!. Report of the FACI Expert Consultation w-r Appropriate use of animal energy
in Agriculture in Africa ami in Asia FAO. Rome.
Ffoulkes D. and Bamualim, 1989. Improving the nutrition level of draught animals usine,
availatrle feeds. In Hoffrnan D., Na.ri J. and Petheram .R.J. (eds), Draught Animais in Rural
Development. Proceedings of an Iternational Research Symposium, Cipansas, lndonesia.
ACIAR Proceedings no 27: 134-I 45.
Fath, D.H. 1990. Fundamentals of soi1 science. 8th edition. Wiley & Sons Publisher. New
York.
Gae, %I.R. 1983. Cu;rent status of research on animal ~traction. n’orld Animal Rer.;iew, 45,
2-17
Gae, M.R. 1987. Animal traction on smallholder farrns in Ethiopian highlands. University
Microfilms International: Ann Arbor, Michigan,
Goe, M.R. 1930. Overcoming constraints to animal traction through a collaborati\\.e
researçh network. In: P.H. Starkey, M.R. Goe and A. F’aye (eds). Proceedings ofthe third
WAATAN. 7-12 July 1988, Saly, Senegal. Animal Trac:tion for Developrnent. CTA.
Goe, MR., and ,McDowell, R.E. 1980. Animal Traction: Guidelines to utilization. Corne11
International Agriculture Mimeo. Ithaca, N. Y. Corne11 University.
Hagard:, :M. 1987. Le parc de mater:iels de culture attelee et des possibilites de sa
maintenance dans le depar-tement de Nioro: resultats d’enquetcs. Document de travail
1987:3. Departernent Systemes et Transfert, ISRA, Dakar, Senegal. 45~.
i
349
Havard M., Le Gal J.P. and Pirot R. 1994. Towards a decision support system integrating
different kinds of knowledge to choose and manage field equipment. U.R. Genie
Agronomique et Mecanisation, CIRAD-SAR,
Montpellier, France.
Havard M and Le Thiec G. 1995. Animal traction in Western Africa and Madagascar.
Influence of the environment on its develoyment and use. Communication in The
challenge of Animal Traction. ATNESA Network, 4-9 December 1995, Kenya. 1 1 p.
Hillel, D. 1982. Introduction to soi1 physics. Academi Press, Inc. New York, USA.
Hopfen, H.J. 1960. Farm implements for Arid and Tropical Regions. Agricultural
Development paper No 67. FAO,’ Rome, Italy.
Jaeger, KW. 1985. Agricultural mechanization: The economics of animal traction in
Burkina Faso. Doctor of Philosophy thesis submitted to IFPRI. Standford University
Jackson, P. 1990. Introduction to Expert Systems (2”d ed.). Addison-Wesley Pubhshing
Company, N.Y.
Jury W.A., Gardner W.R. and Gardner W.H. 1991. Soi1 physics. 5th Edition, Wiley &
Sons. New York.
Kartiarso, M.D. and Teleni, E. 1987. Nutritional study using a treadmill at James Cook
University. DAP Project Bulletin. 3 : 1- 15.
Keith M. 1992. Training Young steers. Tiller’s TechGuide. Tillers International.
Kalamazoo MI.
Kemp, D.C. 1987. Draught animal pow-er: some recent and current work. Worid Animal
Review. 63:7-14.
~LL.awrence, P.R. 1985. A review of the nutrient requirements of draught oxen. In Copland,
’
J.W. (ed.) Draught Animal Power for production: Proceedings of an inte.rnational
workshop held at James Cook University, Townsville, Qld, Australia, JO- 16 July 1!?85.
Canberra: ACIAR Proceedings No. 10, 59-63.
Lawrence, P.R. and Pearson, A. 1985. Factors affecting the measurements of draught
force, work output and power of oxen. J. agric. Sci. Camb. 105:703-714.
Lawrence, P.R. and Pearson A. 1989. hsfeasurement of Energy Expenditure in working
Animals: Methods for different c.onditions. In Hoffman D., Nari J. and Petheram R.J.
(eds), Draught Anima]s in Rural Development. Proceedings of an Itemational Research
Symposium, Cipansas, Indonesia. ACIAR Proceedings no 27: 155- 165.
Lawrence, P.R. and Pearson R.A. 1993, E’xperimental methods in draught animal science:
The need for standardization. In Human and Draught Animal Power in Crop production.
Workshop proce.edings, Harare, Zimbabwe, 18-22 January, 1993. Silsoe Research
Institute. Agricuftural Research Foo’d Council CEE-FAO, Rome 1993, Italy.
Lawrence, P.R. and Stibbards R.J. 1990. The encrgy c.osts of walking, carrying and
pulling Ioad:; on flat surfaces by Brahman cattle and swamp buffalo. Animal Produl:tion,
50: 29-39.
+( Lawrenov P R. and Zerbini E3. 1993. Recent trends in research on draught animal nutrition,
1 JR Huma:n and Braught A~limal qower in Crop production. Workshop proceedings,
Harare, Zimbabwe, 18-22 Januat-y, li993. Silsoe Research Tnstitute-14gricultural Research
Food Council. CEE-FAO, R.ome 1993, ‘Italy.
I,ce J. V. d., Udo W. M. J. and Brouwer B. 0. 1993. Design and validation of an Animal
Traction Module for a smallholder L,ivesteck Systems Simulation Model. Agricultural
Spstems 43 ( 1993) 199-227.
Le Moigne, M. 198 1. Contralintes posees par l’insertion de !!a mecanisation dans fes unites
de production agricole en zone sahelienne. Etude methodologique. CEEMAT. Paris,
France.
Martin, D., and Teleni, E. 1989. Fatigue in buffaloes QII diflèrent work loads. DAP Project
BufPetin, ‘3, 2-6.
Le Moigne, M. 1984. Missio:n d’appui au programme Mijchiisl7le Agricole en Casamance.
CEEMAT. 33p.
Le Thiec, G. 1991. Les hatmachememts pour bovins - “‘De l’Europe a l’Afrique”. CIRAD-
SAR, CEIEMAT, Montpellier, France.
Le ‘IXe,, G. and Havard M. 1995. Quels jougs pour les bovins de trait en Afrique Sub-
Saharienne? Communication in The challenge of AnimaU Traction. ATNESA Network, 4-9
Deçember 1995.
Le Thiec G a:nd Havard M. 1995. Developing suitable ylokes for draft oxen in Sub-Saharan
Africa. Communication in Meeting the challenges of Animal Traction. ATNESA Network,
4-9 December 1995. I Op.
M.A..T. 1,985. Special animal traction. No 91. Juillet-Septembre.
M.A.T. 1988 Activites CEEMAT 1986-87. CEEMAT. Parc de Tourvoie, Antony,
France.
351
Matthews, M.D.P. and Pullen. 1975. Cultivation trials with ox-draw-n irnplernents using
Xdarna cattle in the Gambia. Tropical Agriculturaf Engineering Information, Report series;
Overseas Department. NIAE, Silsoe, Bedford, England.
Mbengue E%M. 1986. LA mecanisation de la transformation des cereales au Senegal:
Aspects techniques et nutritionnels, I§RA CNRA Bambey. DRSAEA. Dakar, Senegal
MDR. 1986. Etude du Secteur Agricole: Plan Cerealier. Ministere du Developpement
Rural, Senegai.
Monnier, J. 1983. Pour un nouveau developpement de la culture attelee en region. Lecons
des experiences acquises au Senegal, Madagascar, Cote d’ivoire. Les roles de la
R.echerche et de la Vulgarisation. Montpellier, France.
Morris J. 1983. The ec,onomics of small farm mechanization in Crossley P. and Kilgour J.
(eds). Wiley and Sons. NY.
Montoroi J.P. 199 1. Etude du comportement hydrique des sols rouge et beige du bassin
versant de Djiguinoum (Basse Casamance - Campagne de mesures 1990). ORSTOM,
Programme CEE DG XII, Contrat no TS2A-0216-M (CD).
Niane A. B. 1884. Etudes cartographiques et agro-pedologiques des sols de p1atea.u de
Basse Casamance. Memoire de confirmation. Cra Djibelor, ISRA. Il0 pp.
Ndiame, F. 1986. Aspects economiques de l’utilisation de la traction animale bovine et de
sa promotion par le biais du Credit Special du PIDAC pour le materiel agricole. ISRA.
DRSAEA. Memoire de titularisation. Dakar, Senegal.
Ndiame, F. 1990. Animal traction in the Lower Casamance: Technical aspects and socio-
economic implications. In: Paul Starkey and Adama Faye (eds), Animal traction for
Agriculture Development. Wageningen: Technical Center for Agricultural and
Cooperation (ACP-EEC Lame Convention).
Ndiame F. and Fall A. 1989. Par quelle cereale remplacer le riz ? Quelques problemes de
la culture du maïs en Basse Casamance. In Economie des Filières en Regions Chaudes.
Actes du X- seminaire d’economie et de sociologie. 1 l- 15 Septembre 1989, Montpellier,
France. Une publication du CIRAD.
Nicou, R. 1975. Le probleme de la prise en masse a la dessication des sols sableux et
sabla-argileux de la zone tropicale seche. In Agronomie Tropicale XxX-4: 325-343.
Nolle, J. 1975. Le tracteur animal. M. A. T. 5 1:46-56
OVeill, D.H., Howell P.J.L., Paice, M.E.R. and Kemp, D.C. 1987. An i.nstr-umentation
system to measure the performance of draught animais at work. In: Proc. Natl. Seminar
Animal Energy Utilization. 53-72. Central Institute of Agricultural Engineering: Bhopal,
India.
352
O’Neili Dl, Hayton S. and Sirns B. 1989. Measurement lofdraught animal performance. In
HolTtnan D., Nari J. and Petherarn R.J. (eds), Draught Animais in Rural Development.
Proceedings of an lternational Research Symposium, Cipansas, lndonesia ACIAR
Proc.eedings no 27:264-27 1.
‘y, Pearson R.A. and Fall A. 1993. R.esearch on the nutrition of draught animals. In IJuman
and draught animal poser in trop production. Workshop proceedings 18-22 Ja.nua\\y,
1993, Ha.rare Zimbabwe. Silsoe Research Institute and Agricultural Research and Food
Council, CEE-FAO, Rome, Italy.
Pelissier, .P. 1966. Les paysans du Senegal. L,es civilisations agraires du Cayor a Ii~
Casan-mce. Saint Yrieux, :France. (rd. Fabreque.
l?eret S., Pirot R. and Razaka B. 1.994. In-fiefd nleasurement of tillage irnplements
pelfi3rl~~~~nces:
Practical applications in soil-tillage relations studies. Paper AG-ENG 94.
Ping.all P., Bigot Y and Binswanger 13. 1987. Agricultural nle.chanization and the e\\olution
of f&ning systems in Sub-Saharan Africa. A World Rank Publication, John Hopkims
Press, Baltimore, MD. USA.
Pirot, R. 1993. Automation of in-feld data recording: Recent research work at CIRAD-
SAR. CIRAD-SAR., Montpellier, :Fra,nce.
Posner, J.L. 1985. Contribution a la connaissance agronorrrique de la Basse Casamance
(synr hese bibliographique). Document de tra\\:aii 1985:3. Deyartement SystenÎes et
Transfert, ISRA, Dakar, Senegal.
Posner J.L. and al. 1988. Les systemes de production en Basse Casamance et les strategies
paysannes face au deficit pluviometriquc. lSRA/?vlSU International Developrnent Papers.
Raman I-l., Mohan S., and Rangacharya N.C.V. 1992. Decigion Support for Crop Planning
during Droughts. In Journal of Itigation and Drainage Engineering, vol 118 no 2 ,229-
241.
RatlifL.F, Ritchie J.T., and Cassel D.K. 1983. Field-meaasured Lirnits of Soi1 Water
availnbility as Related to Laboratory-Measured Properties. In Journal of SSSA, vol 47 no.
4: 770-775.
Rehak, D.R. 1983. Expert systerns for drought nlanagen~ent. 113 proceeding of Conference
on Ernerg,ing Computer Techniques in Storrn. Water and Flood hlanagement W Ja:n-tes
(ed), ASCE. New York, NY, 333-387.
R.itchie, J.T. 1972. Mode1 for predicting evapotranspiration fr#otn a row trop with
incornplete caver. In Journal of Water resources Research October 1972, 8: 1204- 12 13
353
Roose E. 1967. Dix annees de mesure de l’erosion et du ruissellement au Senegal.
Agronomie Tropicale 1967 2”: 123-l 52.
Roosenberg R. 1987. Updating animal-powered mechanization by building on history.
ASAE proceedings: December 15 - 18, 1987 Winter meeting in Chicago IL.
Roosenberg R. 1992. Neck yoke design and fit: Ideas from dropped hitch point traditions.
Tiller’s TechGuide. Tillers International. Kalamazoo MI.
Sebillotte M. 1990. Les processus de decision des agriculteurs. Deuxieme partie:
consequences pour les demarches d’aide a la decision. In modelisation systemique et
systeme agraire. Decision et organisation, J. Brossier, B. Vissac et J.L. Le Moigne, eds ,
Paris, France, INRA: 10% 117.
Sivakumar M.V.K. 1988. Predicting rainy season potential from the onset of rains in the
Sahelian and Sudanian cfimatic zones of West Af?ica. Agricultural and Forest
hdeteorology 42: 295-305
Smith A. J. 1990. Using science to understand the biological constraints that limit work
animal productivity. In: P. Starkey, M. Goe and A. Faye (eds), Animal traction for
Agriculture
Development. Wageningen: Technical Center for Agricultural and
Cooperation (ACP-EEC Lome Convention).
Smith A. J. 198 1. Draught animal Research: A neglected subject. World Anin~al review,
40, 43-48
Sonko, M L. 1990. Disponibilite des animaux de trait et contraintes structurelles en Basse
Casam;ance. In: P. Starkey, M. Goe and A. Faye (eds), Animal traction for 14griculture
Development. Wageningen: Technical Center for Agricultural and Cooperation
(ACP-EEC Lome Convention).
Sonko, M.L. and Fall, A. 1992. Budgets de culture: Recueil de donnees technico-
economiques de production en Basse Casamance. FAO contrat d’auteur no 7 CMT 37503.
Srivastava, N.S.L. and Ojha, T.P. (eds). 1987. Utilization and economics of draught
animal power. Central Institute of Agricultural Engineering: Bhopal, India.
Stapleton, J.H. 1995. Linear Statistical Models. Wiley and Sons, New York. USA.
Starkey P. 1986. Draught animal power in Africa: priorities for development, research and
training. Networking paper 14. FSSP, University of Florida: Gainsville, Florida.
Starkey, P. 1989. Harnessing for cattle and buffaloes: Options and research. ‘In Hoi%nan
D., Nari J. and Petheram R.J. (eds), Draught Animals in Rural Development. Proceedings
of an Iternational Research Symposium, Cipansas, Indonesia. ACIAR Proceedings no
27:280-291.
Starkey P. 1989 Hamessing and Trnplemen~s for Animal traction - AR animal traction
resource book for Africa. GATEIGTZ, Vieweg and s~r‘ts, IWwschweig, Wiesbaden.
Germny.
Surr!ner GI. 1’388. Precipitation: Process and analysis John Wiley and Sons. New ‘York,
USA.
Turban E. and Meredith J.R. 1994. Fundamentals of Marrag,ement Science. 6th Edition
Irwin, Bcxton Ma. USA.
Teleni E ;and Hogan J.P. 1989. Nutrition of draught animais. In Hoffnlan D., Nari J. and
Pet heram R. J. (eds), Draught Animal s in Rural Developmet:~t. Proceedings of an
TteInafkxal Rese.arch Symposium~, Ci!pansas, Jndonesia. .4CIAR Proccedings no 2’7: 1 18-
133.
Tourte, R 1971. Themes legers - Themes lourds: Systc;nes intensifs; Voies difYercnres
ouwtec, a.u Developpement Agriwk du Scnegal, (,Simple technical innovation,
Sophisticated technical innovation, 3ntensive Systerns; Thret- dif??rent Approaches to
Agrkulture Developpement in Senegal). Agronomie Tropiçale, No 26 (5).
‘%\\Il Traverse~ S. 1974. Le boeuf Ndama et la riziculture en Basse Çasamance. Station Rizicole
“x!e Djibelor. ISRA, Dakar, Senegal.
L7padhya’y. R.C. 1989. Performance limiting Gctol s in draught animals: Can they he
manipuhited to improve ouIput? In Hoffma,n D., Kari J. ond Petheram R.J. (eds), Draught
Animal~ in Rural Development. Proceedings of an Iternational Research Symposium,
Cipansas., lndonesia. ACIAR .Proceedings no 27, I66- 173.
IJzure.au, C. 1974. Animal draR in West Afi-ka. \\~‘orlci Crops. 26: 112-l 14
Wanapat,, :M. 1989. Supplementaq fe:eds: Their ttfiects on pwformance of draught animals
and their acceptability to Sout.heast Aslan krmers. In Woffnxx~ Il., Nari J. and Pethcram
R.J. (eds), Draught Animais in Rural Development Proccedings of an Iternational
Research Symposium, Cipansas, Indonesia. ACIAR Proceedings no 27: 146-l 53.
Willocks, T.J. 1979. Tillage energ>/ requirements for semi-arid trop production in
Botswana. In R. Lai (ed). Soi1 tillage and Cq production. Proceedings series NO.~. 30%
323. IITA: Rxdan, Nigeria.
dissertation entitled
METHODOLOGY FOR EVALUATING THE IMPACT OF ANIMAL
TRACTION AT THE FARM LEVEL IN A SMALESCALE
MULTI-CROPPING
SYSTEM
(BASSE CASAMANCE, SENEGAL)
presented by
Alioune FALL
has been accepted towards fulfïllment
of the requirements for
PhD
degree in AE/ATM
Ip,d,MU&
Major professor
o - 1 2 7 7 1
I .._. -_-__--l-.~--...-“--I
---_--
<I--------‘-
i _ _..“- -..-
. .__..” _. “-.
._ _.,-._--.-
.
METHODOLOGY FOR EVALUATING THE IMPACT OF ANIMAL
TRACTIOk AT THE FARM LE-WL IN A SIMALL-SCALE
MULTECROPPING SYSTEM
(BASSE CASAMANCE, SENEGAL)
BY
Alioune FALL
A DISSERTATION
Submitted to
Michigan State University
In( partial fulfillment of the requirements
For the degree of
DOCTOR OF PHILOSOPHY
Agricultural Technology and Systems Management
Department of Agricultwal Engineering
1997 * ‘
/
ABSTRACT
METHODOLOGY FOR EVALUATING THE IMPACT OF ANIMAL.
TRACTION AT THE FARM LEVEL IN A SMALL-SCALE
MULTI-CROPPING SYSTEM
(BASSE CASAMANCE, SENEGAL)
BY
Alioune FALL
A methodology is presented along with an expert system program for evaluating
the impact of animal traction at the farm level in a small-scale multi-cropping system. In
its actual version, the expert system program applies to oxencultivation using Ndama
sipecies in the Basse Casamance region situation. The evaluation is based on a number of
models developed around the draft animals’ daily energy balance. The energy available
for the various field operations is calculated based on (1) the total liveweight of the
oxen(2) the daily amount of feed given to the oxen and (3) the level of draft and power
required to perform the field operation. The liveweight of each ox is modeled by using an
eimpirical equation relating the actual liveweight in kg to the circumference in cm of the
thorax of the ox. The energy is estimated from the characteristics of the field operation
(type of implement, draft, speed, field time) in association with the number of working
days during the rainy season. An optimization module is used to mode1 a cropping system
in relation to the economical environment and endogenous constraints of the farm (land,
la.bor and available energy). The expert system output agreed with current practices and
values found in the literature.
’
------ .
TO my father Macaty, to vieux Adama, to Odette and to Mabeye Sylla, gone too soon
(Peace Uporu Them).
TO my wife: Marne Coumba, my son Ousmane and my mother Fatou Diagne.
. . .
111
ACKNOWLEDGMENTS
1 would like to take this opportunity to gratefùlly acknowledge the assistance and
trust of Dr. R.H. Wilkinson, my advisor and chairperson of my guidance committee.
Also, the completion of this work would have not been possible without the attention and
understanding of the members of my guidance committee, Dr. Burkhardt (AE), Dr. Harsh
(Ag. Econ) and Dr. Harwood (CSS). Their suggestions and valuable ideas helped me
finalize this dissertation and enjoy every moment of the work. Many thanks go also to
Dick Roosenberg and Tillers International.
1 would like to thank a11 the personnel at the ISRA/CRA Djibelor and CRZ Kolda
research stations, for sharing with me ideas and research facilities during my fieldwork.
Special thanks go to Dr. Khouma and Mme Sall, Abdou Fall, Adama Faye, Abibou
Nia.ng, Sabaly, Konte, Bolle, Tuti, Yafaye, Fode, Ba, Andre and Ibrahima Sonko.
1 would like to express my gratitude to ISRVDG, especially to J. Fayes and J.P.
Ndiaye, to the NRBAR/OSU/AID Project, to CADEF, to vieux Diame (Boulom) and also
to EPATA through Mr. Konte for a11 the support provided to make the research possible.
Thanks to 0. Fall, A.M. Mbaye, 1. Diagne, Zeiman, Herman, B. Diagana, C.
B,abou, P. Diop, H. Diakhate, Amadou Fall, Dr. Sall, M. Havard, Sayon K., As Ndiaye,
D. Coulibaly, M.L. Sonko, F. Diame, 1. Thomas, M. Lo, S. Sali, H. Mbengue, S. Kante,
M. Sene, Mme Badiane, D. Sidibe, S. Djiba, 0. Ba, 0. Diop, K. Toure, Ezabele, Brima,
Pape Gueye, Aziz, Lopi’s, and friends, and their respective families for the moral support,
iv
t
~-.--
-.
.- --~---
.~
:
TABLE OF CONTENTS
Pages
LIST OF TABLE% . . . . . . . . . . . . . . . . . . . . . . . . ..<...................... ,.......... . . . . . . . . . . .
xi
LIST OF FIGUREiS.. . _ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
xvi
ABBREVIATION$ . . . . . . . . . . . . . . . . . . . . . . . . . . . .,........................ *.....,.~...*<....
xix
LIST OF VARIAB/LES.. . , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..“.................~........
Xxi
CHARTER 1
INTRODUCTION.. . . . . . . . . . . . . . . . . . . . . . , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . < . . . . . .
1
‘1
1.1. Mechanlzation overview.. _ . . . . . . . . . . . . . . . * . . . . . . . . . . . . . . . . . . . . . . . . . . .
1
l. 1.1, Sub-Saharan Afkica.. . . . . . , . , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2
11.1.2; Senegal. . . . . . . . . . . . . . . . . . . . . ..~.......................................~
6
1.2. Statement of problem.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .,,
8
1.3. Objectivles . . . _. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..<..............<......<~...
9
CHARTER 2
PRESENTATION CbF THE AREA OF STCDY............... ............. <,...........
12
2.1. Presentation of the area of study.............................................
1 2
2.2. Types of production system ...................................................
17
Z!.2.1. Agro-ecosystem 1 (Oussouye). .......... ,, ....... ................
19
Z!.2.2. Agro-ecosystem 2 (RIouf). ....................................
....
2 0
2!.2.3. Agro-ecosystem 3 (Niaguis). ......................................
2 0
2.2.4 .. Agro-ecosystem 4 (Kalounayes). ...................
.............
21
2.2.5. Agro-ecosystem 5 (Diouloulou). ..................................
21
2.3 Croppingj systems................................................................
22
2:.3.1. Groundnut production ..............................................
2.3
2,.3.2. Cereals ................................................................
2 4
2.4. Agricultural productivity. .............................. ,,.................. ....
2’7
2.4.1 .,Land productivity. ...................................................
28
2.4.2. Manpower and labor productivity. ..... ..........................
3 0
2.4. Summary ..................................................................
.......
31
V
.-
.- ._I.. ----~---
-
%.
/
CHAPTER 3
DETERMINANTS OF PERFORMANCE OF DRAFT ANIMALS.. . . . . . . . . . . . . . . . . :
3 4
3.1. Constraints of animal traction utilization......................................
3 4
3.2. Selection of draft animais ......................................................
3 5
3.2.1. Availability............................................................
35
3.2.2. Selection process ....................................................
3 7
3.3. Training of drafl ammals .............................................
-. .........
3 8
3.4. Hamessing systems and team constitution ...................................
41
3.4.1. Types of hamessing system ..........................................
41
3.4.2. Efficiency of hamessing systems. ...................................
4 2
3.5. Working capacities ..............................................................
4 4
3.5.1. Climate effects .......................................................
4 5
3.5.2. Pulling force potential...............................................
4s
3.5.3. Work potential and energy requirements.........................
5 0
3.5.3.1. Instrumentation...........................................
5 0
3.5.3.2. Energy at work.. .........................................
52
3.5.3.3. Energy for animal growth.. .............................
53
3.5.3.4. Energy for maintenance . . . . . . . . . . . . . . . . . . . . . . . . . . ...*...
54
3.6. Feeding systems .................................... ..i ..........................
54
3.6.1. Feed quality and availability........................................
54
3.6.2. Diet for working animals............................................
55
3.6.3. Forage Unit...........................................................
5s
3.7. Implement selèction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5 9
3.8. Soils.. ............................................................................
6 2
3.8.1. Types of soil.. ........................................................
6 2
3.8.2. Mechanical and physical properties ................................
65
3.8.2.1. Physical properties .......................................
65
3.8.2.2. Resistance to traction ....................................
6 7
3.9. Working days ..........................................
..........................
6 8
3.9.1. Soi1 moisture regime.. . . . . ._.............................
,....” . . . . .
68
3.9.2. Weather factors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
71
3.10. Land use and cropping systems .............................................
7 4
3.10.1. Animal traction and cropping pattems ..........................
74
3.10.2. Field performances and trop production ........................
76
3.11. Summary.. .....................................................................
9 4
CHAPTER 4
MATERIALS AND METIIODS ............................................................
8 2
4.1. Research sites ....................................................................
8 2
4.1.1. On-Station ........................... .................................
82
4.1.2. On-Fat-m ..............................................................
82
4.2. On-Farm equipment survey....................................................
8 4
4.2.1. Objectives .............................................................
85
4.2.2. Data collection .......................................................
86
4.2.2.1. Type offarm equipment and draft animal .............
86
4.2.2.2. Management. ..............................................
SS
vi
.-
.-.-_. - .- ~-
4.2.2.3. Utilization and maintenance. ...........................
89
4.2.3. Samplïng techniques. ......................................
........
90
4.2.4. Data analysis ..............................
1':. ........................
9 1
4.3. Estimates of energy requirements ............................................
9 2
4.3.1. Objectives ...............................................
...........
9 2
4.3.2. Experimental design ...............................................
9 2
4.3.3. Draft measurements .................................................
9 6
4.3.4. Power and energy expenditure ........................ ...........
9 7
4.4. Maximum performance ........................................................
9 8
4.4.1. Objectives ............................................................
9 8
4.4.2. Procedures ...........................................................
9 9
4.5. Instrumen/tation .......................................
..........................
102
4.6. Expert System building (see diagram) .......................................
103
4.6.1. Objectives ............................................................
1 0 3
4.62. Rnowledge base design and development ........................
104
4.6.2.1. Presentation of the expert system. ................... .
104
4.6.2.2. Expert system organization ... ........................
106
4.6.2.3. Data processing .................. ,,........ ......... ....
107
4.6.2.4. Program output ..........................................
1 1 1
4.6.3. Program validation .................................................
1 1 1
4.7. Sumn1ary ........................................................................
1 1 1
CHAPTER 5
FARM CHARACTEI@TICS AND SYSTEMS MANAGEME:NT ................
113
5.1.Farmequi’ment ................................................................
1 1 3
5.1.. 1. 1pypes of implements .......................
........ ..............
116
5.1.1.1. Moldboard piow UCF 10”: Tillage ............... ....
116
5.1.1.2. Imported ridger EMCOT ..... ... ................. ., ..
120
5.1 . 1.3. Seeding implement .......................................
120
5.1.1.4. Multipurpose toolbar implements ...................
122
5.1.1.4.1. Multipurpose frame ARARA ...............
1 2 3
5.1.1.4.2. Multipurpose frame SINE 9 ..................
124
5.1.2. Carts ..................................................................
126
5.2. Repair and maintenance .......................................................
126
5.2:. 1. dreakdown fiequencies ................... ........................
126
5.2.1.1. Moldboard plow 10” ........................... .... . .
126
5.2.1.2. Seeder Super ECO ......................................
129
5.2.1.3. Toolbar and EMCOT ridger ...........................
130
5.2.1.4. (lx-cal-t .....................................................
1 3 1
5.2..2. Farm equipment reliability ........................................
1 3 1
5.3. Dali animal management ...........................................
..........
136
5.3.1. Availability of draft animais .......................................
137
5.3.2. Training and working career ............ .........................
137
5.3.3. Sitability ..............................................................
1 3 8
5.4. Land and labor resources ....................................................
142
5.4.1. Land availabihty ....................................................
142
vii
5.4.2. Farm labor ...................................................
144
5.5. Adoption dynamics ....................................................
144
5.6. Surnmary ................................................. ..:: ..............
149
CHAPTER 6
WORKING CAPACITY OF DRAFT AN-MALS ..............................
151
6.1. Environmental conditions ............................................
151
6.2. Maximum performance .................... : ..........................
154
6.2.1. Maximum performance ....................................
155
6.2.1.1. Optimum point ...................................
155
6.2.1.2. Maximum pull force Pf,, .....................
159
6.2.2. Characteristics of power curves ..........................
160
6.2.2.1. Power output mode1 ...........................
160
6.2.2.2. Walking speed (ANSPEED in rn/s). ..........
167
6.3. Draft animals liveweight mode1 ......................................
173
6.4. Management of draft animais .........................................
177
6.4.1. Field work ...................................................
179
6.4.2. Daily working hours .......................................
180
6.4.3. Feeding system .............................................
182
6.4.4. Liveweight changes and body conditions ................
184
6.4.4.1, Body weight losses (LWL in kg/day ..........
184
6.4.4.2. Health and tare .................................
187
6.5. Sumrnary ...............................................................
188
CHAPTER 7
FELD OPERATIONS AND DRAFT REQUIREMENTS .....................
190
7.1. Field operation characteristics ........................................
190
7.1.1. Soi1 characteristics ..........................................
190
7.1.1.1.. Physical properties ..............................
191
7.1.1.2. Water holding capacity ........................
193
7.1.1.3. Penetrometry ....................................
196
7.1.2. Field operations .............................................
196
7.1.2.1. Tillage .............................................
196
7.1.2.2. Seeding ..........................................
199
7.1.2.3. Weeding ..........................................
200
7.1.2.4 .. Harvesting .......................................
200
7.1.3. Soi1 water balance and working days .....................
201
7.2. Pull force and draft requirements ....................................
204
7.2.1. Statistical design ............................................
207
7.2.2. Required pull force ..........................................
208
7.2.2.1. Data collected for UCF 10” plow ..............
209
7.2.2.2. Data for 3-Canadian tines on SINE 9 ...........
212
7.2.2.3. Data for ridger attached to ARARA ............
215
7.2.3. Regression analysis ...........................................
216
7.2.4. Draft requirements ............................................
221
7.3. Implement Field capacities .............................................
225
...
VI11
. .
.,
I
7.3.1.. Fit$d efficiencies ...............................................
225
7.3 .Z!. Fieild capacities .................................................
226
7.4. Summary.. ............................................. :-. ................
227
CHAPTER 8
EXPERT S’YSTEbPVT TO EVALUATE THE IMPACT OF ANIMAL TRACTION
IN THE BASSE C,4SAMKNCE REGION .........................................
230
8.1. Qrganization of the expert system prog:ram ...........................
230
8.2. Database Files structures .................................................
232
8.2. 'L . Faml location and characteristics .............. .,.......... ...
232
8.2.2. Farm implement ............... .............. .................
233
8.23. DM: animais .................................................
234
8.2.4. Feieding system ................ ................................
236
8.2.5. Crops grown at the farm level ................................
237
8.3. Program mqdules ......................................
.......... ........
239
8.3.1” Fi&ld capacities module ......................................
.
239
8.3.12. Energy module ................ ...............................
240
8.3.2.1. Energy doing work .................................
241
8.3.2.2. Energy spent wa1kin.g ..............................
242
8.3.2.3. Energy for maintenance ........ ..................
243
8.3.2.4. Energy fiom LW losses ...........................
244
8.3.3. Feding system module ..................................
...
244
8.3.3.1. Energy tiom the fee,d ............................
245
81.3.3.2. Feed rations ........................................
247
8.3.4. F rm budget module ..........................
....... ......
249
1
8.3.5. 0 timization module ..........................................
251
81.3.5.1. Optimization objectives ...........................
252
d.3..5.2. Constraints and objective functim ................
252
8.3.5.2.1, Decision variables .......................
252
8.3.5.2.2. Objective 1Function ............ . .......
253
8.3.5.2.3. Statements of constraints ........... ...
253
8.4. Optimizati n output ....................................
... ,,..............
257
8.4.1. 1
S rver solution ..................................................
258
8.4.2. Inbelpretation of coefficients .................................
258
8.4.3.
sitivity analysis ....... .................................
..
260
8.5. Program
idation .......................................
.........
...
261
8.5.1.
ample of one session ........................................
261
8.5.2.
cussion on optimization output ...........................
264
8.6. Summary .....................................................................
265
CHAPTER 9
CONCLUSIONS AN-Il RECO~NDATI~SNS.~~ ,........<..,..... * . . . . . . . . *.a... 267
9.1. Conclusion5 . . . . . . . . . , . . . . . . . . . . . . . . . , . . . . . . . . . . ,...,,..,..,,.......I
. . . . . . . . 9 267
9.1.1. Sqrveys , . . . . ,, . < . . . . . . . . . . . . . . . . . . . .,~,.....
. . .
. . . . . . . . . . . . . .
_ 269
9.1.2. FMow-up of field operations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 271
iX
\\,
9.1.3. Maximum working capacity ................................ .,
272
9.1.4. Energy required for field operations ........................
272
9.2. Recoxnmendations ...................... ............:i ....................
273
APPENDIX A ,..._ . . . . . . . . . . . . . . . . . . . . . . ~ .<,.,...” . . . . . . ~ . . . . <.* . . . . . . . . . .
. . . . . I.....
277
APPENDEX B ........ .._ ..................................................
..... ........
285
APPENTIIX C . . . . , . . . . _. .,..........” . . . . *...*. . . . . . . I,.. ._.......... . ,_...........,..
296
APPENDIX a . . . . . . . . , . , . . . . . . , , . . . , . . . . . . . . . . . . . . . . . . . . . . . ..m.........................>.
320
APPENDIX E . . . . . . . . . . . . . . ...<... ~..” . . ..<............................... * . . . . . . . . . . ,,,... 3 2 4
REFERENCES , , . . , . . . . _, . . . , . . . . . , . , . . _ . . . . . . . <
. . . . . . ,
. . . . . . . . ,. . . . . . . . . . . . . . . . . .
346
X
-._. ----
../)UC_.--------I”I_-I
--I
----
--m.-m.--,
LIST OF TABLES
1. Estimated number ‘of draft animals in Sub-Saharan Afiica ............................
3
2. Number of animal drawn impfements in West .Africa .................................
5
3. Land distribution ........................................................................
12
4. Average farm characteristics in the Basse Casamance region ......................
1 8
5. Cereal yields in theiCasamance (kg/ha). ................................................
22
6. Lower Casamance bgricultural production trends in thousand metric tons ........
2 7
7. Reasons for removiing animals from service, .... ............ ........................
3 7
8. Age and weight to initiate training of cattle for work ................
. ............
4 0
9. Environmental factors on draft animals ..................................
..............
4 6
10. Work potential of ban and various animais #.. ......................................
.
4 8
11. Composition and digestibility of feed (DM basis). .......... ............
.........
56
12 Draft requirements! for farm implement (Equatorial Afiica). .....................
6 0
13. Non-working daysi for the West Af?ican regio’n ....................................
73
14. Perceut in yieid indrease with animal traction plowing ............... ...........
78
15. Analysis of variancje.....................................................................
9 4
16. Farm characteristios .......................................................................
115
17. Breakdown frequenci.es on farm equipment ........................................
127
18. Implement repair ahd maintenance. .................................................
135
42. Analysis of variante (LW,, CC). . . . .
. . . r.. , . . . . . . . . . . . . . . . . . .
175
43, Repartition (%) of hraft oxen time., ........................ . ........................
179
44. h/fana,gement and handling of draft animais (@ the fa.rm level.. ....... ...........
181
45. Pasture grass species in zone 4 and 5 ...........................................
183
46. Liveweight (kg) va$ation of pair of oxen.. .
. . . . . . . . . . . . . . . . . . .
185
47. Average liveweightirrariation of Traction Group.. . . . . I . L. , . . . . . ‘
. . . . .
185
48. Liveweight losses abd gained (June to Cctober). . . . . . . . , . , . . . . . . . . . .
185
49. Average soi1 texture of the tria1 site .....................................
. .........
192
50. Annual rainfall (mn$/rank at Ziguinchor (1960- 1996). .............................
202
5 1. Probability (PWD) and number of working da.ys (NIVDA.YS). ...... .........
204
52. Dimension ofyokes used by farmers(survey results). ...............................
207
53. Two+ample t-test for YOKE 1 vs Y0K.E 2., ............................
.........
207
54. Factors’ levels in a hx3x3 Randomized Block design ....... ......................
208
55. Pull Force PF (daNb ,with the UCF 10” piow. ......................................
212
56. Pull Force PF (daNb with the 3-tine SINE 9. ..................
........ ...........
215
57. Pull Force PF (daN) .wi,th the ARARA ridger., ....................................
216
58. Parameter estimates of PF regression ...............................................
219
59. Pull force PF analysisr of varîance ...................................................
220
60. Analys,is of variante (SWC and EQTYPE). .........................................
221
61. Analysis of variante of Specific Soi1 Resistance SSR (daN/cm’). .................
222
62. Field work charactdristics and Field Efftciencies ............ ................. ....
224
63.ANOVAofE(%) .......................................................................
225
64. Average implelment’eflkiency E (%). . . . . . . ” . . . . . . . . . ~
225
. . .
XII1
. ._..
Y
/
65. FW contained in some available feed., ................................................
248
Al. Animal-drawn equipment questionnaire ..............................................
280
BN 1. Maximum draft trial (Gl PI). .....................................................
...
285
B2. Maximum draft tria1 (Gl P2). ..........................................................
286
B3. Maximum draft tria1 (CT1 P3). ..........................................................
287
B4. Maximum draft trial (ci2 PI) ..........................................................
288
B5.Maximumdrafitrial(G21?2) ......................................................
.....
289
B6. Maximum draft trial (G2 P3). .................................................
........
290
B7. Maximum draft tria1 (G3 Pl). ......................................................
...
291
B8. Maximum draft trial (G3 P2). .........................................................
2!?2
B9 Maximum draft tria1 (G3 P3). ...........................................................
293
BIO. Draf’l animals LW and CIRCUMF ..................................................
29 1
Bl 1~ Regression analysis »fthe 18 draft oxen ............................................
293
C 1. UCF 10” Block 1 @SWC-range 6-8% g/g (Extract). ..............................
294
C2. SINE 9 Block 1 @SWC-range 68% g/g (Extract). ............................ ....
296
C3. ARAR.A Block 1 @SWC-range 6-8% ggl (Extract). ............ ..................
298
C4. UCF 10” Block II @@WC-range 8-10% g/g (Extract). ... ........... ........ ..
3 0 0
CS. SINE S) Block II @SWC-range 810% g/‘g (Extract). ...............................
302
C6. ARARA Block II @SWC-range 8- 10% g/g (Extract). .... ........................
304
C.7. UCF 10” Block III @SWC-range lO-13?/0 g/g (Extract). ..........................
306
C8. SINE 9 Block III @SWC-range 10-13% g/g (Extract). ............................
308
C9. ARARA Block III @;SWC-range lO-13% g/g (Extract). ..........................
310
C 10. Output example of the CEEMAT data processing system ............... ........
3 12
.,
;,’
Dl. I;él for maintenanck FM ofworking cattle.. .“, . . .,. . . . ., . . . , . . . . . . . . . I. _..
320
D2. Composition, of so&xr Africain feeds (As-Fed Basis). .......... ................... ..
3 2 1
D3. Composition of sole African feeds (As-:Fed Basis). ..............................
322
D4. Fiekd operation exe/cu.tion tirne., ............ ................ ........................
323
XV
_^. _ _-. .._II_.-. -.--._- -____-
----.-
LIS- OF F JGURE,§
1. Agro-eco:system zones i!n the Basse Casamance region ............................
1 id
2. Is.ainidl distribution fiom k’orth (Bignana) to SOU~~ (oussouye). . , ._, . . . . ~, . .
14
3. Moldboard plow UCF 10” ................................................................
114
4 . EMCOT ridger (Carrbia). ...............................................................
114
5. ARARA toolbar (Ridger). ................................................................
1 1.4
6. Sine 9 toolbar 3-tine .......................................................................
1. 1 4
7. Seeder Super ECO ...... .................................................................
114
8a. Types of implements used by farmers., ................................................
Il?
Ilb. Implements’ manufacturers ...................... ,,.....................................
1 1 7
9. Implem.ents’ working conditions .........................................................
132
10. Acquisition mode. ..... ..................................................................
132
11. DraR animais distribution. ................ .............................................
139
12. Farm ,size (ha) distribution. .............................................................
143
13. Farm labor distribution.........................................................
........
143
14. First year in animal traction......................................... ....................
14s
15. Number of years of experience., ............... .......................................
145
I
16. Temperatures (min and max) at Ziguinchor (1996). ....................... .........
153
17. Relative Humidity (min and max) at Ziguinchor (1996). ............................
153
xvi
-- _ ._-_- -
-<“U”.““l--.“-n-Lq-W”“-
----
m--P
------------
/
.I
1 $a Pc Wet ~CursJe of Tr&tion group 1. .
.
. . . . .. . . .. . . .. . . .. . . .. . . .
.
163
‘_
18b. W,~Iking speed of draction group %
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
‘X63
19a Pclver curve of Traction group 2..............................................
166
190‘. Walking speed of ~Kmion Gpoup 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . .
160
20a. Power curve 0fTr~r:tion group 3..................................................
169
20b. Walking speed of traction Group 3. ...............................
...... ...
163
21a. LW= f(ClRCUM) jregression line (18 oxen). .................................
....
176
21. b. General regressiod line ....................................................
........
176
22a. LW variation of in&vidual ox .....................................................
E84
22b, Liveweight variation of Traction Croups ..........................................
1.86
23. Drained soi1 water dontent (undisturbed cores). ........................
.... . . .
1.94
24. Log-SWC vs time .....................................................................
194
25. Penetrometry Energhr ..............................................................
1137
26. SWC at penetrometby sites ..........................................................
197
2’7a.Monthiyraindistribution(5yearsoutoPIO) ......................................
205
2’7b. PWD at 50% probhbility level.. ..............................................
205
28a. Montly min distribj.rtion (9 years out of 10). .....................................
206
28b. PWD at 90% prob/ability level.. ....................................................
206
29a. PF with UCF 10” at SWC= 6-g% yoke== 90 c:m .....................................
210
29b. PF with UCF 10” at SWC= 8-10% yoke= !30 cm ................................
210
29c. PF with UCF 10” at SWC= lO-13% yoke= 90 cm ..........................
. ....
210
30a. PF with UCF 10” at SWC= 6-X% yoke== 120 cm ...................................
211
30b. PF with UCF 10” $t SWC= 8-10% yoke= 120 cm ................................
21 1
xvii
-...
1,
/
~OC. PF wiith UCF 10” at SWC= lO-13% yoke= 120 cm. . . . _. . . . . . . . . . . . . . . . ~. . . . . . ,. 211
31a.PFwütbSINE9atSWC=6-8%yoke=9Qcm
.,.......... ..-..~ ............LI....I.... 213
3 1 b. PF w:ith SINE 9 at SWC= S- 10% yoke- 90 cm .....................................
213
131~. PF wEth SINE 9 at SWC= 10”13% yoke= 90 cm ...................................
213
32a. PF wltb SINE 9 at SWC= 6-8% yoke== 120 cm.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
213
32b. PF with SINE 9 at SWC= 8-10% yoke= 120 cm ...................................
213
1
3%~. PF wiith SINE 9 at SWC= lO-13% yoke= 120 cm ..................................
213
33a. PF with ARARA at SWC= 6-S% yoke- 90 LL ....................................
217
33b.PFwithARARAatSWC=$-lO%yoke=90cm..
.................................
217
33~. PF with ARARA at SWC= lO-13% yoke= 9LI cm.. ...............................
217
34a. PF with ARARA at SWC= 6-8% yoke=s 120 cm.. , . , . . . . . . . . . , , . . . . . . . . _ . . “. . . 217
34b. PF with ARARA at SWC= S-10% yoke= 120 cm.. . _. . , . . . . . . . . . . . . . . . . . . . “. . . 217
34~. PF with ARARA at SWC= lO-13% yoke== 120 cm... . . I . . . . . . . . . . . . . . . . . . . . .
217
3 5. Animal Traction Eval:uation Mode1 diagram. ............... .... ,,....................
231
C 1. Rairfall distribution at Djibelor Research Station .......................... .........
296
C2. Soi1 water content at Djibeior Research Station.. . . . . , . . , . . . . . . , . . . . . . . . . , . . . I.. . .
296
C3. CEEMAT data recording and processing system., . . . . . . . . , . . . . . . . . . . . . . . . . 319
El-E22 Screen 1 to Screen 22.. , , . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . ..<................. 324-343
/v
. . .
XVIII
lb--
_- ._.- ...---
ABBREWLATIONS
AIAEI.30: .Association des Jeunes Agriculteu~rs et Eleveurs du Departement de Oussouye
(Basse Cas~amance, Senkgal).
AID : Agency for Interriational Development (USAID).
ARC : Agriicultud Research Councii (England)
ASAE: .American Society of agriculturai Engiaeers.
ASCE: American Society of Civil Engineering.
CADEF: Comite d’Acti@n pour le Developpement ,du Fogny (Ba~s~e Casamance, Senegal)
CEEMAT: Centre d’Etude et d’Experimentation en. Machinisme Agricok ‘Tropical
(France).
CIRAD-SAR: Centre de Cooperation 1nterna.tional.e en Recherche: Agronomique pour le
Deveioppement - Departement de:s Systemes Agro-alimentaires et Ruraux.
CPCS: Commission de Pedologie et de Cartographie des Sols (France).
CRA: Centre de Recher/ches Agricoles
CTVM: Center for Tro$ic;al Veterinary Medecine (England).
FAO: Food and Agriculture Organization.
IRAT: Institut de Recherches en Agronomie Tropicale(France)
ISRA: Insti.tut Senegalaks de Recherches Agrkoles (Senegal).
MAC : Mission Agricolk Chinoise (China).
NARS: National A.gricultural Research Systelm.
ORSTOM: Office de Reclherches Scientifiques et Techniques dans les Territoires
d’outremer (France).
X.iX
t
I
l ,
PIIJAC: Projet Integre pour le Developpement de ia Basie
Casamance (Senegal)
SOMIVAC~: Societe pour la Mise en Valeur de la Casamance.
(Senegal).
c
!LX
--
_ _- .-.. -. ..----.
LHI’ OF VARUBLES
AGETRAIN : Training age of draft animais
ANERGY : Energy Dellvered by the Workin,g animals in MJ.
ANIPEX : Draft Animal :Reliability in % or decim:il.
ANNAVAn j. Probability of Draft animal Unavailability in %.
ANPOWER : Power Delivered by the working ianimals in J/sec.
ANSPEED : Draft anima.l Working Speed in m/sec.
ANWKGHR : Numberlof Hours Worked per Day.
AVGLW : Average livbweight in kg.
CI : Cone Index.
CIRCUMF ; Animal th/orax circumference or girth in cm.
CROPRICE : Market price or value of trop per kg.
DE : Digestable Enet-& in MJ.
DM : Dry Matter confent in percent
DN : Soi1 Drainage ini mm.
DEP * depth of penet&&on In the soi1 in cm(penetrometer).
DISTRAV : Distance kraveled in m.
DR : DraR required in d.aN.
DUC, : Drained Upper Limit.
Ed : Penetration Enerbr of the penetrometer a.t depth in J.
xxi
EFC : Effective Field Capacity in ha/day
EFDTM: EtXective field time in sec, min, or hr.
:EPWD: Effective Probability of Working Day in ‘%.
EQTwE: EqUipIIK!nt type
ERE : Energy Required to pull the Implement in Joules (J).
ET : Evapotranspiration in mm per day.
FC : Field Capacity.
FDOP: Field Gperation
FDTM: Field Time (hrs/ha:).
FRFEED : Amount of fie& Feed Required in kg.
Fr : Frequency (decimal number) of precipitation occurence.
FS : Soi1 resistance Force: to the penetrometer bar ,in kgf
FMSIZE : Farm Size in ha.
FU : Forage unit
FUM : Amount of Forage Unit for Maintenance.
NFWKER!3 : Number of Fana Workers.
Gk : Gross Energy in MB.
GSWC: Soi1 water content in g of water per g of soil.
HGT : hei;o;ht of falling weight in cm (penetrometer).
HI : Soi1 Hardening Index.
.
IMPFCGST: Emplement fixed cost.
IMPLEFF : Implement Effoiency in %.
IMPLREL : Equipment reliability in %.
a
XXii
IMFLWDP : implerr@l Working Depth in cm.
IMPLWWD : Implement working width in cm.
LL : SWC Lower L$nit in %
LW : Draft animal Liveweight in kg.
LWL : Liveweight Losses in kg/day
L,WP : Percentage o$LW used for traction in % or decimal.
MAE : Maintenance !Energy in MJ.
MATURELW : Dra& animal mature liveweight in kg.
ME : Metabolizable/Energy in MJ.
NE : Net Energy in h.
PF : Draft animals &lling Force in kgf or daN.
Pr : percentage probability of occurrence.
RF : Amount of raimal in mm.
RO : So~l1 water runoff in mm.
SBD : Soi1 bulk deniity in g/cm3.
SSR : Specific soi1 resistance in N/m2.
SWC : Soi1 Water content in % g/g
SLWGHT : Penetrometer Sliding Weight i.n kg.
TDN : Total Digestible Nutrients in J per kg
VSWC: Volumetric ‘soi1 water content by volume in cm3/cm”.
NWDA’YS : Numbek of Working Days in days.
SWHC : Water holding capacity.
SWP : Wilting Point in %.
a.<
XXIII.
xxiv
l,l. ~~?S~~~~~~i~~ ovt+rview
In eontrast with fa,rrners in industrialized comtries, darxnerrs in de~elc~pkng mmtiies
are gener aily 3rnaBB knld&s who must contend w% increasing ~~~~~J~~~~~~ ~rrxx~re m th:a
land and a !ow levd ofr$eckanizatiotk ta pr,~motz e£âective erop ~intefosification P~~p~~~at~o~
pressure hais skmged m~st of tke traditiorral âamting systems by ~F&%g skort ~C&V!~ RDQ
pennanerd a-opping. Thiese new practices bave led do more soi1 degradation.
in order tu abscirb the impact of yopamla~ion growtk, govemrnent:j irr iJwelo$ng
countries urge farniers to produce more fix>d by improving agrim’ltwaE g;nda;ctivFt;~. Such
improvcments deptxrd irji gcneral,, on tke t-o~nblnâtion of a variety of energy mwces:
mechaniçat brnackinea, $nimals, li:~;r;ans), çfm&xi $ert%zers, pesticides:, ker!Scide~:, ztc; .)
or biolugical (improved~ vanieties, manure, ru:ompmt, etc.). As poinzted OIZ: b:r Esmw and
Hall (19’73), mechanizaiion is one of’the most cri~.ical production inputs 1 o increase cr~p
yields tkrough better cdntrol oftke sai2 watitar I-F: ,;ime (runoff’versus itiltration~, better miF
preparatiori for soi1 aerttion and mtricnts wail;SIity, more &flcient weed wn?rol, and
better timeliness. Tkey dlefined agricultural me: Ihanizatïon as “tke art andi scientific
application ofmec.haniwl aids for increased pmdwtion and presexvation offood and fiber
\\I
?
/’
çrops with less drudgery a:nd increased efftciency’“. Separate studies conducted in most
of the developing countries in Asia, Afirica and South AmeriCa have shown that human and
animal po,wer are probably the most plentifitl source of enere,v (Patrick, 19’93; Campbell,
1990) towards the establishment of a sustainable, mechanized production system. The
switch from hand cultivation to an animal traction based production system is viewed by
Jaeger (1!384) as a Sign&ant step in creating mo:re production opportunities and
increasing retums for farmers through better land. preparation and timeliness. The level of
investment involved in the technobgy is lovv cori..,)G,,. nd to trwtorization. Draft animals are
available and fed locally and thus save foreign exchange (Roosenberg et d, 1987). The
energy provided by the combined two sources (draft animals and tractorization) represent
gO% to 90% of the total eneirgy used in agriculture throughout the world today (Campbell,
1990), The same study repo:rted that there are over 1.5 billion animais used as working
animais in the world compared to 22 million tractors. The species used to generate the
energy needed for agriculture are mainly cattle, water buffalo, donkeys, mules, horses and
camels.
The level of mechanization and intensity of draft animals’ utilization is highly
,variable fi-on1 one region of i;he world to another one. The focus of this research is the
situation in Sub-Saharan Africa and in Sene,gal in particular.
1.1.1. Sub-Saharan Africa
In Sub Saharan Afric.a, animal traction is one of the most important energy SOUK~S
for agriculture (Table 1:). I[n some areas it protides up to 90’5’0 ofthe power required for
trop prodiuction. The level of farmers’ experience varies fronn one country to another. In
-
. -
“**O)UIm--.-
- I I ” - - - - -
--“-
----m-II--
3
Table 1: Estimbted numbers of draft (mimals in Sub-Saharan Africa
RegionKountry
-
-
Cattle
DoRkeys
Horses
- - i
, - -
m--e-
- - - . - _ - - - - - -
West-P&ica
- - - - - - - -r
..-m
Benin
30-40,000
Burkina Faso
7s80,000
Chad
105~130,000
Cote &Ivoire
30-40,000
The Ga.mbia
18”20,000
2S-30,000
s-7,000
Ghana
20,000
1,000
. .
Guinea
100,000
Guinea-Bissau
4,000
Mali
200-3 20,000
Niger
1620,000
Nigeria
1 o’o-200,000
Senegal
130~140,000
200,010O
Sierra Leone
1,000
-
Togo
-
---_.
9- 10,000
-
East Afiica
Central Africa
-
-
-
-
Angola
Cameroon
C. Af. Republique
Zaire - - - -
South Africa
---
Botswana
350-360,000
Lesotho
180,000
Madagascar
260"330,000
Malawi
SO-70,000
Mozarn’bique
100,000
Zambia
180-315,000
Zimbabwe
500~800,000
\\ I
4
/
Botswana, 80% of the fiarmers use animal traction for piowing. Zimbabwe has the largest
draft animal population in th.e region (Starkey and Faye, 1988). Falvey (1986) pointed out
also that one of the most significant impacts of the utilization of working animais is the
increase of cultivated area per active household member: 30 to 40% in Senegal and 40 to
70% in Mali the neighboring country of Senegal.
In ‘n’est Africa, animal traction was introduced before the era of independence in
.i
the early 1960s through the implementation of various agricuItural projects oriented
J
towasds the production and expert of cash crops igioundnuts in SenegaI, cotton in Mali).
The promotion of animal traction for cereal production at that time was almost non-
existent. The introduction of the new technology by the colonial administmtions took
place between 1905 and 1945 in different places stretching between French Guinea,
1
Senegai and Nigeria. The main actions taken were the integration of trop production and
livestock husbandry (mixed farming) in northern Nigeria (192!4), the introduction of plow
farming in northem Ghana (1938) and the implennentation of research on ,animal-drawu
implements in Senegal(1928). In the process, most of the farm implementrs were
originated fiom France or specifically manufactured and adapted for Afiican conditions:
the MANGATM cultivator for Burkina Faso, the A.RIA.NATM, TROPICANATM and SINE?
for Senegal,. the BAJACTM moldboard plow for Mali.
Historical analysiis shows that adopti.on rates by farmers were the greatest in the
first 10 to 20 years after the independence (1960 ,- 1980). This period represents the most
dynamic period in terms of animal traction projects implementation in the regior, (Le
Moigne, l981) (Table 2). This period was also characterized by the creati.on and
development of animal-drawn equipment factories like SISCOMA (1961) in Senegal.
Table 2: Ndmbers of animal drawm impfemen’ts in West Africa
(1957 and 19iW84)
’ MPT: Multi-P’urpoise Toolbars
2 WTC: Wheeied Tooi Carriers
3 Ariana multicuheur
4 Figures from 1974
na: non available
Sourt:e: Extracted fiom Lawrence and IPearson, 1993
.
‘The major research issues addressed during this implementation and adoption
phase to tmprove the technology were as fcbilow:
‘.
- Draft animal :seIection (cattle, horse, donkely)
- Yoke design and fabrication (head, neck-yokes)
- Implement Aection (single or multipurpose)
- Field operation techniques (tillage, weeding,)
1.1.2. Senegal
In the case of Senegal, cattle were the first option for energy supply to agriculture.
The technology was tested in 1925 for the first trime. In the castration and yoking of
ox, farmers were introduced to a new technology that could. generate much greater power
than used previously. The tec;hnology was introduced through the French color-ial research
system RAT. T ‘le main objective was to improve land and labor productivity, and to
ir;.*-case to: 1, cuitivatedl areas and trop yield. The strategy was te produçe enough
groundnuts andIor cotton for expert with regard. to foreign exchange and also to use part
of the production to supply the grain to the emeaging national oil Industry
The major breakthrough in trop production with draft animais came around 1945
when equine and donkey traction were promoted at the farm ieveï to perhorm light works
like cultivatilon and transport. Nong with groundnuts, the cereal production (millet, maize,
and sorghum) reached satisfactory levels satisfying fart fàmily c:onsumption as weli as
supplyin,g local markets with the production SU~~~AS. Progressively, bovine traction was
o..,m~I,m~----..--.,-._ v--m-
------
-----m-œr-
oriented towards the exec@tion of heavier tasks like plotiing in medium and heavy soils,
and carting of bulky loads.
.e
The adoption of dr#& animal technoio&y by farmers was mainly supported by the
rapid increase in areas crogped with cash crops (major source of incorne), the availability
oflocally m’anufactured iqplements and the existence of credit programs to help farmers
purchase their own equipfllent. The animal traction based production system appeared to
be reproduc-ible, as the majior components involved were a11 readily a.vailable to farmers.
Most importantly, drsft ax$mals like oxen were simply borrowed from the household or
village herds.
However, the pro#notion of animal traction by governmental extension agencies
must face regional, cutturti and environmentai diversities throughout the country
Constraints related to thebe aspects have dela;yed for many yea.rs its implementation in the
Casamance area, located Ln the southem part of Senegal.
The latest sutvey çonducted at the national level showed that the draf? animais’
population in Senegal w&; composed of i30-140,000 cattle, MI-1 FLO,OOO donkeys and
200,000 horses (Lhoste, 1988). The Casamance province accounte~ for less than X 5% of
the national totals. Amoig other types of constraints to have delayed the implementation,
were the animal sanitary ic:onditions of the area in relation with the prevalance of the
sleeping sickness. This cbaracteristic has prevented until recently (!in the 1990s) the
introduction of donkeys ia.nd hor,ses for draft purposes.
In the Basse Casmance region, the prevalence of trypfanomiasis disease
transmitted through the tse-tse fly tumed out to be the major barrier as the sleeping
sickness introduces a setious limitation in the working capacity of the draft animals. The
w
/
8
most reliable source of draft energy is represented by the Ndama cattle, &s taurus, (87%
of the total), .while the utilization of horses and donkeys (3% and 10% respectively) is still
marginal. The Ndama breed bas more toleranc;e to ,the disease and is well-adapted to warm
and humid ckmate, The castrated Ndama male oxen are used for traction whiie the cows
are kept in the herd for their milk production. The draft team is always composed oftwo
oxen of simibar weight.
1.2. Statement of problem
TO drlte in the Casamance province, ~gricuhural production based on animai
traction is below expectation. Its promotion .initiated since the early 1960s did not bring
about the expected increase in food production. In fa&, the leveï of adoption of the
technologj by farmers appears to be very low for several reasons mostly sanitary and
economical’. in the process of adoption, fat-mers are very selective in choosing implements
in relation to the level of thel first investment in terms of affordability (Ndiame, 1988). In
general, f&mers justify the use of animal traction for land preparation because it is faster
and presents less drudgeq than using hand tools (FALL, 1985).
Fat-mers in the South did not adopt the technology until some changes were made
to the yoking system. Fa,rmers use different size and multi-purpose yokes for bath
cropping (head yoke) and transportation (neck yotke). Among other factors, the design of
the yoking system was found to affect the working capacity of oxen. In terms of
efEciency, the fitness represents an important criterion when evaluating the work output of
’ The process and level of adoption of improved technologies by farmers in the Basse
Casamanee region is being investigated by a PhD student at Kansas State University.
‘I*.vm%r-,.---“-.- ---a----
---
---I<IuII
I,
9
/
draft anima!s. The yeking system assures cornpatibility In the hnk bettveen the animals
(power source) and the hitched farm equipmcnt (implement irrr cart). A less effective
yoking system is expebted to turn into an appreciable source of potential energy wastage
and tc cause rapid fati!gue if harnessed animais are not working together as a team. A poor
teaming also, afI?ects the quality of the Be1.d operation in terms of timeliness, effective field
capacities: and jpoor ~4ork quality.
Thc real relationohip between the design (width) of the yoke used and the working
conditions (type of im@ement and soil) is not known well enough. to help predict the
amount of pulling forde and energy required to perfarming the diiferent field operations.
In the decision-nnaking process it is important to fit the job to the animnls by choosing the
more suitable team of &en to perform the task How does the size of a head yoke affect
the draft required to pull different types of implement in a sandy textured soif at different
moisture regimes? Thti: determination of this type of relationship is very importa.nt in
evaluating the expectekl role of draft animais in! farmers’ crop,ping system.
1.3. Objectives
The objectives of this study were as follows:
1. TO develop a methodology that Will better assess the Impact of animal
traction at the.farm level.
2. TO develop an expert system computer program to evaluate and prediet
the impact of animal traction at the farm level b;y using a Knowledge
Base; h$anagement System (KBIMS);
1 0
A. significant amount of information on animal trac,tion is available through the
literature but is not systemized or structured into a standard ~&II. The deveiopment of the
expert system a.ppears to be the star-t of such improvement to ease decision-making. In
general, a Kairly Iarge number of technical parameters are involved in the decision making
process. It is a real challenge for decision-makers to integrate local and scientific
knowledge into an accessible structured database. The expert system is expected to
provide the bridge between scientists and the rest of the commumty.
In combining artificial intelligence (AI) and conventional computer programming,
expert systems have been used in a wide range of applications (Raman et al, 1992).
The work presented in this study is organized as follow:
- Chapter 1: Introduc:tion.
The introduction presents an overview of agricultural mechanization in the worid
and specifically in the developing countries and in Senegal in parti,cular, with a focus O~S
sma.Il-stalle farming systems. The statement of problem and the objectives of the study are
afso introduced.
‘- Chapter 2: Presentation of the area of study.
This chapter describes the production and. cropping systerns encountered in the
l
area of study. The notion of agro-ecosystem is introduced and discussed from the point of
view of animal traction utilization.
- Chapter 3: Determinants of performance of draft animals.
The discussion presents different relevant studies conducted on animal traction
a
evaluation in various parts of the world with an ernphasis on Senegal. The main
*a
~**uIII-““,.-
-----.“..---l_l_
I-m--
--
II
11
,’
determinants and parameters of performance are presented.
- C%apter 4: Mat$rials and methods.
The organization of the field data collection aiong with the expert system building
is presented chronological~y with an emphasis on the different statistkal designs.
- C%apter 5: Farnj characteristics and systerm management.
- C%apter 6: Workîng capacities of draft i3Il.imdS.
- Chapter 7: Fielb operations and draft requirements
These t.hree chapters deaf with the presentation and discussi,on of the analysis
performed. In the course if analysis, the resuhs are: compared to existing results in the
literature for first hand validation.
- Chapter 8: Expert System to evaluate the impact of animal traction in the Basse
C a s a m a n c e regqon.
This chapter descrkbes in details the proce ss of building the expert system program.
It shows how the collecteb data and models developed in the analysis part are used. In
some cases, tbe collectedidata are reinforced by relevant secondary data from previous
studies conducted in the &-ea of study by the authcr or by the IS~CRA Djibelor’s
Farming Systems Researdh Team.
‘- Chapter 9: Conclusions and recommerrdations
What are the per(pectives in animal traction studies in the Basse Casamance region
and in Senegsal in generalk What. are the expectations placed on the expert sy!stem
program? Other similar c(uestions are addressed.
Chapter 2
PRESENTATION OF TEIX AREA OF STUDY
2.1. Presentation of the area of study
The Basse Casamance region is located ix1 the southern part of Senegal and covers
7,300 square kilometers (km2), etiending from the Soungrougrou Valley to the Atiantic
Ocean (Figure 1). With population densities ranging from 27 to 60 inhabitants per square
kilometer, there is increasing pressure on land usfe in relation to t:he weather (Table 3).
Table 3: Land distribution
Designation
rea occupied (h
- - -
-
-
-
-
Cultivated areas
109,500
Mangrove (swamp wood)
124,100
Estuary and sedinnent
102,200
Forest
I89,3QO
IResidence area
204,400
- - -
-
‘Tot:al
730,000
-IE-
-m-
-S~u;ce: PIDAC ,Ei by A. FALL (1985)
-
/
'-7I
No a,nimal trac&ion; Wpland
SENk/WL
'.
3: Manding labor organizatlon
b.,,,
Lim.ited animal traction;
-.
Rice. important.
!pAku,
J: Manding ls&or organization
'. -4
Animal traction important;
\\\\
Upland cropa important.
t
k
5: Diola labor organization
\\ "Y =- .----..? *:,.*
ata!drr
b
-+--p,
Anima1 traction important;
- - - J
Uplnnd srops +aad rice.
Figure lr Agro-ecosystem zones in the Basse Casamaniee regiom
i
-I-I.
--CIc-
_..,----
--
‘.
I
- _ .
_ .
. I
_ .
. .
- ”
-
.
-
”
I
. I
-
. _
- ,
.
.
.
.
.
.
.
.
.
.
.
I
.
. ,
.
.
.
1
.
< .
.
.
. ,
.
.
,_
,.
.
.
.
<”
.
.
.
«
.
.
.
.
.
.
.
. ,
.
.
.
.
.
.
.
.
.
.
“ .
.
, .
.
<.
.
.
.
.
. ,
.
.
.
.
.
.
.
. . . ~ . . .
. . . . .
MaY
Jkme
July
m.lg . sept.
act. N o v .
-a-- Normal
* Zigpnfnchor 92
-a- cHxxmly0 92
4Bignona 92
Figure 2: Rainfall distribution frrsm North
(Bignona) to South (Oussouye) and during normal years
The IDiola ethnie group represents 83 percent of the total ipopulation, foilowed by
Bainouck (6%), the Manding (5%) and others (6%).
The climate is sub-Guinean type with a large maritime influence. The occurrence of
one rainy season per year from the month of June to October represents the main
characteristic.
The last two decades bave witnessed a sh’ortage of rainfall throughout the whole
countcy and in the Basse Ca.samance region as well, where the type of landscape has
nJIwr-,.-P..*mm--
_-- -.-._-- -
--m-
--
--Q.
---
induced significant changes in the cultural practices. The amount of rainfall deoreased by
an average of :13X, from 1,500 mm during normal years to 1~~100 mm (39.4 in) (Figure 2).
The lowest amount of railbaIl, 800 mm (3 1.5 in) aras recorded in 1,983 in Ziguinchor,
capital ofthe region.
The topography c$ Senegal is Bat with an average land level elevation of 200 m.
The physical aspect of the Basse Casamance regior: ^ : “aracterized by a dense network of
swamps and srnall stream/s flowing into the Casarnance River. The landscape of the entire
region is d(ominated by a \\Succession of upland and lowland areas. During drought periods,
the smah streams offeredj privilege waterways to :salt intrusion from the Atltintic Ocean
through the Casamance @ver to the iniand valleys following the upstream-downstream
water movement.
With the increase in sait concentration throughout these years, combined with the
low pH of sail and watek, advanced cherni reacrions started takirrg place’in lowland
areas to compromise their agricultural productivity (especially for rice production). Soi1
surveys conducted in the region yie~lded the tremendous figure of I80,OOO ha nosn-
productive, salty soils @art-y, 1986; ORSTOM, 1983). Most ofthese soilis are located
along the CJasamance Dover and its main tributaries (Guidel, Kamoubeul, Soungrougrou,
Baïla).
According to thkir position along the transect, crops are designated as either
upland (groundnuts, maire, millet, sorghum) or lowland crops (mainly rice). AI1 crops are
rainfed.
Several field studlies have been conducted to approach the agricultural and ethnie
diversity of the area. O~IE major finding was the: identification of Gve agro-ecosystem
_P
,’
1.6
zones offering different agricultural production and cropping opportunities. Three criteria
.
were r,tsed to achieve this zoning (Equipe Systemes, 1984):
1. Labor organization at the fkm level between male and female in relation
to the trop to be grown and the agricultural operation to be performed
(tillage, weeding, etc);
2. Ratio of upland and iowland areas in terms of relative importance of
cultivated areas and types of trop;
3. Importance of anime1 traction utilization in the cropping system and the
associated level of rnechanization.
Each of these five agro-ecosystem zones presents different agricultural activities in
reiation to the type of the natural resources avaiiable (type of soil, water quantity and
quality, forest) and the most dominant tthnic group (Equipe Systémes, 1983).
In response to the decrease in Foductivity, farmers developed new strategies to
meet their food security requirements ;Posnler et 4, 1988). The coping strategies to deal
with the, new climatic conditions (drotight, shortening of the rainy season and shortage of
rainfdl) were: primarily based upon a jrogressive transfer of resources from lowland to
upland areas. These major changes ir farmer’s traditional cropping system affected mostly
the use and land tenure at the villag: level. The very first noticeable effect on the
environment was the deforestation, vhich srarted taking place since the early 1970’s to
induce a disequilibrium in the land a!.xation process to different activities. The objective
of the deforestation was to satise tb increasing demand in cropping land. Through this
--
““lll,rl-,.------._..~-..
---
-
-...-..-
3-
--1
--
\\r
/
13
type of management strategy, farmers tried in igemeral to secure production through the
expansion ofupland C$O~S rather than through intensificatio%i.
In an agro-ecobystem characterized by an availability ofupland axeas, the use of
animal traction tumed; out to be the factor contributing most towards the expansion of
fields cx-opped with cabh, trop (groundnuts) to the detriment of cereals such as millet and
maize. The cereals ard now mainly confined in backyards and compound fields. Within a
short period of time (l+s than ten years), th.is type of cropping system afft:cted the
precarious equilibriurn which bas always existed between agricultural and livestock
activities marked by tlie progressive disappearance of long term fallow along lwith village
level grazing areas for! livestock.
In the process,, the meclhanized fieid opemtions on newly cleared 8and brought
about advanced stage4 of soi1 d.egradation to compromise land productivEty. In most castes,
the lack. of good mana/gement practices has led. to alarming soi1 losses.
2.2. Types of Producih system
The agro-ecos@tem zone is defined as a system in which farmers; are carrying out
productive agricultura] activities within their natural environment. The e@vironment is
mainly composed of &ural resources (soil, water, vegetative caver) and: represents an
area homogenous eno$gh in terms of constraints and production opporttinities. The most
Sign&ant observatioi unit to evaluate interacticms and performance of the system is the
fann level (Tabfe 4). $%Lis level of aggregation represents a finalized production system in
itself as it comprises tbe tenter of the decision-making process which identifies agricultural
production objectives ‘and decisions upon resources allocation (Sonko anc! Fall, 19%).
..^_~ . ..- .^ -..
-- .
_--_--_--- -- .
L
d !
-^-_-.
_
_.
18
_-.
--
_...
--_-
_
..--
--
-
--.-
--
2.2J. Agro-ecosyst/em 1 (Oussouye)
Located in the mostjsouthern part of the reg,ion, this agro-ecosystem is composed
of a tigh proportion of lowiand areas. The landscape is domina.ted by swamps, saline and
acid soils (,‘“sol de tannes”)Yiand very few upland fieyds (village sites). The ratio of lowland
to upland areas favors by far lowland areas. Transp~anted rice is the major agricultural
activity (Equipe Systemes, i1984).
Field operations are performed according to the Diola tradition of so&l labor
division by gender: men p&form land preparation while women perform rice nursery
preparation, transplanting bnd harvesting. Women also help in weeding of upland crops
confined iri small fields.
Land preparation i.b manual and starts whenever enoug,h water is stored in the rice
fields. As the cumulative dainfall increases from the beginning of the rainy sehrion, a
network of small dikes difiding the main rice fields into small subplots, 10 m2 to 30 m2 in
area, is used to capture wiater runoff. Farm sizes are generally small averaging 1.3 ha per
household. More than 60% ofthe fields are c:roppc!d with rice.
Ether non-agricuhural acti,vities contribut:: significantly to the total farm revenue:
fishing, prodtrcts from the forest (palm oil ex.trac.:ion for example).
Gvestock activitbs are very important; cattie, poultry, swine are produced. Animal
traction is not used in agricultura~ activities for tiifrerent reasons. The primary constraint is
sanitary with the prevalence of trypanomiasis (G&ssier, 1966; Traverse, 1974). The
alternative to animal traction is to use rentes pcwer tillers fiom non-govemmental
organization (CADEF, &JAEDO). However, \\ery few farm activities are mechanized
through this rental sch&e.
\\,
20
2.2.2. Agro-ecosystem 2 (Blouf)
Located in the northern part of the Casamance River; -agro-ecosystem 2 represents
a continuzrtion of agro-ecosystem 1 but with more available upland areas. More than 30%
of fieldis produce upland crops wbich are mainly groundnuts and millet.
The labor organization (at the household level is the same as in agro-ecosystem 1.
There is no use of animal traction since the big drought event in the late 197Os, which
significantly reduced the avaiiability of animals for draft pur-poses. Farmers are still very
open to the technology (Fall, 1988).
2~2.3. Agro-ecosystem 3 (Niaguis)
Agro-ecosystem 3 is located eastward of agro-ecosystem 1. The population of the
r
area is composed of a multitude of ethnie groups living in different villages, which
r.
continue to practice their original cropping system. Upland areas account for 60% of
cropped fieids.
The introduction of animal traction is still in progress: 30% of cropped areas are
m
mechanicaily tilled with animal-drawn moldboard plows. It is important to mention that
L
faTmers located in this area were the first in the region to experience the use of motorized
equipment for rice land lpreparation with the MAG project (1969- 1973). The
governmental regional extension service SOMTVAC-PIDAC took over the project in 1984
and pursued the rental of mini-tractors and power-tillers to farmers.
I,
. .
-_ --
<~cwY-ll.l~..-<-_
..__- “-
B----,--”
-----
-11--
-----
-.
.,
2 1
/
2,2.4. Agro-ewsystem 4 (Sindian-ECalounayes)
Agro-ecosystem 4 is located in tke northeastern pati: of the region. This area has
been culturally unded Manding ethnie group influences even though the population was’
originally Diola. Theisacial labor division according to gender is organized around the
type of trop to be grbwn. Men are in charge of a11 upland crops which ~cupy more than
84% of cropped areap, .while women are specialize in rice and vegetable production.
The use of animal traction is well implemented and most field operations, from
land preparation to gioundnut lifting, are mechanized. The a.verage fam ,size is around 4
to 5 ha.; rn’ore than S&!! of croplands are mechanically plowed with draft animais,
especially oxen.
0ver the last &y: years, with the emergence of many local farmei-s’ organizations,
mecbanization is beco/ming a major concern ~LYS the population. Farmers iare getting more
involved in the purcba/se of tractors of differernt sizes fused md new) wit&:in a wide pow’er
range (10 I:O 45 hp). The tractorization process is mainly orientecl towards improving the
timeliness of land predaration in the rice fields.
Different econbmical activities are conducted apart from ,the Aeld crops. Livestock
(ca&, sheep, poultry], tree cropping (fruit and palm) and vegeta.ble gar&:ns represent
important sources of *venue.
2.23. Agro-edosystem 5 (Diouioulou)
Agro-ecosysteb 5 covers the northwestern part of thie region. Upland crops are
dominant, and occupy~75% to 80% of a11 crop1:an.d. The labolr organization is Viola type
_ _ ^ .---- -
Y
2 2
/
based on division according to gender in relation to the type of field operation to be
carried out. Men perform land preparation on both the uplatid and lowland areas.
The use of animal traction is also important but to a lesser extent than in the
previous agricultural situation. The level of mechanization barely exceeds land preparation
for 60% of ail cropland.
2.3. Cropping systems
In normal rainfall conditions (1500 mm), the agricultural production of the Sasse
Casamance region is mainly oriented towands riche production. The amount of rice
produced was estimated to be around 75,000 metric tons representing 30% of the national
production. The shortage of rainfkll by 20 to 45%, along with the correlated salinization
process of most rice fieId SC&, has compromised the regional production to the n’oint that
the region has become a net rice importer since 1978 (Equipe Systemes, 1983; Jolly et &
1988). The coping strategies developed by farmers to countleract the decrease in rainfall
were tfo move to more diversification at the regional agricultural production system level.
Ta#ble 5: Cereal yields in Casamance (kgha)
Crops
Millet’Sorghum
.
Maize
Upland rice
Lowland rice
Aauatic rice
S&mp rice
Source: MDR, 1986 “Etude du 5iecteur Agrrcole:
Plan Cérèalier. ”
*‘“sn^la,--.^----.-._r__l--~-.-----
-..- --.-
. ..--.-
-u----.---u-
‘----“N-sll-mwanm,8q~~
.,
/
2:3
A bigger emphasis was put bn the cash trop (groundnuts) while areas cropped in maize
increased hy 19?/0 between 1970 to 1982. Other crops grown ‘&e millet and sor,ghum.’ The
yields measured in farmers bonditions are generally low (Table 5).
2.3.1. Groundnut broduction
The draft animal b+ed cropping system is mainly driven by groundnut production,
the only cash trop support/ed by an organized market system a.t the national l&el. The
cropping ofgroundnuts h& significantly contributed to the extension of croplwnds gained
by rapid deforestation and clearing of new lands: upland crops represent 75O/6 to 80% of
ail cropiands in the regioru/ and fields cropped with groundnuts alone accountis for 45% to
55% of a11 upland crops *ea. The average fa:rm yield is 954 kg/ha (Posner et d, 1988).
In the agro-ecosyhtems loc;ated North of rhe Casamance River (zones 3,4 and 5),
groundnuts are grown eit/her as the first trop in a 2 to 3-year rotation with tiillet!sorghum
production following lan/l clearing or altemating with an annual fallow (2-ybar rotation).
Whereas in the South @@es 1 and 3), the cropping pattern is continuous (I-year rotation:
groundnuts/groundnuts)/as
land availability is a major constraint,
Three varieties o$ groundnuts are grcrwn: 28-206 and 69- 10 1 (110 to 120
days/cycle). These are ijtproved varieties introduced by agricultural researçh institutions
and govemmental extenbon agencies. The third variety is local and is called “Bourkouss”
by the famers. It is a sqort cycle variety (90 days/‘cycle) and Is not suppo&d by the
officiai market system. $‘he local variety is widely used for late secding. The improved
varieties are mostly grobn for their high oil content and are cultivated as cash crops: 69-
101 in the North and 24-206 mainly in the South.
,!
2 4
Groundnuts are the only trop in the cropping system that is highly mechanized
through animal traction. In some agro-ecosystem zones, the level of mechanization
reaches the harvesting operation with the utilization of animal-drawn groundnut diggers or
lifiers. The main difference in the cropping pattem among farmers resides in the type of
land preparation: either flat (with a moldboard plow) or ridged.
if
23.2. Cereals
I !
Cereals (millet, sorghum, maize and rice) account for the remaining cropped area,
45% to 55%, of the total cropped area. Millet and sorghum cari be in either a rotation with
groundnut or altemate with annual fallow. Sorghum is vety seldom in a mono-culture
cropping pattem. It is generally intercropped with groundnuts. Al1 varieties used are local
Millet is important in agro-ecosystem zones with large Manding cuhural influence.
The seeding starts at the very beginning of the rainy season. The varieties used are local
(Sarrio) and have major pest management problems especially with birds and stem borers
(Pentatomidae Coreidae). Millet is generally seeded on ridge plowed land to protect the
seed and Young emerged Iplants against excess moisture. The land preparation is either
mechanically or manually performed in relation to the timeliness of groundnut field
operations.
,, j
Maize production is crucial to farmers as it represents a relay food supply for their
consumption. In general, the harvest takes place during the rainy season and before the
end of the month of September. Maize is mainly grown in agro-ecosystem zones 3, 4 and
5. The cropping pattem is continuous mono-culture (1 -year rotation) in soils rich in
organic matter. Maize fields are mainly located in the backyards of the farmer’s
,-,1-,-1w--
.-<-._ vpn---
--
-w-
IC
-
2 5
compounds on relatlvely small areas (less than 0.5 ha). Lately, there is #ai. trend to cropping
maize out ofthe badkyard on! newly cleared land, but the major concern to farmers’is soi1
fer’tility and the need of field ,protection against pests to secure production.
During the last 10 to 1.5 years, high yielding varieties of 90 day$/growing season
bave been provided do farmers by research institutions and governmental extension
agencies: composite !ZM 10 (yellow grain) and hybrid BDS (white grain). New high
yielding varieties (Synthetic C, Makka, etc.) continue to be distributed.
Maize is the second most commonly rnechanized trop afier groundnuts. The level
of mechanization involves land preparation, seeding and weeding. The intensity of
mechanical cultivatioh is determined by the type of land preparation (flat or tidge).
Ridging ta,kes place gbnerally when farmers are having weed problems ar excess moisture
in relation to the topography. Maize seems to be the trop that offers the largest margin of
progress in terms of ektension of cropped area and increase in level of production
(Ndiame and Fall, 19!!!9).
Most of the ride production is carried out in lowland areas. The cropping system is
rainfed. There are three types of rice fields ranging from strictly raimed (uipland) to aquatic
(lowland). The aquatic rice fields depend greatly on the accumulation of water from runcoff
within the watershed tb low areas inside the vallleys. The third type of rice field is assisted
by the rising water table, which provides enouigh moisture to the rice plant. The types of
soi1 used for rice production vary tiom sandy (upland) to clayey soi1 (lowland). A large
number of rice varieties (local or improved 110 to 120 days/growing season) are used
according to the field conditions in terms of so,il moisture, sahnity, pH, fertiiity, etc.
Tmproved varieties arelwidely used by farmers: DJ 684D, IRAT 133, IR. S., DJ 12-5 19
2 6
Historicahy, the mechanization of rice producti’on in the region started with the
introduction of the first series of power-tillers (DONG FENGTM) fiom The Republic of
China through the MAC project (1969- 1973). One ma:in objective of the project was to
improve productivity throu.gh better timeliness and good quality land preparation (FALL,
1985). The major constraint was related to the size and shape of the plots, which were too
small to facilitate the maneuverability of the power tillers.
The use of animal traction started earlier, in 1967-1968, but was mainly confined
to areas where researchers and extension workers were conducting experimentation
(TUVERSE, 1974). Farmers try to combine animal traction for land preparation and
seeding and tractors for lan preparation whenever possible.
Rice is generally direct seeded by hand broadcasting right after land preparation
except for the aquatic rice grown in lowland areas where transplanting is manually
performed by women when water is still ponding. Rice production requires effective land
preparation before seeding in order to improve weed control and to incorporate organic
matter into the soil. This field operation depends largeiy on the water content of the soil.
”
Other crops are also important at the farm level for their economical and
/
nutritional values. The are,a allocated to these types of crops is generally small: cow-pea,
_j
cassava, sweet potatoes and vegetabies.
The Basse Casamance Agricultural Development Master Plan published the
prediction of production trends for major crops in relation to extreme weather conditions
scemios (drought and normal) (Table 6). The potential increase in productivity to reach
these preciictions is real and depends mainly on how fiirmers Will be able to improve their
capacity for carrying out field activities on time
/ I I
.*UC-,.-.---.-----I
_*_-
- _ - *
__-__
~
..--
--
.-----------
27
Table 6: Lower Casamance agricultural producitiion trends
in thousand metric tqns
-
-
Conditions
-
-
Drought
‘Normal
- -
Year
- - -
1979
1985
1990
2000
1979
1!2%5
1990
2000
1. Senegal
Millet
407
419
430
452
415
497
577
‘779
Rice
6 8
7 0
7 2
7 6
7 7
;93
109
1 5 1
Maize
3 2
3 5
3 8
44
3 4
41
4 7
6 4
Groundnuts
- - -
613
708
798 1015
749
946
1150
1699
-
II. L. Casamance
Millet
2 5
2 6
2 6
2 8
2 6
31
36
4 9
Rice
2 2
23
23
2 4
3 8
4 6
54
74
Maize
6.6
7.2
7.8
9.0
6.8
8.1
9.5
lL!.8
Groundnuts
3 0
35
3 9
5 0
3 6
45
55
82
Source: Master Plan Report, Vol 1 by HARZA (1984).
2.4., Agnicuitural productivity
At the farm level, the increase in productivity to reach a sustainable cropping
systern is a major concern of farmers. A number of decisions are made towards meeting
their production objectives. ‘Ihese decisions are of two kinds (Jouve, 1986):
- gatbering and organizing available resources
- choosing the most appropriate production process and techniques in
relation to the environmental conditions.
According toithe resources available, fat-mers use different strategies to meet their
needs. Hclwever, a number of limiting factors have an impact on their respective
performances and hellp explain why fat-mers do what they are doing.
2 8
2.41. Land productivity
Th.ree major soi1 types, with slopes ranging from 2 to 8%, have been identified
thugh different studies, soi1 surveys and land evaluation in the region (Char-i-eau, 1974).
On upland areas, Oxisol and Ah301 soi1 types dominate whereas in lowland areas, there
are three subclasses: Molisol (rice tïelds without sait), Entisol (Sand deposit in the estuary
of the Casamance river) and peatsol in swampy areas. The Oxisol soi1 type (Ferralsol
according to FAO classification) is sandy-clay, red to reddish in color and characterized by
the presence of kaolinite a.nd a low cation exchange capacity (CEC= 11 rnmol/g). Most of
the upland crops are grown on this type of soil. The Alfisol soi1 type (beige in color) is
mainly located on the transition zones (from upland to lowland) along the transect
between the Oxisol (on top) and the Molisol types. Its organic matter content is low
compared to the Molisol but has a more silty texture than the Oxisol type. It also has a
higher clay content and is recognized by its grey to blackish color.
A number of studies have shown that the increase in land productivity is primarily
bound by the drought situation as farmers do not use minera1 fertilizers due to the
shortage of rainfall (Equipe Systemes, 1984). In general, the fertility level of these soils is
very low because of advanced weathering and lack ofgood management practices
regardless the amount of rnanure provided by the livestock to the cropping system. The
cerelals grown in the farm backyard (maize) and at the village level (sorghum and millet)
usually receive manure and organic matter fi-om different origins, but most of the
groundnut. fields do not receive any type of fertilization. The extensive cropping on newly
cleared lands has generated advanced levels of soi1 degradation. The Sand deposit in the
lowland areas by water runoff from upland fïelds tends to compromise rice productivity.
/ l.a>-/“-l
_-<--.-
.,_._ --.---.
--
_.___.
-.-~
w----
cc
--
2 9
The amount., regularity and distribution of rainfall during the growing season
represent important liimiting factors towards reaching bettef land producstivity. The
decision to start field activities, in terms of working days, for example, is mainly subject to
these twc factors. The date at which the rainy season begins remains the most
unpredictable factor.
The rain disttiibution pattern varies from North to South and fi-om year to year.
The state of drought ihas introduced a significant variation in the amount of rains received
in different localities: 45% deficit in the northern areas and 26% in the southem areas. The
year-to-year variation is the most commonly used indicator by farmers to de’cide about the
cropping pattem (pribrities) and the field operations scheduling towards minimizing risk.
The soi1 rnoisture condition is one important parameter considered by Al-mers to clecide
when ,and how to carry out the field operations. The rainfall pattem is a key factor in the
decision making process to meeting timeliness of farmers’ field activities,
Asronomic tr$als cond.ucted in on-station and on-fat-m conditions have
demonstr.ated the benefits of plowing on land productivity through significant yield
increases (Equipe Systemes, 1984). Usually, land preparation on a11 cropland begins afl:er
.tiater has infïltrated lhe soi1 to a certain depth. The depth of the wetting front at which to
star-t the plowing opdration is mainly dictated by the amount of working power availablle at
the farm level in relation to the soi1 consistency (physical and mechanical properties). On
upland crops, draft atrimals are used when the wetting front is located at around 12 cm or
more aller the useful’rain event (FALL, 1985). In lowland areas for rice production, this
limit is diRerent depending on the type of rice fields described earlier.
/
30
I
2.4.2. Manpower and labor productivity
The availability of manpower throughout the rainy. &ason is the most important
factor for success in carryiing out field activities. The avaiiabiL*y of manpower is mainly
affec.ted by the migration of the Young and active population. Crie way for farmers to
solve this problem is to turn to the mechanization of more and n 3re field operations.
However, the level of access to farm equipment anct gower unitc (animal or tractor) as
factor of production differ fiom one agro-ecosystem to another md fiom one farmer to
another (Fall, 1985; Sonko, 1986; Ndiame, 1987). Xis varia, In in production
opportunit,ies has induced very significant differences in farmers’ performance across the
region. Different studies used the level or degree of mechanization and the availability of
manpower at the farm to divide farmers into three distinct categories at the regional level
(Equipe Systemes, 1985; F’all, 1990; Sonko, 1990):
- Category 1: Non-equipped households (64%)
The manpower (men and women) ranges from 1 to 4. The peak labor demand is
located at the beginning of the rainy season during land preparation and requires 40?/0 of
,total farm seasonal labor demand.
The intensity of the labor constraint is different fiom one cropping system to
anotlner. In agro-ecosystems with important use of animal traction, zones 4 and 5 for
example, some farmers rent a pair of oxen and plowing equipment for a day or two to
meet their fïeld schedule.
For this category of farmers, the area cropped .with groundnuts (cash trop) is small
compared to the cereals: rice fields cari occupy up to 60% of a11 cropped areas.
._,
s _-,.---,-.
-...--m-
_--__--
------.-
--
---
,a--
3 1
- Catbgory 2: Low-levei-equipped households (29%)
Thés categoj of farmers is characterized by the am&int of matipower, ran@,g
between 5 and 8 fati workers. The average is 6 active workers per hobsehold.
Land prepar&.ion is generally the only mechanized field operatibn. The remaining
field operations are rhanually executed which. shifis the peak labor demhnd from land
preparation to weedibg and harvesting. The use of animal traction has a real. impact on the
amount of cropland Iieserved to groundnut production (5 to 8 ha). The’majority of the
!
available Carm workeirs is generally direc,ted to the groundnut fields and penalizes the
cereals. Women invoilved in rice production according to the gender diGsion of labor look
more. and more for oiher alternatives (power tillers or mini-tractors) to’improve timeliness.
- Category 3: High-levei-equipped households (7%) ’
The number cbf farm workers involved in production during the year is large,
rang,ing fiom 9 to 2 1 j persons. These households are well equipped and have 2 to 5 draft
animals.
The area crodped with groundnuts is very important, and avera$es more than 8 ha.
This trop receives thie equipment priority. Farmers are sometimes for&d to use
supplemental extema/l manpower to meet timeliness, as a11 the manpower is not available
during tht: weeding oiperation of a11 cropped land.
2.5. Summary
The identificajtion of the different agro-ecosystem zones at the regional level
represents a major step toward the analysis and understanding of famlers’ specific working
3 2
conditions. The formulation of recommendations to improve their production techniques
has more chance of success ifbased on existing opportunities. The factors involved in
impro,ving productivity in farmers’ conditions are complex.
TO increase productivity and sustainability, it is important to help farmers improve
the soi1 management practices and conservation techniques. Improving tillage practices is
an area to explore in order to reduce soi1 losses from upland areas and to mitigate the
induced and rapid decrease in soi1 fertility. TO meet these objectives, farmers must be
involved at a11 phases of research and development. It is important, for example, that
farmers be able to evaluate the physical properties of soils and Select the right tool and
implernent to perform any given field operation in a timely manner (Henin, 1990). TO
achieve this level of decision-making involving prediction of the power required to carry
out field actïvities on time, the physical and mechanical properties of the soi]: found in the
regïon must be understood in terms of behavior under the influence of various factors.
Farmers in the same agro-ecosystem generally follow similar planting sequences as
a village’s coping strategy against major pests: maize-millet-groundnuts-sorghum with
very little variation in the case of agro-ecosystems 4 and 5 (Equipe Systemes, 1986). The
cropping practices usually dif?er from farmer-to-farmer in terms of type and amount of
agrku:ltural inputs used to carry out the field operations (Ndiame, Coulibaly and Fall,
1988). The degree of combination of these cropping practices also called “itineraire
techniques” explains the diflerences in agricultural production performance among
fkrmers.
The utilization of animal traction by farmers is viewed as an alternative to
improving labor productivity. The potential behind the technology needs to be investigated
.v --r.-s.
-
__--_
_.---
-<----
- - -
__-.__
-
----
w-
more to optimize the output ofthe draft animais. It is important at this point to understand
the difficulty and difference in levels of access to fat-m equipment and draft animals.‘ As a
I
factor of production, the utilization intensity of farm equipment is expect(ed to be different
from one agro-ecosy$em to another and from one farmer to another. The variation in
petiormance among fsrmers belonging to difirent mechanization categories needs to be
better analyzed in ordbr to be able to evaluate the impact of the technology at the fat-m
Ievel and to capture aI1 of the information availabfe in a knowfedge-based system for iater
utilization
Chapter 3
I
DETERMINANTS OF PERFORMANCE~ OF DRAFT ANIMALS
3.1. Constraints to animal traction utiiization
Since the 197Os, development projects involving animal traction have flourished in
.
many developing countries around the world. The utilization of animal traction appears to
- ‘
be very attractive for many small-scale farmers as an alternative way to agricultural
mechanization (Starkey, 1989). A quick overview of the promotion and impiementation of
*
different animal traction-b;ased projects has shown that the utiiization of draft animais is
not a panacea to solve the problem of a11 small-scale fax-mers in the Third World Countries
Eicher (1982) wamed about the “maximum potential benefits” approach used by many
agricultural researchers when carrying out animai traction project evaluation. He pointed
I. I
out the tendency of the approach to inflate the projected retums and long-run economic
I
w.
profitability from the technology when data are mainly collected on experiment stations or
demonstration farms.
The failure of different projects bas demonstrated that the diffusion and
implementation of animal traction have been limited by many factors. The most important
IJ.
are tradition, climatic conditions, diseases, poor infrastructure, lack of appropriate
1.
transportation and communication systems, inefflcient marketing and credit systems, poor
3 4
_-
- _._--
3 5
trainirng, long duration:of the learning process, ineficient extension servides, cropping
systenls types, and inafppropriate veterinary services (Pelissier,.’ 1966; Mum;ciger, 1982;’
Havard, 19:37; Fall, 1900; Ndiame, 1990). Diffi:rent studies have addressed these limiting
factors and their impact on the implementation of animal traction throughout the world.
One exampie reported by Starkey and Faye (1988) in relation to the sanitary constraints
was th.e situation in Caineroon where the use of draft animals was confinad to the
Northern ccltton producing areas of the country with low tse-tse flies infe6tation. Similar
limitations z.re also des@ribed elsewhere: Central African Republic, Conga, Equatorial
Guinea slnd Gabon whdre the use of animal traction is limited to a few mission stations;
less than 1% of the fatiers in Zaire make use of draft animals. In the Basse Casamance
region the diffusion of Animal traction from North to South had to face bath tmditions and
sanitary ‘constraints alofig with the cropping pattem (Pelissier, 1966; Traverse, 1974; Fall,
1985; Sonko, 1990;). Us utilization is mainly confined to the Northem pa& of the region
where 18% (in the Northwest) to 68% (in the Northeast) of the farmers ovjn at least. one
draft animal (Sonko, lg90). These general constraints explain most of the variability in the
geographical distributidn of drait animals in developing countries. A numtier of well-
identified fartors are in+olved in explaining the level of performance of small-scale farms
that ha.ve adopted the tbchnology.
3.2. Selection of draft ianimals
3.2.11. Availabitity
Farmers have a irelatively wide variety of animal species to chose fr,om for energy
supply. Goe and McDawell (1980) in their guid.elines for animal t:raction’s utilization gave
36
the potential of different species ranging from dog to herse. There was a large variation of
dafi potential among species. According to the survey carrüed oit by Schrnitz et a1 (1991)
in 32 countries of Latin America, Africa and Asia, it appeared that cattle were the most
commonly used animal breed fbr draft purposes.
An appreciable amount of quality research is being carried out throughout the
world to increase the work out,put and efficiency of working animals. Crossbreeding
within species is a widely used technique to improve draft potential. Different studies
comparing breeds have shown that it was highly recommended to farmers to use local
indigenous breeds as crossbred animal generally demonstrated inferior working
performance. This was especialky true for tasks of long duration. For extended periods of
time, experience has shown the local breeds always produced the most steady output
(Goe, 1983). Another important advantage of the local species described in .he literature
is theYr availability in numbers and accessibility to farmers, and their adaptability to the
environment. An effective breeding program is always expensive to maintain in order to
produce enough crossbred animais for a continuous supply of draft animais to the farmers.
The supply aspect was found to be very important for the household in relation to the
replacement of injured, dead or sold animais, and for the overall stability of the
technology. In his investigation of the Basse Casamance situation, Sonko (1985) analyzed
the diverse reasons of removing animais fiom service and the impact of the removal on the
household performance (Table.7).
The analysis carried out show& that the technology was unstable for 71% of
animal <traction users who generally possessed one pair of oxen. The loss of one animal
during ,the growing season (2,6%) was likely to compromise the objectives of the farmers’
1 -...rm,--~.--..--~~~
uI,-------
----..
--------.-,
I
37
,
Tdble 7: Reasons for removing animals from service
i
Reasons
Percentages (%)
-
Sales
Death
Trade
Slaughter
ThefI
Rented out
Out for traini/ng
Return to hetd
GifI
U n k n o w n
-
Total
3
/
Sourde: Sonko, 1985
production systems, unless they decided to rent a team of oxen from beh:er off farmers
For .this rcason, he suggested that the optimal. number of oxen at the farrn level is 3 to
secure at ‘east agricuhural production.
3.:!.2. Selectibn process
Specialists bave agreed on most of the criteria used to Select draft animais.
Usudly, draft animalsi for a given breed or species were either selected from the herd of
the household or village, or bought from the rnarket. The conformation (physical form,
neck, legs and feet, hem, etc...), the character, the age, the sex and the weight were the
major criteria for the belection. They also presented the most important fiactors affecting
the work output and working career of a draft: animal. The tractive potential of draft
/
38
,
animals is generally determined by their liveweight LW in kg and their weight distribution.
In comparing crossbred and indigenous cattle, Goe (1983) pointed out that the amount of
‘_
weight distribution over the front legs should enable the crossbred animal to generate
more power. The format and age at which an animal warted its working career were
determinant in its future performance.
The selection must be done from available species in relation to the expected task
to be performed by the animals in the given environmental working conditions. Crosseley
and Kilgour (1983) presented a list of advantages and disadvantages of the most
commonly used draft animals in trop production: horse, ox, donkey and dromedary.
Horses appeared to be very costly to maintain and easy to train while oxen were
characterized as easy to feed and to acquire.
3.3. Training of draft animais
The working capability of animals is determined by in large by the quality and the
level of training. An important number of publications on this topic are readily available to
farmers. These publications are generally used as guidelines by numerous projects during
their ‘early implementation phases. The selection and training of animals are crucial for
their .working career. The rnain objective of the training sessions is to build a working
behavior into the animals in terms of responsiveness and obedience. According to Conroy
(1992!), the trainer and the animais must develop a relationship based on respect and
dominante, as animals deserve feed, rest, shelter and social interaction. During the training
sessions, it is important to Imanage the temperament of the dorninated animals. For this
purpose, animals should not be frightened and negative aspects of the dominante
39
reinforcement must he avoided. One frequent negative aspect seportedl by Keith (199LI)
was to put the animals in discomfort while working which could cause;them to associate
I
the trainer or eventual driver presence with pain and upcoming punishment. The castration
of bu11 is one commdn technique used by fat-mers to calm down the animal and to reduce
their aggressiveness The operation is performed between the age of 1 to 5 years in
relation to the age of! first selection and training.
The age at which anirnals are selected and trained varies amongland within species:
1 to 8 years for buffdoes in Asian countries and 2 to 4 years for cattle ir general (Goe.,
1983:). A good rule oif thumb is to start the trait-ring when the body weight and physical
development of the ahimals have reached 50% to 70% of the matured weight (Goe, 1983,
Lee et al, 1993) (Tadle 8).
Lhoste (1990) reported that farmers with draft cattle in the Groundnut Basin (Sine
Saloum) cf Senegal u!sed different management strategies in relation to the age of train@
and length of working career of the animals on-the-one-hand, and the existing marketing
opportunities at the end of their career on the other. The training ages ranged from 2 to 5
f
years; the ;.ongest carder was 6’ years and the shortest 3 years.
In *:he Basse dasamance region, studies conducted by Sonko (1990) and Ndiame
(1990) shcwed that the training age ranged from (AGETRAIN) 2 to 0 years and the
average working career was 6 years. The training sessions were carried out one month
before the rainy season. Keith (1992) working experience with cattle showed that weaned
calves between the agks of 2 to 6 months were easy to train as they trusted and quickly
accepted the trainer. In such cases, the training would require a firm commitment and
planning fiom the trairher as the use&1 work would really begin at 3 to 4 years of age.
40
Table 8: Age and weight to initiate training of cattle for work
---
Adamawa
Azaouak
-
Maure
2.5-4.0
-
North Soudan
Shorthom
S huwa
N’Dama
Senegal
“,
Fulani
Ankole
Bukedi
Bomro
2.5-4.0
2.0-4.0
-
Ongole Sumba
Khillari
5.0-5.5
2.0-3.0
-
J
Lohani
3.0-3.5
3.0-3.5
-
Valfey
-
Source: Goe and McDowell, 1980.
I
!
4 1
Ah these factbrs are highly interrelated and contribute significantly to building up
the animais’ willingness to work which tumed out to be an important determinant 0fth.e
animals’ working performance. The willingness to work affects the perdentage of the body
weight (PFILW in percent) available for potential power use. The wiilingness affects the
efficiency of the draft! animais teamed through the hamessing system.
3.4. Harnessing systkms and team constitution
3.4.1. Types bf harnessing systems
The main objective of a hamessing system is to allow better transmission of energy
from thle animal to the implement. The ter-m “hamessing” is used in its general sense to
mean yoking as well (Le Thiec, 199 1).
A wide range of hames,ses and yokes are being used around the world. This topic
has been addressed in depth by different resea:rchers (Hopfen, 1969; Starkey, 1989; Le
Thiec:, ‘I 99 1).
Generally, har-nessing systems are divicied into two categories: hom or head yokes
and wither; or shoulders yokes. The tirst type of yoke is tied in fiont or behind the horns.
In both cases ropes or leather straps are used to firmly secure the yoke around the horns.
As mentioned by Le Thiec (1991) and Starkey (1989) a better fit could be guaranteed by
carving the part of the yoke resting on the head of the animal. A head yc&:e must be strong
and light fer better cornfort. Th.e best features of cattle to fit the head yoke turned out to
be a short and strong neck, and strong homs. Impirical experiences have shown that this
type of hamessing syst/em presented advantages of better coordinating the movements of
working animals speci&y when instantaneous maximum draft is required, by allowing a
4 2
rapid adjustment’of the line of draftto reduce the angle of pull. The main disadvantage
found for continuous work was the more rapid state of fatihe induced by the rigidity of
the yoke attachrnent (Le Thiec, 1991). He warned also against the bad quahty of some
ropes used by fanners, which over time tut off the homs.
The shoulder yokes were mainly designed for forward motion. They are adjusted
to pull against the shoulder muscles or the front base of the hump for Bos indicus cattle.
The simplicity of fabrication has caused them to be widely used specially to pull different
types and sizes of carts for transportation. The main advantage of this design allows the
animals to move their heads freely which induces less fatigue than head yokes. Experience
has shown that animals hamessed with withers yokes refused to pull and try to get away
when maximum draft is required.
3.4.2. Eflkiency of harnessing systems
The work effrciency of animais was found to depend primarily on the effectiveness
of the energy transmission. A subjective evaluation technique of this effectiveness by
animal traction users is the classification and codifica.tion of the energy transmission mode
into (1) inefficient, (2) average and (3) efficient hamessing systems. According to Goe and
McDowell(l990), a good and well-design hamessing system would help increase animal
draft capacity by providing at the same time enough comfort to the animal. Starkey (1989)
stressed the quality of a good harnessing system to maximize its positive effects in terms
of mutual encouragement of teamed animais towards achieving a sustained work.
Researchers have been focussing on animal factor ergonomies to improve the energy
_ I
43
transmission for better efficiency. One major step was the improvement of the yoke design
for a better attachment to the working animals (Le Thiec and Havard, 1995; Roosenberg,
1992).
The potential energy and pulling capacity were increased by teaming animals
through a harnessing system which coordinates their movement. The process of tearning
must star-t at the animal selection phase in relation to the task to be pertformed and the
assessment of the expected draft requirements. According to Conroy m.d Rice (1992)
oxen do not need to be identical in order to make up a successfùl working team. The
suggested selection criterian ‘to focus on was animals’ ability to work together. They urged
oxen team trainers to look for simiiar temperament, agility, size, conformation, and speed
This Posit:ion was not far from the recommen,dation made by most animal traction
speciahsts to Select closely identical animals (CEEMAT, 1971; Watson, 198 1). In the case
of anirnals of different size, the strongest should always walk in the fi.rrriow when working
with a moldboard plow in order to balance the draft demand between thc two animals.
The pulling force potential expected fom a team of animals was found to be less
than the cumulative potential of individual animal making up the team. J&tarks (195 l),
cited by Goe (1983) reported the results obtained with different teaming, scenarios. They
showed that the harnessing system induced a iloss of efftciency amounting to 7.5?/0 for a
pair, 15% for triplet, 20% for quadruplet, 3 0% for quintuplet, and 37% for sextuplet. The
researchers of CEEMAT had specificaliy addressed the case of teaming oxea and found
similar results (CEE&T, 1972). Their fmdings showed that the potential of one animal
must be multiplied by 1.9 to estimate the expected potential from the pair of oxen (10% of
efficiency 10s~). Lee et a1 (1993) translated Marks findings into a general formula to
_ I
44
evaluate the efficiency of the team (TE) in relation to the size of the team (TS):
TE = I-(TS-1)*0.07.5
(1)
The average liveweight LW of each animal working in a team cari be determined if
the draft requirement DR for the task to be performed and the percentage of the body
weight (BWP= PF/LW) that an animal could develop into pulling force (PF) were known.
Watson (198 1) proposed another formula by assuming a 15% loss of efficiency for a pair
of oxen:
(2)
LW=
DR
0.85* BWP
The numerical default value of BWP used most oflen for cattle was 0.10 (10%) for
long term traction (CEEMAT, 1975). This value needs to be validated with the local breed
used for traction.
3.5, Working capacities’
In general, fat-mers intend to get the best performance out of their working
animais. The handling and tare of draft animais (health, housing, feed) were found to be
an important consideration in order to keeping them in good working condition. When
conditions were bad or the management poor, animals would become more sensitive to
stresses and diseases.
4 5
.‘3.5.1. Climiate effects
In the hot abd humid climate ofthe tropics, exhaus&d animais will rapidly tend to
lose thenr tolerance ‘to txypanomiasis and mortality Will increase. Acccnrding to Crossley
and %le;our (1983), veterinary services must be made available especially to farmers. They
must look daily for the more! common symptoms related to the appeanance and behavior of
the a.nim,als: urination, mucous membrane, temperature, pulse and respiration rate. Only
healthy animals wer’e found î:o be able to produce the desired power output. Falvey (1986)
defïned health as th$ optimal tinctioning of animal body. This optimu&: could easily be
off:jet by the effects’of stress on the animal?, immune system and by injuries caused by the
harnessing and hitcting system. He warned 1:hat stress could be createql by overworking
animals which are ita a state of malnutrition. This would impair the immunoglobulin
production due to dkficiency of proteins.
Experience dnd tests conducted in different places suggests sttict management
practices to minimize the drastic effects of ciimate on the work output,of draft animais.
Besides the sanitary aaspects, the environmental effects on working anitials’ needs to be
investigated more toi help farmers Select the best animal. Draft animais do react to changes
in working regime ahd envirclnmental conditi.ons by making physiological adjustments to
maintain their body Beat. These adjustments caused thermal stress on thle animais reflected
in their finture behavior (Singhal and Tomar, - ). A few studies have tried to show or
quant@ how much tbe temperature and the humidity really have affected the working
efficiency. These twd factors relate directly to the ability of the animais to dissipate the
heat produced durin$ the task performance. The heat transfer from the animal body to the
environment is a natWa1 phenomenon. The heat exchange is generally conveyed through
/
4 6
conduction, convection, radiation and secretion of sweat. Reported results from Premi
(1979) stipulated that the respiration rate of bullocks doubled ‘with every 10°C rise in
temperature. Goe (1983) reporteci sirnilar results with the Hariana; the respiration rate anc
body temperature increased as the ambient temperature increased from 10 to 28°C. An
increase in relative humidity reported in the same study affected the blood and respiratory
system as it accelerated the puise rate. Zander (1972) gave the range of 30 to 70%1 for air
humidity to be a comfortable climate. The combinat:ion of these environmental factors was
found to play an important role towards the animal’s state of fatigue when at work (Table
9). Animal traction users agree that moming hours represented the best time period to
‘carry out field operations. These important results need also to be validated in different
‘working conditions.
Table 9: Environmental factors on draft animais
Type of
Pulse Rate
- i
Animal
per minute
Respiration
Temperature
under jaw bone
-
Rate per minute
“C
Cattle
40-50
1 O-30
38.5-39.5
Horses
3 6-42
8-16
37.5-38.5
Donkeys
40-60
40-56
36.0-38.0
Source: Crossley and Kilgour, 1983
The most detailed study available which reports the direct environmental factors
affecting physiology deal with the human body (Essay, 1969). Essay developed a linear
,
4 7
mode1 to evaluate a parameter called the Temperature-Humidity Index (TMI) which is
related to the dry bulb temperatu:re and dew point temperature. Gar THI values between
70% and ‘79%, lOo/o of 100 peop1.e were feeling uncomfortable. The THT cari only be used
as an indicator and must *be set for draft animals ,working in a given environment. Other
empirical :methods based on management practices which included rest periods were
recomrnendeci by field researchers to optimize the output of working animal:s in a given
environmental condition. The level of fatigue of working animals was indical:ed through
physiological and behavioral manifestations (Upadhyay, 1989): unwillingness to work,
fiequent stopIj, respiratory problems, etc. In tropi.cal climates, the maximum working ,time
ANWKI-IR in hours must not exceed 8 hrs per da.y. Devnani’s (198 1) investigation showed
that bullocks’ work time during the summer must be kept between 5 to 6 hrs per day and
divided into two sessions (morning and evening). In the Basse Casamance region,
Traverse (1974) suggest$d that the Ndama cattle could be managed in relation to the
intensity 0.f the work:
Light task: 7 hrsVday + rest at mid-day
Medium task: 7 hrs/day + r-est at mid-day f 1 day off/week
EIeav task: 3.5 brs/day t’rom 8:00 to lk:30 am.
This classification was made possible in relation to the draft required to perform
different tasks: transport, harrowing and weeding were considered as light jobs; plowing
sandy soi1 (upland crops):as a medium jobs; and working clay soi1 (lowland rice lïelds) as
heavy jobs. The most recent studies conducted by Fall (1985) and Sonko (1986) have
,
48
shown that in fact, farmers in the Basse Casamance region used draft animals for less
hours per day to perform field work.
3.5.2. Pulling force potentiai
The potential pulling capacity PF in N or daN of a working animal was found to be
mairly determined by the animal species and breed, weight, age, nutritional and health
condition, level of supervi.sion and training, and methods of harnessing. A well
documented research program showed that the pullin,g capacity was directly proportionai
to the animal body weight (CEEMAT, 1974; Goe ami McDowell, 1983; Starkey, 1989;
Campbell, 1990) which in turn was affected by the body condition of the animal and the
harnessing system (Table 10).
The potential was expressed in animal body weight perceniage (BWP in “2). The
. I
average long ter-m pulling force (PF in N or daN) developed by working animals was
1
._
found to be within the ranges of 15% to 20% for donkeys, 12% to 15% for horses and 9?/0
/
_ I
to 12% for cattle (CEEMAT/FAO, 1972). Goe (198’7) mentioned a draft potential of 18YO
to 23% of the body weight from Ethiopian draft oxen
-
t
Table 10: Work potential of man and various animais
I
.a.
I
_.. I
Item
I Man
Horse
0x
Donkey
Weight (kg)
8 0
400-700
300-900
100-300
Pull (N)
100
500
500
400
Speed (m/s)
1.0
1.0
1.0
1.0
Power (kW)
0.1
0.5
0.5
0.4
Daily work hours (h)
6
6
5
4
Daily work output(MJ)
2
10
9
6
Source: Crossley and Kilgour, 1983
Research findings showed that during their growth phase, the closer the draft
animais’ laveweight approached their mature, or genetic weight, MATUREL,W in kg, the
heavier the type of work they could perform. This point has been emphasized by
Upadhya:J (1989) who stipulated that heavy animais possessed more pulling capacity than
light animais but were less suited to speedy work. In evaluating this potential and taking
into account the morphological aspects, Goe (1983) and Lee et d (1993) suggested a
body weight parametier called the condition score CS. The condition score excludes from
tract.ion any animal storing less than 1 (afIer rounding off). This corresponds to less than
5QX of the body weight ratio depicted in the following equation where LW in kgfand
MATURELW in kgf were respectively t he animal live and mature weight:
The liveweight LW could be measured by using proper scaling devices generally
available at the veterihary semices or the governmental extension agencies. In areas
lacking infrastructure and effective technical support services to farmers, researchers bave
developed different models to help estimate the weight of cattle based an the
morpholoçical features of the animal (CEEMAT, 1974; Watson, 198 l):,
(girth $“* L
LW -= -
300
5 0
Where the &th (in inches) represents the circumference (CIRCUMF) of the torso
at the point of heart and L (in inches) represents the length of’the animal from the point of
shoul’der to the rump. However the applicabiIity of this formula was limited to non-
castrated oxen. Cows were also used in areas where they were culturally accepted. The
pulling capacity of cows was found to be approximately 5% lower than for male animais
(Lindsay, 1986). Also cows needed special tare and more attention than maies in order to
maintain their reproductive capacity and milk production. In the Sine Saloum region of
Senegal (known as the Groundnut Basin), survey repo,rts indicated cows were 26% of the
draft cattle and were used as a second or third team member to support the already
existing draft oxen (Lhoste, 1990). In the Basse Casamance area, only male cattle were
used for traction (Sonko, 1985).
3.5.3. Work potential and energy requirements
3.5.3.1. Instrumentation
The evaluation of work output and energy expenditure was found to require from
“!
simple to highly sophisticated instrumentation in relation with the type of parameters to be
included in the energy determination. Research efforts were mostly centered around the
development of different sensors to help measure integrated parameters as heart beat,
pulse rate, respiratory rate, draft, speed, oxygen consumption, body temperature, etc.
l-1
Substantial resources were put into the development of new methods integrating
I
physiological and physical parameters. At the same time, the classic approach using a
simple dynamometer and integrating distance traveled DISTRAV with animal speed
ANS’PEED measurements was still widely used because of its simpiicity compared to the
---
I
51
new methods. The neiv methods were divided into direct and indirect measurements.
These two techniques were well documented by Lawrence arid Pearson (1989).
The direct method involved the measurement of heat output dissripated from the
animal’s body. The he;at generated during the task performance was produced by the
oxidation of carbohydrates, fats and proteins to carbon dioxide, water aad urea. The
method presented some serious limitations. The animals must be kept ericlosed in a
chamber. The theoretical and practical aspects, of the measurement techhque need to be
investigated more in oirder to make the results usable in field operations.
The indirect miethod took the measurements one step further by aiddressing not
only the heat production but also the amount and the quality of gaseous exchange
involveld. The technique was based upon the calculation of the heat dissipated by the body
using a measurable parameter represented by the amount of gaseous exchange. The
underlying principle was that no heat was produced without gaseous exdhange. Oxygen
(OI), ca.rbon dioxide (CO4 and respiratory quotient were the parametersi :mostfy
concerned. Two types of apparatus were used by the Center for Tropical Veterinary
I
Medicine (CTMV) at the University of Edinburg:
- Classic Open Circuit System for use on the animal at wdrk and at rest.
The system was not portable and the use of a treadmill was necessary to Ienable the animal
to work while staying in one place.
- The P!ortable Rreath-by-Breath Analyzer in the ,form of an airtight
facemask fitted with a !low meter was to measure the volume of each breath. Samples of
i
5 2
I
expired air were to be analyzed in the laboratory with the application of correction factors
.
related to temperature, atmospheric pressure and hurnidity. The apparatus could not be
I
‘.
used in the field because of the laboratory intervention, representing a serious
disadvantage.
The AFRC instrumentation package was also developed for continuous
* i
/
measurement of both physiological and mechanical work performance variables (O’Neil et
al, 1989). The data stored in the data logger during the work session were: draft force,
an,gle of pull, speed, heart rate, breathing rate, body temperature and stepping rate. The
only disadvantage resided in the non-availability and high cost of sensors used to collect
the data. The CEEMAT package put more emphasis on the mechanical variables.
Other methods to estimate the different levels of energy required of working
animais are indicated in the hterature.
These levels of energy have been designated as maintenance, work and liveweight
gain by many authors (CEEMAT/FAO, 1972; Goe, 1980; Lawrence, 1985; Falvey, 1986;
Pearson, Lawrence and Ghimire, 1989; Lawrence and .Zerbini, 1993).
3.5.3.2. Energy at work
FAO (1972) reported the energy requirements ,for ruminants at work to be about
2.6 fold the maintenance energy. Pearson and Fall (1993) believed that the FAO method
overestimated the energy required. For the specific case of cattle, the values published by
Lawrence and Zerbini (1989) were the most recent. Thle technique used the factorial
method and was found to be feasible for long measurements in the field. It was based on
4
_
.mmmamqm-.III-IC-----
-.
-
---
.,.
I
5 3
determining the extra dnergy E used by working animals of liveweight LW (in kg) when
walking on a horizontal distance HD (in km), carrying loads LC.(in kg), doing work W (in
W) while pulling loads, and walking uphill a vertical distance VD (in km):
The coefficients given by Lawrence and Zerbini 1989) were mainly fo,r cattle:
A = 2.0 kJ/kg of LW per km of ItD
k = 3.0 kJ/kg of LC per km of I LD
G = 3.3 kJ/kJ of W
1) = 0.033 kJ/kg of LW per km of VD
The values of the coefficients were obtained through laboratory experiments,
which represented a major disadvantage compared to measurements perfomed in the
field. The coefficients need also to be evaluated for local breeds.
3.5.3.3. Energy for growth
IMost draft animals reached maturity during their working career Energy was
required to foster their growth while at work. The level of energy was found to vary in
relation 1:o the management practices and the task to be performed (worktonly versus
work and fattening). Goe (1983) wamed that the energy used by working Ianimals for
growth needed more investigation as the values, published in the literature ipresented
important di,screpanciei. He presented research results obtained in West Africa reporting a
rate of growth of 1 kg/day for cattle trained at 0.5 to 3 years of age. Experiences
conducted in fat-mers0 condition in Senegal showed that the oxen used in the Groundnut
*.1i
5 4
Basin areas gained an average liveweight of 33% per year or 0.16 kg/day, 25% per year or
. . i
0.11 kdday and 17% per year or 0.05 kg/day. This was obse&d for draft animals tramed
!
at age 2-3 years and weighing 170 kg, 3-4 years and weighing 220 kg and 5-6 years and
weighing 300 kg (Lhoste, 1990).
3.5.3.4. Energy for maintenance
The energy required for maintenance has been investigated by numerous scientists
and appears in most animal traction manuals and guidelines. Lawrence and Zerbini (1993)
presented MAFF (1984)‘s contribution in evaluating the energy required for maintenance
MAE (:in MJ/day) for cattle ofliveweight LW (in kg):
M.4E = 0.36 * (L w y
(6)
The working capacity and the source of energy required for draft animais at work
l.i
were described to be largely dependent upon the availability of quality feed and the
efficiency with which this feed was transformed into potential energy .
3.6. Feeding systems
3.6.1. Feed quality and availability
An important amount of information was published in the literature to present
research carried out towards :improving draft animal feeding systems. The rationale behind
providing working animals access to better feed was the maximization of the amount of
work ‘obtained from the limited available resources. Feed is the main source of energy for
II*UIïULU,U-.“..-,-l---
--
---
111-----P
I
55
animcils’. One; important aspect of that research was to increase the nutritional value in
order to provide the nutjrients required for the execution of vari&us physiological fimctions
as work. AS presented aarlier, working animals rely on ,feed to provide energy l‘or
maintenance, growth anti work. ‘The necessary a.mount of feed for a workihig animal could
be calculated if a11 the required energies were known. The only limiting fa&or identified in
the literature was the non-availability of a reliabl:e feed supply (Eicher, 198:!). This was
found to be <.I major technical constraint as feeding prac,tices commonly ustd by
smallholder farmers were simple and dependent on naturally available feed resources:
grazing a.nd utilization of seasonal trop residues and by-products. But moW: trop residues
were regiarded by nutritionists as, poor quality feed (Wanapat, 1989). Ffo&es and
Bamualirn (1989) suggested that the quality of straw feed must be improved by the
addition (of urea. A limited number of natural fo:rages species were found to be high
protein,: Leucaena, Cassava, sweet potato and banana (Table 11). Beside the mine,ral
requirements, draft animals need high enerby feed mainly based on glucose and fat to
satisfi the energy demand of difl?erent physiological functions as sustained muscular
activity. .Aceording to Teleni and Hogan (1989), work was achieved through the
combination of various systems ‘based on adenosine triphosphate (ATP) gtneration,
respiratory and cardiovascular fimctions.
3.6.2. Dlet for working animais
Current research results on nutrition are yet to be transferred to fdrmers. In the
West AfYican region wihh a long dry season, reports show that farmers are still
experiencing a shortage! of energy as land preparation occurs at the end of the long dry
season when animals are in their poorest condition. For most farmers, the diet ofworking
Table 11: Composition and digestibility of feeds (DM basis)
-
Minerals
(db DM)
Types of feed
DM
CP
CF
Fat
DMC
%
%
%
%
%
- - -
-
Ca
- P
- N i
-
S
-
Crop residues
Rice straw
33-95
4-6
33
1
1
38-55
Maize stover
85
6
32-46
0
2
54
Sorghum straw
91
4-8
33
1
.
42-48
Sugarcane tops
2 8
6
35
1
4
6 2
Cassava leaves
2 6
2 0
21
1
2
72
l
Banana leaves
2 4
12
23
1
2
45-66
la
Potato forage
1 1
1 3
14
<l
3
93
Tree legumes
.A
/
Leucaena
3 0
2 9
21
4
1
4
51-60
Glyricidia
2 7
2 4
18
4
1
3
6 7
l
Concentrates
IRice grain
8 6
10
10
2
1
3
1
63-95
Rice bran
8 6
10
2 0
1 l-22
1
18
1
2
43
Maize grain
86
10
2
5
0
3
1
1
65
Soya bean meal
9 0
37-48
5
4
3
6 <l
3
25
” t
Kapok seed meal
9 0
31
3 0
8
5
1 3
50
Cotton seed meal
8 6
4 4
14
8
4
11
28-50
I
Molasses
75
4
0
0
11
1
9
9
100
Tapioca waste
9 0
2
3
4
6
2
6 9
Palm kemel cake
8 9
19
1 3
5
3
7
2
2
= = r =
DM: Dry Matter; CP: Crude Protein; CF: Crude Fiber;
DMD: Dry Matter digestibility.
Source: Extracted from Ffoulkes and Bamualim, 1989
1. IIV.IIIU
___.
--~
. ~ . _
-.-
-I
--
_ I
57
animals is mainly based on roughage and browse. The loss of weight during this dry period
represents a serious limitation to performing heavy fi,eldwork at the beginning of the
cropping season (Starkey, 1989). As pointed out by Wapanat (19893, ver-y few
experiments on feed rations were conducted in on-farm conditions, in order ‘to evaluate
the feasibility and acoeptance of the rations by the farmers.
Th.e literature is documented well in methods and techniques of evaluating the
amount of’ feed necessary for working animal:;. In general, a11 the authors take into account
all forms of energy involved in satisfying the demand of physiological fujnctions. Lawrence
and Zerbirti (1989) included in their estimation the energy for milking CO~S a.t work based
on h44FF (1984)‘s model. Crossley and Kilgour (1983) partitioned the energy content of
feed into gross energy (GE), digestible energy (DE) and metabolizable energy (ME). It
was only at the stage of metabolizable energy that the animal could really use the energy
to perform work. Al1 these levels of energy were interconnected:
A4E =iT8*DE
0
For poor feed (straw for example):
DE =:O.-15*GE
(8)
For good feed (grains for example):
DE =0.85*GE
I
5 8
The energy content of feed was generally determined on a dry matter basis (Db{)
and largeiy depended on th.e percentage of digestible dry matter (DMD). Crossley atid
‘.
Kilgour (1983) indicated in their analysis that the approximate maximum amount of DM
called the Appetite Limit (AL in kg of DM per day), was related to its liveweight:
AL =O.O25*LW
(10)
3.6.3. Forage Unit
Goe and McDowell (1980) reporting different nutritional research activities
conducted around the world used the concept of total digestible nutrient (TDN) and
forage unit (m) to convert the amount of feed into GE and the ME’. For the West
African region, most of the research activities have been performed by the French
institutions CEEMAT and JEMVT. A comprehensive list of nutrient sources for working
I
animais and their energy content using the feed unit (FU) concept has been published
(CEE.MAT, 1975). Average estimates of feed unit for maintenance (FUM in kg)
published by CEEMAT (1975) could be modeled (R2= 0.999) into a relation between the
liveweight (LW in kg) and the number of forage unit FU to provide the required level of
energy2:
FIJM’ = 6.321* IO” * (L w -- 0.7089
’ TDN = carbohydrate -+- protein + 2.25 * fat
1 kg of TDN = 4.409 Mcal or 18.447 MJ of DE
= 3.615 Mcal or 15.125 MJ of ME
1 kg of FU = 0.74 k:g of Barley
” The raw data used to develop the linear mode1 are presented in the appendices.
J
5 9
The mode1 output gave the same TDN results as the values published by Goe and
‘.
McDowell(1980) and Goe (1983) for West Afiica.
’
3.7.. Implements selection and characteristics
Implement selection described by different authors was reveaied to be as important
as the selection of draft animais (Le Moigne, 1981; Starkey, 1989). Foir many farmers, at
the earlier stage of adoption, the selection could be very complicated, a:5 a variety of
equipment designs are available on the market. The introduction of farfn equipment in the
cropping system by development agencies w(as not neutral and was done as a means
towards achieving production objectives. The major challenge in the piocess was to fit
most of the tasks to the draft animals by selecting the right implements, Once the
implement was selected, its utilization would likely predetermine the e4t:cution mode of a11
the following field operations.
Animal-drawn implements were generally designed with ergonomie considerations
for comfort and efficiency. They were mainly composed of three majors parts: working
components (moldboard, ridger, tines, and shares), frame or beam (equipped with or
without uheels) and handles. ‘The shape and sharpness of the working components were
mostly responsible ofthe draft: required (DR in kgf or daN) to move the implement in the
soi1 in the horizontal direction of travel. Each farm impfement had its own requirements in
terms of draft and energy in relation to the type of soil. A number of draft tests conducted
in difI’erent areas of the world have been reported in the literature (CEEMAT, 1975; Goe
and McDowell., 1980) (Table 12).
6 0
Table 12: Draft requirements for farm implements (Equatorial Africa)
-
-
Implement
-
-
I’low
Indigenous
Moldboard
Disc
Disc harrow
single action
double action
Rotary tiller
Harrow
spike or peg
spring tine
I
Roller or puddler
Lfeveler, float
Row-trop planter
Grain drill
Transplanter
=
Source: Goe and McDowell, 198 1
61
For giwn working conditions, the draft required was found to depend on the type
. .
and condition of soil, and the angle of pull (y in degree angle).’ Al1 farm equïpment
involved in trop production were mainly characteriz.ed by
working width (IMPLWWD in cm) and depth (IMPLWDP in cm). If I!F in daN is the
animal pulling force, the draft required DR in daN cari be expressed as:
DR = (Pi?* COS(j)
D!fYerent studies have shown the importance of the angle of pull ‘y and its
!
dependenzy on the type of draft animal and the harnessing system (Dibbits, 1993; Le Thiec
and Haval-d, 1995). The optimum angle of pull was found to give the truc line of draft.
The line of draft or traction was defined as thle straight line connecting the point of draft
located at the hamess level (yoke clevis) and the
tenter of soi1 resistance on the working component of the implement (Watson, 198 1). Le
Thiec and Havard (1995) found the angle of pull for Ndama cattle to be between 15” to
18” and 19” to 23” for Zebu. The variable horizontal and vertical hitch points located on
the implement hitching bar are used to adjust ,the line of draft and the puli.ing angle in
order t,o balance the soi1 forces acting upon the working components and the weight of the
implement .
The utilizatiorj. of animal-drawn implernents requires skilled operators to minimize
eneraT los,ses. Differeht studies found farmers differing in skills and years of experience.
Eicher (19’82) reportitig Barrett et 4 in a study conducted in Burkina FasI; pointed out
i
that there was a slow leaming process.for farmers who were using donkeys or oxen for
_ /
the fir:st time. It took about three to four years before a farmer-knew how to use a
complete package of equipment. Implements were provided with different control systems
I
and adjustment devices. Besides the horizontal and vertical adjustments previously
mentioned to adjust the working width and depth, the wheel was designed to provide
betner stability. Adjustments by triai and error were commonly used to reach the point
where the wheel was not pressing too hard on the soi1 surface and cutting into the ground
to increase the soi1 resistance (rolling resistance and bearing friction) or not lifting off the
soi1 surface (Watson, 198 1). IIandles were also provided to help steer and tilt the
implement to correct the line of draft while in motion. A good adjustment was achieved if
no force was felt at the handles to prevent the operator fiom steering the implement and
contrcrlling the working animals (Crossley and Kilgour, 1983).
3.8. Soils
3.8.1. Soi1 types
The draft DR in daN is mainly defïned in relation to the resistance of the soi1
against advancing implements’ working components (WKGCOMP). In other words, draft
is the force required in the horizontal direction of travel: as stipulated by ASAE Standards
(1984) to initiate and sustain movement. In trop production, the physical characteristics of
the soi1 were found determinant in the evaluation of the amount of force required for
pullïng a given implement.
In the IJS system, soils are generally classified into textural classes or orders.
There are 12 major textural classes according to the size and distribution of the primary
. -.,4~~-*““-I-I~----.-
-
--.
--
-p--e
6 3
partides (percentage of sand, silt and clay). Soi1 structure relates to the arrangement of the
_ .
partkles into peds or aggregates (shape and size) (Foth, 1990). The different soi1 classes
are summarized in the commonly used textural triangle. Within the categories of soi1
taxonoqr, soi1 orders represent the highest category. There are 11 ordkrs divided into
subgroups, family and series.
As pointed out by Charreau (1974), the soi1 survey and classificiatiora of the West
African region had been made possible by a g,roup of more than twenty French researchers
(LRAT and ORSTOM) who helped train Afiican soi1 scientists during a twenty-year
period. They used the French system elaborated in 1967 by the CPCS to develop the Wesr
African scil map. The classification was made of 12 classes subdivided 11nto subclasses,
groups and subgroups, families, series and types. The family was composed of soils
formed fiom the same kind of parent material. The series related to the position on the
toposequence and the types relate to the texture of the surface horizon. They also
presented 5ve vast zones originated from the following parent materials:
” a. In the northern part of Senegal, Mali, Niger and Chad: sands ,lnd sand Idune:s
b. In the Niger Rive?s arc and a great part ofthe Chad basin: alUuvia1 deposits
ranging from pure sand to fine clay
c. Westward, south-westward, and eastward of the Niger Rive?s arc in Burkina
Faso, in two-thirds of Senegal and in Southem Chad are found the Terminal
(Jontînentaliformations. . The materials generally went through a strong
ièrralitic althration.
d. In the southern part of Niger River$ arc, in the central part of Malï, and in the
western part of Burkina Faso is a vast area of Co:mbro-Ordovician sandstones
overlanded by an ancient “lateritic” iron pan . . .
e. In the southem part of Mali, in Burkina Faso, in the central part of Chad:
crystalline shield made up of plutonic rocks, :etr.Torphic rocks and volcanic
rocks. . . . ”
The soi1 map showed that large areas were occupied by grey and yellow-to-beige
I
1
ferrugineous soïls, and red fèrralïtic types of soil. The soils were also characterized by a
I
I
,j I
field capacity (SEC) of 15 to 20% v/v and a wilting point c !VP) of 7 to 9% v/v.
The oxic horizon of the ferrugïneous and ferralitic sons are mainly made of a
mixture of three elements: kaolin, amorphous hydrated oxides (iron and/or alumirum), and
quartz. The amount of kaolinite is highly variable. The soils are desaturated and
characterized by a low cation exchange capacity CEC due to the presence of kaolinite. As
described by many soi1 scientists, they have almost reached the end point of weathering in
their evolution (Charreau, 1974).
These two types of soi1 are dominant in the southem part of Senegal (Ducreux,
1984). The gray ferruginous soils (“Beige soil”) presents a sandy to coarse loamy texture
in the Upper horizon with a subangular blocky structure and a -fine loam to fine clayey
texture in the deeper horizon with angular blocky structure.
The ferralitic soils also called “Red soil” (due to the presence of amorphous iron
oxides) present some Alfic characteristics with subsurface horizons of clay accumulation
and medium base supply. The texture and structure are similar to the grey type.
65
3~.8.2. Mechanical and physical properties
3.8.2.1. Physical properties
Nicou (1975) and Ducreux (1984) preseinted, in their respective studies, the
ferrugineous and ferralitic soils to derive their physical characteristics and mechanical
behavior from the predominance of kaolinite and sandy to sandy-clayey texture of the
surface h,orizon. The dry climate through the wetting and drying cycles had significantly
affected ,the aggregates and structure stability. The stability refers to the ability of the soi1
aggregates to resist the disintegrating effects of water and mechanical manipulation (Jury
et al,. 1?91). The low clay content in the Upper horizons (8 to 12%) and the presence of
kaolinite were described to confer to the soi1 surface a massive to structure:less behavior
when diy. Ir this situation, the a;ggregates tended to harden and a cementation of the
whole soi1 mass takes piace (C%arreau, 1974). This process of increasing cohesion called
“prise en masse” during the drying phase of the cycle is a well-known soi1 characteristic of
the dry tropical regions. The phenomenon has been studied mainly by penetrometry by
differerrt researchers, in relation ‘to the soi1 texture, water content and porosity. The soi1
cane penetrometer was designed and validated as a standard device used to measure the
penetration resistance of soils called the “cane index” (CI in kgf or daN per cm’). The
cane index CI is used to compare mechanical properties of different soils, and to develop
performance and prediction relationships. ASAEI Standard (1984) has recommended two
cane base diameter sizes: 20.27 mm with 15 9 mm shaft diameter for soi? soils, and 12.83
mm with 9.5 mm shafi diameter for hard soils. The hand-operated cane peretrometer use’d
by Charreau (1974) Nicou (197 S) and Ducreux (1984) is described as a simple graduatecl
and Sharp metallic shaft (15 to 20 mm in diameter) with a 60” cane and a shding weight
I
6 6
(SLWGHT in kgf or daN). The sliding weight falling from a given height (HGT in m>
drives the shaft into the soi1 to a certain depth (DEP in m) (Ducreux, 1984). If SR in kgf
or daN is the soi1 resistance to the shafi penetration, then the penetration energy EDEP
corresponding to the depth DEP is given by:
II/. I
1
The penetration forces were found to be 5 to 10 times higher when the soi1 was
dry than wet. At various moisture regimes above the wilting point, field observations and
L !
soi1 manipulation in the laboratory showed that the sandy texture of the surface horizons
I
l
tends to give to these tropical soils a non-sticky, non-plastic and friable state of
consistency (Charreau, 1974; Ducreux, 1984). The mechanism involved in this hardening
process during the dry season and alter intermittent rainfalls followed by dry spells during
the wet season is still under investigation. Nicou (1975) and Ducreux (1984) defined a
textural index called the “hardening index” HI to characterize this physical behavior. The
index was expressed as the ratio of the clay content over the coarse fraction (coarse silt +
coarse Sand). They found tha.t there was a linear relationship between the hardening of the
soi1 Upper horizons and the index HI for soils studied in Senegal and Niger: the higher the
index, the higher the tendency of the soils to harden during the drying cycle. The
phenomenon was found to be important in relation to trop production (tillage and root
penetration). For land preparation, Char-r-eau (1974) wamed that during this hardening
process, the draft required by these ferrugineous and ferralitic soils appeared to be too
high for draft animals found within the area of study (donkeys, horses or cattle). A draft
-w-
,------
-y-------
--
P-m-B.8
6 7
value of 240 kgf or daN from a plow working at a depth of 15 cm in a ferrugineous soi1
‘_
was recorded in the southem part of Senegal.
5
3.8.2.2. Resistance to traction
The most frequent parameter used to characterize the mechanical itlehavior of
different soils is the specific soi1 resistance SSR in kgf or daN/cm2. It represents the sum
of the soi1 and trop resistance and the implement rolling resistance. The draft per unit of
fùrrow cross-section given by the ASAE Standard (1984) relates high-speed (greater than
3 km/h) field operations with implement types and soi1 types. With animal power,
experience bas shown that draft depends mainly upon the soi1 moisture conditions and th.e
type of implements usad.
The specific resistance was found to be highly variable among soils. Smith and
!,
Mullins (1991), cited by Lee et a1 (1993), classified the soi1 specific resistiance in relation
to their textural characteristics. They used the clay content as a criterion to divide the soils
into 5 classes of specific resistance, ranging from 30,000 N/m2 for coarse sandy soils (O?G
to 8% of clay) to 130,000 N/m’! for clayey soils (50% to 100 % of clay). ‘In these
condit.ions, the draft required (DR in daN) to pull an implement with worik:ing width
(IMPLWWD in cm) and depth (IMPLWDP in cm) in a soi1 of known specific resistance
(SSR in daN/cm2) is as follows:
DR,,,,! = (A4PL WfZ’,’ + (IMPL WDP) * (SSR )so,,
w
t*1
6 8
The main limitation of this approach was found to be that the soi1 specific
resistance was not correlated to the soi1 rnoisture regime. Diffcrent studies on monitk-ing
field activities revealed that farmers carried out field operations at non-optimal soi1 water
content in order to meet their production objectives through better timeliness (Charreau,
1974; Fall, 1985; Lee et al, 1993).
The occurrence of the first useful rain event is usually a signai for farmers to
schedule their field activities according to the number of working days available during the
cropping season. The number of working days was considered to be highly determinant in
* !
the process of evaluating how farmers achieved their production strategies in relatio:: with
the ciimate and soi1 type.
3.9. Working days
I
3.9.1. Soi1 moisture regime
I
The soi1 moisture regime described by many authors as the main criterion is used
to evaluate a working day in relation to the type of soi1 and weather. It is generally
described by soi1 physicists to be highly dependent on temperature through
eyapcrtranspiration (ET in mm/day), rainfall (RF in mm), slope of the soi1 surface through
runoff (RO in mm), drainage characteristics (DN in mm) and type of field operation to be
perfolmed (AME Standards, 1984). For every soil, a characteristic curve cari be
developed to relate the soi1 matrix potential yrn in MPa or bar and the soi1 moisture by
volume I& in %v/v. The soi1 water content cari be determined either by in-situ
measurements or estimated b:y using prediction models.
,r . --.- . ..-- -- --._.-__
x*
._I ____“_““__ -_
,
6
9
Thle first techniqwe described in many procedures of determining soi1 water content
in-situ involves the use of either soi1 moisture sensors (tensioniete;, porous blocks,
psychrometer, TDR) for volumetric soi1 water content (VSWC in % cm3icm3), or soi1
sampies with laboratory facilities for gravimetric soi1 water content (GSWC in % g/g>.
These methods are complementary. Once the soi1 particle density (SPD in gkm’), the soi1
bulk density a(SBD in g/çm3) and the water density (WAD in g/cm3) were known it
becomes easy to determine the other related soi1 parameters (Hanks, 1992).
The second technique used by trop modelers was mainly based on equations (soi1
water balance) to predict on a daily basis the soi1 moisture status. The prediiction equations
were generally derived from the hydrologie bala.nce which states that the water stored in a
given volume of soi1 (root zone) is equal to the difference between the amount of water
added (riain and/or irrigation) and the amount ofwater withdrawn (evapotranspiration,
runoff, drainage) (Hillel, 1982). In absence of irrigation, Hunt (1986) proposed the
following equation to determine the soi1 water content (SWC in mm) during day n (the
other variables have been previously defined):
s@C, =:(SWC,.,)-(ET,)+ (R%)-(ROn,h(LK)
(15)
TRO main characteristics of the soi1 water system given as boundxy values, are the
field capacity (SFC in % or mm) and the wilting point (WP in % or mm). The field
capacïty is defined in fthe lïterature as the soi1 water content, sometimes talled drained
Upper limi:: (DUL in % or mm), reached in a few days after wetting of tht: soi.1 layers
located near the soi1 surface (Ritchie, 1972).
70
At field capacity, the soi1 water drainage DN from the lowest soi1 layer of the root
‘.
zone is generahy considered negligible. Jury (199 1) argued that’ a true field capacity. does
net exist, as water Will always continue to drain under gravity reaching an insignifrcant
level after a few days (two days for rnost sandy soil). The water content at field capacity
corresponds to a soi1 matrix potential ranging from Y,-,,:= -0.1 bar (-0.01 h4Pa) to \\Km== -
0.33 bar (-0.03 MPa). During the desiccation process, the wilting point is described as the
Jower Emit (LL in % or mm) of soi1 moisture at which water is not available to plant. Only
adhesion water around the soi1 particles is retained by the soi1 matrix (Foth, 1990). The
soi] matrix potential at this stage of desiccation is Ym= -. 15 bar (- 1.5 MPa).
The soi1 water balance cari be determined with the measurement in-situ of the soi1
volumetric water content at DLL and LL. The soi1 water holding, capacity (SWHC in % or
mm), sometimes called extracrable water is defïned as the difference between the soi1
water content at field capacity (SFC) and the soi1 water content at wilting point (W)
SWHC’ = (SFCT -(HT)
(16)
The average extractable water in the root zone is 13% for most minera1 soils. This
is reiated to the texture, as plant water uptake also takes place during the drainage process
(Ritchie, persona1 communication). The soi1 water content at which the execution of
different animal traction field operations are recomrnended lies within the soi1 water
holding capacity range in the Upper 30 cm of soi1 and 15 cm.
*.MUl-mI-LIIIIIUC
,_-_ --.-
--
--
--v
--UII-*
7’1
3.9.2. Weather factors
.‘- ,
The weather related factors induce a probabilistic aspect into the evaluation ofthe
number of working days. The technique developed was based on determining the
precipitation probabilities (Prah in percent) or frequency (Frai,, in decimal). The probabilities
of occurrence was calculated by using the em.pirical equations adopted by the American
Society of Civil Engineering (ASCE) and reftxred to as the Gumbel’s equation (AID,
1977; Schwab G.O., Fangmeir D.D., Elliot W.J. and Frevet R.K., 1993):
(17)
P ra,n = 100 *(I- F,,,,i
(18)
Where m is the order assigned to the precipitation event ranked in ascendiny order
and n is the total num.ber of data points, The associated return period is the reciprocal of
the fi-e!quency Frai,,.
The probability values for working da:ys (PWJI in decimal) during chosen period
intena.ls (weekly, bi-weekly or monthlv) have been calculated and tabulated according to
difTerent geographical locations. The probabilities are used to help estimate the number of
working days (NWDAYS in days) a specific field operation cari be petibrmed within the
time inter;als at a cefiain level of confidence.
Le Moigne (198 1) had carried out a similar analysis for the West Afi-ican region
He proceeded by det&-mining the number of non-working days. For monthly period
72
inter-vals, he based his analysis on hypotheses built around soil/tool relationships. He
identified four large zones characterized by one rainy season w’ïthin Senegal, Mali, Niger,
Burkina Faso and Chad. For each zone he used the number of rainy days per month
(Table 13) with the following hypotheses:
Scenario 1:
He considered two predominant types of soil: sandy and clay soils. The sandy soils
are located at the highest position of the toposequence and are mainly used for upland
rainfed crops. Their physical characteristics are the same as the ferrugineous and ferralitic
soi1 types described earlier. Generally, these sandy soils have a good drainage systems.
They drain well and fast. In contrast, the clay soils are found in lowland areas located
within the watersheds along the main West African rivers (Niger, Senegal, Volta, Lake
Chad, Casamance, Gambia). They cari stay saturated for longer periods of time compared
to the sandy soils and prevent any tillage operation.
Scenario 2:
He identified the different field operations and cropping practices that are
significantly affected by the amount of rainfall in terms of soil-implement working
component interactions and ground trafficability.
In sandy soils, a11 fields operations involving the utilization of cultivator tines were
classified as quasi-impossible to carry out for rainfall events amounting to more than 50
nun/day. For plowing and ridging, the execution was difficult for rainfall between 30
mmlday to 50 mm/day. Seeding and weeding required 1 day of drainage after rainfall
Table 13: Non-working days for the West African region
:- *
--
!-----II
Zones (rain) /Field operations
Sandy soils
- -
-
-
-
F
-
h?
- 1
A
-
M
<== 350 mm
Plowing
0
0
0
0
Tines land preparation
0
0
0
0
Seedhg
0
0
0
0.5
Weeding - Ridging
0
-
0
-
0
-
0
350r:o600mm
Plowing
0
0
0
0.5
0.5
0
Tines land preparation
0
0
0
0
0.5
0
seeding
0
0
0.5
1
2
0.5
Weeding - Ridging
0
-
0
-
0
-
0.5
- -
1.5
-
0
.-
600t.0800mm
Plouling
0
0
0
0
1
0
Tines land preparation.
0
0
0
0
0.5
0
seeding
0
0
0.5
1
2
0.5
We&ing - RidRing;
0
-
0
-
0
-
1
1.5
-
0
_-
800 to 1200 mm
Plowing
0
0
0
0.5
1
0
Tine.s land preparation
0
0
0
0.5 0.5
0
Seeding
0
0
3.5
1
2.5
0.5
WeedinS - Ridging
0
0
- 0
- 0.5
0
L-
2
-
-
Clay soils
~ -
-
-
-7
- -
<=:35omm
-r
Plowing
0
0
0 0
0.5
1
1.5 3 2
1
0 0
Tïnes land preparation
0
0
0
0.5
1
2 1.5 0.5
0 0
seedirlg
0
0
0.5
1.5
3
5.5 3.5 1.5 0.5 0
Weeding. - Uidging
0
0
- - -
0.5
1.5
3 5
3.5
1.5 0.5 0
-,- - -_-
3.50 to 600 mm
Plowing
0
0
0
0.5
1
1.5
3
4 3.5 1.5
0 0
Tines land prepar.
0
0
0
0
0
0.5
1.5 2 3
1
0 0
seediig
0
0
0.5 0.5
1.5
3
6 7.5 6
3
0.5 0
Weeding - Ridging
0
0
~ - 0.5 0.5
1
2
4
7,5
2
0.5 0
-
,- -
600 to 800 mm
Plowing
0
0
0 0.5
1
2
3
5
4
1.5 0.5 0
‘Tines land preparation
0
0
0 0
0.5
0.5
1 2 1
0.5
0 0
Seed.ing
0
0
0.5 0.5
1.5
4
7.5 9 6.5
3
0.5 0
Weeding - Ridging
0
0
- - ,0 0 . 5
1
- - 3.5
A
5
6.5 5.5
3
1 0
-
800 to 1200 mm
Plowing
P
0
0
0.5
1
1.5
2.5
6 5
2.5 0.5 0
Tines land preparatidn
0
0
0
0.5
1
1.5
:
3 1.5
1
0 0
seeding
0
0
1
2
4
5
8
11 9
5
1 0
Weeding - Ridging
0
0
E Z 0.5 1
.3
3.5
5
8 6
4
0.5 0
z- L
-
I
a
- = -=
-
- - E
Source: Extract froc Le Moigne, 1981
events between 10 to 50 mm. In clay soils, he found that trafficability was greatly limited
by the level of soi1 wetness near saturation. For fïeld operatior$ the amount of ail rainfall
,..
event thresholds found for the sandy soils were shified upwards by 10, 20 and 30 mm.
The major limitation of his work resided in the fact that detailed studies on rainfall
incidence on the execution of field operation did not exist. He did not consider the
limitations induced by the soi1 desiccation and hardening processes near the lower limit of
soi1 water content (LL in ‘Si). The 30 to 50 years rainfall data collected in research stations
throughout the region have never been analyzed in terms of linking amount and
distribution of rainfalls to limitations in using the different farm implements present at tiie
household level.
3.10. Land use and cropping systems
3.10.1. Animal traction and cropping systems
The main objective of the implementation of animal traction projects around Third
World Countries was to increase agricultural productivity through yield increases and
labor savings. The challenge facing many agricultural policy makers was to produce more
than subsistence production levels. The transition from hand hoe to animal-drawn
‘?<l ,
I
implements has led to an increase of farming intensity. New cropping scenarios and
agricultural production opportunities have been created as more energy has been made
available to farmers through the utilization of draft ani,mals. At the same time, timeliness of
field activities has improved. Farmers view the labor savings objective in another way, a
means of reducing drudgery (Campbell, 1990; Ndiame, 1988). Ndiame (1988) reported
75
that the majority of farmers (79%) in the Southern part of Senegal were using animal
traction for plowing because it was casier and faster than hand lools.
Ammal traction has made significant progress and has produced changes in
cropping systems driven at least by one cash trop. The fast spread of animal traction in
different parts of the world was mainly due to the introduction and expansion of cash
crops (Eicher, 1982). In the Sub-Saharan region most of the cash crops bere introduced
and developed by the colonial system for expo:rt: groundnuts in Senegal, cotton in Mali,
Burkina. Faso, and Ivory Coast: etc... These cash crops represented the main source of
revenue to farmers involved in these cash trop oriented cropping systems.
Also, animal traction has brought about signifkant changes in cropping pattems.
Two non-desirable major effects have been the decrease in fallow through the extension of
croppedl areas and the acceleration of land degradation process through erosion. The
extension cf cultivated areas was made possible by the capacity to do more work in the
time available (Crossley and Kilgour, 1983). In. order to use animal-drawn implements on
new fields, not only must the land be cleared, but stumps and big roots also must be
removed which resulted in accelerating water and wind erosion (Pingall et al, 1987). More
detailed studies on the impact of animal traction utilization on the environment need to De
made.
The literature is rich in experiences sho-wing that working animal3 ‘were involved in
different field operations at the farm level: primary tillage, ridging, seeding,, weedimg,
harvesting, transport, etc. Farmers have found the utilization of animal traction for field
activities very attractive. Spencer (1988) pointed out that the degree of adoption by
7 6
farmers was limited by the availability of the different components making up the package
. . . \\
and by the production costs involved in their utilization.
_
TO carry out all field operations required to grow a given trop, farmers must make
strategic decisions in allocating available resources in combination with a variety of
implements present at the farm level to achieve a finalized production system (Equipe
Systemes, 1984). Sebillote (- 1983) addressed this question in terms of cropping practices
and introduced the terminology of “itineraire technique”. The “itineraire technique” was
defined for a given trop cycle as the combination of different cropping practices executed
in sequence, from tillage to harvest. Two “itineraires tecihniques” were considered differenr
if a change was made in the mode of execution of any field operation. An “itineraire
technique” involving the use of a ridge or moldboard plow for land preparation with the
remaining post-tillage operations manually executed would be different from another
“itineraire technique” using a moldboard plow and a seeder, or a moldboard plow with a
seeder and a cultivator (Equipe Systemes, 1984). The level of energy and amount of labor
used were found to be significantly different among “itineraires techniques”.
3.10.2. Field performance and trop yields
Different freld studies have tried to correlate the ievel of energy used for trop
production and yield. The power used per unit of cultivated area was a common indicator
used to demonstrate the benefit of animal traction over hand tools. Morris (1983) argued
that the yield increase induced by more energy per hectare was mainly due to improved
timeliness and better quality or precision of the field operation execution. Most
agricultural policies and decision-makers have always looked at mechanized cultivation as
77
an effective way to significantly increase the income of the household. TO this perspective
many African governments have encouraged the permanent occupation of land by farmers
but prevented at the same time any for-m of private ownership of the land. As pointed out
by Charreau (1974), the Government of Sene:gal, for instance, passed a legislation to
declare the Law of National Domain stating that any land in the country is nationally
owned. The main objective ofthis govemment legislation was to give incentives to
farmers involved in the national agricultural program of intensification tbrough animal
traction utilization. They urged farmers to rernove a11 the stumps in their fields in exchange
of 400 kg of rock phosphate (produced in Senegal) for each hectare cleared.
Agronomie research irrstitutions have carried out field investigations to establish
and quantiti the benefits of the technology. The introduction of the animal-drawn plow for
land preparation had given more opportunities to farmers to perform deeper tillage to a
depth of 10 to 20 cm compared to the super5cial traditional hand cultivation. The yield
increases brought about by this cropping technique were mainly due by the combination of
different factors. These factors are related to the improvement of the soi1 structure for
better water infiltration and root development: better weed control and c’rop residues
incorporation into the soil, better seed bed preparation for good conditions of plant
germination and emergence, and better plant use of water stored in the soil. A fairly large
number of technologies have been developed and, as shown by the percentage of yield
increase over traditionai practices, and most of them make significant contributions to
productivity. Jaeger (1985) presented a summary of station fïndings wid~ely used to
demonstrate these beneficial effects (Table 14).
7 8
Table 14: Percen.t yield increase with animal traction plowing.
Sources
Millet Sorghum Maize
Cotton Groundnuts Rice
Senegal
Charreau and Nicou
3 0
3 0
3 0
2 7
19
50
In Sergent (198 1)
Nicou
19
2 4
5 0
17
2 4
103
In Le Moigne (1979)
Charreau (1971)
2 4
25
35
25
23
Ramond and Tournu (1973
50
130
Tourte (1971)
- Plowing only
- Plowing + OM
25
4 4
43
41
2 0
4 9
4 2
85
3 9
9
Mali
SRCVO (1978)
--T--T--T-l---
In Sergent (1981)
-Donkey
- Oxen
iource: Jaeger, 1985
Iu_I/
---
q-,*m----
-
,,*a-
‘The benefits gained from the utilization of the plow have been demonstrated in
different parts of Africa. Especially in French-speaking coun&es of West Afiica, pl& use
appears to be highly variable depending on the type of soil, the year and the type of trop.
The depressive effects of plowing were mainly explained by poor plowing,.
.Field experiments conducted in farrners’ conditions in the Casamance region for 8
to 10 years showed a rnodest trend of benefits (Equipe Systemes, 1984) when plowing
was cornplemented with different levels of mechanized post-tillage operations like seeding,
weeding, and harvesting. For groundnut production for example, the combination of
mechanized plowing, seeding and weeding (“Itineraire 4”) gave 59% more yield than
plowing (“Itineraire 1 or 2”) alone while mechanized plowing and seeding (“Itineraire 3”)
gave 13% more. These yield improvements were essentially due to better timeliness of the
field operations, the efflcient eradication of weeds and the physical stirring of soi1 for
better aeration and water distribution in the root zone. In this process, labor input per
hectare was also reduced by, as much as, 20% to 25% (Sonko and Fall, 1,992).
3.11. Siimmary
The exhaustive review of literature pre:sented on the determinants of performance
of animal traction’s utilization for trop production has shown a variety of research
activities carried out in different parts of the world Most of the well-published research
activities and results found in the literature were from French (CEEMAT, IFUT), English
(ARC, AFEIC, CTVM] and Gerrnan (GTZ) institutions, American univettsities
(Department of Animal Science at Corne11 University) and NARS in Third World
countries. Investigations were conducted in dilTerent directions without really following
8 0
any predetermined standard or procedure. ,The area of research appears to be new and
‘.
open to more investigations. Lawrence and Pearson (1993) have addressed the issue of
standardization in terms of (a) length of pre-experimental periods to allow draft animais to
reach a ‘steady state’ (training, fitness, diet and climate) before measurements cari be
performed, (b) number of animals needed in an experiment for level of statistical
significance ofthe results, (c) work and performance tests to evaluate continuous and
instantaneous maximum efforts, (d) instrumentation and units to be used, and (e)
physiological tests to understand the difference in draft animal performance when at work.
The lack of standards in the measurement procedures introduced a factor of variability to
explain the discrepancies found in published research results. The nature of the study
involving living species made the task difficult for a11 researchers to agree upon the defauit
values of different indicators used to predict working animal performance.
The main constraints limiting animal traction utilization ar ,i impleme:ltation around
the world has been presented under different aspects, from climatic conditions relrired to
the cultural behavior of target groups. The difficulty in laying out experimental designs
involving weather as a variable made it quasi-impossible to rigorously compare draft
animals working in different environmental settings. Along with weather, local sanitary
conditions were described as a major issue in promoting animal traction. The prevalence
of trypanomiasis in many wet and warm areas of the tropical region for example, has been
considered by many authors to be a significant factor which slowed down the ditision of
the technology. The recommendation gained from the hterature was to help farmers
develop well-organized and feasible work plans to avoid over working their animals. More
important was fitting the job to the draft animals’ capacity in relation to the draft
-
81
requirements and working conditions, the type of impiement, the efficiency of the
hamessing system, the type of soi1 (SWC) and its resistance to’traction, and the
probabilitiems of working days.
The new methods of energy evaluation presented have been maimy impaired. by the
laboratory determination of important parameters. This fact was described to represent a
major disadvantage in the evaluation of some components of the energy balance in
comparison to field measurements. For non-fielld measurable amounts of e:nergy required
from working animais, these parameters cari be used to calculate the default values until
the development of new, more realistic methods.
The quantity and quality of feed given to animal must meet the draft animal
nutritional needs. In general, the right amount and composition of feed stuff was ofien
reported not available to farmers. Most of the feed was mainly obtained through grazing
on natural pasture. As poin.ted out by Ffoulkes and Bamualim (1989), when browsing,
animals naturally tended to Select feeds that were likely to enhance their body condition
Nevertheless. additional feed must be given on working days to avoid weight loss during
peak work demand.
The use of draft animals for field work (plowing, seeding, weeding and harvesting)
has demonstrated opportunities and potentials to generate substantial revenue to farmers
through better productivity (yield increase and labor saving). This is made possible
through better seed bed preparation, weed control, timeliness of field operation,
improvement of soi1 moisture regime, reduction of labor.
9
,”
‘
Chapter 4
. . :
MATERLALS AND METHODS
;
4.1. Research sites
This study was conducted in distinct phases including both on-station an 1 on-farrn
i
sites and focussed on the use of a pair of oxen as draft animals. Par-t of the human
.IB
I
resources (animal scientist and soi1 scientist) needed to conduct the field work were
i
provided by the Senegalese Agricultural Research Institute (MA) through its local
i
Agricultural Research Center (CRA) of Djibelor located 7 kms South of the regional
11.
capital cit! Ziguinchor.
4.1.1. On-station
“1. I
The on-station facilities were mainly used for the evaluation of the draft
requirements of different field operations (plowing, ridging, seeding and weeding).
The Djibelor Agricultural Research Station (ISRAKRA research ‘tenter) is the
major research station in the southern part of Senegal. Until 1979, research activities were
mainly oriented towards the development of rice production techniques in relation to the
amount of rainfall, the type of relief and the importance of hydro-agriculturai production
potential of the region. Since the events of the drought situation (197%82) research
CI
-. . .
83
activities were adjusted to reflect the new production strategies developed by fat-mers to
. .
coumeract the effects of the shortage of rain. Multidisciplinary research ‘leams were
constituted at the station level to approach the diversity of the productioin systems found
in the region from the perspectives of integrated agricultural and livestock activities, an.d
mixed cropping systems. Following the new d,ynamic, research operations on upland crops
(cereals and cash crops) and animal traction were given more emphasis ait both on-station
and on-farm levels.
Djibelor research station has several research facilities: soi1 and w’ater laboratov,
equipped farm shop, experimentation plots, pest management and plant genetics
laboratories. The experimental units of the research station are located aiong the
toposequence. The lowland plots are mainly used for the development of rice cropping
techniques and the upland plots for the improvement of the rain-fed cropping systems
(maize, millet, groundnuts., and sorghum). For this study, the on-station’s research
activit!:es were conducted on the upland areas in the way animal traction is mainly used by
farmers’ circumstances (Posner, Kamuanga and Sali, 1988). The soils are “red soil” of the
ferrallitic type, representative of most upland soils found in the region.
41.2. On-farm
Two research activïties were programmed on-farm: survey on animal-drawn.
implements and test for maximum pulling perfclrmance along with a follow up of the draft
animaXs’ act.ivities during the rainy season.
The on-farm taaget group was represented by farmers located in the two agro-
ecosystem zones 4 and, 5 in the not-them part of the Ba.sse Casamance region where animal
traction is well established. A sample of fat-mers using animal traction for cultivation from
four representative villages (two from each zone) was drawn for’the field research
activities. An important database about the target group is readily available which consists
of information collected by the Farming System Research team (Equipe Systemes) of the
Agricultural Research Center (CRA) of Djibelor during several years since 1982. Data are
mainly related to field operations, production techniques, cropping systems, and calendar
of agricultural production activities (Posner et al, 1985; Ndiame, 1990; Equipe Systemes,
1983). This database Will be used as a reliable secondary data source.
The villages of Boufandor and Bougoutoub in agro-ecosystem zone 4, and
Kagnarou and Sue1 in agro-ecosystem zone 5 were chosen for this on-farm study
consisting of equipment survey and maximum performance trials
4.2. On-farm equipment survey
As part of the methodology, an animal traction survey was conducted at the farm
level (unit of observation). The farm is defined as an autonomous family composed of
I..
members who produce and consume together (Jolly, Kamuanga and Sali, 11988). The fat-m
cari be made of different dependent households. The tenter of decision making process in
terms of management and resources allocation is located at the main farm level.
The focus of this study was put upon different aspects related to the use ofanimal-
drawn equipment at the fat-m level. This type of survey could easily fit into a Rapid Rural
Appraisal @WI) approach (PRAAP, 1992). It was important to complete this survey in
order to identify the range of implements used by farmers before any draft measurements
and other energy estimation was performed.
*.
i
-
--
-
_
-------
85
The enumerators to conduct this survey had to undergo technical training to be
able to identify the implements currently used by fat-mers.
“* .
4.2.1 Objectives
The objectives of the survey were:
1. Identify the types of equipment actually involved in field operations for
which more detailed technical information was needed towards energy
requirement estimates.
2. Evaluate fat-mers’ learning process for animal traction use.
The nul1 hypothesis (Ho) was that there was no difference in fat-mers’ level of
equipm’ent A farmer was considered to be equipped as long as he owed at least one farm
implement (moldboard plow, ridger, cultivator, seeder, cart, etc.). The level of equipment
encompassed two aspects of the mechanization process. The first aspect d.ealt with the
total number of implements possessed by farmers and used to carry out field operations.
The second aspect specified the possible combinations of implements used to perform field
operations from tillage, seeding, weeding to harvest. The higher was the number of
different types of implements, the higher the level of mechanization. Fat-mers with higher
level of equipment were expected to perform better in the production system in terrns of
quality and timeliness of field o:peration.
8 6
m,
!
4.2.2. Data collection
. .
‘_
The questionnaire was designed to help collect thrce categories of information:
type, management, and utilization of animal-drawn implements present at the farm level
(Questionnaire in Appendix Al).
4.2.2.1. Type of farm equipment and draft animals
The main characteristics of animal-drawn implements retained in this survey were
the type and shape of the working components. The configuration and aspect of these
components were generally described to be responsible for the draft required to pulling the
implement in given working conditions (type of soil, moisture regime).
The most popular working components expected to be found at the farm level
..,
were the moldboard plow, ridger with adjustable wings, fùrrow opener, cultivator tines
i
equipped with various size of sweeps, duck foot shovels, or chisel points, and groundnut
I
lifiers.
Multipurpose farm ecluipment is also available in the market for fiumers. These
were designed and used for different field operations f?om tillage to harvest (groundnut
lifting). This was made possible by simply changing the type of working component
attached to the tool-frame. These multipurpose types were also sized to fit the job to the
draft animals through the selection of suitable working components. In general, they
require more ski11 from farmers- compared to an implement with a simple frame design.
A second important c:haracteristic was the type of traction used to pull the
implement (pair ofoxen, horse or donkey). The stability coefficient of the technology
(STAEKOEF in percent) under farmers’ conditions was determined by evaluating the
87
I
difficulties felt by farmers who have the minimum number of draft animais for field
operations and in particular when these animal are oxen. The technology is consideredto
be unstable for farmers with only one pair of oxen, as the loss or unavailability of one
animal Will compromise the production objectives. The stability coefficient cari be
lexpressed as one minus the probability of farmers of having only one pair of oxen
(PBPAIR. in decimal):
STABCOEF,, = [l - @‘BPAIR)]
The readiness of draft animals at work time was considered as a determinant
indicator for better timeliness of field operation.,.~9 Draft animal reliability ANIREL in
percent was introduced at this point to report thle probability that the draft ianimals were
not available to petiot-m field operations on time. It was computed as one minus the
probability of draft animal unavailability, which is the probability of farmers of having one
ox only at the beginning of the rainy season (ANNA.VAIL in percent) when both
probabilities were expressed in decimal form:
Ah’IREL = [I - (ANNA UIL)]
(20)
In general, farm implements are designed an.d built with ergonomie considerations
to fit height and draft requirements for the working animals. Fortunately for the farmers,
most implelnents are interchangeable among draft-animals species through simple
8 8
adjustments of the harnessing system. Switching from a pair of oxen to a single draft
‘”
animal (donkey for example) requires the use of an evener.
’
The types ofhamesses used by farmers to hitch the implements for their fieldwork
were charactenzed in terms of attachment and size to help analyze the efficiency of the
energy transmission mode.
The third characteristic was the type of the design and the pur-pose of the
implement. Animal-drawn implements are generally built to fùlfill specific Ipurposes like
1 .
:
land preparation, secondary Mage, seeding, weeding, cultivation, groundnut lifting,
transport, etc. The utilization of one type of implement usually predeterrmnes the
execution mode of the following field operations. For example, if the land preparation is
done with a ridge plow, it is likely that a11 remaining field operations (seed.ing, weeding,
‘etc.) Will be manually executed. It is important at the farm level to identify the function of
each piece of equipment.
. ”
/
/
4.2.2.2. Management
The status and utilization rate of each animal-drawn implement found at the farm
level was diagnosed to help understand its working conditions. The farm structural
characteristics in terms of family composition (number of farm workers NFMWIERS,
household dependency) was expected to dictate the type of ownership and explain how
priorities were set up in the process of allocating available farm implements to different
fields and crops.
The mode and date of acquisition in relation to the benefits generated were
e:xpected to give indications of the number of years of farmers’experience.
,I
8 9
4.2.2.3, Utilization and maintenance
‘The level of use of farm implements during the grotiihg’season was genera&
reflected by the wear and the number of maintenance performed every yea:.r. From this
infcmnation it was expected to be able to determine the frequency of equipment
breakdown or the probability of equipment failure or the reliability (IMPLREL in percent)
for each type of implement. The farm equipment reliability is defined by ASAE (1984) as
the “statistical probability that a machine Will fimction satisfactory under specified
conditions at any given time. It is computed as one minus the probability of a failure”
(PBRKl~OWN in decimal):
IMPLREL = [i - (PBRKLIO WN ,)mp,ement /
Farmers’ skills in adjusting and using the implements, and the physiical aspect and
nature of the fields (presence of roots and stumps) are described in the literature. These
working conditions shorten the life expectancy of farm equipment in fatiers’ condit.ions
(FAKMLIFE).
13lacksmiths established in the villages were also contacted to evaluate their
capacities for maintaining implements used by farmers in good working conditions.
It was important at this point in the survey to have the farmers’ opinion about the
efficiency of the technical support provided by the extension service and fzlrm equipment
dealers in terrns of traihing, spare parts availability, and access to any type of credit.
9 0
42.3. Samplling techniques
The first step in the process was to determine the sampl’ê sïze n in order to have a
good estimation of the population mean CI. This must be done in relation to the amount of
errer d that cari be tolerated and the level of probability that this error was expected to
hold (Bhattacharyya and Johnson, 1977). Farmers equipped with animal-drawn equipment
represented the target population. The sample was drawn at random in the four
representative villages of agro-ecosystem zone 4 and 5. The sample size n for the survey
was determined by using the concept of standard error (S.E) and length of confidence
intervals (C.1). At this level of planning, the sample size must be calculated in order to
achieve more precision in the est,imation of the different statistics. In this procedure, it was
kept in mind that in general, a survey using a large sample size was more costly to
implement and much more time consuming in terms of data processing. Without loosing
toc) much precision in the results, the sample was sized large enough for the Central Limit
Theorem to apply. The theorem states that, if n is large enough the distribution of the
sample mean is approximately normal. In the case of non-normal distribution:5 because ,f
small n, the Chebyshev’s inequality is still applicable. It stipulates that the probability that
the difference between the population mean p and an observed random variable X mean is
less than a constant d, is greater or equal than 1 minus the variante of the observed
random variable mean divided by the square of the constant d.
_
_
..-
c.
,.,cIUI*-“-II,-,.-,------
.___-
---
‘--WI-*’
9 %
If the levei of confidence is lOO( 1 -a)%, then the upper bound of the sample size is:
,l zz: (-$
(23)
In bhis present study. the standard deviation CT of the population was determined
fiom previous studies condu.ctt:d in the same area (FALL, 1985; Sonko, 1986). The
amount of error d that could be tolerated was set to 10% or 0.1 as suggested by Le
Moigne (1984). The probability that the error of estimation would not exceed 0.1 was put
to 90?6 level for a= 0 10. Since the population variante 0 was equai to 0 2, one
calculated upper bound of the sample size was:
2
,, ZZZ
o.2
= 40
o.r*(O.I)I
TO implement the survey, n (total numiber of observations per variable) was
equally divided into th’e four villages: ni= 10 for i= 1 1.0 4 (i= village numt:ler)
4.2.4. Data analysis
The data collected in this survey were analyzed with the package MIN1T.A.B
Statistical Software of Release 8 PC version. MIXIT.AB is a registered tra~demarkT~l of
Minitab Inc. Basic statistics were performed on the data collected regarding each type ot’
equipment: frequency and distrilbution, mean, standard deviation and staridard error, 90°i
t-confidence Interval for the mean (Bhattacharyya and Johnson, 1977).
The factor leveis b (number of implement types for EQTYPE) and c (number of
different yoke sizes for YOKE) were determined from the où’tput of the smvey. The factor
level a (number of soi1 water content levels for SWC range) was :Function of the rainfall
profile and the soi1 water status. The soi1 water content was determined by the gravimetric
rnethod (GSWC in % g/g) and was converted into the volumetric soi1 water content if
necessary (VSWC in % v/v). Different studies on monitor-mg of field opera.r:ions have
shown that farmers never carried out field activities at optimal soi1 water content. Land
preparation was generally performed at the beginning of the rainy season just afier the first
useful rain. Three levels of factor a were arbitrary chosen within the range jof a working-
day’s soil-water content during the period allocated to land preparation and weed control.
These three leveis of soi1 water content corresponded to three days of fieldwork: one day
was chosen in the month of June and two days in the month of July.
The analysis of variante was performed to determine the significance of each
fiictor and combination of factors on the pull force PF (Table 15).
The overall F-test was performed to detect treatment differences alcng with lOO( l-
a)?/0 simultaneous Confidence lntenal (CI) to compare specifrc pairs of treatments.
1
I’he data analysis was oriented towards approaching the variability among farmers
in terms of the different types of possible implement combi&ons to conduct field
operations from land preparation to harvest
4.3. Estimates of energy requirements
43.1. Objectives
The objective of this research activity was to estimate the average energy required
in carrying out field operations.at the farm level. It was performed through the
measurement of the amount of draft needed to pull an implement in a given work ~g
condition. The draft evaluation was carried out at different soi1 moisture regimes.
The nul1 hypothesis (Ho) was that the amount of draft required to pull farrx
implements in given working conditions wa$; mainly affected by the soi1 moisture regirne
and did r1.x depend +3n the type of energy transmission system used to hitch the impiement
(widt’ of head yoke).
43.2. Experimental desigu
For this purFose, a 2 x 3 x 3 factorial. in randomized blocks desilgn was la::d out to
help determine the cDmbined effects of three factors: .A= soi1 water content SR’C range,
B= type of implement EQTYPE and c‘= size of YOKE on the amount c’f pull force PF in
daN measured. The ;hree factors had respectively a, b and c levels and m obsenations
(number 3f repetitions) tak.en for each combination of the three factor levels, for a total of
abcm ob jervations.
9 4
” I
I
Table 15: Analysis of Variante - .
Source
Sum of
Mian
ii
of Variation
Df
Squares
Square
F
/
I A
(a-1)
SSA
SSAIDf(A)
MWMSE
I
J3
(b-1)
SSB
SSBIDf(B)
MSB/MSE
C
(c-1)
s s c
ssciDqc)
MSUMSE
AB
(a-l)(b- 1)
SSAB
ssAB/DqAB)
MSAB/MSE
AC
(a- l)(c-1)
SSAC
SSAC/-Df(BC)
MSAUMSE
BC
(b- l)(c- 1)
SSBC
SSBC/Df(BC)
MSAUMSE
ABC
(a- l)(b- l)(c- 1)
SSABC
SSABCDf(ABC) MSABUMSE
ERROR
abc(m- 1)
SSE
S SE/Df(ERROR)
Cor-r. Total
abcm- 1
//
Mean
1
/I
II Total
abcm
TSS
II
.1
The general linear statistical mode1 associated with this factorial experimental
design was stated in the form of? (Stapleron, 1995):
I,-
(27)
and Eijti are random errors independently distributed: &ij~~N(O,~).
F o r i=l,...,a
j= 1, . . . . b
k= 1, . . . . c
I= 1, . . . . m.
9 5
It was also assumed that the observations Yijkl (PF in daN) were normally
.
distributed as Yijp N(pijk, 02) and independent.
YijM represented the pulling force PF in daN or kgf and p the expected average
draft in daN or kgf. The parameters a, p, and y were unknown parameters called the
interact.ion effects of A, B and Ç respectively The first and second order interactions
among factors were also included in the formulation of the regression mode1 as well.
The general mean p was calculated.
(29)
(30)
There were a total of 18 different combinations applied each to a 200 mc to .jOO
m2-experimental unit.
Besides this experiment, additional draft measurements were performed to
determine the power requirernents and energy expenditure for other post-tillage field
operations: seeding, weeding.
9 6
43.3. Draft measurements
The measurements performed on each experimental unit were:
Weight of each animal LW in kgf,
Arnount of pulling force PF in daN or kgiT,
. Angle of pull y in degrees (“d),
. Oxen working time or field time FDTIME in s,
. Oxen working speed ANSPEED in ms,
Distance traveled DSTTRAV in m,
Implement working depth IMPLWDP in cm,
Implement working width IMPLWWD in cm,
Soi1 gravimetric water content GSWC in % g/g.
Soi1 penetrometer Cone energy EDEP in J/s.
The pulling force (PF) was measured at the hitch point on the implement and
represented the a:mount of force developed by the pair of oxen to pull the implement on a
continuous working rhythm. For each experimental plot, the range and the average pulling
force were observed. Not a11 the pulling force developed by the draft animais was used to
do usefùi work. Only the component in the direction oftravei was converted into tractice
force to pull the implement in the soil. The amount of tractive force, also called the draft
required (DR) by the implement, was expressed as a fùnction of the angie of pull y in
‘degree:
9’7
DR -= Pr‘ *(Cos y)
.
(31)
This required draft DR was expected to be highly correlated to the effective
working width (IMPLWWD in cm) and depth (IMpL.WDP in cm). The e:lTective working
width (more or less than the measured implement working component width
WWKCrCOMP in cm) was the width over which the implement actually worked.
4.3.4. Power and energy eupenditure
The energy fùrnished by the working animais to perform the field operation was
represented by the amount of energy needed to pull the implement and the distance walked
from he beginning of the operation to the end. The energy delivered to pejform the work
ANERGY in MJ was estimated by multiplying the pulling force PF in daN by the distance
traveled DSTRAV in rn:
ANERGY = PF * (DISTRA V)
(32)
The energy required to pull the implement ANERGYRIQ in MJ was calculated as
the draft: required DR in daN multiplied by the {distance traveled DISTRAV in m:
(33)
9 8
Another approach to the energy evaluation was derived from the power concept
defined as the rate of doing work. The animal power ANP&J%R was calculated by
multiplying the pulling force PF or the draft required (DR) and the animal field speed
,a.
l
ANSPEED in ms. The field speed was defined as the average rate of implement travel in
the field during an uninterrupted period of functional activity (AME, 1984). ‘The rate of
travel was evaluated as the distance traveled (DISTRAV) divided by the fieid time (FDTM
in sec) that the working component of the implement stayed in a non-stop working
position.
ANERGY = k * (ANPOJER) * (ANWKGHR)
PJ)
The energ? spent by the draft animals walking across the field while per%rming the
work must be added to have a more complete formulation of the total energy (ZIquation
5).
The energy required and delivered to perform a given task was used to determine
l^
i
nutrient requirement by converting energy expenditure into amount of availabie feed to be
supplied to the working animals.
IV.
/
4.4. Maximum performance
4.4.1. Objectives
The average available power was expected to increase as the body weight of the
,s.
;
animals increased during the rainy season resulting fiom the availability of more grazing
i
,.,
areas. The objective of this experiment was to measure the maximum effort developed by
I
i
7
*
‘99
a pair of oxen and to determine the efficiency of energy utilization by comparing rhe
energy used and the total energy available.
‘. 3
4.4.2. Procedure
The available animal power was measured at the beginning of the rainy season
during the period allocated to land preparation. The power measurement was used to
determine the maximum draft that a pair of oxen could develop during th;e growing
season. As mentioned by Goe (1983) farmers are facing a problem of animal feed at the
beginning of the cropping season when work labor demand is at its peak.
The research activities were divided into two parts:
- Measurement of the maximum pull force at the beginning of the rainy
season;
- Follow-up with the selected draft animais to monitor the type of
management used by fat-mers during the growing season (see Appendix
A2 for follow-up guidelines).
The measurements were performed in each of the four previousiy mentioned
villages located in the agro-ecosystems zone 4 and 5 and on-station. Two pairs of oxen
were randomly selected per village to make a total of 9 pairs of oxen ( 18 draft oxen) to be
used for the test and the follow up during the cvhole rainy season (June to Oct.ober).
100
- Pulling trials forcontinuous maximum effort
The experiment consisted of pulling triais on soi1 re&iy for land preparation by a
‘_ \\
pair of oxen with a progression of known loads placed on a sledge. The variables observed
were:
Age of each animal AGE in years (yrs),
Weight of each animal LW in kgf,
Pulling force PF in daN or kgf,
Angle of pull y in ‘d,
Distance traveled DISTRAV in m,
Working time PULLTIME in s,
Working speed ANSPEED in m/s.
For thc data analysis, the following graphs were analyzed to determine the
optimum power (AXPOWER,,,t) delivered by the draft. A general 1inea.r modeling
technique was used:
.
Power output ANPOWER:
ANPO WER = f(PF)
(33)
. Specifïc power output ANPO~~IXSPEC:
(36)
The same analysis was conducted with the speed ANSPEED in m/s:
ANSPEED = f(pF
(37)
(38)
Two identical sledges were locally fabricated for the pull trials.
- Follow up of draft animais
Mer the pulling trials., a follow up was agreed upon with the famers to monitor
a11 the activities performed by the 18 draft animals during the rainy season. Information
collected were (,Appendix: AZ):
. Liveweight (LW in kg) at the end of every month.
Type of activities canied out during the given period.
. Feeding system (type of feed and amount given)
Handling and tare.
The foilow-up was performed on a dai1.y ba.sis by two of the survey personnel who
were in continuous contact with concerned farmers.
102
4.5. Iusf rumentation
‘The instrumentation used in this study was developed by CEEMA’T and was
designed for automatic field data collection (CIRAD-SAR, 1993). The anj,mal-drawn
system was composed of two basic elements: a data logger 2 1 X (Campbell Scientific, Inc j
and a series of sensors to be mounted on the implernent ancl connected to the data logger
The data measured by the sensors were made on real time basis through the interna1 clock
of the data logger used for their storage. The main sensors used in Fhis study were
500 daN-dynamometer sensor for measuring the amount of pulling
force (PF),
Dickey John (registered trademark) radar to measure the working
speed (ANSPEED) in m/sec,
Tntcrnal clock of data storage unit to measure the working time
(FDTM in s)
For later data processing, the data logger was downloaded to a computer with the
support software EDLOG via a SC32A RS-232 interface using the software TERM or
SC532 RS-232 interface from an auxiliary storage unit using the ShlCOM The RS-232
interface was to convert the CMOS logic levels of the data logger to the RS-232 levels of
the computer. The SC32A and SC532 provided an effective isola.tion between the
computer’s and data logger’s electrical system, protecting against normal static discharge,
and noise (Campbell Scientifïc, 1992).
103
An additional mechanical dynamometer was also used to supplement the electronic
system in the case it failed to operate properly. The dynaniometer was a spring-loaded
type made by PIAB’M and had a maximum force of 500 daN.
The additional field equipment used to carry out the field measurements were:
I-land held stopwatch,
Measuring tape,
Rain gages,
Soi1 penetrometer,
AMS Soi1 sampiers (Ben Meadows Company).
Electronic weighing scales (Bario Electronic ScaleYM hIode 2 100).
The data collected during the season were used to develop a mode1 to evaluate
draf? animais liveweight.
4.6. Expert System building
46.1. Objectives
The expert system mode1 was built to help achieve an efflcient utilization of animal
traction for trop production through choice of animais and implements to perform specifk
field operations. It was designed for researchers, extension agents, farrners and policy
makers ‘to simulate different scenarios of animal traction utilization at the farm ievel.
13y introducing a confidence threshold level set by the users for “fbzzy reasoning”,
104
the expert system mode1 based on Rnowledge Base Management Systern (KBIvfS)
program was expected to yield more realistic results in farmer’s’ conditions.
._
4.6.2. Knowledge Base Design and Deveiopment
4.6.2.1. Presentation of the expert system
The expert system was developed with the LEVEL 5 Object program (for
Microsoft WindowsTM Standard Edition Release 3.6 Copyright@ 1995 Information
Builders, Inc. NY). The selection of a suitable shell was considered to be the first step in
developing an expert system.
LEVELS OBJECT is a software development tool kit. The program shell is able to
activate external conventional programs called within the expert knowleaclge base and was
able to perform mathematical operations, incorporate uncertaintv and partial id: rmation.
and explain its decision logic. The program was also user-friendly and composed of a
built-in screen editor, a compiler for the inference engine and an interfacing mechanism
with program written in TURBO PASCAL ar:d Microsoft FORTRAN.
The program development was made by using a combination of clecision-makin-
rules, methods and demons with a high-level knowledge engineering laquage called PRL
(Production Rule Language) based on inferences in linguistic form and written in IF-
THEN rules.
Rules are used to express logic, cause-and-effect relationships, between the facts
and conclusions, while a demon proceeds by testing for patterns in data based on
conditions and then executes an action based on the conditions. The method is similar to a
,..k,m.
I
:105
macro with a number of commands ‘associated with an attribute. It defines the attribute’s
behavior and executes a series of actions when the attributë.yAe changes.
Rehak (1983), cited by Raman et a1 (1992), reported that there were three types of
knowledge that could be integrated into an expert system: Heuristic (from experience),
conventional (regarding facts) and inferential knowledge (study of results). They found
that inf’erential knowledge was encountered ün most of the engineering applications.
The representation of knowledge requires an efflcient organization of the
objectives or goals of the expert system and the supporting structure of information.
The inference engine of the shell evaluates the knowledge captured in the rules,
demons and methods by exploring the problem space represented in a form of an IF-
AND/OR-THEN structure.
The program was developed using a backward chaining control structure to
connect the individual rules and forward for the methods. In the backward reasoning, the
reason.ing is performed starting from what the program wanted to prove towards the facts
that are needed rather than beginning from the facts. This system of reasoning described
by Jackson (1990) was more focused than forward chaining, because only potentially
relevant facts are taken into consideration. An example of such reasonimg:
Rule for draft animal selection:
RULE for Draft animal selection
IF the animal is trained
AND ihe animal liveweight > (Mature weight2)
THEN the animal is a draft animal
RULE for trained animal
IF the animal is 2 years old
AND the animal has learned basic instructions
THEN the animal is trained CONFIDENCE SO
106
CONFIDENCE is used to quantify the degree to which one cari be confident in the
‘_ .
accuracy of the conclusion or to give an indication ofthe likelihood of occurrence. The
.. .
lowest confidence level for LEVELS to reach a conclusion or goal is set up through the
THRESHOLD statement.
With the command CHAIN, multiple knowledge bases cari be linked if the domain
of investigation cari be partitioned into separate smaller knowledge bases. This procedure
allows the building of programs of a virtually unlimited size program. It also improves the
readability of the program by enabling the primary knowledge base to activate the
necessary sub-knowledge base depending upon the context.
4.6.2.2. Expert system organization
The mo’del was built around three major components.
- Field capacities evaluation
- Animal energy balance ay,d feed ration
- Annual farm budget and Optimization
The values of the input variables required to t-un the mode1 were either developed
previousiy in the field experiments or taken from a reiiable secondary database
The data input system is composed of disk files and operator entered data through
the user interface system. The mode1 was designed to use different databases related to.
Farm: farm size, number of farm workers,
Farm implements: TYPE, WORKING CQ$@ONENT, PF,
Draft animals: SEX, LW, AGE,
Available feeds: TYPE, NCTRIENTS, FU,
Crops grown: TYPE, VARIETY, YTELD.
Validated default values are used to supplement for missing parameters. The data
entered by the user was made possible by the capability of the expert sys,tem to use prompt
boxes for a11 the information required fiom tbe user. The user could also query the
progra.m to know why a given information was needed and get the explanation,
estabii.shing very interactive communication.
4.6.2.3. Data processing
- Draft and power estimates, and field capacities
The statistical models developed in the second part of the methodology were used
to evaluate the amount of traction delivered, the traction required, the energy delivered.
and the energq required.
The pulling force PF in daN and draft DR in daN required were evaluated for
evety working day in relation with the type of implement used.
In relation to the pull force PF or draft required DR the field capacities EFC in
haiday werr evaluated along wiith the number of working days NWDAYS.
The soi1 water balance was used to estimate the number of working days
NL+DA‘YS affected with a level of probability (5 years out of 10 and 9 years out of 10).
108
The number of working day’s per month was evaluated by subtracting the number
of non-working days from the total number of days of the, &responding month. .The
_ .
execution deadlines of specific field operations were taken into account. For each field
operation, two soi1 moisture limits were considered: drained Upper limit. (DUL) and lower
limit (LL). The difference between DUL and LL, was used to determine: the soi1 field
capacity and trafficability in relation to by the amount of rain. The lower limit was main11
indicated by the level of hardness of the soi1 during the drying process. .4t this level of soi1
water content, the implement worlcing components were barely able to lpenetrate the soi1
surface.
The probability of a working day (PWD in percent) was calculated as the ratio
between the number of working days (NWDAYS) over the total number of daq’ (NDAYS)
of the chosen period (month= 30 or 3 1 days, week= 7 days and SO on). In this study the
probability of a working day was evaluated on a monthly basis:
(39)
A coefficient called effective probability of a working day (EPWD in percent) was
introduced and defined as the PWD weighted by the draft animal reliability (ANIREL in
decimal) and the farm implement reliability (IMPLREL in decimal):
(40)
109
- Energy and nutrient requirements
The nutrients required to carry out the field activities‘were evaluated fi-omthe
energy balance equation or factorial method developed by Lawrence and Stibbards (1990)
(Equation 5). It is quasi-impossible to determine directly in the field the total amount of
energy used by a working animal (Falvey, 1986). The main components used for the
enerjgy balance were the net energy for the work performed by the draft animals and the
net energy used for walking throughout the field.
The calculated energy represented the extra energy on top of the maintenance
energy (h4AE), supplied by the draft animal’s body to perform the work required. In the
process, the digestive system of the working animais converted the feed into digestible
energy (DE). The part of the digestible energy used to perform the work is cailed
metabolizable energy (ME), and is measurable in forage unit (FU). As defined by many
researchers, the feed or forage unit represents the net energy of 1 kg of barley with no
losses in the animal feces, urine, gas, etc. (CEEMAT, 1974; Goe and McDowell. 1980);
Watson, 1981). The number of forage unit per dav for maintenance was calculated by
using Equation 11.
The program was designed to compose the ration from the variety of feed available
to farmers, which could satisfy the daily total amount of forage unit. The amount of feed
requir’ed FRFEEDDREQ in kg of fresh feed per dav cari be calculated by converting the total
FU into fresh feed (forage, grain, bran, straws) on a basi.s of 1000’0 dry matter content
(DM) and intrinsic value (FU) of each type of feed.
110
100
TotalFIl
.
FRFEEDm, = (----
.w
DM) * ’ FiIJFzED )
‘. ,
The upper Emit of the amount of feed an animal could take per day on a dry matter
DM basis was set equal to the Appetite Limit (AL in kg of feed/day) for the corresponding
animal of liveweight LW (Equation 10).
- Production costs and Optimization
Cropping system techniques (“Itineraire t,echnique”) were evaluated in relation
with the level of energy used. The cost associated with animal and implement utilization.
and man-labor used was included in the production cost evaluation with updated prices.
For the benefit calculations, the average yield of the 10 past years’ experience was used for
each cropping technique and level of energy used for four “Itineraire techniques” using
different implement’s combinations.
It was assumed that the cropping system was driven by a cash trop (groundnuts)
and marketed rhrough the officia1 network to help farmers caver a11 the production costs
The optimization was formulated in terms of:
Objective function: What is the trop mixture and hectarage that
would maximize farmers’ profit.
Constraint fùnctions: land availability, labor availability, available
animal energy and required cereals consumptio,n based on FAO
standards,
111
4.6.2.4. Program output
The output of the: expert system was (oriented towards different aspects of farm
equipment selection and management. and decision making process.
- Field Capacities
- Labor balance per ha (available vs required)
- Animal energy balance
- Feed ration
- Complete annuat farm budget
- Optimization solution.
4.6.3. Program validation
The expert system was built around different statistical models and simple
equations to evaluate or calculate variables involved in the decision-making process at the
farm level in terms of resources allocation. The quality of the output is mainly relat’ed to
the validity of the different equations used.
4.5. Summary
The methodology developed in this stuldy was mainly oriented towards the building
of an ex.pert system that would really evaluate the impact of animal traction at the farm
level. The evaluation was centered on trop production activities. The methodology \\Vas
divided into three phases:
112
- ‘survey: the objective of the farm survey through a simple questionnaire
was to identify ail types of farm implements really involved m’crop production at the farm
. v
level. ‘This investigation technique offered an oppot-tunity to Select the implements along
with the working animais on which more detailed technical information vdere needed in
terms of power and energy requirements.
- Energy estimations: the energy estimations were performed on use field
implements and working animals only. The pulling force delivered by the draft: animais and
measured with a dynamometer represented the tore of these estimations. Al1 the
measurements were made in real working conditions in order to include as many variability
factors as possible beside three major ones: type of implement and yoking syst.em, and soi1
moisture regime. It was important to place the animals in their environment to avoid an>
extrapolated data. The mode1 developed would be only applicable to the current soi1 and
weather situation. The expert system program would offer possibiiities to use the methods
in other environments.
- Expert system building: the expert system was designed to integrate a11
aspects dealing with the technology evaluation at the farm level. The pro:gram cari be used
for calculating or predicting the draft requirements, the number of working days during
the growing season, the cost of production and also for estimating the feed rations needed
by the draft animais.
The validation of the expert system was mainly performed through the validation
of the different models used to evaluate various aspects of the technoiogy at the preset
confidence level. The major outputs of the program are expected to compare with actual
/
recommendations made to farmers in terms of resource management and allocation.
‘.
.
Chapter 5
FARM CHARACTERISTICS AND SYSTEMS MANAGEMENT
5.1. Farm equipment
The process to identifl a11 the different farm implements used by farmers took two
fi111 wet:ks to inspect each item surveyed in detail. It required quite an expertise and
patience to evaiuate the quality of the impIement> the effects of utilization and maintenance
in famlers’ condition, and to get the needed information from the users.
A total number of 126 implements belonging to 40 farmers located in agro-
ecosystem 4 and 5 have been surveyed. Al1 the implements involved in trop production
were walking equipment with handles (Figure 3 to 7). At the 9006 confidence level, the
total nulmber of implements consisted of the following (Table 16): 23% moldboard plow
UCF (0.73 $I 0.19 per farm), 22?/0 ARARA and EMCOT ridgers (0.70 L 0.16 per farm),
9% cultivator Sine 9 (0.28 F 0.13 per farm), 19/0O/ SUPER ECO seeders (0.60 2 0.16 per
farm), :and 27% ox-cart (0.85 & 0.16 per farm).
The distribution of each type of implement per agro-ecosystem shows that most of
the UCF msldboard plows were located in zone 4 with an average of 1.25 per farm (CV =
43.2%) against 0.20 per farm in zone 5. On the other hand, most of the ridgers were
confined in agro-ecosystem 5 with an average of 1.10 per farm (CV = 40%) against 0.30
113
IJCF Moldboard Plow
BBG Ridger
Figure 3: Moldboard plow UCF 10”
Figure 4: EMCOT Ridger (Gambia)
Houe She 9 Cultivatx
Figure 5: ARARA toolbar ridger
Figure 6: SINE 9 tooibar 3-tines
Super Eco Seeder
Figure 7: Seeder SUPER ECO
115
in zone 4. The same analysis shows also that the seeders SUPER ECO and cultivators
SINE: 9 were mainly used. by farmers
.
located in agro-ecosystem 4 respectively withi an average of 1.10 (CV = 40%) and 0.50
(CV -= 118%) per fat-m. 0x-carts were mainly
Table 16: Farm characteristics
Farm Fat-m Size Plow
Ridger Weeder Seeder 0x-CartI
Statistics
workers (ha)
UCF
ARARA SINE9 SUPER
i 0x
ECO
Mean.
10.53
4.03
0.73
0.70
0.28
0.60
0.85
2.23
Std .Error
0.71
0 26
0.11
0.10
0.08
0.10
0.10
0.16
Median
9.50
4 00
1 .oo
1.00
0.00
1.00
1.00
2.00
Mode
9.00
4.50
0.00
1.00
0.00
0.00
1 .oo
2.00
Std Dev.
4.51
1.64
0.72
0.61
0.5 1
0.63
0.62
1.03
Variante
20.36
2.70
0.51
0.37
0.26
0.40
0.39
1.05
Kut-tosis
-0.26
0.41
-0.89
-0.54
2.02
-0.54
“0.35
0.96
S kewness
0.50
0.63
0.36
0.25
1.66
0.56
0.10
0.87
Range
19.00
7.00
2.00
2.00
2.00
2.00
2.00
5.00
~finimum
2.00
1.00
0.00
0.00
0.00
0.00
0.00
0.00
-Maximum
21.00
8.00
2.00
3.00
2.00
2.00
2.00
5.00
Sum
421.00
161.35
29.00
28.00
11.00
24.00
34.00
89 00
Count
40.00
40.00
40.00
40 00
40.00 40.00
40.00
40.00
C.l(O.90)
1.17
0.43
0 19
0.16
_~-
0.13
0.16
0.16
0 27
Zone 4
Mean
0.00
0.50
Std dev
0 30
0.59
Zone 5
Mean
1.10
0.05
0.10
Std dev
0.44
0.22
0.30
116
used for transport especially in agrorecosystem 5 with an average of 1.05 (CV == 56?d, per
farm against 0.65 (CV = 88%) in zone 4.
‘_ .
- ,
It appears that farmers in the area of study were more concerned with the
acquisition of tillage tools (5706 of the total) (Figure Sa). This was cons’idet-ed as a
priority by farmers for a number of different reasons that previous studies have tried to
show. Three main reasons need to be mentioned: the improvement of field operation
timeliness, the extension of cropped areas, and the reduction of drudgery (Ndiame, 1980)
The geographical distribution of types of implements is an indica.tor of différent
cropping techniques used in relation to the resources available to the farmers (land,
number of farm workers). The type and number of implements that exist at the farm le) rl
is also related to the agricultural practices. In other words, two farmers l’ocated in the
same area may be expected to use different cropping techniques to produce the same trop
51.1. Types of implements
Most of the implements are Eocally manufactured by SISCOMA ‘(35OG) created in
1963 which became SISMAR in 1984 under a new management staff (FN-L. 198 1). Since
that time SISMAR has manufactured 49% of the implements actually used by farmers
The few implement imported (16%) are mainly found in agro-ecosystem 5 and originatzd
from Ciambia (Figure 8b).
5.1.1.1. Moidboard plow UCF 10”: Tillage
In the Basse Casamance region, a moldboard plow is one of the most important
implements used for land preparation (Figure 3). Besides breaking the soi1 for better
117
.
-.----
30
20
œ
Implenent types
-l_-- --.- --~-- ---- --------I_
Figure 8x Types of implement used by farmers
I.- .-_---
60 ------. _.-_ .______ _-- .._--._ .__-_-_
-- ..-.---.
50
0
s1sMA.R
IMPORTED
Manufactures3
---------.---
- .--.--.-. ------.-__.-__- __--.-.-_-.-. ~--
Figure 8b: Implements’ manufacturer
aeration., its main purpose is to plow. under both weeds for better control, and manure for
rapid decay. The better the bottom scour and turning the soil; the better 1::he moldboard
plow.
Designation: Moldboard plow CFOOOP LJCF type.
Hitching system: chain.
Traction: oxen.
Stability: single wheel for longitudinal stability.
Reversibility: one way,, right-hand plow.
Single bottom moldboard plow.
Nominal cutting width: 254 mm.
Total weight: 38 kg.
Working depth: 17 to 20 cm.
- Beam
The beam is made out of a 50x20 mm flat mild steel. The standardization test has
shown resistance to effort as high as 950 daN. The maximum effort reached in farmers’
conditions never attains this level, even when taking into account the presence of stumpj
scattered in their fïelds. In term of hardness, the value of the Brinell hardness test is 55
kg/mm2.
- Share and landside
The share and landside are made of the same mild steel and tested 95 l@nmL. The
general acceptable working component hardness recommended by research is wirhin the
119
range of65 to 70 kg/mm*. The share is a straight: edge type and its main role is to tut the
soi1 shce or furn~~ before sending it over the curved surface-of the moidboard in order to
,
be twisted and inverted or pulverized. The landside while riding along the Furrow wall
counteracts the throw or side pressure to help keep the plow steady while working.
- Moldboard
The moldboard is a universal type or cylindricol-helicoidal in design. This part of
the bottom of the plow serves to break or invert the fùrrow slice. It plays an important
role t,owards the quality of work to be done. It is also made of mild steel. The rnaterial
used to make the moidboard is of great importance, the better the soi1 scour on its surface,
the better the material.
These three working components along with the beam are bolted on top ofa mild
steel piece of irreguiar shaped metal called a fi-og.
- Hitching sytem:
The hitching system directly affects the maneuverability of the implement, as the
handling and the draft depend to a great extent upon correct hitching adjustments. The
hitching system consists of horizontal and vertical regulator to adjust the working depth
IMPL’WDP and the width MPLWWD It is important that the point of hitch be on a
straight line with the center of draft or resistance and the tenter of power or yoke hitch
point.
120
‘fhe implement is maneuvered by the plowman with the handles weEdeld on the rear
of the beam. The manufacturer SISMAR indicates an average theoretical field capacity of
‘. .
0.45 to 0.50 ha/day in an 8 hr-day.
5.1.1.2. Imported ridger (EMCOT)
This farm equipment is imported from Gambia and is equipped with a bottom
ridger (Figure 4). The implement is also called a middlebreaker. The bottom is built with a
double moldboard plow to invert the turrow slice, half to each side on every pass.
Completing the operation in a-round-trip results into the formation of a ridge f’or planting
crops. The winged designed moldboards are adjustable to control the height RIDGHCT
and the spacing of the ridges RIDGDIST.
The implement is similar in design to the UCF plow described earlier except that
the beam is made with an 1-beam instead of flat steel and -bat the share is in one piece for
the two moldboards.
The implement settings are performed through a clevis equipped with vertical and
horizontal hitch adjustment. The adjustment rules are the same as the UCF 10” plow for
obtaining a correct line of pull.
5.1.1.3. Seeding implement
The success to growing crops depends on the conditions the seed is pla.ced in the
soi1 in relation to different agronomie parameters along wixh the seeding equipment
features: type of seed distribution and dropping mechanism, furrow opener andl seed
coverage system (Figure 7).
121
Designation: Super Eco seeder
One-row and multi-purpose seeder.
‘. ..
Hitching system: chain with or without evener.
Traction: pair of oxen, horse or donkey.
Nominal cutting width: 254 mm.
Total weight: 38 kg.
- Frame
The frame is made ou’t of two 30x7 mm parallel flat mild steel runners, 1062 mm
long and braced in the middle. The frame opens wide progressively towards the front end
to allow the attachment of the hopper (galvanized zinc to protect from rust and
corrosion). The runners a.re joined together at the forefront to form the hitching system
made out of a 30x12 mm flat mild steel.
.4t the rear of the parallel runners two welded vertical standards: /one on each
nmner., are used to fix and adjust a press wheel (two point-to-point conical elements)
which firms the soi1 around the seed. The press wheel action is to complete the combined
action of the seed shoe (furrow opener) and the set of two sweeps (30x12 mm flat. IL!
solid steel) used to caver the seed with soil. The furrow opener and the sweeps are borh
clamped on the runners and are both adjustable in depth. At the very rear end, the runners
are firmly joined by an angle iron brace used also to support the row marker.
1 2 2
- Ilandles
The user guides the seeder with a pair of handles, adjutiable in he.ip,ht fix better
.
comfort. The handles are made with a 25x7 mm flat 1/2 solid steel.
- Metering wheel
The seeder is equipped with a pair of 400-mm diameter metering wheel made from
a 40x4~mm 1/2 solid flat band- steel. When rolling on the soi1 surface, the wheels drive the
distribution mechanism by means of an aile used as driving shaft. The distribution system
is composed of a system of two gears (driving and driven) and a plateau located inside :he
happer. The plateau is equipped with two pins which help hold the seed plates made of
aluminum. Two types of seed plates are more common: Ml-drop round-hoIle and flat-drop
types.
5.1.1.4. Multipurpose toolbar implements
Toolbar implements have been developed by French agricultural research
institutions (RAT, CEEMAT, MOUZON Mfg) working in West Mica to help farmers
increase their field operational capacities (Figure 5 and 6). The strategy behmd the
toolbar is to buy a single multipurpose frame and different attachments corresponding to
specific tasks (tillage. weeding, groundnut harvesting). There are a large variety of teolbar
impiements available to farmers in the whole West African region.
--W
.wœ”“.---
_- _--_
---- - --.
.,XWUU*IU.IIYUIIP-s---‘--
123
5.1.1.4.1, Multipurpose Carne MUR.4
.
Designation: ARF&4 toolbar
1
Utilization: Tillage, weeding, ridging, lifting.
Traction, oxen, horse and donkey for light soil.
Weight: 3 1 to 46 kg with the atta.ched working component.
Tillage attachments:
10” bottom
8” bottom
6’ bottom
Ridging attachments
.250 mm fixed wings
350 mm fixed wings
Adjustable wings
Cultivator attachments:
3 fi-111 and half sweeps (left and right)
5 full and half sweeps (lefi and right)
Optional: double reversible points
Groundnut lifiing attachments
200 mm sweep
350 mm sweep
500 mm sweep
124
- Frame
The main beam consists of a straight squared shaped tube made of mild stee,l. The
. .
stability of the beam in the working position is provided by a wheel. The wheel standard is
clamped at the front of the beam through a welded vertical adjustment device on which a
vertical 5-hole hitching system is also welded.
A pair of 1/2 solid steei braced-handles is bolted at the other end of the beam to
help the user guide the implement.
- Equipment attachment
The standard for either moldboard or ridger bottom and groundnut lifier is
at:ached to the beam by bolts in a special groove located between the handles attachment
and the rear end of the beam.
For the attachment of the other types of working components (spring tooth shanks
or straight-st nmed), two cross bars are used to increase the versatility and working
capacity of tire implement. Up to 4 tine standards cari be damped on : le (cross bars and
one under the main beam.
5.1.1.4.2. Multipurpose frame Sine 9
Designation: Sine 9 Cultivator or hoe
Utilization: secondary tillage, weeding, ridging, groundnut lifting
Traction mode: oxen, herse and donkey on light soi1
Weight: 30 a 45 kg with the attachment.
1
1
125
- Frame
‘.
The main beam of the frame is made out of a 40x20 mm flat milld steel on which
cross bars of the same marerials cari be added to increase its working capacity. The
attachment system is very similar to the ARARA multipurpose toolbar described earlier.
The handles are made out of2Ox27 mm flat mild steel and designed wide enough over the
frame to help its ease in use.
The stabilizing wheel standard is fïxed on the main beam by means of a clamp
equipped with a vise for easy continuous height adjustment.
- Attachments
10” plow equipped with a special shank and standard.
Reversible point mounted on a 25 mm square straight-stemmed shank.
SCanadian spring tines for weeding and stirring the soil.
The shovel and sweep attachment system is composed of one 160 mm centrai
sweep and 2 half sweeps :for weeding. Better action of S;oi1 stirring is performed with a
systern of 3 160 mm-füll sweeps.
Ridger: equipped with two adjustable or fixed wings with 2.50 m.m and 360 mrn
options. The attachment to the main frame is made with a special standard.
Groundnut M?er: The Mer knives or shovels are available in 3 sizes according to
the traction mode and the soi1 tvpe: 300 mm. 350 mm and 500 mm. The
attachment is made with a standard constructed of matrix steel to better
1 2 7
’
Table 17: Breakdown frequencies on farm equipment
Agro- Location
Implement
Repair and maintenance
Freq
Average
eco
type
type
(%)
distance
zones
il,- \\
126
5 . 1 . 2 . C a r t
Designation: 1.500 kg Grand-plateau cart
- .
Traction mode: pair of oxen
Plateau: 2.40x1.50 m
Base: 1.40 m
Pneumatic 145x 14 tires
Me with conical ro.ller bearings
Unladen weight: 220 kg
5.2. Repair and maintenance
The quality of repair and maintenance represents a key factor to t e stability of thc
technoloWq in farmers’ conditions. The survey has revealed that the level of expertise of rhe
local blacksmith was directly related to the farm implement working conditions and
reliability. Most of the blazksmith were not equipped enough to salve a11 the maintenance
and breakdown problems encountered by farmers on their piece of equipment (Table 17)
52.1. Breakdown frequencies
5.2.1.1. Moldboard plow UCF 10”
The straight edge share supplied by the manufacturer (SISCOMA and SIS;L’I.AR!
to equip the plow lasts no more than two years. The locally manufactured share made of
flatten truck Springs by the blacksmith needs to be changed almost every year. Plow shares
r
are d,amaged mainly by the presence of roots and stumps in newly cleared fields as farmers
* l
have the tendency to extend cropped upland areas. In agro-eco system -1 the high rate of
1127
.
’
.
.
Table 17: Breakdown frequencies on farm equipnent
3 Bougou,toub / Cropping Working components: frog
88
8
shares, landside,
sweeps.
Wheel bushings.
Transport Tongue welding
l:!
,-.-‘79
Tube and tires
- -
-
-
-
-
-
-
5
Kagnarou
Cropping Working component: Wings
64
On-site
shares, landsides.
Wheel bushings.
Adaptation and wing
welding on EMCOT ridger
1
Transport Tongue welding
36
On-site
Tube and tires
1 5
-
-
-
- -
5
Sue1
Cropping
Working component : Wings
12
8
reversible shares,
1
and
Wheel bushings
;
On-site
Adaptation and wing
45
welding on ERICIOT ridger
Transport Tongue welding
13
1 7
Tube and tires
-
128
utilization of moldboard plows explains the high maintenance frequencies of 60 and
88%. The landside has a longer life than the share. In gener& 1Aacksmiths buy the raw
‘- .
material used to make the share and landside from junkyard retailers.
The share plays an important role in the amount of draft required to pull the plow,
the penetration, and mostly the stability of the plow in. working position. Different studies
has showed that 75% of the draft is centered along a line passinr across the plowshare and
localized 113 from the landside.
The frog represents *the central piece of the UCF plow. It holds t
(share, lanside, moldboard) and the beam together. It needs to be protected from any type
of wear by replacing worn share and landside on time.
The majority of the implements that are in bad and very bad conditions (Figure 9)
are made of implements with worn frogs. These implements need to be discarded and
replaced as it cari be more costly to fix a plow bottom at this level of wear than to
purchase a new and complete implement.
According to the manufacturer SISMAR, GCF 10” plows are built to last at leasr
for MWLLFE= 5 years against FARMLIFE= 15.30 years in farmers’ condition. The stu&
shows that farmers are still using plows more than 20 years old (Table 18). If bad and
very bad plows are removed from the analysis because of the worn frogs, the life
expectancy of UCF plow, in farmers’ condition becomes FARMLIFE= 8 years. This
parameter is important in the process of evaluating the total number of plows for a given
time period needed by farmers in order to replace existing plows. The implement’s life
expectancy in farmers’ condirion (FAELVLIFE in years) cari also be used to estimate the
129
annual implement utilization fixed cost (IMPFCOS~T in local currency) when performing a
. .
farm budget analysis.
.
5.2.1.2. Seeder Super ECO
The seeder Super ECO requires more expertise from local black:smiths to service
because of the complexity of the seed distribution system. The mechanism housed in cast
iron r,equires little servicing. The manufacturer SISMAR urges farmers to verify the
quality and level of the anti-friction grease in the mechanism and to provide grease to gear
systefnn if necessary. The system is built strong enough to last the seeder’s lifetime.
Most of the other seeder”s working parts is serviceable by local blacksmiths. These
parts include the fùrrow opener, and the two f31 sweeps used to caver the seed with soil.
The weak point of the seeder resides in the plateau which is unservicealble by local
blacksmiths. The plateau made of brass is located inside the hopper and is part of the
distribution system. Tts role is to turn the seed plates held in place with a round brass nut,
a spring and two anti-slide short pins. Two major problems are noticed: the nut and the
spring are ofien lest or the pins are broken. An equivalent nut cari be found in rnarket
places located in nearest biggest town; but the pins, when broken, are difficult to replace.
In terms of breakdown frequencies, the seeder mainly contributed to the overall
breakdown and maintenan.ce frequencies of the shares and sweeps.
In relation to the level of utilization, seeders in working condition were ;found in
the area of study older than FARMLIFE= 11.9 1 years. The manufacturer3 specifications
indicat.e: a lifetime of klANLIFE= 10 years.
!
130
5.2.1.3. Toolbars and EMCOT ridger
There is little breakdown and maintenance on toolbars: The breakdowns identified
in the course of data analysis are classified in the same category of breakdown as the
attached working component. In ail cases, farmers in the area of study considered the
toolbars te, be identified with the working component they came with originally.
words, a multipurpose ARi& or SIXE 9 equipped with plow bottom is no more than a
piow, and the same for those equipped with ridgers. In the long run, the toolbars lose theil
original multipurpose characteristics. In generai, an ARARA toolbar is considered to be a
ridger as most of them are equipped with a ridger bottom and most of the SINE 9 tooibars
are considered to be cultivators or weeders, generaily equipped with 3-Canadian %Il
sweep.
The breakdown rate of ridger bottoms appeared to be closely related to tbc type of
wings they are equipped with. The E,MCOT ridger from Gambia tends to break at the
fiuture level between the wing and the share. Most of these ridgers need welding to repair
the damaged wing fixtures and account for 459’ o of the breakdown in agro-ecosystem 5.
On the average, the lifetime of EMCOT ridger is FARML.IFE= 10 47 years (the
manufacturer given lifetime MXNLIFE ivas not available). If bad and very bad implemenrs
are not çonsidered usable in the analysis, the lifetime decreases to F.AU!LIFE= 8 years.
The ridger bottom attached to the ARARA toolbar faces problems of wear on the
reversible point and breakage or loss of the wing nuts. The wing nuts are important for the
adjustment of their height in the process of monitoring the geometry of the ridges. This
type of failure contributes to 32’io of breakdowns mainly registered on ridgers in aura-
ecosystem 5. The ARARA toolbar with its attachment has a farm lifetime of almost
C
FARMLlFE= 14.60 years and 8 years when the bad and very bad toolbars are not
counted.
The sweeps used on the SINE 9 toolbar are also subject to the physical conditions
of the tields in relation to the presence of roots and stumps. The rate of sweeps wear is
similar to the plow share or ridger reversible point. The lifetime of toolbar SINE 9 in
farmers’ condition is around FARMLIFE= 6 years.
5.2.1.4. 0x-cart
O?c-carts are used year round by farmers. They mainly face problems of tires and
tubes due to the poor quality of the fat-m roads Flat tires on the road cari be changed in
the nearest town. They account for 1 2 ‘?O and 1 j”G of the total breakdowns respectiv,elv in
aoro-ecosvstem 3 and 5.
2
The prim,ary failure of the ox-cart is the break age of the tongue. u;hich generallv
occurs at least once in its Uifetime. Tangue breakage in combination with rhe tire problem
account for between 36 to 4004 of the breakdo\\vns. ‘The liferime in farmers’ condition is
close to FAR!KIFE- 14.67 vears in total and 10 vears when poor quality caris are net
counteci (Figure 9).
5.2.2 Farm equiprnent reliability
The reliability of farm equipment is closely linked to the existing maintenance
network and to the quality of technical services provided to farmers. The studv sho\\+s that
farmers rely on rhree categories of service providers: local blacksmiths. advanced shopr,
located in nearest big towns and farm equipment dealers.
132
,
10
---
---F-
Good
Average
rl
0
Bad
Very bad
State
Figure 9: Implments’ working conditions
60
50
10
Credit
Purchase
Trade
Gift
Acquisition mode
Figure 10: Acquisition mode
133
Until 1980, only govemmental extension services were involved in. making farm
‘_ .
implements available to farmers through the ONCAD and SONADIS network. There
were no private retailers involved in the process. One rnajor advantage of :his :system was
the quality control performed on implements and spare parts, but a11 the cost of
distribution had to be supported by the Govemment. The system did not work well as
fax-mers were net getting the spare parts on time.
The situation became worse alter 1980, when the distribution system administered
by extension agencies carne to a complete stop: no more dealers or spare parts retailers.
The otdy open alternative for the Govemment was to use the newly launched relgional
agricultural projects to support the distribution of implements and spare parts to farmers
located in their geographicai coverage. The PIDAC “Special Program” fmanced by
L?AID in 1979-80, was the only source of credit to farmers in the area of study. This new
iine of credit accounted for at least ZO?/O of the UCF plows, 64% of the seeders and a little
more than IJ”G of the CR-ca-ts (Figure 10). At the same time PIDAC tried to improve
repairing capacities of selected blacksmiths willing to undergo some training and to
purchase a set of tools at the end of the sessions, Currently, the local blacksmiths
represent the central piece of the network of the farm equipment servicing by carry& out
more than SOO’O of the repair and maintenance.
In this context, farmers do not rely completely on their farm implernents. Hand
tools are still read:y to be used to compensate for any implement’s failure.
In these conditions, the most often asked questions by animal traction practitioners
are: How long will it take ta fix a breakdown? How reliable is the technology to fat-mers”
1 3 4
There is no need to put stress on the importance of the reliability parameter when
farmers are in the process of adopting a new technology. The‘parameter is determinant in
- .
evaluating the performance of households using animal traction implements for trop
production. Timeliness is critical in the short period allowed for field operations.
The reliability of each farm implement or equipment (IMPLREL in percent) has
been evaluated and tabulated (Table 18). Two types of reliability coeffkients are
calculated, one on the yearly basis and the other on the implement lifetime basis. On the
implement lifetime basis, two reliability coefftcients are calculated, one on the
manufacturer!3 given life expectancy (IMPLREL man in percent) and the second on the reai
fart-n lifetime (IILPLRJZL rarm in percent). In any case, the reliability is given by folfowing
equations:
(42)
The real implement reliability per year IMPLREZL in percent is expressed by the
combined reliability of the two:
&CPLREL = ( 13fPLREL,,o~) * (LMPLRiq>)
A
I
?
3
-
-7
22;
-=-
-
CI
IC.
--I
N c, 5
X --C
: x
3 -
2
‘I :
136
Among the most used fat-m equipment, the ox-car-t appears to have the lowest
overah reliabihty lMPLREL= 58.2 l?/o. The observed breakdotins are frequent and
.
complex (pale welding, pneumatic problems) and require more time to repair in
comparison with cropping implements (66.07% for UCF plow, 68.75% For AKW,
70.20% for EMCOT ridger, 50.89?/0 for SINE 9 and 69.37% for Super ECO seeder).
According to the level of utihzation, breakdowns on newly purchased farm equipment
happen more ofien afier 2 years of use for cropping impiements and afier 3 to 5 years for
ox-car-t, evaluated from their first date of service.
The longer the idling or immobilization time for breakdowns, the lower the
reliability and therefore the higher the cost of utilization per year or per hectare
5.3. Draft animal management
A total of 89 oxen were counted in the study for the 40 househoids. They
represented more than 80% of the draft animals used in trop production. Al1 the oxen
belong to the Bos taurus breed commonly called Ndama. Small in size and
trypanotolerent, they are physically characterized as humpless and have an optimal or
mature weight MATURELW between 280 and 4 10 kg (Goe and McDowell, 1980) .A
mature weight value MATLRELW= 360 kg is often used for practical applications
The average number of oxen per farm is (2.23 2: 0.27) @ 90% confidence intena
The distribution is slightly different betkveen the two agro-ecosystems with an average ot
1.95 oxen per faml (CV= 47?/0) in zone 4 against 2.50 oxen per farm (CV- 4 106) in the
agro-ecosystem zone 5.
137
5.3.1. Avaiiability of draft animals
Oxen used by farmers were either locally withdrawn,f<om the farmer’s own herd or
purchased (50%) in the village from other farmers. Beside the withdrawal fiom the herd,
the other SO?/0 comprised exchanges against equivalent goods and gifis. I:t appears that the
availability of oxen, for draft purposes is not viewed by farmers as a significant major
constraint to adopting the technology
In the selection process, only male oxen were selected as social and cultural
standards of living prohibit the use of female animais for any physical activity
5.3.2. Training arnd working career
Training of draft animais for teamwork as well as training of farmers to better use
and adjiust farm implements was revealed to be a serious concern to farmers. The only
forma1 training sessions mentioned by very few farmers, occurred in the late 1970s. The
sessions took place at the Guerina Training Center located South of Bignona. The tenter
*as created in 1964 as a pilot unit in the Basse Casamance region, fmanced by the I,XDP
Ils ma.in objective was 10 train 32 farmers in the span of 9 months each year to acquire
skills in using draft animals and the variety of implements. The training \\V~S stopped in
1980 when the administrat.ion lacked financing to support the activities. Xowadays,
farmers are lefi to themselves to learn by doing Common farming practices were to train
animais at a much younger age. Most of the draft animals in the area ot‘study ivere trained
between the ages (AGETRQX in years) of 3 t.o 3 years. On some rare occasions. older
animais (5 to 6 years) were trained. The older the animal, the more dificult to be train&.
Staning one month prier to the beginning of the rainy season, the draft animals ivere gi~en
!
i
138
names, ad then taught basic moves through pulling and working as a team. During the
cropping season, animals are literally forced to work togetherqas a pair. It is only.during
their long working career ,that the pair of oxen begins to demonstrate more willingness to
work. A minimum number of 3 years (YRSEXP) is generally accepted in the area as
needed for a pair of oxen to respond correctly to commands. The average working
experience of’individual draft animal was found to be YRSEXP= 3.86 :? 0.79 years @
90% confidence inter-val. The longest working career was 9 years.
Farmers are facing more technical learning problems. An early study conducted 0‘
the Farming System Research team of ISRA’s Agricultural Research Center of Djibelor
(Basse Casamance) reported in comparing farmers with and without animal traction chat it
took 5 to 6 years of experience for farmers to master the technology. The mastering is
necessary before steady yield improvements becomes apparent. This includes implement
hitching and adjustments, adjustment of line of draft, keeping plowshare Sharp and
moldboard in good scouring conditions, ox-team driving, resources allocation and choice
of appropriate cropping techniques.
5.3.3. Stability
The key factor to a successtùl growing season is to maintain draA: animais in good
working condition. Stability of the technology from the draft animal perspective is what
reliability is to farm implements.
A close look at the data shows that 13?/0 of the farmers had only one draft animal
I
(Figure 11). They needed to borrow another animal to constitute a pair in order to car-
out their tïeld activities. This situation is risky, especially if the transaction takes place at
--- --
139
,
---- ----. ----_--.-.-___---.-..---~-..-._
-.------_-__--.
70 -------_--
- .____- -_---- -.-_^ -.--~-. _----. --.
.
60 -r
250 -
-
-
$40 -
d?
(r0E 30 -
u
-
b
L 20 -
10 -
---m
2
3
1
O -
--.,.-v-1.-
4
a
Number of draft animais
Figure 11: Draft animals’ distribution (zones 4 qnd 5)
140
the beginning of the rainy season. Areliability factor on the draft animais’ availabiiity
(ANIKEL in decimal) is therefore attached to this situation. Using Equation 20, the value
‘. .
of AIVIREL is equal to:
ANIREL= 0.87
The next group of farmers (66%) disposed of one pair of oxen. This repsesents
also a risky situation. If one animal happens to be unavailable because of in-jury, disease.
death, thefi or some other reason, then the production objectives Will be seriously
compromised.
For these two groups of farmers (79%) the technology is not stable, associating a
large amount of risk in the res’ource allocation process at the farm levei. The coefficient ot
stability using Equation 19 is equal to:
STABCOEF= 0.21
The third group of fat-mers (3?,$) showed more stability with 3 draft animais The
third animal was used to replace any unavailable one. The animal must be trained to work
on both the left and right side. This aspect was the most difficult for farmers to achieT+.e, as
working draft animais need to be familiar to one another for better and sustained
performance.
The fourth group (15%) was more concerned with continuity than anything else.
Replacement of the actual working pair of draft animals was a concem and planned well
ahead of time. With two pairs of oxen (4 draft animais), the second pair was trained to
111
replace the first aging one. The average ages of rhe first and second pair of oxen were
given. as follow (Table 19):
.; ’
Agro-ecosystem 4
.lst pair of oxen
---A---
Right side: 6.79 years with a CV= 3 1%
Left side : 5.80 years with a CV- 33%
2nd pair of oxen
-
-
Right side. 4.33 years with a CV= 44?/0
Left side : 3 .OO years with a CV== O?/o
Agro-ecosystem 5:
1st pair of o‘ten
Right side: 8.15 years with a CV== 309’0
Left side : 8.16 Yeats with a CV= 3796
2nd pair of oxen
Right side: 4.20 years with a CV= 51s’;
Left side 4.40 years with a CV= 4796
In any case, the second pair was not expected to yield the same productiv.ity as the
first pair. Switching to the second pair would s,low down the working speed. cau.sing both
fieid capacity and quality of work to decrease.
142
Table 19: Average age of 1st and 2nd pair of oxen
The fifth group of farmers (3%) had 5 draft animais to make up two pairs and one
ready for replacement. They planned for both the future, by training younger animais and
were ready for the replacement of any animal unable to work in the first pa.ir.
The stability analysis shows that for the majority of the farmers the technclogy is
unstable. In many cases they depend on their neighbors to borrow missing components or
on community work organized at the village level to achieve their production objectives. lt
was not the focus of this study to analyze the alternatives but previous studies conducted
in the same area have showed that in most cases, they still rely heavily on 4:raditional hand
toois to complete the unfinished work of the draft animals (Fall, 1985; Ndiame, 1986).
This fact demonstrates the importance of the presence workers at the farm level for animal
traction users until stability or reliability is improved.
5.4. Land and labor resources
5.41. Land availability
The average farm size in the area of study in terms of land area plowed by draft
animais was FMSIZE= 4.03 5 0.43 ha @ 9006 confidence inter-val (CI) (Figure 12). The
143
.
30 .--_--.-_
----_..~-----
25 -
20 _-
15
10
5
0
1
2
3
4
5
6
7
8
Fann size (ha)
Figure 12: Farm size (ha) distribution
--------_
--. --.-._____- _~-~ - ___.__ - ...-___.^ --.- --..--..- --.. ~- -- . . . .._ -..-
50 -I---~-.--_I--_----.--__-_-~-----.-_ ~---~- _-
40
1 to 4
5 to 9
-110to-----141
0
15 to 19
20 to 24
Nmnber of farm workars
Figure 13: Farm labor distribution
144
average farm size is slightly different from agro-ecosystem 4 with FMSIZE= 4.26 ha
(CV= 36%) to agro-ecosystem S with FMSIZE= 3.8 1 ha (CV= 44%). The potential for
- .
land extension is limited by the presence of valleys and lowland areas generafly decoted to
rice cultivation whenever possible.
5.4.2. Farm labor
Most fat-mers were equipped with tillage implements (57041 of total farm equipmenr
surveyed). The lack of post-tillage implements has relegated draft animals to a low level of
utilization during the growing season.
The average number of fat-m workers NFWKERS present at the fa.rm level during
this study was NFWK.ERS= 10.53 + 1.17 @ 90% probability level (Figure 13) The
difference between the two agro-ecosystem zones were not significant with an average of
NFWKERS= 11.60 fat-m workers (CV= 360/0) for zone 4 and NFWKERS- 9.45 farm
workers (CV= 4104) for zone 5, The availability of the farm workers throughout the
growiny season is crucial to meet the production objectives. The temporary absence of
active family members was pointed out to be a constraint at weeding time
5.5. Adoption dynamics
‘ .
As expected during the implementation stage of many agricultural development
projects around the world, three phases have been identifïed in the process of animal
traction adoption in the Basse Casamance region (Figure 14):
/* :
*
145
m
*
?
0 ! m
? ? ? ?
? ? ? ? ? ?
? ? ? ? ? ? ????? ? - __.-- ----_-.~.-
7??
75
80
85
90
95
Year
-~--~
_-
.~---.--
Figure 14: First year in animal traction
-. ----~.
-.-- --_- .-..
100
.
. .
T
L
rn 80
E
f
. .
c
kW
= 60 -
5
,
. .
0
h
/
. .
0 40
i
j-
.-
I "
z
+
I
3
. .
E 20
1
E
.
6
. . . .
.
0 -..-a222--- __ - --_- . . .._. ._. _.-. -._--- . ..--. _
0 5 10
15
20
25
Number of ‘eut-s of experience (years)
-- -. __--.
Figure 15: Number of years of experience
Early adopters:
This phase extended from the end of the 1960s to 1974.. Animal traction adopters
were made of farmers who acçepted the challenge of securing agricultural production
through diversification and extension of upland crops, mainly on groundnuts (cash trop).
They represented no more than 1 5 ?/o of the farmers in the area of study. F’armers located in
agro-ecosystem zone 4 were the first to adopt the new technology.
Adopters and innovators
The better performance of households using the new technology through the
increase of gross ‘cash trop production generated a lot of interest from manual farmers.
The adoption rate was the highest between 1976 and 1980 when no less chan 509.0 of the
farmers became involved. This cari be explained by two decisive factors. First, the
incentives were put on groundnut production at the governmental level through the
application of the National Agricultural Policy.
The policy was characterized by an easy access to credit and organized markets for this
trop. The quasi-total groundnut production was used to supply the national oil indust-,
Second, the progressive encroachment of drought in the area forced farmers to look for
other alternatives to better secure production at the household level. Rice production \\~as
reduced by the decrease in the amount of rainfall (30% to ~0% in eariy 1980s) to rhe point
that rice was no longer driving the farmers’ production system in the area of stud?
147
a d o
Late p t e r s
Tbe decrease in the rate of adoption starred in the ear& 1980s and went tiom .~-~“o
to an insigificant lesel of 2.0 in the early 1990s The National Agricultural Poiicy
administered through the Lo\\:ernmental agexies was ended in 1980 at the national 1~x4
for many reasons The main reason was rhe inability of farmers to psy back the medium
rerm (5 years) credit loans on fartn implements afier the Ion3 lasting droug:ht of 197S-
1982. The farmers’ primnrv source to acquire farm equipment was put to an end ar.d as
was pointed out by ?Jdiame i 19%) it is unlikely that a similar program w41 be initiated in
the filture. Following this,, che credit systems fosrered by the Government and polie‘
makers to finance agriculture Lvere merged into the collaboration with private banks anci
newiy implementing rural development projects Along with other reasons. the conditions
for eligibiiity ofthe Kxmers to be par-~ ofthe newiy implemented credit systems iLere net
eupect~rd :o increase t.he nunber ofnew animal tract::on adopters.
The dynamics ofadoption appeared to be i.ery sompiex in regards to thr: \\xriabi?s
in\\,ol\\.ed As the study sho\\\\;ed. almost 70°,ù ofrhe farm implements ac:uaII~: used lx
farme:-s lvere financed through a credit system (Fipu.re 110). From 1930. mcst oìrhe
farmerr, who benetïted. fro:m the neivlv formed credit lines. tvere already animal rraction
users The rate of new adopters slo\\ved doivn for différent reasons that included :he C~G
of transactions invoived:
- Awareness ofrhe credit systern: farmers with a past credit systcm’s
experience L\\-ere able to reduce the cost of transaction by getting correct information El+
access tu information helped thym Lvork thelr l.vav through the svstem. Besides having
148
easy access to information, they had scrong connections for rapid administrative paper
work and were also more aware of ways to avoid sanctions .or,penalties for no annuity
.
payment
- Members of farmers’ production group or fax-mers’ organization: There is
more and more tendency for farmers’ organization to guarantee the credibility of their
members. They carry out preliminary negotiations on behalf @fat-mers to speed up the
process of acquisition. Farmers’ organizations administer have their own sanctions to
default repayment in accordance with the organization byIaws. In the area of study and
since 1988, the CADEF farmers’ organization guaranteed 5 1 % of the implements acquired
by fat-mers.
- Conditions of eligibility: the involvement of private banks increased the
cost of transaction. Interest went up and farmers were asked to constitute a down
paument in order to be eligible for the credit. The National Agricultural Credit Bank of
Senegal (CNCAS), the main actor in the credit system, was asking for a 20% down
payment and 14% of interest on the capital. With these conditions, farrners generally
believed the cost of money was just too high.
- Non existence of second hand implement markets: Onl:y 26% of the
implements used by farmers were purchased in cash (Figure 10). These farm implements
were mainly bought from out of state markets, across the Gambian border (16%). These
implements imported by Gambian dealers were mainly ridgers fabricated by the English
8.
i
-
. .---
*RINI-,-’
.
149
manufacturer EhKO’T Local markets, especially those located in North and South
Central regions of Senegal accounted for nearly 10% of thé $urchases. Xt appears that
there is a general tendency for farmers to keep their implements on the farm (19%j until
they were completely lworn out, out of service and finally discarded (Figure 9). The ver-y
few occasions whereby farmers got rid of their implements were only when they traded
them off against other equivalent goods (504) or gave them away as gifts to relatives and
fi-iends ( 1%).
- Learning period: It required a long time for fat-mers to master the new
technology (5 to 6 years). In agro-ecosystem 4 and 5, 10 to 15% of farmers were
classified inexperienced (Figure 15). In these conditions, a new adopter experienced
mixed results 0fusin.g animal traction. Animal cultivation cari be ver-y fi-ustrating and using
farm implements cari appear to be a surn of tough skills to master. Lack of working skills
cari prevent fat-mers from performing proper adjustments. and identifving the signs of
inefficiency.
Farmers located in the area of study do not have a long history of animai traction
The average number of years of experience was (16.43 :: 1.5 1) years @ 90?/0 level of
probability with a maximum of 25 years.
5.6. Summary
This broad overview of animal traction in the Rasse Casamance region shows that
many factors were involved in the implementation of the technology.
150
The quality of the implements made available to farmers is of major importance in
relation to its life expectancy. Quality control is a must. The design also is ,a.s important.
The simpler the implement, the better as farmers are not generally skillful in the handling
and management of complex toolbars. The analysis shows that toolbars were generally
,.
identified with the working component that they were originally equipped with. The
introduction of toolbars must go along with specific training to show their flexibility and
to demonstrate how they cari improve the overall farm working capacity.
The quality of maintenance and servicing is important to the viability of the
‘” ,
technology. The skills and expertise of local blacksmiths must be improved through
!
i
training and better tool supply. Increasing reliability of farm implements is hkely to
decrease production cost in the short r-un.
There is a need to improve performance by facilitating the adoption of
complementary implements. This Will help to avoid the side effects of the shifting of
bottleneck to fil1 the labor deficits. TO meet this need it is crucial to reduce tlne learning
period by better training in order to achieve rapid returns on investments that cari be
reinvested.
In order to tie these diRerent aspects together, it is important to carry out an
analysis of actual and potential use of the technology. A correct evaluation of the impact
of the technology Will help farmers to have a better understanding of the effects of animal
traction on the achievement of their production objectives. It Will also help differentiate
between yield increases and land extension effects. A good starting point is the evaluation
of the working capacity of draft animals in the farmer’s condition.
h
Chapter 6
WORKING CAPACITY OF DR4FT ANIMALS
Two pairs of oxen per village and one pair on-station for control were used. A
total number of 9 pai,rs were classified into 3 traction groups according t:o the pair? total
liveweight LW:
Group 1 : 6 0 0 t o 7 0 0 k g
Group 2: 5 0 0 t o 6 0 0 k g
Group 3: JO0 t o 5 0 0 k g
The 18 draft animals were monitored separately throughout the growi.ng season:
workllng schedule (type and time spent), liveweight follow up, disease, and feeding system
6.1. Ebironmental conditions
In the Basse Casamance region, farmers are urged to carry out their fïeld
operations as early in the morning as possible to avoid heat stress and to rest rhe animais
according to the level of intensity of the fieldwork. The heavier the work., the shorter rhe
time alliocated to the task and the longer the rest.
151
152
Accorchg to the weather data collected during the study, it shows that drafi oxen
in the Basse Casamance region faced the weather related fatigue between midJuly and the
end of September when the average minimum relative humidity RHmin approached 70%
(Figure 16 and 17):
- From onset of rain to July 14
. Minimum Temperature Tmin:
24.21 “C
Maximum Temperature T,,,:
33.73 “c
: Minimum Relative Humidity RHmin: 52.56
. Maximum Relative Humidity FUI,,,: 94.77
- From July 15 to end of September
. Minimum Temperature Tmin:
23.88 “C
Maximum Temperature Tmar:
31.04 “c
: Minimum Relative Humidity FELin: 67.82
Maximum Relative Humidity REL: 99.49
The difference in RE%,, shows that the oxen have a narrow windcsw of working
days before the draft animals entered the uncomfortable working period of the year. This
situation dictates a working strategy based on the number of animal working hours
ANWKGHR per day (in hrs/day) allocated to field operations: more working hours before
mid-July and less working heurs for the remaining of the growing season.
It must be noted that both minimum temperature TA~ in “C and maximum
temperature TmX in “C were on the decreasing phase from the onset of the rainy season
and reach the minima towards the end of the year. Minimum temperature ‘I’ti was
decreasing from 27.00 “C and maximum temperature T,, from 43.80 “C. Records show
that the highest temperature has always been recorded during the month o-fbfay and the
lowest in December-January. The effects of temperature alone need more imvestigation.
i
/
i
1
15 ~~~li:i~“.‘:
M
J
J
A
s
0
N
Months
,.-~_- _--,
/
1
I_ Tmin -.Tmax
---~
[-__- - - - - - - - -~-----
__-_..- -- -.-.-..
Figure 16: Temperatures (min and m.ax) at Ziguinchor (1996)
-~---~-------.__
__- .--.-.-.--_--._------ ~_~--_-_-- -.- -.-
-- Rh min -- Rh max
Figure 17: Relative humidity (min and max) at Ziguinchor (1996)
1 5 4
The main objective of the recornmendations on working strategies is to optimize
the energy available from the draft animais to yield a better’&ciency. At this stage of
investigation it is important to evaluate the level of animal energy output ANERGY in
h43’day that fat-mers cari Count on, especially at the beginning of the rainy season when it is
needed the most.
6.2. Maximum performance
The maximum performance test was conducted on-farm during morning hours.
The test took place on already cleaned fields and ready to be plowed for the cropping
season. The time constraint was such that only two villages out of four could be visited
per day.
The pair of oxen from each traction group used the same loaded sledge for each
puiling triai. The pulling distance or distance traveled DISTRAV in m (100 to 200 m) was
sometimes adjusted to fit the geometry of the fieids. Each tria1 lasted approximately 2
hours starting with a warming up of the pair of oxen yoked to walk together without load
for 10 to 15 min. The head yoke type was used for the trials. The characteristics of the
draft animal’s performance are given in Table 20 and 21.
Different adjustments by triai and error were performed at the sledge hitching point
to reach an acceptable line of draft. Two variables were recorded during each repetition:
the pulling force (PF in N) and the time (PTIME in s) to travel the given distance
(DISTRAV in m). A regression analysis was t-un to mode1 the power output of each pair
(ANPOWER in J/s) in relation to the required pulling force PF in N to move the loaded
sledge. The same analysis was carried out on the average performances.
-- Y’--’
1
I
155
6.2.1 Maximum performance
i
.
6.2.1.1. Optimum point
‘_
-- Optimum power (ANPOWEK,t’)
’
!
The optimum poser ANPOWER,pt in J/s is used to compare the performance of
each pair of oxen. The power is calculated by multiplying the PF (N) and ANSPEED
(~/S:I. A one-way analysis of variante is conducted with each of the factors contributing to
Table 20: Performance of team of oxen
/LIV Speed Optimum Max PF
(mis) p o w e r
PJ)
(JN
1267.49 353 1.60
-
1796.70 34, 0.00
-
-
-
0.90
1241.27 2256.30
1282,.29 3256.00
1.09
1603.94 2950.00
-
-
933.64 1962.00
817 68 2158.20
1377.81 2530.00
-
-
( 1) P stands for pair of oxen
(2) The pair of oxen showed unwillingness to pull
Table 21: Average performance of team of oxen
l
II
3 ( 4 4 8 . 3 3 <I 1058.34/--J 0.98--j 1 0 4 3 . 0 4 ( 2 2 1 6 . 7 3 11
I
( )The pair of oxen P2 of Group I was not included in the average
1 5 6
the puwer: the pull force PF in N and the draft animais qketl ANSPEED in m/s (Table 2f.
and 21).
The analysis of variante on the optimum power ANPOWERopt Hn J/s shows that
the difference among the 3 traction groups is not ver-y significant (Table 22). The
arithmetic difference arnong the groups cari help evaluate the energy balance involved in
order to determine the exact amount of feed required for a given group.
p,
Table 22: ANOVA of the Optimum Power (3s)
l
SOURCE
DE:
S S
M S
F
-
w
P
-
Group
2
322285 161142 2.05
0.224
E R R O R
5
393748 78750
TOTAL
7
716033
1
,.* l
INDIVIDUAL 95% CI’S FOR MEAN
BASED ON POOLED STDEV
I
l
LEVEL N
MEAN
STDEV ____- + ________- + ---_---__ + --___-___ +-
Group 1 2
1532.1
374.2
( -m-----m--- * --“v----“m1- 1
Group 2 3
1375.8
198.6 -..e----em * -- -w--mve-
(
1
Group 3 3
1043.0
295.7 (-------- * ---- --..-)
----* + ..--* -----+ ---- ‘. --_- + -------__ f-
lr
POOLED STDEV = 280.6 800 1200 1600 2000
I
- Optimum Pulling Force (PFQ,~)
The arithmetic difference in optimum power ANPOWER,,t is mainly due to the
significant difference found between the pulling force PF,,t of the different traction groups
at the optimum point. The 95% confidence level analysis shows that difference in PF,,,t is
!J
157
signifcant between traction group I/group 2 and traction group 3. The diflerence is not
significant between traction gsoup 1 and 2 (Table 2;3).
‘_
.
,_ ,
The ratio of pulling force to liveweight (PF/LW)o,t at the optimum point is the
same for the 3 traction groups. The difference is not significant (Fisher-F= 0.92 and p=
0.45) and the average for the 3 groups is (PF/LWJ,rc= 0.24 with a CV= 15% with pulled
Table 23: ANOVA of PF in N at Optimum point
- - -
-
-
-
SOURCE
D F
S S
MC;
F
P
Group
2
32 1668
160834
8.19
0.019
E R R O R
6
117780
1.9630
TOTAL
8
439449
----
-
-
-
INDIVIDUAL 95% CI’S FORMEAN
BASEDONPOOLEDSTDEV
LEVEL
N
M[Em
STDEV w___.._ + _____..___ -/- __-______ + _________ +
Group 1
3
1489.7
113.0
( -----.- * ------- 1
Group 2
3
1419.9
47.1
------- * ------ 1
Group 3
3
11058.3
209.6
POOLED STDEV := 140.1
1000
1250
500
1750
standard deviation Stdev= 0.036. Monnier (1965) and Nourrissat (1965) repot-ted in
different field triais that the Ndama cattle could pull continuously 10% to 14% of their
body weight for a long period of time. Field operations requiring pulling 20% or more of
the body weight were considered to be from medium to heavy work and could not be
sustained for a long time. According to Goe and McDoweil(1980), citing Swamy Rao
I[ 1964), an improved harness yielded a pull to weight ratio of 24%.
158
- Draft animais’ speed ANSPEEDopt
The speed ANSPEEDopt in m/sec is the most important factor in the power
.. .
estimation. The significant difference in the pulling force PF is rapidly buffered by the level
of low speed of the pairs of oxen across the groups. The analysis of variante shows that
the speed of travel ANSPEEI& is almost the same for the traction groups (Table 24)
Table 24: ANOVA of the walking speed (misec) (Optimum point)
SOURCE
D F
SS
MS
F
P
Croup
2
0.0018
0.0009
0.07
0.935
ERROR
5
0.0666
0.0133
TOTAL
7
0.0684
INDIVIDUAL 95% CI’S FOR MEAN
BASED ON POOLED STDEV
L E V E L N
ME)Jj
STDEV
mm__ + ______.._m + _____m___ + ____ _ me_.. -+-es
Group 1
2 1.0050 0.1385
(----_ ----* --*..--------)
Group 2
3
0.9667
0.1069
(__-______-- * ___--I_____-)
Group 3
3 0.9767 0.1041
( c~~~~~~~~-*-- *-------- )
---- + -------mw f.. -------- + --------- j--a,
POOLED STDEV = 0.1154
0.84 0.96 1.08
1..20
Considering the level of pulling force PF required for fteld activities, it would be a
slight improvement to accept the level of significance of the optimum output. This cari
help to better allocate the length of time to perform field operations and also, to evaluate
the amount of feed required for draft oxen to carry out specific fïeld activities
-cm
-. ~-
v--
-,,yY,u-II~--’
..***.-YY(I*YUYI
159
6,.2.1.2. Maximum Pull force RF,,,
The maximum pulling force PF,, in N represerits the va%ue for whi.ch the
. .
pair of oxen had difficulties in walking while pulling. It corresponds to the animais’ speed
ANSE’EED almost equal to 0 m/sec. In other words, PIF,, represents a continuous
maximum pulling effort in comparison with the instantaneous maximum pull
The analysis of variante shows that the difference is significant between the groups
at p= 0.038 (Table 25). The 95% confidence interval separates the means and shows that
the maximum pull PF,, of group 1 is significantly different from group 3.
Table 25: ANOVA of maximum pull force (N)
- - -
-
-
---~
SOURCE
DF
S S
A?s
F
P
v-------
---.-
Group
2
1892816
9416308 6.76 0.0
<
ERROR
7 0 0 2 2 3 14004,5
TOTAL
;
259304 1
---...---p_
- - - ~--
INDIVIDUAL 95% CI’S FOR MEAN
BASED ON POOLED STDE%
.-----
_
-
- - - - -
LEV -_ N
MËAN S T D E V
-------+----e-.--e + --e---w-- + e---e--<.-
Group 1 2 3465.8
93 1 (
--L---m--
*
-------.-)
Group 2 3 2820.7
512.4 m------ * o------ )
Group 3 3 2216.1
288 fi (__--___ 5; ____.___)
-.. ----- -i- -....--4-.-- + --------- + ---------
POOLED STDEV := 374.2
2 100 2800 3500
The maximum pull force PF,, cari be translated as the capacity of draft oxen to
overcome the peak draft errcountered in the fields while continuousiy working. The
average ratio of pull force to liveweight (PF/LW),, at this point is 0.5 1 with CV= 14?$
and Fisher-F= 0.08 at p= 0.923. At this stage of the triai, the pairs of oxen were asked to
give the maximum effort which yielded this high level ‘of pull ratio PFILW. The ratio
I
1
1
II
/
160
corresponds to pulling force PF,, tif almost half of the body liveweight LW. In any case.
it is not expected in the reai working situations that draft oxen be submitted to this Ievel of
. .
pull on a regular basis.
Values found in the literature shows that the tria1 results for the Ndama breed are
in the expected range. Goe and McDowell(I980) reported a ratio of 0.5 1 or 5 196 during
tests conducted in India with different pairs of oxen.
6.2.2. Characteristics of power curves
A regression analysis was carried out to mode1 the power output ANPOWER in
J/s of each pair ofoxen. The results are presented in the Appendices. A regression analysis
I
WI
l
was also performed on the average performance of each traction group. The different
models are used in the expert system computer program to help evaluate and refine the
energy balance at the farm level.
6.2.2.1. Power output mode1
The analysis was carried out on two determinant parameters directly involved in
the performance of the cropping system: the power output ANPOWER in. J/s of the draft
animals (pair of oxen) and the level of speed ANSPEED in rn/s at which the task was
performed. Firstly, the parameters were estimated for the general linear regression,
secondly an analysis of variante was performed, and thirdly, a look at the cor-relation
coefficient R among variables to help explain the relationshïp. The RZ is also important as
it helps to determine the amount of variation explained by the regression lime. The best tit
was obtained with the quadratic form of PF using the variables (PF/lOOO) and (PF/lOOO)’
161
A general comment: most of variation accounted in the linear regression was
.
mai:nly due to the border values at:
PF = 0 and PF = PF,,,
At PF == 0, the pair of oxen was walked without load to caver the path. Farmers
had different feeling of walking with the draft animals in terms of how fast ta perform the
task. .At the other end ofthe trial, pulling the maximum load depended on the willingness
of the pair of oxen to do the job. It did not matter h.ow well the oxen were trained. the
greater the amount of draft required, the less the willingness of the oxen. to pull. As a
consequence, the pulling of the maximum load was put to an end to manage the
temperament and excitement of the animals. In such. cases, the value recorded for the
maximum Pull force PF,,,, would be underestimated. The analysis is presented by tracrion
group’
- Traction Group 1: 600 to 700 kg of :LW
The average performance of the Traction Croup 1 is summarized in Table 26 The
regression equation is:
1 6 2
“/
Table 26:‘Average performance of Traction Group 1
PF/LW,
0.00
0.04
0.08
0.14
0.16
0.19
0.21
0.23
0.24
0.28
0.31
0.35
0.37
042
0.52
. 1
.
. 1
Table 27: Parameter estimates (Group 1)
Predictor
Coef
Stdev
t-ratio
P
-
-
No constant
(PF/I 000)
44.350
2.120
20.92
0.000
(PF/ 1000)’
-12.6875
0.8319
-15.25
0.000
-~
s = 4.454
R-squared = 0.88
Table 28: Analysis of Variante (Group 1)
- -
SOURCE
DF
SS
MS
F
P - -
Regression
2
13687.5
4843.8
344.96
0.000
Errer
13
257.9
19.8
Total
15
13945.4
--
163
i
‘2. ,,
?
m œ
?
N
.
m
?
?
.,.m... .,.
T
I
l
0 &--.-------------~-,-,,k----+-
1
0
5 0 0
1 0 0 0
1 5 0 0
2000
2500
3000
3500
Puli Force (N)
_._--~-_-_l_-.--.-----
---
Figure 18a: Power curve of Traction Group 1
1.4 T------
r
~\\
1.2 .i
’
. . .
?
?
m
? ? ? ??
1 .-
‘\\ . a
j.
z
,-L
2 0.8 - -
/
\\
?
?
??
?
3 0.6 --
OF
cnt
*
?
?? ? ?
? ? ? ? ?
?
m
Y
?
/
0.2 -
c
1
\\
I
-\\
0 2 --7--------l-‘
-.-- _...*_ -.*--_
0 0.1 0.2
0.3 0.4
0.5
0.6
Pull Force iliveweight
-- Fitted line
Figure 18h: Walking speed of Traction Group 1
164
,* \\
The fitted linear regression equation R2 explains 88% of the total variation, which
I
,
is ver-y significant (Table 27 and 28). Most of variation was mainly introduced by the,
border values at PF = 0 and PF = PF,, (Figure 18a). Accord& to this model, the
unusual observation at PF,, suggests that the pairs of oxen in the Group 1 çould have
pulled more that was recorded. This influence is not negligibl’e in the amount of variation
translated into the value of the R2.
In the region of the optimum power ANPOWER+ the amount of pull applied
PFCJpt = 1750 N started to become unsustainable and to induce the unwillingness to pull
mentioned earlier. This fact indicates the needs for more studies to improve the harnessing
systems.
- Traction Group 2: 500 to 600 kg of LW
The average performance of the Traction Group 2 is given in Table 29:
-1
The regression equation is:
I..,
1
I
(46)
The fitted regression equation shows a good coefficient of detetmination R’= 0.93
which is very significant (Table 30). However, the analysis of the data shows two
singularities. Firstly, according to the model, the amount of maximum pull PFmas= 20 13.9
N developed by the pair of oxen was also underestimated (Figure 19a). Once again, the
unwillingness to walk: at this level of pulling induced part of the variation. Secondly, the
value of the maximum pull force PFmx, I‘nfluenced the trend of the curve.
165
Table 29: Average performance of Traction Group 2
-
-
-
-
I?F (IN)
Speed
PF/lOOO
(PFA 000)2 P o w e r :. (Power)‘” PFLW
- - -
(mkec)
-
-
(N) CN)
(J/s:)
(Jhec)
-
-
-
0 . 0 0
lu32 0.00
0 . 0 0
0 . 0 0
0 . 0 0
0 . 0 0
8 6 8 . 9 0
1.25 0.87
0 . 7 5
1086.13
3 2 . 9 6
0.16
9 2 9 . 7 0
1.04 0.93
0 . 8 6
-966.89
3 1 . 0 9
0 . 1 7
1209.90
1.11 1.21
1.46
13442.99
36.45
0 . 2 2
1285.l.O
0.89 1.29
1.65
1143.74
3 3 . 8 2
0.23
1409.1.0
0.91 1.41
1.99
1282.128
35.81
0.26
1471 .io
0.99 1.47
2 . 1 7
1456.79
3 8 . 1 7
0.:27
1917.410
0.76 1.92
3 . 6 8
1457.22
3 8 . 1 7
0.35
1970.10
0.75
1.
.97
3 . 8 8
1477.57
3 8 . 4 4
0 . 3 6
2013.90
0.35 2.01
4.06
7 0 4 . 8 6
2 6 . 5 5
0 . 3 7
2412.10
0.29 2.41
5 . 8 2
699.51
2 6 . 4 5
0.44
2452.50
0.42 2.45
6.01
1030.05
32.09
0 . 4 5
3103.00
0.00 3.10
9.63
0 . 0 0
0 . 0 0
0 . 5 6
CI==
-
-
-
Table 30: Parameter estimates (Group 2)
l’redictor
Coef
Stdev
t-ratio
P
Noconstant
(I)F/ 1000)
48.167
2 . 0 0 9
2 3 . 9 7
0 . 0 0 0
(PF/1~OOO)2
- 1 5 . 3 3 7
0 . 8 7 2
-17.58
0 . 0 0 0
s = 3.604
R-squared := 0.93
Talble 31: Analysis of Variante (Group 2)
---y-
-
-
-
SOURCE
D F
S S
MS- F
P
-
--
-
-
-
-
Regression
.2
- 12505.1
6252.6
481.32
0 . 0 0 0
Error
11
142.9
13.0
Total
1 3
12648.0
- - - -
-
-
-
-
166
e----p -. -_.
-~- --_--
50 -j
--.
I_-
/
.- .
/
t!
I
?
?
i
?? ?20 4I
&
i
10 1...
l
l
I
/
oh-
:
/-+--+-m--i
0
500
1000
1500
2000
2500
3000
3500
Pull Force
--
--.
Figure 19a: Power curve of Traction Group 2
1.6 -
-
- -
-
,I
----\\.
1.4 fy-,...
.._
.__..__
1.2 7..
:r-,.,
!!
:
I.I
t
.
I
oc
/--
--,
-_
0
0.1
0.2
0.3
0.4
0.5
0.6
Pull Forceaiveweight
Figure 19b: Walking speed of Traction Group 2
.--
.‘mwa-----
167
- Traction Group 3: 400 to 500 kg of LW
The average performance ofthe Traction Group 3 is summarized in Table 32.
The regression equation is:
P F
P F
(;4NPOR’ER /+== 56.1 I * ( -/ - 75.63 * I- Y
1000
1000
(47)
The parameter estimates and the analysis of variante show a very goodl fit with an
R2 =: 0.91 (Table 33 and 34:). The draft animals had a tendency to walk fast in the area of
lower required PF a.t the beginning of the trial. The unwillingness to pull appeared in the
same region of the graph and. mostly beyond the optimum power point (Figure 20a).
Most of the variation was introduced once again by the maximum pull force PFmaX,
underestimated according to the model.
6.2.2.2. Walking speed (ANSPEED in m/s;)
The walking speed ANSPEED in m/s while performing a task is the second
important parameter towards the evaluation of the power output ANPOWER. :Regression
analysis has also been performed to give more insight in order to explain the accounted
variation in the power output models. The analysis has been conducted for each; Tra.ction
Group. The best fit was obtained by plotting the speed ANSPEED as the dependent
variable and the ratio of pull force PF over the average liveweight of the group PF/L,W
The speed plotted against the PFLW yielded a better cor-relation than the speed plotted
against pu.ill force PF.
168
Table 32: Average jkformauce for Traction Group 3
PFfLW
“--
0.00
0.04
0 . 0 8
23.92 0.12
0 . 1 6
0.21
0 . 2 8
0.31
0.33
0 . 3 8
0 . 5 0
Table 33: Parameter estimates (C;roup 3)
f
-
_--
Predictor
Coef
Stdev
t-ratio
P
-
Noconstant
(PF/ 1000)
56.111
3 . 1 0 6
18.06
0 . 0 0 0
(PF/ 1 ooo)*
- 2 5 . 6 3 2
1.816
-14.11
0.000
-s= 3.417
R-squared = 0.91
Table 34: Analysis of Variante (Group 3)
-
_--
SOURCE
DF
S S
M S
F
P
----
-
Regression
2 -
5 3 6 2 . 4
268 1.2
229.66
0 . 0 0 0
Errer
9
105.1
11.7
Total
11
5 4 6 7 . 5
169
50
7
‘_
‘.
, ’
oa----------
------ --..--_m
i
0
500
1000
1500
2000
2500
Pull Force (N)
------p_
- - - .~.-
--v.---
Figure 20a: P’ower curve of Traction Group 3
-~--
-.
1.6 --p
---.--_--~---_~.--
i
,
1.2 f. -..hw,
û
j
‘-m .._.
1
‘.. m
!
,$
7’ .’ ... ..
. . . .
+
-+
E0.g ;.
‘.\\
,,
FI
I
8
00.6
_-.
“. .*.
rn
i
0.4
1.
m -.
4
1,
/
‘.._
Os2
7
.’
+. . .-.
i
‘..
i
t-J
GIC---
--.4-.-
-1
0
0.1
0.2
0.3
0.4
0.5
Pull ForceILiveweight
- Fitted line
/
- - - - I
- - _ l _
- -
Figure 20b: Wialking speed of Traction Group 3
170
Generahy, the fitted straight lines with their respective negative slopes show that
the speed ANSREED of the pair of oxen was decreasing with higher required pull force.
The higher the pull force PF, the lower the speed level. This sceéd reached the value of
zero at the maximum pull FF,,.
- Traction Group 1
The regression equation is:
PF
ANSPEED =1.40-2.53*(Lw)
(48)
The fitted straight line to mode1 the speed is highly significant as the correlation
coefficient R = 0.94 (with R2= 89.2%) is very close to 1 (Table 35 and 36). The
relationship between the pull force PF and the speed ANSPEED
is ver-y strong and in opposite direction (Figure lgb). The straight line accounts for 89%
of the variation. One observation at PF = 1618.7 seems to be high and introduces part of
the total variation. The high value of the speed at this point gives the impression that the
pair of oxen could have develop more power than actually delivered. The unwillingness
facxor induced by the amount of pulling seems to be the only reason they did not.
,
171
Table 35: Parametet; estimates of walking speed (Group 1)
.
--<
Predictor
Coef
Stdev
t-r&
P
Constant
1.39930
0.06719
20.82
0.000
(PF/Lw)
-2.5326
0.2450
-10.34
0.000
--
s= 0.1326
- R-sq == 89.2%
R-sq(adj) = 88.3%
--
--
Table 36: Analysis of Variante of walking speed (Group 1)
--
-
-
-
SOURCE
DF
SS
MS
F
P
Regression
1
1.8795
1.8795
106.86
0.000
Errer
1 3
0.2287
0.0176
Total
14
2.1082
- Traction Group 2
The regression equation is:
PF
ANSPEELI =1..53-2.56*(~w)
(49)
The parameter estimates and the analysis of variante shows that the iit is good
with a negative slope and a high correlation coefficient of R = 0.94 (with R2= 88.5%)
(Table 37 and 38). The mode1 explains more than 88% of the variation.
172
Tqble 37: Parameter’estimates of walking speed (Gronp 2)
-
Predictor
Coef
Stdev
t-ratio
P
. .
-
Constant
1.52727
0.09084
16.81
0.000
-2.5580
-
_--
s=O.1421
R-sq = 88.5%
R-sq(adj) = 87.4%
Table 38: Analysis of Variante of walking speed (Group 2)
-
---
SOURCE
DF
SS
MS
F
P
-
Regression
1
1.7040
1.7040
84.40 -
0.000
Error
11
0.222 1
0.0202
12
The low speed ANSPEED = 0.29 m/sec recorded at PF = 2412.1 N was :not
singled out by the analysis even though it introduces part of the variation. This low speed
is also due to the unwillingness to pull under heavy load regardless of the quality and level
of training (Figure 19b).
- Traction Group 3
The regression equation is:
P F
AAWEED = 1.15 - 2.95 * (---$
(50)
I
l
173
1
Table 39: Pariameter .estimates of walking speed (Group 3)
Predictor
Coef
Stdev
t-ratio
P
:* ..’
. .
Constant
1.44785
0.04920
29.43
0.000
(PF5W)
2.9540
0.1862
15.86
0.000
s -0.09128
R-sq = 96.5%
R-sq(adj) = 96.2%
_----
-
Table 40: Analysis of Variante of walking speed (Group 3)
---
SOURCE
DE
SS
MS
F
P -
Rekession
1
2.0969
2.0969
251.69
0.000 -
En-or
9
0.0750
0.0083
Total
10
2.1719
The best fit line for the Traction Group 3 is shown in Figure 24. The parameter
estimates and the analysis of variante (Table 39 and 40) shows a very good cor-relation R
= 0.98 (with R2= 96.5%) to translate the close relationship between the amount of pull
required a.nd the speed. There is still a certain amount of unexplained variation related to
the unwillingness factor under heavy load, which leads to more energy losses.
6..3. Draft animal liveweight mode1
The liveweight LW in kg of draft animais is a required input in a11 models
developed to this point. It is important to be able to evaluate this parameter before using
any of these models. The simpler tbe way to do this the better. The liveweight of draft
1 7 4
animals is generaily u’nknown ta the majority of farmers. During this 6eld study none of
,
them was able to give the liveweight of their draft animals. Also, weighing scales are not
-* ,
generally available to farmers to help them determine the weight of their animals.
In order to develop the mode1 at this stage of the research, more draft animais
were included in the sample to bave enough point to generate a better fit. The extra draft
animais used belonged to the farmers in the 4 villages visited during the maximum pull
trials. The simplicity and speed of the measurement technique allowed this increase in
number of oxen, from 18 (original sample) to 73 cases. The technique consiisted of
measuring the circurnference at the point of heart (in cm) of individual animal and
determining the liveweight LW with an electronic weighing scale (Bar10 Electronic ScaleL:
Mode1 2 100).
The regression analysis performed with the first 18 draft oxen gave ai coe&ient of
determination of R2= 0.87 which is an acceptable level of significance. The cor-relation
coefficient between liveweight LW in kg and circumference CIRCLJMF in cm was R =
0.93 to demonstrate a close relationship between the two variables (Figure 21a)
The regression equation is:
L W = -367.80 + 4.78 * (CIRCUMF)
(51)
The parameter estimates and analysis of variante are presented in th,e Appendices.
Only, the complete statistical analysis of the broadened base sample (73 oxen) is given.
-
1 7 5
The corresponding regression equation is:
.:-
.
.
L W == -515 + 5. II * (C’IRCUMF)
(52)
T:he broadened base sarnple gave a better cor-relation R= 0.97 and expiained more
of the variation observed. The coeffkient of determination R’ = 0.94 shows a better fit
(Table 41). Most of the variation, is due to the errors
Table 41: Parameter estimates for liveweight vs circumferenee
Predictor
Chef
Stdev
t-ratio
P
-
Co&tant
515.13
23.88
-21.57
0.000
CC (cm)
5.10
0.15
33.41
0.000
-
--
--
-
-
s = 14.97
R-sq = 94.0%
R-sq(adj) = 93.9%
Table 42: Analysis of Variante (LW, CC)
-
SOLkCE
-
-
-
DF
SS
MS
F
P
-
Regression
- - 1
250037
25003 7
1116.48
0.000
Error
71
15901
224
Total
7 2
265938
1 7 6
400
350
2
i300
5
g 250
z
.C
4
200
150
135
140
145
150
155
160
165
170
li’75
Circumference (cm)
- Fitted Line
l
i
I
1
Figure 21a: LW= f(CIRCUM) regression Iine (18 oxen)
/
450
I
I
400
!j- 250
200
150
130
140
150
160
170
180
1190
Circumference (cm)
- Fitted Line
Figure 21b: General regression line (73 oxe.n)
177
committed when performing the measurements of the thorax circumference (Table 42).
Draft animais do net appreciate being touched by outsi.ders. The measurem.ent had to be
made with the collaboration of the farmers. The best solution was to put them under a
yoke before making the measurement. The coefficient of correlation R= 0.97
Sdemonstrates the positive close relationship between the two variables.
Fiigure 21 b shows that a couple of points are unusual observations. They deviate
fiom the line and present a fairly large standard residuals. Those points are represented at
location (CC LW) = (140, 238) (157, 255), (159, 264), (155, 306), (166, 364). Two
other observations need to be pointed out at (CC, LW) = (188, 442), (189,445) as they
are located at the extreme end of fïtted regression line. They have a real influence on the
line even with their small standard residuals.
The fitted regression line works well in predictirrg liveweight of humpless Ndama
breeds. Caution must be taken when performing the measurements. It must be donc as
early as possible, in morning hours. It is better to do the measurement before the draft
animais glo for grazing, to avoid introducing an extra source of variation. The more feed
they take before the measurements, the more the variation is introduced in the model.
The liveweight LW mode1 developed in this study needs to be validated for :specific
applications. The mode1 Will be very useful in future calculations of draft animal energy
balance ancl feeding system evaluation.
6.4. Management of draft animals
According to their production objectives, the farmers’ systems management of
draft animais was mainiy oriented towards minimizing risk and securing cash trop
,” I
178
production. Farmempointed out that thé strategy of using draft oxen on different crops
depended greatly on the timeliness of the rainy season onset. Two main allocation patterns
.
./.
. . ,
are used:
- If the rainy season is on time: The. cereals (maize and millet) will receive
a ,
:
the draR animais first. Afterwards, they are allocated to the cash trop (groundnuts) for the
time remaining. Their utilization on sorghum and rice fields is occasional depending on the
pattem of rain distribution and on the number of working days.
- If the rainy season is late: The cash crops (groundnuts) are given priority
in using the draft animals. The remaining crops (cereals) are usually manually cultivated.
,*
,
In general, the amount of area for cereal production is fixed compared to the cash
i
,* 1
crops. The total cereal production is mainly used for the farmers’ own consumption. On
,..
the other hand, the amount of land allocated to the cash crops depends on tlhle size of the
i
farm in hectares. In agro-system 4, for instance, the groundnut field cari occupy 65% to
70% of the farm (Equipe Systemes, 1983; 1986; Fall, 1986)
I.
The management of the draft animals during the rainy season deals mainly with the
following questions: What type of field work cari be performed?; How many days per
I . .
t
I
week cari the draft animais be used?; What type of feeding system is used?; How are the
problems that affect the animal’s health resolved?
l
The type of management has an impact on the liveweight and the working capacity
,.
ofthe oxen during the rainy season. The working schedule of the 18 draft oxen was
179
,
monitoredi throughout the rainy season on a daily basis t.o better understand the type of
management practices.
‘- .
6.4.1. Field work
During the rainy season (June to October), the draft animals were mainly used for
three maijor field operations (tillage, seeding and weeding) bksides transport. For the rest
of the season (November), the draft animals were idling or used occasionally for transport
They were not used for harvesting as groundnut diggers were almost non existent at the
fiirm level.
Table 43: Repartition (%) of draft oxen time
Field Operations
June- July
August
September
October
-
-
-
Tillage
83.33
66.72
22.22
0.00
Seeding
38.95
lP.13
0.00
0.00
Weeding
5.56
33.33
0.00
0.00
Harvesting
0.00
0.00
0.00
0.00
Not usled
0.00
22.22
72.22
77.78
-
-
-
Transport
27.78
22.22
5.56
22.22
It appeared that most of draft animals working time was used for tillage activities,
especially in the months of June amd July where 83.3% of the oxen were used for seedbed
preparation purposes (Table 43). Alter plowing, seeding (38.9%) was the second most
important field operation carried out by farmers. The concomitante of these two first field
1 8 0
II
activities was such. that fat-mers with more than one pair of oxen were better off. In
contrast, weeding was not as mechanized (5.5%) in favor of transport (27.8%).
‘. .
During the month of August, more weeding took place with 22.2% but plowing
remained the most important activity. Alter the cash trop, sorghum and rice fields were
the crops to receive the oxen. Starting the month of September, most of the draft animals
stopped being used for trop production as 72.2% were just idling, 22.2% used for tillage
.,*
on other secondary crops and 5.6% for transport. When the growing season reached the
month of October, more oxen were idling (77.8%) and the rest were used for transport
6.4.2. Daily working hours
The l&el of intensity of draft oxen utilization was very high at the beginning of the
ramy season, during the months of June and July. Farmers were working the draft animais
for an average of 7 hours per day with a minimum of 5 hours and a maximum of 9 hours.
This number of hours included the time spent to travel to and from the field location and
the time spent yoking the animais and hitching the farm implements. This working
intensity was repeated 6.47 days per week with little rest, 0.53 day per week. (1 day
maximum) (Table 44).
An average field time of 6 hours per day were used by farmers to perform field
operations. Field time was adjusted according to corresponding field operation efficiency
E in percent (plowing, seeding, weeding). A little time was left for other activities like
transport.
V:i-.--
--
--W
--
-.-‘--
--
-.
__-_
__
-_,_
.
.-
-...
-
--
-
Table 44: Management and Handling of Draft Animalr @ the farm leve!
Days restedIweek_June-July
Days rested/week-August
Grazing hrs/day-June-July
Graz@ hrs/day-August
Grazing hrs/day-September
F. Days with Diseases
7.00
7.08
5.63
1.45
0.00
2.00
1 8 2
6.4.3. Feeding system
TWO types of feeding systems were identified: one corresponding to the rainy .
._ ,
season and the other, to the dry season. The main focus was on the growing season’s
feeding system.
According ta farmers, the feeding system of the oxen during the dry season was
mainly extensive grazing, along with the whole farm or village herds. A complement of
feed based on trop residues, especially groundnut hays was given late in the afternoon.
During the rainy season, draft oxen were kept separate from the herd at tlhe farm
level. TO complement the natural grazing, during the breaks oxen were given a daily ration
of cereal grain (maize, millet) or cereal by-products (stover). Farmers across agro-
ecosystems 4 and 5, used the same strategies as the resources available for fized were the
same: 2.5 kg of corn per animal in June-July and part of August, and 2 kg of millet for the
remaining of the month of August. The amount of complementary feed decreased along
I
with the decrease in amount of work to be performed towards the end of the rainy season,
1
I
as natural grasses were more abundant. The natural grazing played a key rolle in the
.r.
I
feeding system of working animals. The grazing time passed from 5.58 heurs per day at
th’e beginning of the growing season to more than 9 hours per day towards the end. The
most important grass species found for grazing are given in Table 45.
In addition to the grass species, a number of tree leaves were also used to
complement the grass:
” I
- Vene (pterocarpus eraniceus)
- Bu pumbapumb (calotropis procera Ait)
-
1 8 3
fable 45: Pasture’grass species in zone 4 and’5
:=
Gass local names
Corresponding Scientificlnames
.-
Elimpey djikel
cyperus rotundus
Edjilangaye
panicum maximum
Eboussaye
digitaria cilia.ris
Enoname
bracharia plantaginea
Elbongaye
chloris pilosa
Kohing kohing
desmodium hirturn
Essabay
urena lobota
Ebucay
bracharia pallide-lùsca
Essoulangay
pennissetum pedicellatum
Baarafita
Oryza species
Essiteye
Acanthospermum hispidum
33
=-a
In general, the quality of feed given to draft animals is ranked according to the
energy content expressed in forage unit (FU). The total forage unit FU is converted into
energy b:y the draft animals digestive system, based on the t.otal digestible nutrient (TDN)
c,ontent of*the feed. Pasture grasses are generally considered as poor feed. Based on the
work carried out in Senegal by CEEMAT (fiench research institution), grass tut has only
0.14 FU whereas feed used for complement have a higher FU: groundnut hay = 0.35,
maize grain = 1.10, for example.
In ,the energy evaluation, the number of FTJ used to provide the energy required for
the maintenance of the draft a.nimal were used to evaluate the energy balance in relation to
the intensity of the task to be perfbrmed (Equation 11). The lack of the energy required
to petiorm the task Will impact growth, health and liveweight LW of the wolrjking animais.
T:he health problem is translated into the loss of resistance to sleeping sickness.
1 8 4
6.4,4. Liveweight changes and’ body condition
In order to evaluate the impact of the intensity of work on the draft animal body
‘- ,
and health, the liveweight of each draft animal of the 18-base sarnple was monitored by
”
:
means of weighing with an electronic scale (Bar10 Electronic ScaleZ Mode1 2100). This
measurement was made at the end of every month throughout the growing season.
The measurements performed on the pairs of draft oxen are shown in Tables 46
and 47.
6.4.4.1. Body weight losses (LWL in kg/day)
The analysis of the variation of the liveweight shows that both individual ox and
pairs of oxen across the traction groups lost weight during the period of intense work in
June-July. The recovery period started at the end of August and correspondled to the
decrease in field work demand. The coefficient of variation CV ranged from 2% to 9% for
C?oup 1,2% to 6% for C+roup 2 and 1% to 3% for Group 3 (Figure 22a amd 22b).
The average liveweight variation LWV in kg and in percent per Traction Ciroup
was calculated and summarized in Table 48.
From June to July, each pair of oxen went through high level of utilization. The
amount of body weight lost during the process was converted into energy (15 MJ/kg of
LWL given in the literature) towards the accomplishment of the job. This means that the
execution of most field operations required more energy than available from the feeding
system. Thés effect of work on body weight has been reported in many publications but
needs to be evaluated for different species.
.
,-: II
_-
1 8 5
,-
Table 46: Live&igbt (kg) variation of pair of oxen
Month (
Gl
G 2
._._, <
G3
"
Il
l Pl P2 P3
Pl P2
P 3
Pl P2 P3ll
June ' 683
690
648
535
600
517
469
435
441
July
699
673
583
522
544
499
438
427
419
Aug.
713
755
605
567
571
523
423
441
461
Sept.
717
742
635
580
600
537
472
456
487
oct.
718
753
668
598
605
538
486
466
505
Table 47: Avcerage Liveweight variation of Traction Group
Month 1
431
G 2
G 3
Il
AvIs
std
Avg
std
Avg
st:d
(kg)
(kg)
0%)
-
(kg)
(kg)
---
Ju ne
673.67
18.37
550.66
37.96
448.33
14.82
July
651.67
49.70
521.67
18.37
428.00
7.79
Aug
691.00
63.18
553.67
21.75
441.67
15.52
Sept.
698.00
45.70
572.33
26.28
471.67
12.66
oct.
713.33
34.88
580.33
30.07
485.67
15.92
Table 48: Liveweight losses and gained (June to 0ctober)l
Period
I
Gl
G 2
G 3
I Lw-v
%
LWV
Oh
LWV
%
0%)
--
(k%)
(kg)
i
-
--
June-July
-22.00
3.27
-28.99
5.56
-20.33
4.53
i
July-Aug.
39.33
6.03
32.00
6.13
13.67
3.19
Aug.-Sept.
7.00
1.01
18.66
3.37
30.00
6.79
Sept.-Oct.
15.00
2.15
8.00
1.40
14.00
2.97
186
500
400
300
200
100
0
Figure 22a: Liveweight Variation (LWV) of individual ox
l
1
I
4
/
j
!
!
j
4 0 0
- -
J
i
A
S
0
Months
-A-- Group 1 -C Group 2 -+ Group 3
L__~~~
--
--_-
--.----
.
Figure 22b: Liveweight Variation (LWV) of Traction iGroups
“. B
l
I
. ,
1
:
1 8 7
The average number of days worked during this period was’19.47 with a CV=
44% (Table 44). Based on this number of worked days, the liveweight Iosses LWL in
. . .- ,. .’
kg/day per traction group were:
- Group 1
LWL= 1.13 kg/day
- Group 2
LWL= 1.47 kg/day-
- Group 3
LWL= 1.04 kg/day
In accord with a nurnber of cases observed throughout the world, animal traction
researchers agree on the fact that even a good feeding :system would not prevent draft
animals from losing weight during this peak working time of the season. The body weight
losses converted into energy
:must be taken into account when calculating the daily energy balance of draft animais
6.4.4.2. HeaI,th and tare
I:n the majority of cases, farmers benefited from ,the veterinary services from the
local governmental extension agency. Vaccines were given twice a year to protect oxen
from losing tolerance to sleeping sickness and treatments were used against other different
parasites Cases of diarrhea have been mentioned by farmers in relation to thle changes of
tlhe feeding system from the dIry season to newly growing grasses after the first rains. This
never stopped fat-mers fiam using the animals as cases were not very serious and were
treated quickly.
188
,
6.5. S u m m a r y j,
This study showed with detailed calculations that the d@rence in ,optimal
performance an-mg the weight classes of draft Ndama cattle L&e not ver-y significant. A
further analysis would have showed that the coefficients or parameters estimated for the
diff+erent mode1 developed were aiso not very significant. It means that one equation cari
be used to mode1 the working output of Ndama cattle in the Basse Casamance region
environment. The level of significance cari be used, if needed, to differentiate specific
behavior between traction groups in order to make marginal improvements Ion the energy
balance calculations. This cari be done with consideration given to the cost and amount of
human resources available to conduct the study. The nature of the problems may play a
role in keeping the three groups separate. The difference in farmers’ circumstances
required to satisijr both animals needs (feeding and tare) and production objectives may
justify the treatment of each pair of oxen as belonging to a traction group that bas specifïc
needs in relation to their body weight.
The Ndama is a well-adapted species to the Basse Casamance region and needs to
be properly managed in order to satisfy the level of energy required. It is important to
monitor the feeding system and the amount of work demand. Draft animals need at least 1
day’off per week for better efficiency. Body weight losses LWL per working day are
expected from working draft animals but the amount in kg/day cari be kept to a minimum
through good management. Farmers need to develop expertise in fitting the job to the
animais by being more aware of their maximum working capacity. This objec.tive c.an only
be achieved if they know their animals well. One major improvement would be for a
farmer to follow up on the liveweight of his draft animals throughout their working career.
“_ ._ _ _ ___ - ._ . . .
7
I
189
The means of doing this follow up using weighing scales is not always possible in the local
areas, however other means cari be used. The mode1 developed.iq this study is significant
. .
.
and cari be used by any farmer equipped with a simple measuting tape.
In addition to the liveweight, it is important for farmers to develop strategies and
set up priorities in the execution of the field operations in relation to the specific ener+q
requirements. This cari be achieved by planning activities during the working days and in
selecting the appropriate implements to carry out the field operation.
-
‘.
.
‘.
.
Chapter 7
FIELD OPERATIONS AND DRAFT REQUIREMENTS
The evaluation of the pull force PF in daN and draft DR in daN was conducted on
station to better control the variables involved. The technique of land prepar,ation used
was the same as fat-mers’ and the size of experimental plot was in the range of field plot
unit delimited by farmers for plowing.
7,,1. Field operations characteristics
7.1.1. Soi1 characteristics
The soi1 type is the primary factor in the process of evaluating the draft required to
move different farm implements through the soil. The physical and mechanical properties
of the soi1 are the most important parameters involved in its behavior at different water
content levels (SWC). These two characteristics were used in selecting the sites for the
pull force tria1 and draft measurements. The physical properties were: soi1 texture, bulk
density, and soi1 moisture. The soi1 resistance to cane penetrometer was also used to
I
rninimize variability among plots.
A field of 2 ha was cleaned of natural vegetation composed of grasses (Pennisetum
Polvstachvon, Eragrotris tremula) and small trees. The area was characterized for
l
;
*.i I
1 9 0
-1
192
” Table 49: Average soi1 texture of the tria1 site
1 Soi1 depth (cm)
16 - 58 cm
% Clay < 2 um
7.6
116.7
?? ? ?
% Silt 2 - 20 prn
3.5
8.0
i
% Very fine sand 20 - 50 um
6.2
6.5
% Fine sand 50 ?? 200 urn
47.9
38.3
% Sand 200 - 2000 um
34.8
30.5
% Organic matter
0.9
0.8
Source: Extract fiom Niane A., 1984
In the American soi1 classification, the very fine sand is called coarse silt, combi.ned with
the silt to represent 9.7% and 7.6% for the clay. In the USDA textural tria.ngle, the type of
soi1 corresponds to Sand. The high soi1 groundmass which is the ratio between coarse to
,,.
fine particles shows that the soi1 matrix tends toward low soi1 water holding capacity and
hîgh permeability (7 crnihr).
The clay minera1 (1: 1) is kaolinite type and is mainly responsible of the s’oil red
color. The presence of kaolinite and Fe sesquioxides act as bridges between adjacent
particles and play the role of cernent in the soi1 aggregation processes. The level of organic
/
matter is too low (0.9%) to provide enough cernent to the particles. These linkages are
weak in general and confer a massive character to the soi1 structure that becomes hard
when dry and very malleable when wet. The wetting-drying cycles in relation with the
/
‘1 ,
tiequency of dry spells during the rainy season play an important role in the dete:rmination
<”
of the number of working days for different fïeld operations (FDOP).
I
193
IBeside the s&l texture, the soi1 bulk density SBD in g/cm3 of the tria1 site was also
determined with the aid of a tore sampler. Several sets of three undisturbed cores were
-:
‘. .
sampled fiom the field in the first 10 cm-horizon, weighed, oven dried and weighed again.
On a d:qr weight basis the a,verage soi1 bulk density SBD in g/cm’ was calculated
SBD= 1.40 gkm” std dev. = 0.01, CV= 1%
The average value of SBD found by Montorio (199 1) in the Djguinoum watershed
(agro-ecosystem zone 4) was 1.45 g/cm3. He used in his investigation a gamma-ray
attenuation densitometry (CAMPBELL type CPN 507-l. 5)
7.1.1.2. Water hoiding capacity
The soi1 water holding capacity SWHC in % g/g was determined at the
ISRA/Djibelor station soi1 laboratory. The values obtained from this experiment were used
only for indication purposes. The real soi1 water regime must be evaluated in-situ for
better results. For the laboratory determination, three weighed undisturbed soi1 tore
samples were taken from the tria1 sites in the first 10 cm-horizon and broug;ht to saturation
by capillarity. Alter a11 the initial weights were measure:d, the soi1 samples were lefi to
drain for a number of days. ‘The ,weight of each tore sample was determined on a regular
basis (every hour, every two-heurs, every day) until the sa.mples were completely air-
dried. The recorded data we:re used to calculate the soi1 gravimetric water content SWC in
% g/g. Ti’ne SWC was plotted against number of days (NDAYS) to have a drainage curve
(Figure 23 and 24).
194
I
-- ..-~
-.--.
20
.a.
.
5
/
/
4
t
1
I
!
I
I
/ +..++-~
/
I
8
I
1
1
0
1
2
3
4
5
6
7
8
9 10 11 12 13 14 15 16 17
Days
’ -I- Core 1
+ Core 2
-c core3
- Average :1
L-
-
_-.---~..-
Figure 23: Drained soi1 water content (undisturbed cores)
100 i
- -
-~-
t
I
<*
3r
-si
-3
= 1
E 10 I ..-.-.
_.
,... .
..I
w m
!
?
1
ii
i
?
?
m
I
?
i
i
?
1
/
?
/
L-
-+- - ..+A
--.--.- /I
0
5
10
15
20
Time (day)
Figure 24: Log-SWC vs time
1 9 5
The analysiç ofthe graphs shows that the 1/2 ha plot was uniform as regards to the
drainage profile and soi1 water content SWC, One soi1 tore showed a preferential flow at
.
the beginning of the experiment but quickly followed the two other cores drainage pattem.
The SWC in % g/g at characteristic points were calculated:
@ Saturation (MT)
: SWCsal= 23.69% (CV= 5.28%)
@ Upper Limit (DUL) : SWCDCL= 20.00% (CV= 4.08%)
@ Wilting point (L-L) : SWCLL= 6.00% (Charreau, 1974)
@ Air-dry (AD)
: SW&-,= 3.00?4 (CV= 13.60%)
T’he DUL in % g/g corresponding to a point close to the field capacity SFC in %
g/g was reached about 1. day after saturation while the air-dry moisture level under
ambient air temperature was’ reached afier 14 days. As defined by Ratliff et a1 (1983) the
DUL is the highest field-measured soi1 water content after being saturated and allowed to
drain until drainage becames, practically negligible. The soi1 water holding capacity SWJX
is the diiyerence between SWCDUL and air-dry SWC)m.
The SWC at which, fieid operations EDOP are performed is an important factor
towards the evaiuation of the amount of energgy required for the task. Fat-mers are
expected to schedule field operation SO soi1 is worked ait a moisture close to the optimum
soi1 water content SWC,,,.
196
7.1.1.2. Penetrometry
The penetrometry described earlier was the investigation- tool used to evaluate the
mechanical properties of the tria1 site. The dynamic method was’ukd whereby a known
weight falling from a predetermined height hammered the 50 cm long cane penetrometer
rod into the soil. The depth of cane penetration was measured and the associated energy
calculated using Equation 13.
Three locations were chosen at random along one diagonal of the 112 ha plot
Along with the soi1 resistance to penetration measurement, the SWC profrle of the spot
was also determined at the same time.
The graphs of the Cone energy in J against the depth in cm and the SWC profile
” f
helped describe the soi1 resistance to penetration at the given SWC levels (Figure 25 and
I
26). It appeared that the 1/2 ha plot was uniform enough to minimize the amount of
variation expected in the pull force PF and draft requirement DR trials.
I,.
7.1.2. Field operations
In relation to the monitoring of the utilization of draft animals (previous ch,apter), a
number of field operations performed by farmers to meet their production objectives were
characterized. Among those, plowing was the most important followed by seeding.
7.1.2,1. Tillage
Two types of implements were used for seedbed preparation: moldboard and
ridger plows. In the Basse Casamance region, plowing was recommended as a met.hod to
control and slow down weed pressure at the beginning of the rainy season.
197
--
--
1 0 0
2 0 0
300
400
5 0 0
600
Cone Energy (J)
e-Site1 +Site2 + S i t e 3
-
-
Figure 25: Penetrometry energy
0
--m
,------ - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
-5
- \\
Ê-10 't -------
_"----------_--_--_-----------------
0
c
k-15
- - - - - - -
---_-_-__-__---_-----------------
8
a3-20 ------,-- ----_-_-__------.- - - - - - - - - - - - - - - .---
!
:I
-25
-- -.---.--
t
ila--_-___ I -__-.--_-- *.--------------. I--
-30 jv
+
-
-
t
-
v
-
t
-
-
-
t
-
-
-
t
-
0
5
20
25
-
I
?? Location 1 V Location 2 + Location 3
.
Figure 26: SWC at penetrometry sites
l
1 9 8
“‘s<
- Flat land preparation
The most common method of flat land tillage performed !y farmers witih the
*.
‘_ .
moldboard plow (IJCF 10”) is to divide the field into small plot units of 0.6 ha in average
size and to plow each plot until the total field is done. The technique also known as the
Fallenberg method consists of starting at one corner of the delimited plot and of plowing
around the perimeter with the pair of oxen. In the process, the fùrrows are tumed outward
1
i
on one side ofthe delimited plot and adjoined side by side until the tenter of the plot is
reached. The majority of the farmers prefer this technique because it is faster than other
techniques like plowing in lands or by starting from the middIe of the plot.
The average plowing depth in farmers’ field was 1 O- 12 cm and the width 23 cm.
These two parameters depended greatly on the pulling capacity of the pair of oxen and on
the level of expertise of the farmers to properly adjust the farm implements.
- Ridging techniques
The second most common technique of land preparation is ridging. The data
analysis shows that farmers used two techniques of ridging. The first technique was
performed with the ridger equipment attached to the ARAR4 toolbar or with the EMCOT
ridger imported from Gambia (Figure 4 and 5). The technique of ridging is similar to the
plowing in lands technique with the difference that equi-distant ridges parallel to ‘one side
of the plot are made.
The second ridging technique was executed by means of UCF 10” moldboard
plow. In this case, the making of a good standing ridge required two furrows placed back
to back to create a crest of soi1 narrow enough to be planted with a single row trop. This
199
,,
/
method seems to be‘too slow to farmers who are concerned with timeliness. The common
practice Twith a moldboard plow is to make one-way ridges instead. of a back to back .’
:
ridges. This technique is faster as only half of the total field is plowed. With this practice,
the ridge quality is very poor as ,the height subsides faster with rainfall and water erosion.
The characteristics of the ridges made were highly variable depending on the piece
of implernent used and also on the ability of the farmers to properly use the pair of oxen.
The average depth and ridge height, and between ridge distance were 12 cm, 15 to 18 cm
and 50 to 60 cm respectively.
7.1.2.2. Seeding
The seeding equipment that farmers used was the Super ECO one row seeder
!
((Figure 7). The seeder was designed for flat land preparation. This represents a major
hmitation in its utilization. Al1 the fields that have been ridged were manually seeded.
It appears that seeding was not felt as a major constraint as the surplus of labor
generated by the mechanization of the land preparation was mainly used to hand seed
plowed areas. During the rush pe.riod in the months of June and July, only 38.7% of the
pairs of oxen were used for mechanical seeding a,nd it shows the importance of hand
seeding. hJost of the seeding was performed on the cash trop fields (groundnuts) which
occupied more than 50% of a11 cropped land. Al1 the Super ECO seeders were equipped
with at least one groundnut seed distribution plate and 59.9% were equipped with
groundnut seed plates only. Afler groundnuts, maize was the second most mechanized
trop with 3 1.82% ofthe seeders equipped with maize seed plates, followed by sorghum
I
( 18.18% of the seeders) and millet (13.64%).
f
IIl
200
7.1.2.3. Weeding
Mechanical weeding was not found to be as important.as.plowing and seeding. The
‘_
.
cultivator SINE 9 was the only implement used by farmers for weeding purposes. Al1 the
SINE 9 toolbars were equipped with 3-Canadian fYi.tll sweep tines. For better efficiency, the
type of sweep used to equip the cultivator must be able to stir the soi1 surface at a certain
depth to destroy Young weeds in favor of trop growth. The average depth in farmers
conditions was around 8 cm. The weeding period generally statts about 20 days alter
seeding (DAS) and constitutes a peak activity for the available farm labor. The extension
of cropped areas through mechanized plowing and seeding has created a labor bottleneck’s
shifiing situation that needs to be properly managed. The analysis shows that the weeders
owned by farmers were mainly used on part of the cash trop fields (groundnuts) while the
rest of the crops were manually weeded. The level of utilization of the weeders was not
enough to complete the field operation on a11 crops. The low level of utilization was
mainly due to a lack of expertise and to a fear of damaging the crops at advanced growing
stage when the weeding operation was not scheduled on time.
7.1.2.4. Harvesting
Farmers were not equipped with animal-drawn harvesting equipment. Harvesting
of’all crops is still manually performed. The only available animal-drawn harvester is the
groundnut digger built by SISMAR. The timeliness of the field operation in ter-m-; of
satisfling the labor demand is not critical according to the farmers’ point of ,view. Farm
labor is generally available in numbers suf?icient to carry out the field operations by hand
in time for upland crops. Only rice harvesting is felt to be a problem.
201
7.13. Soi1 water balarrce and Working days
Since 1960, the Sahel is globally experiencing an overa’l $ying of the climat?.
.
Following the two major droughts in 1970s and 1980s farmers’ behavior in resources
allocation have shown that the year-to-year variation in total amount of precipitation is an
important and determinant factor in planning field activities. Under these conditions over
the years, fat-mers have developed a risk aversion attitude towards investment.
The weather in the area ‘of study is mainly characterized by the seasonality of the
precipitation. During the wet season (June to October), rainfall does not occur every day
making daily rainfall highly variable,
The intra-annual variation of precipitation dictates in general the farmer’s strategy
to decide on the types of crops to grow and the corresponding field operation to perform
in relation to the types of implement available at the farm level. The analysis of daily
precipitation amount is important to soi1 water balance calculation to heip predict runoff
and drainage and to help eva.luate the field trafficability and determines the suitabihty for
performing daily field operations.
The most important factor in determining the number of working d?ys (NWDAYS
in days) is the soi1 moisture regime. This factor significantly influences the timeliness of
performing field operations. The soi1 water content is generally influenced by highl:y
stochastieal weather variables such as rainfall. relative humidity, temperature, wind speed.
solar radjia.tion, etc. The main effect of the weather variables is to associate a certain level
of uncertainty in the process of scheduling farm operations
The method applied in this analysis is based on the approach using probability
levels The probability level represents a degree of certainty. Two probability levels are
b-
202
considered 90% (9 years out of 10) and’ 50% (5 years out of ten). A database of 37 years
(1960 to 1996) of daily rainfall for the region of Basse Casamanqe. was used (Table 50).
The data was pulled from the Ziguinchor Weather Station databàse which is part of the
National Weather Stations Network. The town of Ziguinchor is 50 km away from the on-
farm research sites. The Gumbel(l941) approach is used (Equation 17 and 18)
Table 50: Annual rainfall (mm)/rank for Ziguinchor (1960-1996)
--_
Y ear
Amount
Rank
Year Amount Rank
Year
Amount:
Rank
(nm
(mn-9
w-w
1960
1274.60
1 6
1972
9 5 1 . 8 0
3 1
1985
1379.90
10
1961
1549.30
5
1973
1289.40
1 4
1986
975.20
2 9
1962
1567.50
4
1974
1240.40
1 7
1987
1042.6~0
28
1963
1429.40
9
1975
921.90
32
1988
13 10.60
12
X964
1222.80
2 0
1976
1297,lO
1 3
1989
1175.40
2 4
J. 965
1756.60
2
1977
7 9 0 . 3 0
3 6
1990
l’I14.30
2 5
1966
1603.80
3
1978
1512.10
6
1991
1223.00
1 9
1967
2006.50
1
1979
1187.30
23
1992
967.10
3 0
1968
8 8 2 . 5 0
3 4
1980
6 9 8 . 5 0
3 7
1993
1481.70
7
1969
1460.70
8
1981
1221.50
2 1
1994
1204.20
2 2
1970
1282.00
1 5
1982
8 9 7 . 8 0
33
1995
10’95.40
2 7
1971
1098.60
2 6
1983
8 1 7 . 2 0
35
1996
1310.90
11
1984
1236.00
1 8
TO determine the number of working days, the days suitable for field operations
are counted for every month of the rainy season for each of the 37 years. The counting is
manual and complex in relation to the main characteristics of a working day (see AME
203
StandOard D230.4,~Agricultural Management Data, Section 8-Working Days, Timeliness).
In addition to the scenarios introduced by Le Moigne (198 l), the following empirical
._ ,
‘. .
assurnptions for the Bass#e Casamance region were used:
Data collected are norrnally distributed for the long period of 3’7 years.
Onset of the rainy season is “the date afier 1 May when rainfall
accumulated over 3 consecutive days is, at least 20 mm and when no dry
spell within the next 30 days exceeds 7 days” (Sivakumar, 1988).
Plowing and Ridging are difficult to perform afIer a lO-day dry spell.
Seeding and Weeding are difhcult to perform afIer a 2-week dry spell.
The pair of oxen is at rest 1 day per week.
The confidence limits were used to estimate the range for the 90%, and 509’0
probability levels. Assuming a normal distribution (mean annual rainfall = 1229.08 mm and
standard deviation= 274.77 mm), the 9 years out of 10 for Ziguinchor lies withinf the limits
given by the mean plus and minus 1,6440*standard deviation (777.36, 1680.88) mm. The
5 years out of 10 limits are within the mean plus and minus 0,6745*standard deviation
(1043.88, 1414.27) mm.
The number of working days NWDAYS and their associated probabilitier; (PWD
in percent) (TILLWDY for tillage and SWWDY for seeding and weeding) are summarized
in Table 51
i
204
Table 51 Probability (PWD) and number of working days (NWDAY’S)
%Rain
Tilldays
TILLWDY
S@ays
SWWDY
fYr
(days)
e4
. - @vs>
pi)
- - - -
SO% Confidence level
8 . 2 0
12.18
0.41
9 . 1 8
0.3 1
2 6 . 4 4
25.53
0 . 8 2
16.88
0 . 5 4
3 2 . 1 8
2 6 . 1 2
0 . 8 4
15.59
0 . 5 0
2 5 . 1 6
2 6 . 8 2
0 . 8 9
17.76
0 . 5 9
7 . 0 7
18.88
0.61
15.88
0.5 1
90% Confidence levei
June
7.85
11.59
0 . 3 8
8 . 8 2
0 . 2 9
July
25.03
25.85
0.83
17.09
0 . 5 4
A.ugust
3 1 . 0 2
25.41
0 . 8 2
14.88
0 . 4 8
September
2 6 . 7 4
2 6 . 7 6
0 . 8 9
17.03
0.513
Bctober
8.38
19.94
0 . 6 4
16.26
0 . 5 2
The probabilities are monthly averages. TO adjust for the 1-day rest per week. the
PWD must be multiplied by an average coefficient of 0.86 (6 days per week). In the Basse
Casamance region, the rest day is usually on Fridays.
The difference is not very significant between the SO?& and 90% confidence levels
in relation to the rainfall distribution and intensities during the rainy season (Figure 27
and 28).
7.2. Pull force and draft requirernents
The total draft animals’ liveweight used for the tria1 was 683 kg with an average
age of 7 years and 5 years of working experience. The pair of oxen belonged to the
Traction Group 1.
-.
205
.- .
35 _-~--.---_l_ll__l---..
---.-
.
- - -
:
z
30:
.-
\\ '\\
$
:
m
025 -
.
hi. ,_..
R
*
izo*
II
‘\\,...
t:
.g
1
,,,’
‘~,,:
!Xl5 -
.\\
&'
\\
10:,I m
5
---_--_-_l___-.-_l_l--
---
J
J
A
S
0
Months
Figure 27a: Monthly rain distribution (5 years out of 10)
+
-
J
J
A
s
0
Months
,-- Seeding - Weeding
- - -
~.--_-.-..-
Figure 27b: PWD at 50% probability level
c
2 25
8
u
g20
c:
+--
J
À
s
0
Months
~-
--l_.-. -
Figure 28a: Monthly rain distribution (9 years out of 10)
.,
13
-
/+
I
i-‘.
30.6
P
3 0 . 4
!...
/
/
l
!
1.
:
; 0 . 2
!-_
,.I
/
;
!
l
0
i
J
J
A
s
Months
l-e Seeding - Weeding
~-~
---.---
-~ .-
Figure 28b: PWD at 90% probabilit! level
207
‘7..2.1. Statistical design
The 3-way factorial statistical design (2x3~3 Factorial in Randomized Blocks) .was
.A .
.- .
laid out within the 1/2 ha plot already tested for homogeneity in soi1 moisture regime and
mechani.c.al properties.
‘4 certain number of studies bave indicated that choosing the right type of
harnessing system could improve the pulling capacity and .the energy output of draft
animals. The head yokes used by farmers for their pair (of oxen cari be classifïed into two
<groups according to the lenggh: 90 cm and 120 cm (Table 52).
Table 52: Dimension of yokes used by farmers (survey resullts)
.-----
--
- -
-
-
-
Yoke type N
MEAIN
MEDIAN
STDEV
s E ME A\\
-
-
-
-
-
-
- -
-
-
-
-
-
YOKE 1
16
0.9694
0.9500
0.1080
0.0270
YO.KE 2
2 9
1.2069
1.2000
0.1043
0.0199
- - - - - -
- -
- -
-
-
-
Table 53: Two-sample t-test for YOKE 1 vs YOKE 2
- - - -
-
-
-
Yclk:e type
N-
MEAN
STEV
SE MEAN
- - - - - - -
- -
--~--
---.-
YCKE 1 16
0.969
0.108
0.027
YOKE 2 29
1.207
0.104
0.019
--I_--~-
-
-
-
_--
- - - -
95% C.1 for ~YO~I - ~‘You~!: ( -0.305, -0.170)
T-test ~YOKEI = ~YOI.E (VS N-E):
T= -7.15
P=O.OO
DF= 30
--~- .-...---_
-
-
-
-
-
~
I
208
The average dimension of YOKE 1 (96.9 cm with CV= 11.14%) was fcwnd to be
slightly different from the 90 cm-head yoke recommended by Research over the years.
‘_ .
The comparison between the number of the two sizes of yokes (90 cm and 120
cm), using the t-test distribution has showed that the difference is highly significant (Table
53).
The factors and their levels used in the 3-way factorial experimental design are
summarized in Table 54.
Table 54: Factosrs’ levels in a 2x3x3 Randomized Block design
swc
Implement type
w Eh)
-
-
.
6 to 8%
UCF 10” plow
8 to 10%
3-Canadian tines on SINE 9 Tooibar
! 10 to 13%
Ridger equipment on ARARA Toolbar
A total number of 18 treatments were blocked by SWC in % g/g. Each treatment
was performed on a 200 m* size plot. The whole experiment was conducted on 4590 tn’
with alleys between elementary plots included.
7.2.2. Required Pull force
The pull force PF in daN was the observed or dependent variable measured by
means of 500 daN dynamometer (PIABTM).
209
_,
7.2.i.l. Data collected for IJCF 10” plow
The distribution of the data with the 90 cm-yoke at SWC.= 6-8% (avg SWC=
.
7.01% gjg) is slightly skewed to the lefi (Figure 29a) co:mpared to the distribution at
SWC= 8-10% (avg SWC= 9 34% g/g) and SWC= 10..13’% (avg SWC= 11..97?/0 g/g). The
SWC= El- 10% distribution ;presents two peaks, the first at PF= 100 daN and rhe second at
PF= 160 daN (Figure 29b). The first peak (less than 5% of the data) is not as signifïcanr.
The distribution at SWC= 1 O-l 3% seems to be more normal than the first twio (Figure
29c).
With the 120 cm-yoke, the tendency in the data. distribution looks the same. A
slight skxwness to the right at SWC= 6-8% (Figure 3Oa) must be noted. The most normal
distribution was achieved at SWC= 8-lO?/o (Figure 301)) and the most skewed at SWC=
1 O- 13% (Figure ~OC).
The data collected fclr the moldboard plow UCF 10” are summarized in Table 55.
The avera.ge PF values shows that the lowest required PF for the UCF 10” plow was
obtained within the 8- 10% SWC range. The 90 cm-yoke gave slightly higher values, about
5.29 % more than the 120 cm-yoke type. The 90% Confidence Interval for the average PF
required by the UCF 10” moldboard plow is:
PFITFIW* = (146.98, 168.03) daN
The corresponding implem,ent average working width and depth were respectively
IMPLWWD= 23.17 cm and IMPL,WDP= 10.62 cm.
Pull Force (daN)
Figure 29a: PF with UCF 10” at SWC=G?% Yoke= 90 cm
I
4o i
/
90 100 110 120 130 140 150 160 170 180 190 200
P u l l F o r c e (daN)
-_
Figure 29b: PF with UCF 10” at SWC=S-10% Yoke= 90 cm
-<
. /
j
/
/
I
o 30 ;
!
1
/
I
k
1
I
oh
1
5 20 -
s f
E=3y():./
!
,km
1.11,
100 110 120 130 140 150 160 170 180 190 200
PuIl Force (daN)
‘Figure 29c: PF with UCF 10” at SWC=lO-13% Yoke= 90 cm
211
”
130
140
150
160
1’70
MO
190
200
Pull Force (da.N)
‘Figure 30x PF with UCF 10” at SWC=6-8% Yoke= 120 cm
40, - -
- - . - - - - -
/
I
30
;.
.s
à--
0
/
52(-J-
.-.
5
4
e
;
10
+
____
(
/
1
0 '
UJ-LL'
110 120 130 140 150 160 170 180 -
Pull Force (daN)
Figure 3Ob: PF with UCF 10” at SWC=S-10% Yoke= 120 cm
50 -----.-.-.__l--~-..-/
!
i
410
_
/
/
g
I
/
g* 30 I
-'.
5
5 20 --
II
k .
10
+
OL-IZ
100 110 120 130 140 150 160 170 180
Pull Force (daN)
-
-
-
-----.-_-.I_-
.-.
‘Figure ~OC: PF with UCF 10” at SWC=lO-13% Yoke= 120 cm
iI
2 1 2
Table 55: @Il Force PF (daN) with the UCF 10” plow
7.2.2.2. Data for 3-Canadian tines on SINE 9
The distribution of the data collected with the 120 cm-yoke is more nornnal than
the 90 cm-yoke. However, the variation between the graphs is not significant. CMy two of
the six are negatively skewed (Figure 31a and 32~) while the other three are slightly
positively skewed (Figure 31 b, 31c and 32a). The distribution at SWC= 8,- lO?G with the
120 cm-yoke is normal (Figure 32b).
The statistics on the data collected with the SINE 9 cultivator are summarized in
Table 56. It appears from the data table that the lowest PF required correspond to the 8-
10%~SWC range for the 90 cm-yoke type. For the 120 cm-yoke, the difference betw-een
the 8- 1004 and 1 O- 13% SWC range is not significant. Higher SWC has the tendency to
generate more pulling effort as the wet soi1 adhered to the sweeps. The 90% Confidence
Inter-val of the average pulling force, required by the 3-tine SINE 9 across the SM’C (6
13%) is equal to:
PFsIs~~ = (102.84, 135.20) daN
213
._.
30 J
_
i
,
ô
I
/
&
1
-
.
!
8
520;
3
8 .
r= 10 -
n .-ILI--
”
110
120
130
140
150
160
170
180
Pull Force (dz&)
Figure 3111: PF with SINE 9 at SWC=6-8% Yoke= 90 cm
---~--~___----~_-
--_I
40
----.~--------
-
30 - . .
h
<
I
e
3 20 -
z-
/
y”
80 90 100 110 120 130 140 150
Pull Force (daN)
Figure 31b: PF with SINE 9 at SWC=&lO% Yoke= 90 cm
40 __---_--_--------_ .-.- ---^._I_--
3 0 -
ô
o\\
Q
= 20 -
0
ii
-b ’ 1
k 10 I
,IL!
o.mm.
-
100
110
120
130
140
150
Pull Force (daN)
-.------- .---. -.-.-- _.._- ~---.--- -.-- ---~
Figure 3 Ic: PF with SINE 9 at SWC=lO-13% Yoke= 90 cm
214
-~
_-
----__-
30 ---
1
. . .
/
[
T 20 t
0,
7’
+-
l
0
110
120
130
140
150
160
170
Pull Force (dam
Figure 32a: PF with SINE 9 at SWC=&S% Yoke= 120 cm
--
/
70 -
/
J
!
60:
1-m m-
80
90
100
110
120
Pull Force (daN)
Figure 32b: PF with SINE 9 at SWC=8-10%
Yoke= 120 cm
35
30
0
80
90
100
110
120
130
Pull Force (daN)
Figure 32~: PF with SINE 9 at SWC=lO-13% Yoke= 120 cm
_.- i-
215
Table 56: Pull Force PF (daN) with tbe 3-tine SINE 9
-=
-
--
Yoke (cm)
9 0
‘.‘_ .*
120
-
-
-
-
-
SWC-range
16-S%
8-10%
10-13%
6-8%
8-10%
IO-13%
SWC(%gig)
7.01
9.34
11.97
7.01
9.34
11.97
- - -
-
-
-
-
Mean
149.09
111.91
130.45
124.91
100.00
97.76
S.E
1.93
1.97
1.58
2.63
1.31
1.51
Median
150.00
105.00
130.00
120.00
100.00
100.00
Mode
150.00
100.00
140.00
120.00
100.00
HOO.
Std dev.
14.3 1
19.08
14.92
19.52
9.38
13.12
Min
110.00
70.00
100.00
100.00
80.00
60.00
Max
180.00
150.00
200.00
170.00
120.00
130.00
m
-
-
The average implement working width for this PF range was IMPLWWD=- 52.08
cm and the depth. IMPLWDP= 9.60 cm.
7.2.2.3. Data for ridger attached to the ARARA
A similar analysis of the graphs shows that most of observed data dis,tribution are
slightly positively skewed (Figure 33a, 34a and 34b). The skewness coefficients af the
distribution corresponding to the IO- 13% SWC range are positive and greater than 1 (1.18
for 90 cm-yoke and 1.57 for 120 cm-yoke) to witness higher frequency values towards the
low required pull force PF (Figure 33c and 34~). Only one set of data is slightily
negatively skewed with the 90 cm-yoke at 8-10?/0 SWC :range (Figure 33b).
TIhe lowest required PF with the ridger working component attachment was
recorded within the lO-13% SWC range different from the IJCF 10” plow and the 3,tine
SINE 9 (‘Table 57). The 90% Confidence Interval for the average required PF in daN
within the ó-13% SWC range is equal to:
PFARA.RA = (109.57, 135.29) daN
Table 57: Pull Force PF (daN) with the ARARA ridger
The average corresponding implement working width and depth were
IMPLWWD= 22.59 cm and lMPLWDP= 10.44 cm respectively.
?.2.3. Regression analysis
The general linear regression mode1 for this analysis uses different levels of
qualitative independent variables: equipment types, soi1 water content ranges and yokes
type. The following dummy variables were used to describe these levels: SWC2 (1 if SWC
range is 8- 10% and 0 otherwise), SWC3 (1 if SWC range is 10- 13% and 0 otherwise), HS
(1 if the equipment type is SINE 9 and 0 otherwise), ARA (1 if the equipment type is
ARARA with ridger and 0 otherwise) and YK120 (1 if the YOKE is 120 cm and 0
otherwise).
- . ..-_
7
0
100
110
120
130
140
150
160 170
180
Pull Force (daN)
Figure 33a: PF with ARARA at SWC=4-8% Yoke= 90 cm
80
90
100
110
120
130
140
150
,-.-.---.---
.__-
Pull Force (daN)
Figure 33b: PF with ARARA at SWC=E)-10% Yoke= 90 cm
~ ----.---. --_--~.-----_. -.- . ..- - .-.---_--..~
.--,
l
(j(J _----_--_---
,
/
i
100
110
120
130
140
150
Pull Force (daN)
Figure 33~: PF with AURA at SWC=lO-13% Yoke= 90 cm
218
-
-
-- - - - - - -
100
110
120
130
140
150
160
170
180
Pull Force (daN)
-~-
--------_.
Figure 34~1: PF with ARARA at SWC+8% Yoke= 120 cm
100 110 120 130 140 150 160 170 180
-!'ull
Forfia)-. ..__ ._. ~ .._. ~_.
Figure 34b: PF with ARARA at SWGS-10% Yoke= 120 cm
/
” *O go 100 110 120 130 140 150
Pull Force (daN)
I-.--_---_-._.
Figure 34~: PF with ARARA at SWC=lO-13% Yoke= 120 cm
.-.-
‘
.-,rllUY-~---
219
- Full mode1 with interactions
PF = 177 - 23.1* (SWC2) - 17.5 * (SWC3j - 35.5 * (KY) - 38’.2 * (MA) - 9.99 * LX1 20)
- 8 0 * (SWC2KY) -L 16.‘9 * (SWcJI2ARA) - 2.0 * (SU’C3HS) - 11.0 * (SW~:3ARI)
(53)
The regression equation explains a significant part of the total variation (:$=
88.5%) (Table 58 and 59). Most of the non-explained variation cornes fi-om the lack 0.f
normality in the distribution of data collected within some elementary plots. One part of
the variation is also induceId by the variation of the different implements’ working width
(LMPL,WWD) and depth (IMPLWDP). The low values of the skewness coefficients, less
than or equal to 1 in absolute value for 88.89 % of the observed distribution, did not
strongly indicate a. need for data transformation. The assulmption of normal distribution
Table 58: Parameter estimates of PF regression
----~
- - - -
---.-.
Predictor
Coef
Stdec
t-ratio
P
-1-
-
-
-
-
-
Co:nstant
1’77.45
- 8.77
2o.r
0.000
SWC2
,-23.08
11.77
-1.96
0.086
swc3
-17.54
11.77
-1.49
0.174
HS
-3 5.46
11.77
-3.01
0.017
ARA
-38.19
11.77
-3.24
0.012
M(l20
- 9.99
5.55
-1.80
0 109
SWC2HS
- 7.96
16.65
-0 48
0.645
swc2ARA
1.6.89
16.65
1 .Ol
0.340
SWC3HS
- 2.00
16.45
-0.12
0.907
SWC3ARA
-10.99
16.65
-0.66
0.528
s= 11.77
R-sq = 88.5OX,
R-sq(adj) = 75.6?,0
Table 59: Pull force PF Analysis of Variante
--
----<
SCWRCE
D F
S S
M
S
““F
P
-
--~
Regression
9
8561.0
951.2
6 . 8 6
0 006
Errer
8
1108.6
138.6
Total
1 7
9 6 6 9 . 7
-.
- - - -
Variables
D F
SEQ SS
-
_~
s w c 2
1
336.5
swc
1435.8
HS
1708.0
3 9 3 6 . 0
YK120
449.3
SWC2HS
2 9 2 . 9
swc2ARA
3 3 4 . 0
SWC3HS
8.2
swc3ARA
6 0 . 4
holds as the central limit theorem allows when the number of observation n from the
random variable PF in each elementary was large (n= 5 1 to 149).
The probability values in Table 58 of testing the hypotheses that the interaction
coefficients are equal to zero are quite large. This implies to remove the interaction terms
as they are not ver-y important. The same analysis for the yoke suggests its removal from
the model.
- No-interaction mode1
I’F y 173 - 20.1* (iwc2) - 21.9 * (SiK’3) - 38.8 * (HS) - 36.2 * (AM)
(54)
2
2
1
This new Xinear equation simple to use explains 76.7% of the variation and has
slightly higher probability l’evels (Table 60). However., it also has a slightly larger standard
.
errer.
Table 60: .Analysis of Variante (SWC and EQTYPE)
-----
----
-----
SOURCE
1%
SS
MS
F
P
Regression
4
7416.3
1.854.1
10.70
0.000
Ekror
13
2253.4
173.3
Total
17
9669.7
Variables
DF
SEQ SS
WC2
1
336.5
swcri
1
1435.8
HS
1
1708.0
ARA
1
3936.0
The PF averages corresponding to the SWC-range and EQTYPE factors are of
particular interest i:n the evaluation process of the PF required to perform a given field
operation.
‘7.2.4. Draft requirelments
Another approach to the evaluation of the working capacity of draft animals is to
star-t from the draft requkements DR in daN of a given field work (Equation 12 or 31).
The draft is defined as the force needed to move the farm irnplement in the direction of
1:ravel. The conversion factor between the pull force PF :in daN and the required draft DR
222
in daN is the angle of pull y at the implement hitching point, During the trials, the angle of
pull, which gave the best line of traction, was maintained throughout the experiment. The
average y was 22.22 degrees with a CV= 1.15% and a 95% CI off22.09, 22.35) with a
coefficient of traction COS y = 0.9257 (Table 62).
For the type of sandy soi1 in the Basse Casamance region, a more interesting
parameter in relation to the working width (IMPLWWD in cm) and depth (IMPLWDP in
cm) is the draft required per unit cross-section of furrow slice at low speed This
parameter is called the specific soi1 resistance SSR in daN/cm2 (Table 61).
Table 61: Analysis of variante of Specific Soi1 Resistance SSR (daN/cm’)
_ - I _
-
SOURCE
DF
SS
MS
--
---_---
-- swc
2
0.10629
0.05315
EQTYPE
2
1.62410
0.812OC
ERROR
13
0.03464
0.00266
TOTAL
17
1.76503
---
-+
Individual 95% CI
-_--
s w c
Mean
-+ m-d-c ----f---- __-__ f _ec- *--“-i--------
6-8%
0.704
(mm--..-* _-e..m-)
8-10%
0.502
(
*--....--)
m---w
lO-13%
0.526
(”
___- * -wC-er)
0.490
0.560
0.630
0.700
EQTYPE
Mean
-+ _________ + ____--_-- + ------- --+-------
UCF 10”
0.613
( 3 - )
SINE9
0.222
t-*-j
l4RARA
0.957
t-*-J
-+ _________ + -_------ “+ ---- -----+-------
0.200
0.400
0.600
0.800
-..
-
7
223 .
The difference in SSR in daN/cm2 at the higher SWC ranges is not significant, 0.56
daN/cm* for the 8- 10% S’WC versus 0.53 daN/cm2 for the 1 O,lJ% SWC range. The
difference is highly significant between the lower and higher SWC ranges.
The implements show also a highly significant difyerence between their respective
draft requirements per unit area within the whole 6-13% SWC range. In relation to the
characteristics of the field operation performed, the ridger attached to the ARARA toolbar
has the highest draft requirements, followed by the moldboard plow. The sweeps attached
to the SINE 9 toolbar required less draft. A good estimate of the SSR daN/cm2 fi)r
commo:n use when evaluating the draft required DR in dtiN or the required pull fcbrce PF in
daN during an average working day (8-13% SWC range) is:
Moldboard plow: 0.55 daN/cm’
Ridger plow
: 0.89 daN/cm2
Full Sweep
: 0.20 daN/cm2
From this perspective, it Will be important to know ahead of time the implement
working width IMPLWWD in cm and depth IMPLWDP in cm in order to use Equation
14. The value of the pulling angle y Will also be needed. mer calculating the
corresponding traction coefficient COS Y, it Will be possible to determine the required PF in
#daN to be developed by the pair of oxen (Equation 31). Other important field
performance characteristics must also be taken into account in order to achieve a good
evaluation of the draft animals working capacity.
- --. - -- -_ --
_ -
l.l
-
Table 62: Field work cbaracteristics and Field eff~ciencies
ARARA with ridger
120.00
22.50
10.10
480.00
11.33
9 10.00
690.00
75.82
22.00
3-tine SINE 9
ARARA with ridger *
120.00
23.00
10.00
1238.01)
92O.W~
7
225 .
7.3. Implement Field capacities
7.3.1. Field effkiencies
‘_ .
The field efficiency E (in percent) of a farm implement is defined a.s the ratio of the
implement’s effective field capacity (EFC in ha/day) to its theoretical field capacity (TFC in
ha/day). An equivalent approach is to determine the ratio in percent of effective feld time
(EFDTM in hrs) to total field time (TFDTM in hrs). The effective field time of an
implement is the time spent performing a fimctional activity.
(55)
Table 63: ANOVA of E(%)
-----
--~--
---~-
Source
DF
Seq SS
Adj SS- Adj MS
F
P
-PV--
----
WC
2
378.08
378.08
189.04
--~-
4.74 0.030
EQTYPE
2
571.56
571.56
285.78
7.16 0.009
YOKE
1
480.09
480.09
480.09
12.03
0.00s
Errer
12
478.86
478.86
39.90
Total
17
1908.58
Table 64: Average implement efficiency E (%)
.-----
---
EQ’TYPE
N
MEAN
MEDIAN ST?&
SEMEAN- C’G(O4,) -
t-6 10”
6 ---
61.0’7
60.93--- 6.87
2.81
--~
11.35
74.54
72.63
11.43
4.67
15.33
65 19
64.69
9.46
3.86
14.51
4..
t
226
The summary in Table 62 sugge’sts that the field effkiencies E depend no’t only on
the type of implement but also on the type of yoke used. The analysis of variante shows
‘_ .
that the difference in E is highly significant (Table 63).
Field efficiencies are best given in terms of field operation. It makes more sense to
use the eficiency attached to the type of implement to perform the field activities (Table
64).
The effective field time (EFDTM) to perform a field operation is affected by the
losses of time when turning at the end of the furrow and when the implement is grounded
because of continuous high draft requirements caused by the rooting system of tree stumps
scattered across the fields.
Another source of variation was introduced by the reluctance to pull on a
continuous basis of the pair of oxen and their. They ofien made frequent stops to tope
with the draft requirements close to their optimum draft performance, with an average
PF/LW in the following ranges:
UCF 10” plow
: PF/LW = (0.21, 0.24)
3-tine SINE 9
: PF/LW = (0.15, 0.20)
ARARA with ridger : PF/LW = (0.16, 0.20)
7.3.2. Field capacities
Based on the field eficiency E in percent, the effective field capacity EFC in
ha’day is equal to the theoretical field capacity TFC in ha/day times the efficiency E!,:
2 2 7
ANWKGHR
EFC = (WWKGCOMT * ANISPEED) * (E) * (
utn,
j*30 *m4
(56)
‘_
.
Where the WWKGCOMP in cm is the width of the implement’s working
component; ANISPEED in m/s is the animal walking speed and ANWKGHR in hrs/day is
draft animal daily field working time. The coefficient .36*‘10-” is used to convert the units
into halday.
The field capacity depends significantly on the number of hours ANWKGHR the
draft animais are effectively used in the field to perform a task. The time spent traveling to
the field and yoking the pair of oxen are not included in the number of animal-work-heurs.
The total area cropped in ha during the season is calculated using the EFC’ (haiday)
and the number of days worked in relation to the probabihty of working day (PWD in
percent).
7.4. Summary
14 number of decision variables are involved in farmers’ daily activities during the
rainy season. Some of the variables are weather t-elated, therefore out of the fàrmer’s
control. The soi1 moisture rlegime SWC (Oh g/g) remains the most determinant faclor when
it cornes to performing farm activities. From the onset to the end of the rainy season.
farmers rely on their good judgement to decide when to perform a field operation in
relation to the available resources at the farm level.
The number of days suitable for tillage and other tïeld operations is dictated by the
type of sd along with the configuration of the rainy season. One of the best approaches 10
predict the number of working days NWDAYS is to use the level of probability or level of
1
,I
228 .
certainty. Two levels are generally used the 50% or 5 years out of 10 and the 90% or 9
years out of 10.
.- .
It is unfortunate that precipitation data are not always normally distributed. The
duration of the period to analyze is important to draw acceptable statistical inferences: the
shorter the duration, the less the likelihood of a normal distribution. Long duration periods
are preferred as the World Meteorological Organization WMO recommends a minimum of
30-years data. Even with this range of data, annual rainfall data are still skewed
(positively), with fewer totals significantly greater than the mean. Sometimes the use of
the median instead of the mean cari provide a more realistic and usable results especially
for the 50% level of occurrences (Sumner, 1988).
During a suitable day for fieldwork, the level of intensity of the draft animals
utilization represents a determinant factor in the evaluation of their field working capacity.
At the beginning of the rainy season, the work demand is always high and farmers
consequently try to maximize their utilization for greater output. The number of working
hours per day ANWKGHR along with the speed ANISPEED to perform the work will
determine the farmer’s capacity to meet his cropping objectives. It is important to choose
the correct combinations of implements, which Will manage the available working capacity
ofthe draft animals. As demonstrated in previous chapters, farmers do not have a wide
range of implements to Select from to carry out the tield operations. Only the expected
draft required DR in daN from the working component-soi1 relationships would help make
a decision. The amount of draft found in this study shows that farmers normally use the
draft oxen within the area of their optimum energy output (PF/LW= 0.20). This level of
utilization corresponds to medium-to-heavy work intensity. Working an average of6
A
1
229 .
hours/day in these conditions reinforces the management strategy of 1 day-rest per week.
It is important to fit the job t’o the draft animals in ternis of work,load to improve their
eficiency and productivity.
The use ofdifferent size head yokes does not seem to have a Sign&ant effect on
the pull force PF developed by the pair of oxen. It is a good management p:ractice to look
aRer dra.A animal’s comfort by fitting the yokes to the animal as energy is converted to
usefiA work through the yoking system. A poorly fitted yoke Will reduce power output.
On the other hand., the type of yoke has a significant effect on the field efficiency 13. The
use of a larger yoke has the tendency to increase the working width without really
performing a better quality job. In order to use realistic parameters, the use of the
implement field efficiency (highly significant) is recommended in the process of field
working capacity evaluation.
‘.
.
Chapter 8
EXPERT SYSTEM TO EVALUATE THE IMPACT OF ANIMAL
TRACTION IN THE BASSE CASAMANCE REGION
8.1. Organization of the expert system program
The expert system developed in this study is aimed at evaluating the working
capacity of a pair of draft oxen in the Basse Casamance region conditions. The program is
built with Rexibility in order to be expanded in the future to include other draft species and
other geographical regions.
The program is organized around different modules (Figure 35).
- Field capacities module
- Energy module
- Feeding system module
- Farm budget module
- Optimization module
The main module (Energy module) is used to calculate the traction potentiai and
the traction delivered in terms of energy.
230
231
r-------‘
/
/
Default Values I---------b’
CALCULATIONS
Traction Potential
~--_
ANIMALS i+ -_.-_ --e-d
i-.------
I Traction Delivered
I_-----.
..~
Traction Required
/
OUTPUTS
------l--- .--. -_ __. 9 Nutrients Requirements k-b
rSOIL FILE /--7
Field CapacQ
Power Requirements
_-----_
Feed Rations
WEATHER 1
Farm budgets
WORKING DAYS
LYE-__d i-
ry--------
/
l
Recommendations
/ PRODUCTION -+
FIELD
/
COSTS
/ ()pERjjTI()NS k----------- .------- +!
i
-17-r
i
- costs
l OPTIMIZATIQN ~
Figure 35: Animal Traction Evaluation Mode1 Diagram
7
232 .
The program was built with the LEVELS Object Oriented 3.6 for Microsoft
Windows Copyright 1995, Information Builders. The language is a high-level knowledge
engineering language called Production Rule Language (PRL). The syntaxes are organized
within Methods (When changed and When needed), Rules, and Demons which contain
eomrnands that tel1 the inference engines to initiate one or more actions at run time.
The program at run time needs a certain number of external files ma
DBASE III program and graphie files made with the Paintbrush software of the Microsoft
Windows 3.1 package.
Most of the information and default values supplied to the program were data
gathered from the field or models developed in the previous chapters.
For the Optimization module the program is set to use the Optimizer tool of an
external Windows spreadsheet server program like Microsoft Excel, Lotus 123 version 3
or later versions, and Quattro Pro.
8.2. Database Files’ structures
8.2.1. Farm location and characteristics
Barne of the file: FARM.DB3
Fields Name:
COUNTRY : Senegal.
COUNTPIC
: Country picture (Senegal.bmp).
REGION
: Basse Casamance.
REGPIC
: Region picture (Basse. bmp).
SITE
: Southwest (Geographical Location).
NZONE
: 5 (Number of Agro-ecosystem zones).
ZONE
: Agro-ecosystem zone
VILLAGE
: Village name (optional).
FMSIZE
: Farm Size in ha.
NFWKERS
: Number of Farm Workers (active workers).
MCASCROP : Groundnuts (Main Cash Crop).
WPESOIL
: Sandy clay (Type of dominant upland soi]).
2
3
3
It is assumed that small-scale farms’ production systems are in general driven by at
least one main cash trop (MCASCROP). The information is needed for the program to
execute the energy evaluation module and the optimization
Two types of information need to be input in the p:rogram at this point by the user:
- The types of crops grown at the farm level which make up the cropping
system (Groundnuts, Maize, Millet, Sorghum, Cotton and Rice).
- The First year the farmer started using animal traction. The name of the
fa.rmer is optional (I-irst screen of the program).
.8.2.2. Farm implement
Name of the file: EQUIP.DIB3
Fields’ Niame:
NAME
:UC.F 10” (Current Name of the implement)
WKGCOMP
: Moldboard (Working component)
NWKGCOMP : 1 (Number of working component)
WWKGCQMP : 25 cm (Working component Width)
WKGDEP
: 12 cm (Avg Working component depth)
,QVGPF
: 145 daN (Average Pull force required)
RETPRICE
: Retail market Price
IMPLPIC
: Implement picture (* .BMP)
MANLIFE
: Years Life according to Manufacturer
F.ARMLIFE
: Years Life in Farmer’s conditions
IMPLEFF
: Field Efficiency (in percent)
IMPLREL
: Implement Reliability (in percent)
The program Will use the given average pull force (AVGPF in daN). The mode1
developed in the study (Equation 54) is used to supplement a default value for AVGPF
The efficiency IMPLEFF in percent and reliability IMPLREL in percent were
Idetermine:d previously. New data entered by the user Will overwrite the current ones.
234 .
Information on the price of the implement (RETPRICE in local currency) cari be
supplemented at run time.
.
A bitmap filename with extension .BMP for the implement picture IMPLPlC is
needed to be loaded for display to the user.
8.2.3. Draft animals
Name of the file: DRAFTANLDB3
Fields’ Name:
TYPE
: 0x.
BREED
: Ndama.
SEX
: Male.
AGE
: Age of the younger animal if in pair.
CIRCUMF
: Circumference at point of heart in cm.
AVGLW
: Average Liveweight in kg.
MATURELW : Mature Liveweight in kg.
NDRFTANI
: Number of draft animals.
MKTPRICE : Market price.
MATPRICE
: Mature draft animal’s market price.
AGETRAIN : Age trained.
YRSEXP
: Number of years of experience.
ANWKGHR
: Animal number of field work hours per day
ANIPIC
: Animal Picture (*.BMP).
- Readiness for traction
The liveweight LW of draft animals are not generally known or readily available
from the majority of farmers. The mode1 developed previously (Equation 52) is used by
the program to estimate the value of the liveweight LW in kg. The information neelded is
the average value of the circumference CIRCUM in cm, measured at the point of hean of
each draft animal The average liveweight (AVGLW in kg) is used by the program to
evaluate the pulling capacity of the draft animals. For the specific case of cattle, different
studies have shown that oxen with a liveweight less than half ofthe value of the species’
235 .
optimal or mature liveweight (MATURELW in kg) must be excluded from traction (Lee
et al, 1993). When a pair o:f oxen is rejected on the basis of that rule, the program Will
. .
display a message and Will mot proceed fùrther:
Not ready for traction
- Age versus Years of experience
The program Will us’e the AGE in years of the younger animal (if a pair of draft
animais is used) to test the readiness for traction in terms of capacity to perform field
activities. The average training age (AGETRAIN in years)‘s threshold of draft animais of
current qpecies are usually well known to animal traction users. In the Basse Casamance
region, oxen are trained when they are two years of age. To be ready for field work an ox
must be a.t least more than 2 years. The threshold cari be overwritten.
The number of years of experience (YRSEXP in years) is used to rate the field
Yworking capacity of the draf? animais. In the Basse Casa.mance region 3 years of
experience is an accepted number ofyears to carry out good quality fieldwork. The
number of years of experience and the age for training alre compared to the age o:f the
animal to verify the consistency of the data entered in the session. The program. willl reject
inconsistencies and ask for new data. If a draft animal with correct liveweight has less
experience than required, the program Will inform the user through a displayed message
that:
Need more traiuing
236 .
The program Will proceed from’that point but Will treat the draft animals as
inexperienced.
.
- Working potential
Draft animals with correct liveweight and adequate working experience, are
characterized by the program as:
Good for Field work
The average number of working hours (ANWKGHR in hrsfday) expected from the
draft animals in relation to their pulling capacity is needed to evaluate the effective field
capacity EFC (Equation 56).
Information on draft animals’ market prices MKTPRICE and MATPRICE in local
currency are used in the cash flow analysis in the farm budget module. This information
cari also be supplemented at r-un time.
The draft animal’s picture ANIPIC is a bitmap file type with .BMP extension.
8.2.4. Feeding System
Name of the file: FEED.DB3
mds’ name:
MAINFD
: Main feed (Pasture grazing, tut grass, .)
MFDQUANT : Quantity of main feed in kg
DMCMFD
: Dry Matter content of the main feed in %
ERGCMFD
: Main Feed Energy Content (in MJ NE/kg DM)
SUPLFD
: Supplemented Feed in kg (Grain, stover, ,..)
SFDQUANT : Quantity of Supplemented feed in kg
DMCSFD
: Supplemented Feed Dry Matter Content in %
.
237 .
ERGCSFD
: Supplemented Feed Energy Content in MJ/kg DM
S F D P R I C E Price of Feed per kg
In farmers’ conditions, the main feeding type (MAIlWD) during the rainy season is
the naturall grazing system. T’he quantity and quality of feed (MFDQUANT in k:g) fiom the
natural grazing is never enough at the beginning of the rainy season to provide enough
energy to the draft animals. Farmers use different sources for feed supplement (SUPLFD
in kg) like maize grain, sorghum grain or stover. The groundnut hay is generally used
during the dry season along with the other trop residues. Fe:ed information during working
days is required.
,
The price of the supplement feed (SFDPRKE in local currency) is used to
evaluate the cost of the daily maintenance of the draft animais
8.2.5. Crops grown at the farm level
Name of the file: CROPS.DR3
celds’ naEg:
TYPE
: Type of Crop (Groundnuts, maize, millet, . .)
VARIETY
: Variety of trop grown (69-101, ZM 10, . ..)
CYCLE
: Number of days from seeding to hawest
SDRATE
: Quantity of seed in kg/ha
SD LINE
: Distance in between rows in cm.
RIDGDIST : Distance in between ridges in cm.
FTRATE
: Fertilizer rate in kg/ha.
SDILIMIT
: Date limit of Seeding for potential yield
SDPRICE
: Seed Market price per kg.
FTIPRICE
: Fertilizer Market Price per kg.
CROPRICE : Crop h4arket Price per kg.
YIELD
: Avg Yield without animal traction in kg/ha
LAIBCDAY : Farm labor cost per day
II
238
,.. &‘.. .
CMy information related to thc types of cr:;?
:rown at the fam level are entcrcd
in the current session. The pro,. r-am Will not recognize $3 trop that has not been selected at
the beginning of the session.
1st of the variables are used for the elaboration of the farm
budget.
Somc variabIcs listed in the file are net activated I* ,* ’ current
of ,the
version
program but wili bc in the r
:turc. Variable such as I-
:
‘I’ (date limit of swdirig
fc)r a givcn trop) are used t
.valuate t he timeliness cost,
-,ariable FTRATE
(fer-tilizcr rate) to include farmers that use fertilizers. The 11.
.f farmers in the Basse
CJasamancc. rcgion do ,t use ? l-tilizers. Different studies ha\\
,,,,
. ed that the actual
aseragc rate of fuliliz(.
under 10 hg/ha of NPK (8: 18:27)
Thc YIELD in L requircd by the program correspol
JC Icvel ofyield
observed before the uti!
dion of animal traction. In thc pmct :s of evaluating the yie?d,
tlic program assumes th
.u-mers with less than 6 years of expcrience or using non-
experienced dra.ft animai.\\ il not benefit from the yield increase induccd by the
mechanical plo\\ving. The yield inçrcments givcn to experienc.c.d fxmers come fro!n the
average rcsults of past experienccs 05 aitir,, .: : raction induc.ed yield increases over m;:.nual
farming. A percentage increase on : id i:.
.! for diffcrent level of mechanization for
cxh trop grown at t.he farm level. ‘The yi~,~~ ~;icreases generated by the mecharr::-ation of
land preparation were found to be most significant and consistent according to ‘i’ourte
(‘197 l), Nicou (1379) Clcarreau and Nicou (1981) cited by Jaeger (1985) (Table 14).
Results fi.on field work carri<,: out by thc FSR tcam (Equipe Systcmcs, 1982-90) in the
Basse Casamance region werf ,-‘w used. Rice is the trop with the highest yield increase
(S@.OO%), follom,ed by coti~ j_, i .:, .lOOh), maize (46.71?/0), Sorghum (34.60%), groundnuts
adjusfed by a coeficient (0.86) for the 1 day of rest per week.
trop and for each field operation at given probability Icvel: 5 years out of E 0 md 9 ycars
out of 10. The LABOIIBALh, in mn-days/ha is the diifercnce between thc labor Sait&Eable
(AVI..A.BOR.t,, in mn-days/ha) r$nus the labor requit-cd (T,ABORE@, in mm-da:~~%a) to
The output on the fart labor shows the yotential trade- offs and thc Mm enjxts
uf‘each levd of mechanization for each trop. The farmer must choose among thc t rades-
of’f, tnwnrds meeting his production objectives. Qne major strategy oAen perfomc-d by
fxrrwrs is to use the iabor surplus generated by the FmxhaFizatiorl of land preparation to
perfCm~ seeding oyeration instead ofusiug seeders.
8.32. Energy module
fhe energy balance uses the rnost recent scientiGc work on animal emrgy
evahtation. The factorial method (Equation 5) is used to perforrn the difYerent energy
culc~lations. The coeficients used in the equation for this study are mainly pulled fi-or-n the
dissertation work carried out on cattle in the West Africm region by Abdou FALL (1995)
(59)
i
244
8.3,2,4. Energy from L,W bsses (LWLERG)
The beginning of the rainy season represents a peak energy dcmand period because
of intense iield activities. Thcse activities especially tillage, take place within the numbcr of
warkings days in the months of Jrrne and July. The level of draft oxen utiliz.ation is orten
SO intense that draft oxen uses their body reserves to mcet energy needs, resulting in a loss
ofbody weight (LWL in ke/day). ,AnimaI traction researchers in general, bave reported
nc:gative LW change,s during this time ofthe year. Even well fed draft animafs were found
to Jose weight (Goe, I983).
The analysis of field activities in far’mers!.I conditions have shown in yrevious
chapters that draft oxen in the Basse Casamance region lest an average of LWL - 1 2 1
kdday. Lawrence and Becker (1994) suggested the use of a coeficient a’f 15 MS NE/kg
to convert every kg of LWL into usable energy by the draft animais. The en~gy
(I..WI.,ERG in h4J hFlday) is calculated as followed:
The LWIXRG is expressed in MJ NE/day and t.,WL~fiotip in kg’day.
8.3.3. Feeding system nodule
The calculation of the feeding system is based on rhe ratio k of the energy used
during every working day (daily maintenance + work donc) over the daily :majntcnarice
k, 0.019 * {xj -: 0.503
where x in MJ/kg IX.4 is the equivalent cnergy content of the total diei.
246
- NE available for worir. frcm the feed ~~~~~~K~~)
Only part of the energy made available from the metaboliz&le enetgy ME is tised
to perform work and tlie rernainirrg part used for maintenance pur-poses. The net energy
available for work from the feeding system (FDWRKNE in M.J NEiday) is calculated by
subtracting the requir-ed energy fGr maintenance per day MAE from the total net available
eFlCX&$
(h$NWRKNE
in NJ NE):
According to the fa]-mers’ systems management of draft animais, the pair of” oxen
maly be used 6 days a week. The amount of energy available for work M’RKNE in MJ
NE/day is evaluated for the working days only and adjusted to rake into accourt tfie
energy available fiom LW losses LWLERG in kg/day:
This level of energy is available and expected ta he expended from tl-te pair of
oxen. It is important tu mention that the quantity of feed used to carry nut a11 these
c,a!culations corresponds to the total amount given to bath animais workiog in the team
k
248
A fist of available feed stuffs along with their FU cC)ntent per kg of DM (Table 65)
is presented for ~he user to select and make up his own ration FDRATION exyrcssed in
FI-J.
=:-.-..I -_._-
-1------
----.-
I_-_-~
--
.--vv.
---
~kemn&----.
Airailable Feed
FU/kg
___--
~~~
_---_- -_--.,
Sorghum grain
0.90
Millet grain
0.70
Ma& grain
1.10
Cow peas grain
1 .oo
Cotton grain
1.10
X.
Groundnut cake
1 .oo
Millet/Sorghum stover
0.85
Cut grass
0.14
Pasture grazing
0.06 - 0.19
Good quality hay
0.29
Groundnut hay
0.35
R.ice straws
0.26
------
m7=l=:lP-----
Sowce: CEEMAT, 1975.
Most of these fecds are available on a yearly basLc’ It is important to select the feed
avaifable during the intense working period, mainly June, July and August.
The program w-il1 present to the user a calculator ;i)~pe of display to sum the N-J of
the selected feed.
0‘
(7’7)
~~gj,_l variable cost wo labor (TVCWOLAB in CFA.4ta)
TVCWQLAB = SDCQST + FERCOST
Total fixed cost
--_---..-_
(TFCOST in CFA/ha)
TFCOST =i JMPFCOST + ANJFCOST
~~~OZ&CI~~QSJ (TCWLAB; TCWOLAB with and without tabor in CFNha):
‘TCWLAB = TVCWLAB -t TFCOST
TCWOLAB = TVCWOLAB + TFCOST
GIMWLABha = PRQDVAL - TVCWL,AB
(82)
CMWOJ,ABha = PRODVAL - TVCWOLAB
C8.3)
l~&tkjaygin NM per ha (NMWJ,ABha; Wh’fWOLABha w and wo labor in CFA/ha)
MiWLABha = PRQDVAL - TCWLAB
w
K.MWOJ~ABha = PRODVAL - TCWQJ,AB
(85)
.I.
P--1-....---
..~-,-^--..~--.--L=--
----n-
.msœweewIm--
(88)
252
‘E-E objectives of the optimiza~ion problem GUI be expressed as followcd:
8.3.5.1. Optimhtion objectives
Goal 1: Maximize revenue from the cash trop
Gai 2: Produce enaugh cereals for farmer’s own consumption.
Instead of using a goal progr-amming approach, the second objective of the
problem is considered as a constraint in the search of solutions to producing cereals
corresponding to no less than the farmer’s needs.
In small-scale farming system, farmers’ objectives are not alwaps monetary.
Trades-off are sometimes possible in relation to the current economfcal production
environment. A major adjustment performed in the Basse Casamance region was to
sacrifice the cereals’ production in general and rice in particular, in favor of the groundnut
cash trop. Rice was kept to a minimum and was allocated the least resources compared to
other upland crops. The main reason was the risky wealher situation characterized by a
persistent shortage of rainfail. Among the cereals produced by farmers to meet their oyVvn
consumption, rice production always plays a role of regulator and oc,cupies a fixed
percentage of land in the low land areas acc,ording to the spatial distributbon of crops.
8.3.5.2. Constraints and objective fonction
8.3,5.2.1. Decision variables
The decision variables used in the program, in the form of area allocated to a trop
(CRCIPAIEA in ha) are defined as follow:
Land constraints
&,,, (LARUREQ,, * CROPAREA) < AVLABOR
,e.
C-W
The available farm labor (AVLABOR in man-days) is evaluated in relation to the
number of active farm workers NFWKERS at the farm le\\el multiplied by the number of
working days during the growing çeason.
,,
The labor required per L~O~Qi-,, in mari-days’ha corresponds to the total
amount of man-days per ha needed to conduct the trop from tillzge to harvest.
Animal energy constrairrts
The animal energy must be evaluated for the time the fïeld operation FDOP Iasts
and for as many field operations FDOP as performed (tillage, seeding, weeding). If the
pair of oxen is used for plowing only then t.he competition betueen different crops
bccomes a constraint in relation to the total energy available for the number of’working
days (NWDAY S in days) for tillage field operation.
The available animal energy (AVANERG in MJ NE) is equal to:
AVANERG = (k,,,)*(MAE)*(~,,,,, NWDAYS)
(92)
f
3’1~ consomption ccrnstraints are evaluated usin@, the FAO standards for food
dried, thrcrshcd, de-hulled and milled. AfIer proccssing, only approximatcly 65% of the
procc:ssi;~g from harvest to the final stage for cC)ï:Iî1IrrlPti(311 The cotton is net edible (094,)
team bave repo1-te.d that most farmers provide more than 60% of their groundnut seed,
arxxnd 20 to SO kg per ha, ‘The total household ~SL’ for groundnuts is estimated to be 5%
of the harvest per year.
l’he constraint on consumption cari lx formulattx! as follow: The sum of
ho~xxho’lcl total consumption needs (CONSNEED in kg:).
256
C’ONSNI:GD = 200 * (NFW)
The consumption constraints cari be written as follow:
Y
~~r.R,zp (CONSVAL * CROPAREA) > CONSNEED
(94)
Cropping system constraints
The cropping system constraints are expressed in terms of minimutn area of the
farm land to be reserved to specific c.rops. Rke production in the Basse Casamance region
car-ries a heavy load of cultural behavior a.s the traditional cropping systems in the past was
mainly centered around this trop. A number of social scisnce researchers refer to the
Basse Casamance population as the “Rice Civilization” people (Pelissier, 1966).
For the Basse Casamance region, the percentage of land allocated to rice
production (RCPCENT in percent) varies on the basis of type of agro-ecosystem. In agro-
ecosystem zone 4, for example an average of 15% of the farm land FMSIZE are used for
rice produc.tion against 20% in the agro-ecosystem zone 5.
The value of RCPCENT in percent is expected to be different for individual
farrners but the present version of the program is designed to use a default value pcr agro-
ecosystem z.one.
_ .
._,._.
“...
-
____
__-__._
-
-.-
-
-.
-_., <,
_~
”
.--
-“.- ‘.~,.l...‘: . . .-*-! -
. ..“..Tll / 2, .j ..W./” Xe“ ^A .-. . . .: I _,
:
-.*a
.__.<
.
.,_
The cropping systcm csnstraint is net includcd in the table of constraints prrxnted
‘Il~e nn~ount of land resouxes akcatcd to a trop (CRQPAREA in ha) ~nust be
positive or equal to zero. W’hen using the CROPRREA amiable in the eyuation, C‘ROP
has IO be replaced by GN fo,r groundnuts, MZ for ntaize, ML for millet, SG for sorgh~~rn,
CT for cotton and RC for rice. This general condition on the decision variable also called
the normegativity restrictions will guarantee a rcal value to the solution
3.4, Opltrn&ation outyut
The optimkation oul:put is bound by the amount and quality of information the
server is ahle to process. The Sofw- tool froc Microso~R Excel cx Optimizer from Quattro
i
258
Pro for Windows have their own technique to handle such problems. The current version
ofthe expert system program is designed to fit the Quattro Pro for Windows’ forrnat
8.4.1. Server solution
It is important to mention at this point that the solution to the optimization
problem refers to the combination of the different decision variables and not to the
maximum value of the net profit (NPRO’FlT in CFA). The main concern for the user is to
determine the trop mix to be grown and the area of land that will maximize the profit,
usjng the available resources.
The solution is given in ha of cropped land for each trop. A zero value means that
the trop must not be included in the cropping system. As the decision-maker, the user
must weigh ah the information given by the optimisation output before fc~rmulating, a
decision. The cropping system chosen should reflect the specificity of his farrn. Qualitative
variables not used in this analysis cari help reach a better viable solution.
8.4.2. Interpretation of Coeflkients
A number of information items are given along with the solution to the problem.
Thc information items are presented to the user on the same display. The main çoeffkients
given are:
- SlacWSurplus
The coeflkients represent the difference between the level of utilization of a given
resource and its availabihty.
considered a “free good”. A free good means that an increase in its availability ,wvill not
increase or change the }>tofits. The slack must aha~ays be zero or positive. Slack. is better
equations.
In surplus situations, the difference bctween the Ievel of utihzation of a resource
and a minimum requirement i s positive. This difference is used to measur-e how much the
requirements are exceeded.
Dlfferent names are used to represent the same coef!kient, like opportunity cost,
margina! or dual values in relation to the \\~;ty the optimkation problem is stated. Shadow
prices are applied to Jesources to estimate thcir opportunity cost when not one unit of tbe
conespc>ndding r-esource is included ~I!I the. solution. Jn other words, it expresses the change
in the value of the objective func.tion when Ione additional unit of the resource is added to
261
8.5. Pragraïkl vnfidwtion
In the proccss ofc.ontinuous validation, the expert system program is built to \\+Bork
with val:‘c!afed input data. Ruilt-in :ncssagcs guide the execution of thc program. ‘fhc ert-or
routines execute and vcrify at the same titr\\e tbe quafity of t he da.ta entered. In case of
/nconsistencies, the program will prompt the user to enter correct data. The amoi~m. of
TWO modules need to be testtd and validated against other methods of c\\~alllatiom:
260
any othcr \\wiable enfers the basis, it cm only cif her decrcases thc profit or ieave it
- Weduced cost
The reduced cost for the decision variables reprewtts the amount by which the
objecl-ive tinction will be reduced if the variable is introdtIced into the optimal solution. In
other words, it expresses the penalty imposed on the objeclive function if one unit of the
variable net currently accounted is entered into the optimal solution
A fully used rcsource is “hinding”. It constitutes a bottlcneck for any improvernent
or reorganization of the environrnent. The Kuhn-Tucker conditions stipulate that ifthe
shadow price is nonzero then thc resource is fully utitized mcaning there is no slack OI
wplus (Turban and Meridith, 1994).
8.4.3 Sernsitivity aaalpsis
An explicit module to run a sensitivity analysis is mt included in this version. ‘I’he
progmm musl. be re-run in order to change the parme.ter and observe its efiect on the
optin;al solution. The purpose of the sensitivity analysis is tu give to the user additional
decision making information.
261
: 4 years
262
- Prqpm? sutput
Effective fieid cqacities CMdayj
G N
M Z
ML
S G
R C
l’illage
0.18
0 . 1 8
0.18
0 . 1 8
0 . 1 8
Seeding
0.04
0.07
0 . 0 5
0.07
0.05
Weeding
0.03
0.03
0 . 0 5
0 . 0 6
0.05
</.
Harxsting
0.05
0.08
’
0 . 0 8
0 . 1 0
0.02
R C
.
T’illage
39.69
WcedFng
<.
.,.
-
*.
Maintenance (h4J NE/day) = 37.02
~1s-g evaluatjon rnod~~.&~-W!?&
G N
MT2
ML
S G
R C
‘TII age
223.78
223.78
2 2 3 . 7 8
223.78
223.788
W’eeding
+.
Seeding
I-farvesting
Maintenance (MJ NEYha) = 208.76
..-..___
._ --.-
.m”e..-...-..-“.-
I<oI---
-11111
.*-.-.1-m
Q~ti!~~~i~3-~!~t_lll--
Land (!m)
Reduced cost (CFA/ha)
4 . 7 5
0 . 0 0
a.35
0 00
0 . 0 0
-34264 3 1
0.00
-43607 64
a.90
0.00
Slack
Sh2ldOW prices
0.00
108154 90
0 . 0 0
509.5 1
1 OI 4.40
a 30
44 24
0.00
0 . 0 0
-,54468.06
consiraints (ha, mm-day, MJ NE, kg equi\\xletzt-cereal and ha of rice in the ordel- written).
dL---
-- -
Field
canacities
In this farmer’s situation, only land preparation is mechanized with the UCF 10”
moldboard plow. AH the post-tillage &Id operations are manually performed.
According to the tillage effective field capacity EFC in halday, it will require from
the pair of oxen an average of 5 days (6 hours/day) to plow one hectare of land. It is
a ~r.umed that flat land preparation is executed with the moldboard plow.
In comparison with the values pubiished in the literature, the program output for
ca& oper ation is within acceptable range. For example, L)e Moigne (198 1) gave the I ange
of6 to 8 days for a pair of oxen, \\vorking 4 to 5 hrs yer day to complete: the task.
Draft animais Energy balance
The validation of the animal energy balance is more dif’frcult in terms of fInding
similar utilization environment. For testirtg the consistency of the information given by the
program, the output is compared to the experiment repor-ted by Falvey (1986) in terms of
procedure. Al1 the steps were followed to reach the em gy balance for ea.ch field
operation. The value given in the output is mainly refated to the level of confidence given
to the raw data and to the validity of the diflerent equations used.
Qptimization output
The solution of the optimization rccommends that tbe user grow ‘79% of the farm
with groundnuts which is consistem with the practices in agro-ecosystem zone 4. In terms
of cash slow, it was expected from the. program to allocate more land resources to the
cash crl~p. In a small~ scale farming system, the x,alue attached to the cash trop :qresents
the drivin~ force for trades-of’fin relation to thr: pricc levels of goods in the officia1
mar-ket Su& decisians in turn cari reshapc the type ofcrapping system to the point that
thcy become the determinant factor in the evcrlut icsn of the global production sl’stem.
getting îhe rwded information to run the program. As this sttrdy has show, the
ewluation of the dif%wnt variables a.rBd yarameters rquired conriderable ti:.ne atxf human
266
resources. The good news is that once the data are validated and tested for stnbility, they
remain consistent and ready to be used.
Expert systems in agriculture represent a valuable rool to conduct farming system
diagnostics, field operations scheduling and management. The existence of such programs
gives opportunity to farmers, govcrnment servants, non-gsvemmental organizations, to
access technical information and to interpret simulation in relation to the socio-economiçal
The organization of the program in modules makes information flow and outputs
easy 10 interpret. The energy evaiuation module uses in its important routines, animal
x
energy equations developed elsewhere by other researchers. This approach is aimed to
stimulate the continuity of the research on that topic. The factorial method developed by
Lawrence (1985) to evaluate the extra energy used by draft animals to perform work is
actually widely used around the world. An effort has been niade in this study to use
coefiçients vaiidated in the West African region with cattle.
The annual farm budgeting and the linear optimization are classic tools used in
decision-making problems. The program output was designed 1.0 fürnish as much
decisional infomlation as possible to the user.
As a general comment, this version ofthe program needs to be used for more
validation and improvement. Some actions are going to be taken in the near future to
expand the versatility and the geographical coverage to other regions of Senegal and,
possibly to other countries.
--
._--
-.__-.
-.
.
__.
.-
_.--
..-.-
.-
‘~l-~llr-.F.------
#,,-..-.“.
e.--u
-r.IIIIIu---
?
9.1. corscllusions
Aninul traction is providing more !han 80% ofthe energy needed in the
agricu1tura.l activitie s, in %ne&d in piifliCtJli4r. l’he animal species used for traction ai-e
rnainly catl:le, horse and donkey. In the Southern part of Senegal, in the Basse Cassmance
rlegicn, cattle are the llia;n source of power. The Ndama brecd is \\,vell adapted in the
environment characterized by harsh sanitary cunditiorrs fix working animais. The level of
~tilizik~~ is highly variable across the region, and yet it needs to be evaluated at the farm
level Tht: purpose of the evaluation is EO estimate its potcntial and cwrent impact. Tt
267
268
represents an important factor of evolution of the cropping systems and towards farmcrs’
achievements in building up a sustainable production system.
In this study, the evaluation technique used was built around an expert system to
simulate different levels of utilization of animal traction in differcnt agro-ecosystern zones
in the Basse Casamance region. The expert system was structured according to modules
linked by routines and subroutines using a high-level language cafled PRJ, (Program Rule
Janguage). The PIU language is similar to natural langu.age (like natural English) with
. “ <
interactive editors for more flexibility and accuracy in building applications. The LEVELS
OHJIXT~~ (Release 3.6 for Microsoft Windows, Copyright@ 1995 by Information
J3uilders, Inc.) shell was u.sed as the main programming environment and development rool
kit. As any expert system derived fiom the artificial intelligence (AI) domain, it needs a
considerable amount of quality information in order to be used for decision making. ‘The
i.nformation supplied to the expert system cari be performed through a well-organized
knowledge base management system (KBMS).
The knowledge gathered from animal traction utilization in the area ofstudy was
processed into 5 modules:
- Field capacities c.alc,ulations
- Animai Energy evaluation
- Feeding system
- Annual Farm budget
- Optimization of farm resources’ utilization.
-.v
.-,x--.*-s
“,.s.“ma
----
Y
269
Information kd in the program has betn collertecl directly from the fields in
farmers conditions and from reliable sources in the. existir!g scientitic literature. A
trials. TO carry out the fkld activitjes, the farm was wnsitlered to be the unit of
obser\\;ation ar!d was viewed as the basic ele:dnent of the prod~~ction system
ofthc: diffcrewt types of farm implements cwrcntly used by farmer-s for trop prodwtion.
The second N~S a follow up ofdifkrent kld operations ]xrf?nrmed with the draft ~~X~II.
The field triais were desi,gned 1.0 evahatc two things: (1) The amount of poser
a..nd ewrgy required to pelform difkent ficld operations; oncl (2) TO meawre the
rcaximum and optimum poser that the draft oscn ~wuld ~~evclop at the beginning ofihe
rainy season.
9.1.1, Sumeys
- Farrn hplenrents
It V:~S possible in a relatively short pcriod of time (two weeks for 40 farmers) to
idcntify all the implements actually used by farmers. The C’ata collected ranged from the
type and number of implement s, the acquisition mode, the system management, the repail
and mainterxtrxe problems, the type and number of dl aft cnimals to the local b!acksn-ith
activities. I%~~X data were ana.lyzed to better understand ihe agricultural cropping system
alternatives available to fumers in relation to the opportw~ities ofEred by the type of
agro-ec.osystem %OIles.
270
It appears that the range of implements is not very large. Only one mode1 per type
of implement was used: UCF 10” moldboard plow was the only type available; the same
for the cultivators including the SINE toolbar; also one type of one-row seeder (Super
I-ICX)). Only the ridging implements were available in two versions. One was attacbed to
the ARARA toolbar and the second, imported from Gambia.
AN the farm implements experienced several breakdowns. This was associated with
the quality of the materials used to make the working c.omponents and the conditions of
use. Most of the soi1 working components were victims of the natural wear and ofthe
presence of roots and stumps scattered in farmers’ fie&. The problems were aggravated
by the absence of technical support. The local blacksmiths were not able to handle some
types of breakdown. The unavailability of required tools and good quality raw materials
prevented timeliness of the services. In these conditions the reliability, evaluated for each
piece of farm equipment was generally found to be low.
- Draft animals
Even though oxen were available to farmers, the distribution was found to be
highly variable. The main source of draft animais remained the herds in the village. The
s:~alysis shows how unstable the technology was as the majority of farmers refied ou only
one pair of oxen. They represented more than 70% of the farms. One pair of oxen wifl net
protect farmers from production Eosses in case of animal injuries, diseases or deaths
tn
such cases the only alternative is to borrow another animal, if available, or to rent for a
day or two, or to simply use hand tools. The majority of the farmers in that situation use
hand tools. For some fax-mers, this situation cari last a Iongtime before they find a
271
replacement. The frequency of these types of event placed a certain Jcvel of unre~iability in
the utilization of dl-ait animais for a given year.
How the technology was helping fa1 mers to fi.&3 their production needs WC+ the
main conccr~~ ofthe follow up. The aspects ineluded in the daily follow up were: type of
field operations pcrforrned, hoxw it MJ~S performcd in relation to the type and ar~~~rrt of
resourcx:s Ernrolved, the health and tare of,the animais and, the fceding system uscd to
maintain r:heir working cxpacilies. The draft animals ~~ere \\veighcd every n~onth to
evaluate the liveweight L.W variation during the rainy season.
TO gcrt the best perfol mance from the animal, fa]-,iners had to manage the oxcn
working time in relation to the climatic conditions. They mainly used their animals during
morning heurs. An average o,f 6 to 7 heurs of field\\i~nrk per day put a lot of physical stress
on the animaks, to the peint Ihal. they lest weight to covcr the le\\el of energy demand. The
main feeding system through rlatural graxing aAer work or during breaks was net Ienough
to s~;pply 1:he enerby necessary i-or the ta&:;. Feed x5’as supp~emcnted early in the nrc~rrzing
on a daily basis. The supplement las mainly made of rnaizc, millet, sorghos grain or
stover.
The bveweight of draft animais is a :necded paranreter in the process ofevalu;~ting
tke impact of animal traction. ‘The nrodel dcvelopcd herei:r cari he nsed to estinsate thc
liveweight through the measurcrnernt of the circumference or girth ‘The circumference and
the liveweight are highiy correlated.
212
9.1.3. Maximum working capacity
The maximum working capacity n’as evaluated through the amount of power
developed to pull different loads. The pull force PF deLeloped was measured along with
the speed used to pull the load. These two variables were combined to evaluate the pouer
output of the draft arnmals at different level of utilization intensities. Drafl animais were
classified by liveweight into traction groups and the power output of each group modeled
using a linear regression. The results showed that there was no significant dif3erence in the
power output among the t.raction groups, therefore one general mode1 could be used to
mode1 the power output.
Two interesting points on the power curve were anaIyzed to compare the groups’
performance: the optimum and maximum power. The amount of pull force PF
corresponding to the optimum power revealed the existence of a range of forc.es too
dif%cult to sustain. NI pairs of oxen used in the pulling tria1 showed a certain level of
unwilhngness to pull in that range and induced an underestimation of the corresponding
power output. At the maximum point, it was found that a pair of oxen could develop
around 50% of its liveweight into force. This level of pull was an extreme and was not
expec,ted on a tegular basis in farmers’ working conditions.
9.1.4. Energy required for fie!d oyerations
The difTerent implements in the repertoties at the far-m level were tested to evaluate
the level of energy reyuired for their daily utiiization. The amount of pull force generated
for moldboard and ridger plows corresponded closely to the optimum area of the power
.
_ __ -_._ .- -- -
-
273
curve. The bel ofworking intensity could be c~assifkd in the range of mediurn to heavy
job
‘The effective field capacity 1342 ofeach field operation was calculated a!ong v,zith
the number of working days during the rainy. Tsvo lcvels of probability were used: 5 years
out of 10 and 9 years out of 10.
‘The mount of energy delivered by the pair of oxen was evaluated using the
factxial mefhod with rh use of local cctef?icit:~~~fs, The atiimal energy balance was
detsr-mined by taking into account the encrgy used for working, for t+xlking, for
liveweight losses, and for the daily maintenance.
A main output of the energy balance was the calcu1ation of the feed ration fiorn
fized availabfe to fanners.
This methodology will nee:d more refinernents in the friture to better evah~atc the
infomation needed to build up the KBMS.
‘The first reconmendation is tIo improve the ~mthodc~logy of coflecting information
fi-or-n farnters, in par~iwlar, da.ta. on anjmal traction. The technology is wjdcly used in
developinq~ comfries and need more attention. ‘Thc other aspects to investigate furrher are
as follow.
The technology has brought about significant changes ta famers cropping
systems. l’he levei of utjlkation is still low as many famers me draft anirnals fx land
274
preparation only. There is room for improvement by designing better equipment to carry
out post-tillage field operations. Weeding is a serious problem which farmers are faced
every year due to fluctuation of available family labor during the season.
The use of draft animais for fieldwork is very attractive to farmers as their
utilization cari generate substantial revenue. By raising the level of utilization of the draft
animais, it Will be possible to increase the rate of adoption of the technology. One aspect
that will increase adoption of animal traction is the reduction in production costs resulting
fi-0111 mechanization.
- Draft animal selection
The environmental effects on working animais need to be hetter addressed to help
filrmers Select the right format of draft: animais and to fit the job to their pulling capacity.
The stress generated by the local climate has an effecf on their effrciency. The only existing
recommendation is to work morning heurs but farmers’ working needs go be)lond morning
hours.
The format and the age to star-t working drafl anima& are determinant for future
performance through better training. Farmers are not generally consistent on these two
factors because of lack of technical information.
. ”
- Feeding system
Meeting the feeding requirements of their working animais at the begInning of the
rx.iny season is a serious concern for fiirmers. The state of the art shows that agro-
_-. ._-..
ec.osystem zones offer w-y limited possibilities. It is important to focus on what is feasible
in farnws’ conditions. More research is nccded.
Different triais dto\\~~ed tllat draft animais are wtsting considerable amount of
energy in relation 1:o the high level of pulling intensity. Regular field working c.onditions
\\vill require pull force around the optimum power out@ Ievel. One aspect to comider in
the futcre is the improvemeut of the yoking systcm to haw animais pull in a more
cc-mforMAe attd &Sent way The loss ofthe willingness to pu11 by the ox-team does help
the farmer or improve the quality of the ficvldwot-k.
The major challenge is to fit the job to the tizorking animais. This cari be done. by
desi:gning and adapting imptements to bettcr fit the wor&.ing conditions. This must be done
in consitieration with the ergonomie factors intolved in 1 he implement desigII fo a.cc;:ount
for -!he animal format, the soi1 characteristics, and the u!:er.
- Learning pwiod
It is crucial to find ways to reduce the Iearning period. Fal-mers do not see the
bcnc:f3s ;If using animal traction because they do net master the technology. The lack of
skiils shows that most of Ihe advan2ages of using animal traction are offset. One
alterrrative is to reinforce the capacity of the farmers’ organization to train farmers on a
APPENDICES
i. .,
i
Ekh---
..__
_-.
276
regultir hasis. The classic approach of regional training ceriters depcnds to a large exilent
ou exm-nal fhding smmes
The evaIuation of work output and energy espenditure stilf requires highly
and labora~~:ory mcasurements is almost the only icchnique used. Tt should be possible to
develop pc,rtable heupensive semors to meatwre integrateti paranieters used in the tzrrergy
evaluation: draft, spced, heat, temperature, alid healt rates. ‘The instrumentation developed
by the French research institution CEEMA’T PWds to be completed with a computer
software capable of analyz.ing data recorded in the data lol;!,ger anti c.arrying out basic
power atld cnergy calculations.
- Ccbmputer progrnms
More computer programs dealing with animal traction management need to be
de\\ eloped. This Will errable information to be shared by difTèrent users through a common
database and to better simulate czxnditLxx of utilir,ation. The present expert system is
expected to be improved, to be ,tra!idated, and to be cstended to other regions in Sencyal,
and to 1 he M’est Afiican region.
AWENDJX A
i,k
c%J*ctives
-
-sx.-m-I-
A s p a r t o f t h e met:hod01.~3gy, t h e a n i m a l t r a c t i o n
qucstj onnaire i s conducted at t h e household l e v e l .in order
to evaluate t h e t y p e and number of farm e q u i p m e n t avai! able
to farmers. T h i s sulvey m u s t be comp.?eted b e f o r e any draft
a n d other energy evaluat.ion .is perfo:r-med. %-JC? maj n focL;s o f
t h e q,lest.ionnaire cleals with different tecIlni.cal aspects
re.lated ‘to t h e utiliza.tion c f an.imal drawn-equipment at t h e
housel-lold l e v e l . T h e o b j e c t i v e s of this survey are:
1. TO iC!entify t h e t y p e s o f equiprnent actually
i n v o l v e d i n ficld. operations for which m o r e d e t a i Ied
investigations are needed to determine power
.reqi2irements.
2 . TO evaluate the leaxning process for animal
t r a c t i o n u s e b y correlating perfor:mance a t t h e frirm
‘1 cvel against years of exper i.ence.
Mel:hQ(.1l31c>qy
--- .. . . .
-.-~.I~-I.LI~c
‘Cne t a r g e t grou;p j ncl.ude faxmers equj.pped w i t h 3nd
usj ng an:-mal tractio’n implemcnts.
T h e enumerator for thi s
survey must u n d e r g o technical traj ning i n i à e n t i fyj ng t.he
differ:ent: pieces o f Iequipment 2nd their u s e b y farrnerz.
The questionnai.re addresses a 1. imited number of
v a r i a b l e s :
1
-. . Agro-ecosystem zone :
There C?xiS’:S f ive agro-
eCOsÿsten1 z o n e s i n t h e area Iof s t u d y and formers locnted i n
two of them a r e considcred e f f e c t i v e a n i m a l t r a c t i o n users.
;!. T,ocation: B y l o c a t i o n i t i s understood v i l laqes
wi t h i n the i d e n t i fied ,agrrc--ecosystem :<one.
-1
- ,P kmseho3d:
I d e n t i f i c a t i o n o f the chief o f the
hoc sehol d.
4 <I fhmber of EamiLy workera: r-iunher of persans
belonging t o ‘the househo.3d and carrying out f i e l d
activi t:ies (labor f o r c e ) .
_
_
-.
_
_
--..ee-
_
-
-. -._.
_
_
_
-..-_.
__.-
5. Eq-uipment type: equipment must be identified
properly. A variety of equipment is being used by farmers.
T'he most common items are:
. Moldboard plow UCF 12,"
. Multipurpose beam .?W4RA
. MUltipurpOSe beam SINE
. Ridger EMCOT
. Seeder SUPER ECO
. Transport cart
6. Field aperation: this variable relates to the piece
of: equipment and its use in the field for land preparation,
weeding, etc...
7. Working tools: this variable is typical of the
multipurpose tool bars which cari be equipped with
moldboards, ridgers or cultivator tines.
8. Traction mode and age of the animal(s): Three
species are used: donkey, horse and cattle as source of
energy.
9. Yoke type: in relation to the field operations, the
activity and the species, yokcs are derigned to provide
efficient transfer of energy from the animal to the
implement . The type Will be described by its name and
length.
10. Acquisition of the equipment: information is
collected as regard to the Date of acquisition, Mode
(credit,
cash payment, gift, etc...), State (new, used),
Price if bought.
11. Origin: farmers use equipment of different trade
names (locally made or imported).
12. Management: what is the status of the equipment at
the household level: does the farmer own the equipment? 1s
he co-owner with a neighbor? 1s he borrowing, renting, long
term leasing, etc . . .
13. Actual condition af the equipment: 1s the
equipment in good condition (working parts must be
checked), does it need repair and maintenance ? Should it
be discarded ?, . . .
14. Repair and maj.nl:enancs: what repair has heen
performed and at what cost? Who took carre of the
maintenance?
15. Obse:rva t-i QX1S : important infoxnatj.on corning up
during t h e s u r v e y
-
.__.
..-_-l-_-_.-___-__------“-
--..--.
---_-
..-
-II
..
_ -
.”
_,-..
-._
_
---,m
,“*
I,.
.3$iaMmm
1
1
i
:
---I”m---
N 0 c
Qf
.
_--
0”
._
_.
-.-
-
f
282
APPENDIX A2: Draft Animal8 Follow-Wp Guidelines
1. Draft animals utilization
-
1. Utilization Timeframe (heurs per day)
1.1" Field work
- Tillage
- Seeding
- Weeding
- Harvesting
1.2. Transport
- Destination
- Distance
- Rate (paid or non-paid)
1.3. Custom work
- Number of days
- Type of field work
- Site
- Rental rate
2. Draft animals feeding system
2.1. Dry season
- Types of feed
- Daily Quantity (average)
- Origin
c
2 8 3
2.2. Rainy season
- T;ypes o f
feed
- - Jxily Querit,ity (average)
-- ‘Supplement (type, quant.ity)
-. IOrigi:n (Fl2r natural g r a z i n g , sample t h e
type of grasses a n d hays f o r i d e n t i f i c a t i o n
a n d eva.l.uate i ts prcr;3ortion i n t h e feeding
SySt:ep) ,
I I . Health
.e-.,--
1 . Health Proble~s antl diseases
- Dry season
2. Manifestations / Xdentification / Freqvency
3. Length of irmmobi3..izI-tt.i.on
4 . Gare
.5. Housing
--l.-l .- .---_ .-_-_.< --
8
III. ADDITIONAL OBSERVATIONS 284
Table Bl: Maximum draft tria1 (GI PI)
/June 25,1996
LOCATION :
Ijibelor Station
1
RAFT ANIMAL NAME :
Etoile Double Les
Right
Left
LENGTH (cm) :
85.00
82.00
166.00
162.00
357.00
326.00
8.00
7.50
XPERIENCE (yrs) :
5.00
5.00
200.00
IlAvg. Draft STD S.E
cv
Min
Max
Spe@d
Powcr
fY%-
Draft/LW
Powcr/LW
N
4r
N
N
ds
J/a
N/N
J/s-N
I/
0.00
0.00
1.28
0.00
0.00
0.00
46.89
11.37
18.06
196.20
294.30
1.17
303.81
0.04
0.05
259.67 517.58
50.82
9.44
9.03
392.40
588.60
1.14
590.04
O.ûR
n na
“SU,
9îû.37
62.90
;ü.ôû
6.a5
76‘4.ti‘u
98i.00
1.09
1003.21
0.14
0.15
/!
1086.95
91.53
16 . iô
8.44
951.00
1275.30
1.02
1108.69
0.16
0.17
/e 1409.32 1270.49
83.78
13.08
6.59
1079.10
1471.50
0.92
1168.65
0.13
0.17
65.92 i 3.83
4.68
1275.30
1563.60
0.90
1267.49
0.21
0.19
1531.44
66.51 I 9.05
4.34
1373.40
1667.70
0.75
1148.58
0.23
0.17
1866.84
155.59
27.08
8.33
1667.70
2452.50
0.53
989.43
0.28
0.15
2138 -09
145.97
23.37
6.83
1962 n n
.Y”
2452 -50
0.49
1047.66
0.32
0.16
2344.59
68.67 I 21.72
2.93
2256.30
2452.50
O,?O
79?.16
0.35
O.iî
2777.01
351.39 i 97.46
12.65
2452;50
3433.50
0.34
944.19
0.41
0.14
3531.60
I1
3531.60
3531.60
0.00
0.00
0.53
0.00
286
-
??
?
I
I
3
0
0 0 -
r-!
r-7
3 _-.l
m 1
??
3 3 ---
30000
3
2dmc\\10
^-=-ï-=1-===7=~---i-;====r:=~:---.-
?
.__,
?
?
.__-_--.I-._-^_-.”
?
???
?
_---_.__,
.
?
?
0
w
?
, . ri
?
?
r-c\\iI-l
???
r-mw
InLcr
07
??
??????
??
???
cn 0
c-i r-l r‘ r-
--..
OI 0
ri
WI--W ??
??
w w
??
??
.
??
?
?
?
w w
*
?
??
mm ??
????
??
m 0 -----.-.-
???
r-l
Dl
w 0
??
?????????????????????????????????
u-1
TJ’
IT‘I
w
l-4
??
r-i
Cl
??
?
II
?
0
, m ?
?
?
?
.
??
“l-----ll.-.ll-
0
-l._----.--.---
II
.
?
?
?
??
Ln 0
r-i
?
??
???
cn
PJ
0 0
CO
w CO
r-i
?
m
??
?
??
?
??
??
0
0 ??
0
0
??
??
r-4 0
C\\I
w
8-i 0
0
0 0
I
I
0
.
.
?
--
.
..--.
-------1
-....m---aE--^--.-.--
~_I
.
--
..-.
.
-.,.1.-1-.-1
<_-.“--m.-_-
-
---.-.--
_-.-,
-_--.-.sl.
--.----.--
--
__II,.
--_
i
Table B3: Maximum Draft tria: (Gi P3)
>ATRE :
;OCATION :
!ONE :
1RAFTANIMALNAME:
Amerique Gaston
SIDE :
XNGTH (cm) :
:IRCUMF'ERENCE (cm) :
7EIGHT (kg) :
iGE (yrs) :
EXPERIENCE (yrs) :
'ath Length (m) :
1.16 11386.26
Lt c;In
l-cl
ti
4.J
0 0
0 0
0 0
c c,
0 0
0 0
..“_.._
-----_”
.<“--
--“-,,--
v; hl
ci*
----
p
._-_
01 Q
to
0
-.1..-----.
0
00000
--“_.-I-~-.--<---l.l_
r’
0
0
-“--^------.
0
0
or-r-r0
-.-ll”.-~-~“-.l-~-.-^-~“----
<
.
0
0
0
CO
03
ru3
r-lwr-l
P
Cd
.
.
0
Lt-l
0
cn
0
r-lr-l
N
t-4
h)
.
*
0
0
0
u-l
w
0
P-l
u-l
.
.
_.--~
0
CT
0
0
0
0
0
.
.
---.--“<..-<
-.--__
._.
-_.-.
---
-----
l-
--
----
--
4 Y
QZ k ti
m
w
5
c 5
B
.
x
&È (u d
*
2
2
289
-L
;
:
I
;
:
: 3 t0
; )
; :
; n r:
-
, , 43cnl-icooo0U-Jl-i&030 I
, 34 300000,
(
, 34 0, --
r
OI VI
m
61 OI
‘9
a w
N
u-l 0
0
I
0 0
-ll--__-_
---.~_
.
.
__i u 2 r; Ef f2 z
r-i s)
0 0
r-4 Fi?
0 0
.
.
CJ :
0 0
4.
C\\I cs 0 0
.
.
7-I 0
0 0
0-r
0 0
u-l
0 0
cs
0 0
.
.
?
-.~..<-__._-I~-
??
1P.l u
?
?
.._
?
--_-
?
???
UJ 0-l
????
LIT
?
?
?
?
---.----
?
????
???
i‘-
???
r
r-i v-i
?
?
?
??
??
????
CO
???
r-l
0 m
??
?
??
?
??
?
0 N
?
r In
?
?
?
?
??
?
?
?
w as
=cr N
?
?
.--l---.l._l-_.-
?
0 0
?
?
-.I,---.--_-_
--
Table B7: Maximum draft tria1
(GB PI)
JATE I
L'3CATION :
ZONE :
JRAE'TANIMALNAME:
SIDE :
;ENGTH (cm) :
~IRCUMEZRENCE (cm) :
GIGHT (kg) :
iGE (yrs) :
3XPERIENCE (yrs) :
== v: a . .
-.A- z E jj r*
-xIi- Hfd u x- E-4 ?? ??? . .
????
---. ----
--
2
rb ti
ca 2:
-
:
292
3 3 ----_--._---.
?
?
c-r) r-t
?
?
w r
?
?
0
CU
?
?
r r-
?
u-l
?
0
?
??
w 0 ---
-__-___
j
.
1
ui w
293
i
:
r)
;
n
t
3
-
0
0 0
.
0
(v 0
*
0
M 0
.
(3
P-l m
*
0
m 03
.
----l_
Table B10: CiraEt animais LW and CIRCUMF
.---
----. --..-..
, -*_,- ..__~-
_---.--
lec.
-.--II
LW
ieng th
:1KIc1s
bx! .
IN
,ength
~IRCWMF
#
.~__
km)
(cm)
#
- - -
(kg)
-._- -__.
km)
( cm)
_-_--
(kg')
--<.--- .----_- -_---
1
166.00
96.00
135.00
38
273.00 loo.00'
152.00
2
173.00
90.00
135.00
39
277.00
95.00
159.00
3
186.00
89.00
137.00
40
278.00
101.00
15'4.00
4
186.00
94.00
138.00
41
279.00
99.00
158.00
5
190.00
90.00
1.39.00
42
282.00
95.00
159.00
6
.207.00
88.00
145.00
43
286.00
103.00
155.00
'7
208.00
99.00
141.00
44
290.00
99.00
16:3.00
3
212.00
95.00
144.00
45
294.00
98.00
15r3.00
3
213.00
93.00
142.00
46
297.00
103.00
160.00
1 0
1213.00
91.00
140.00
47
300.00
102.00
1 5 "7 . 0 0
11
215.00
89.00
140.00
48
300.00
99.00
158.00
1 2
222.00
91.00
146.00
49
301.00
92.00
162.00
1 :3
227.00
88.00
148.00
50
304.00
96.00
164.00
1 4
228.00
91.00
148.00
51
305.00
101.00
158.00
1!3
233.00
89.00
145.00
52
306.00
103.00
155.00
1 6
233.00
95.00
148.00
53
3 0'7 . 0 0
99.00
1 5 9 . 0 0
1'7
234.00
96.00
148.00
54
31?.00
97.00
1 6 0 . 0 i3
1 8
234.00
84.00
148.00
55
311.00
106.00
1 5 6 . 0 0
1 9
235.00
94.00
1.50.00
56
322.00
104.00
166.00
2 0
238.00
90.00
140.00
57
325.00
103.00
1 60 . 0 0
2 1
2 3 8 m 0 0
96.00
150.00
58
.32 6 . 0 0
82.00
1 162 * 0 0
2 2
2 4 0 ,, 0 0
9 3 . 0 0
149.00
59
.3 3 15 . 0 0
100.00
1 6 -? . 0 0
2 3
2,4 3 " 0 0
95.00
152.00
60
:3 3'7 . 0 0
100.00
1 6 9 . 0 0
2 4
2 ,4 4 " 0 0
98.00
150.00
61
:3 4 4 . 0 0
99.00
1'7 2 . 0 0
2 5
2.51 1 0 0
91.00
150.00
62
:3 4 6 . 0 0
97.00
1 6 9 . 0 0
2 6
2 .54 <I 0 0
92.00
l!jO.OO
63
:3 4 8 . 0 0
107.00
168.00
2 7
2 .5 5 a. 0 0
100.00
157..00
64
:3 5-7 . 0 0
85.00
166.00
2 8
2 5 6 m 0 0
95-00
1 5 3 m 0 0
65
:3 5 8 . 0 0
108.00
172.00
2 9
257.00
Ill.00
ltj2 ,, 00
66
3 6 4 . 0 0
105.00
1 6 6 . 0 0
3 0
2 61 . 0 0
93.00
154,.00
6 -7
3 7 :1 . 0 0
107.00
1-71.00
3 1
2 6 4 < 0 0
95.00
159,, 00
68
3 7 6 . 0 0 105.00
1 '7 Cl , 0 0
3;'
2 65 . 0 0
100.00
152.00
69
3 7 6 . 0 0
105.00
1-7 4 . 0 0
3 3
L! '7 1 . 0 0
l 0 0 * 0 0
156,,00
70
3 7 9 . 0 0
108.00
17 0 . 0 0
3 4
2'71 . 00
:1 0 2 * 0 0
156.,00
71
4 0 2 . 0 0 118.00
181.00
3 5
2' 7 2 . 0 0
97 < 00
153..00
72
442.00
114.00
188.00
3E
272.00
94.00
151.00
73
4 4 5 . 0 0 119.00
189.00
3-Y
- -
273.00
----,m
87.C!O
---.
152.00
- -
-_.l--ll_
--. -.,-
295
_.
Table B11: Regression analysis of the 18 draft oxen
--
.--.-
LW
JLENGTH
--m CIRCUMF
LW ks CIRCUMF
(cm)
139.00
20'7.00
88.00
145.00 Constant
-407.8(
213.00
91.00
140.00 Std Err of Y Est
19.4i
222.00
91.00
146.00 R Squared
0.8'
234.00
84.00
148.00 No. of Observations
18,01
238.00
90.00
140.00 Degrees of Freedom
16.0(
256.00
95.00
153.00 X Coeff
4.78
261.00
93.00
154.00 Std Err
0.45
264.00
95.00
159.00
272.00
97.00
153.00
279.00
99.00
158.00
290.00
99.00
163.00
301.00
92.00
162.00
304.00
96.00
164.00
326.00
82.00
162.00
344.00
99.00
172.00
357.00
85.00
166.00
376.00
- - -
105.00
170.00
APPENDIX C
SUMMARY OF DATA COLLECTED DtJR.iNG ON-STATION TRIALS
100
80
8’
20
!j/IiiI!l/ !/ii
0
Ii<,
4-n
M
J
J
A
S
0
N
Month
/
_.~ ._... --- ____ -.---_._-..-.--.. .----_-.-. ..- .-... .~. ---.-.-..
_-
-...
Figure C.1: Rainfall distribution at Djibelor Research Station (1996)
14
u 8
& 6
150
200
250
300
350 I
Date (Julien)
= O-2 cm
??
2-7cm
+ 7-15cm
3 15-26 cm
_~__-.-.--- _________ --~-- _-_._ _.- .._.____ - _____ -_-_. ._
Figure C.2: Soi1 water content at Djibelor Rescarch Station (tria1 site)
296
297
1m:plerrent
P1.o~ UCF
Wiclth (cm)
23.00
23.50
Depth (cm)
9.150
9.00
Total time (s)
li366.67
840.00
Effect. time (s)
'740.00
460.00
Efficiency (%)
54.15
54.76
Angle of pul:L
22.30
22.00
Yoke (cm)
90.00
120.00
Draft (ciaN)
PU 1. 1
ReCJ,
E)u11
Req.
130
120.53
160
148.35
180
166.89
160 148.35
150
139.08
160 148.35
120
111.26
150
139.08
190
176.1.6
170
157.62
170
1. 5 -7 . 62
150
139.08
210 194.71
140
129.8:L
190
176.3.6
170
157.62
180 166.89
180 166.89
190
176.1.6
150
139.08
180
166.89
160 148.35
200 185.44
180
166.89
170
l. 5 -1 * 62
180
166.89
190
176.16
150
139.08
180 166.89
150
139.08
160 148.35
160
148.3'5
150 139.08
170 157.62
160 148.35
160
148.35
180
166.89
150 139.Of3
190 176.16
160
148.35
190 176.16
170
157.62
190
176.16
190
176.16
210 194.-71
170 157.62
210
194.'71
160
148.313
210
194.71
170
157 .62
200 185.44
150
139.013
160 148.:35
180 166.8!3
190 176.16
170
157.62
220 203.98
170 157.62
170 157.62
160 1413.35
298
(Table Cl continued)
160
148.35
170
157.62
21 0
194.71
190
176.16
160
148.3.5
170
157.62
170
157.62
160
1.48.35
160
148.35
170
157.62
180
166.89
170
157.62
200
185.44
180
166.89
190
176.16
270
157.62
170
157.62
180
166.89
170
157.62
160
148.35
150
139.08
170
157.62
190
176.16
170
157.62
180
166.89
160
148.35
140
129.81
150
139.08
180
166.89
190
176.16
170
157.62
160
148.35
150
139.08
180
166.89
160
148.35
160
148.35
1'10
157.62
160
148.35
170
157.62
170
157,62
190
175.16
170
1137.62
200
185.44
170
157.62
190
176.16
170
157.62
210
194.71
170
157.62
200
185.44
160
148.35
Avg (daN)
178.91
162.92
166.00
153.91
Std dev.
21.38
29.47
11.05
10.25
CV(dec.)
0.12
0.18
0.07
0.07
Max (daN)
220.00
203.98
190.00
176.16
Min (daN)
120.00
0 . 0 0 1 4 0 . 0 0 1 2 9 . 8 1
299
Table Ci2 : 'SINE 9 Bioek 1 @ SWC-range 6 - 8 % g/g (Extract)
I:mpl.ement
:3-.tine SINE 9
Width [cm)
s 0 . 0 0
Cl1 . 5 0
Depth (cm)
9 . 3 3
9.00
Total time (s)
7 7 4 . 7 5
520.00
Effect. time (s)
4 9 0 . 0 0
350.00
Efficiency (%)
6 3 . 2 5
67 , 3 1
Angle of pull
9 c
L. L .‘50
22,50
Yzke (cm)
9 0 . 0 0
120.00
Draft '(kg)
Pu1 1
Req.
Pull
Req.
160
147.82
1. 0 0
92.39
1 1. 0
101.63
17 Cl
157.06
150
138.58
1.2 0
110.87
1 4 0
129.34
1. 5 0
138.58
160 147.82
1. 4 0
129.34
1 4 0
129.34
1 3 0
120.10
140
129.34
:1. 1 Cl
101.63
1 3 0
120.10
140
129.34
150
138.58
1. 5 0
138.58
160
147.82
110
101.63
160 147.82
1OC
92.39
140
129.34
12Cl
110.87
17 0
157.06
120
110.87
150 138.58
100
92.39
180
166.30
140
129.34
130 120.10
130
120.10
150 138.58
100
92.39
140
129.34
130
120.10
1 3: 0
120.10
100
92.39
140 129.34
130
120.10
160
147.82
120
110.87
130
120.10
:1 3 0
120.10
170
157.06
:1 1 0
101.63
150 138.58
:1 3 0
120.10
150
138.58
:1 2 0
110.87
140
129.34
:1 00
92.39
130
12c.10
:L 00
92.39
150
138.58
140
129.34
150 138.58
120
110.87
180 166.30
120 1 1 0 . 8 7
300
(Table C2 continued)
150 138.58
130
120.10
150 138.58
110
101.63
150 138.58
160
147.82
140 129.34
170
157.06
150 138.58
120
110.87
140 129.34
100
92.39
160 '14'7.82
120
110.87
160 147.82
160
147.82
130
120.10
100
92.39
130 120.10
110
101.63
160 147.82
150
138.58
150 138.58
110
101.63
150 138.58
110
101.63
170 157.06
140
129.34
170 157.06
120
110.87
150 138.58
100
92.39
130 120.10
140
129.34
150
138.58
150
138.58
140
129.34
140
129.34
170 157.06
110
181.63
160 147.82
130
120.10
130 120.10
160
147.82
150 138.58
110
101.63
160
147.82
120
110.87
160
147.82
120
110.87
Avg (daN)
149.09
137.74
124.91
115.40
Std dev.
14.18
13.10
19.34
17.87
CV(ciec.)
0.10
0.10
0.15
0.15
Max (daN)
180.00
166.30
170.00
157.06
Min (daN)
110.00 101.63
100.00
92.39
301
Table C.3 :
'ARARA Block 1 @ SWC-range 6 - 8 % g/g (Extrac:t)
Impl.ement
Ridger ARARA
Width (cm)
23.00
22.50
Depth [cm)
10.00
10.10
Total time (s)
'7 1 0 . 0 0
480.00
Effect. time (s)
3 9 0 . 0 0
340.00
Efficiency (%)
.54 .,!33
70.813
A-rgle C>:E pull
2 2 . 0 0
22.00
Yoke (cm)
13 0 . 0 0
120.00
D:raft r:
kg)
Pull
Req.
PU11
Req.
12 0
lli.26
13 0
120.53
120 ll.l.26
:1 7 0
157.62
130 120.53
:i40
129.81
100
92.72
l 2 0
111.26
100
92.72
1. 1 0
10 1 . 9 9
120 111.26
1. 1 0
101.99
140 129.81
1. 2 0
111.26
100
92.72
1. 4 0
129.81
130 120.53
1. 3 0
120.53
11.0 101.99
1. 2 0
111.26
140
129.81
100
92.72
130 120.53
1 2 0
111.26
120
111.26
130
120.53
150 139.08
14c
129.81
120 111.26
180
166.89
100
92.72
160
148.35
110 101.99
:170
157.62
130
120,53
:t 2 0
111.26
160
148.35
:170
157.62
110
101.99
:1. 3 0
120.53
150
139.08
:i. 6 0
148.35
130 120.53
260
148.35
140 129.81
120
111.26
150
139.08
130
120.53
150
139.08
1.30
120.53
130 120.53
1. 50
139.08
150
139.08
3. 7 0
157.62
160 148.35
1.30
120.53
150
139.08
1.50
139.08
130
120.53
1 4 0
129.81
302
(Table C3 continued)
100
92.72
150
139.08
120 111.26
170
157.62
100
92.72
140
129.81
150 139.08
130
120.53
130
120.53
110
101.99
110 101.99
140
129.81
130 120.53
110
101.99
130 120.53
140
129.81
130 120.53
150
139.08
120
111.26
140
129.81
140
129.81
130
120.53
170 157.62
140
129.81
110 101.99
160
148.35
130
120.53
110
101.99
160 148.35
140
129.81
160
148.35
160
148.35
110
101..99
130
120.53
150
139.08
160
148.35
120 111.26
15 0
139.08
130 120.53
120
111.26
120 111.26
180
166.89
180 166.89
160
148.35
120 111.26
130
120.53
150 139.08
110
101.99
110
101.99
100
92.72
Avy (daN)
130.18
118.55
138.36
128.29
Std dev.
19.49
24.01
20.61
19.11
CV(dec.)
0.15
0.20
0.15
0.15
Max (daN)
180.00
166.89
180.00
166.89
Min (daN)
100.00
0.00
100.00
92.72
303
Table CU: 'U"S' 10” Hock II @ SWC-range 8- 10 % g,'g I(Extract)
Impl~?rnerlt
I?l.ow 1JCF
Width (cm)
23.43
24.44
Depth (cm)
12.11
9.33
Tcltal. time (s)
2050.. 00
1033.00
Ef feet:. time (s)
1374.00
710.00
Efficiency (3)
67.02
68.73
Angle of pull
2.2 II 50
22.50
Yoke (cm)
9 0 ., 0 0
1.20.00
Draft (daN)
Pu :Il
Req.
Pull
Req.
160.00
147.82
120.00 110.87
160.00
147.82
150.00 138.58
150. 00
138.58
160.00 147.82
170.00
157.06
170.00 157.06
1 s 0 . 0 0
138.58
160.00 147.82
160.00
147.82
150.00 138.58
150.00
138.58
160.00 147.82
150.00
138.58
160.00 147.82
150*00
138.58
140.00 129.34
150.00
138.58
140.00 129.34
160.00
147.82
150.00 138.58
200.00
184.78
140.00 129.34
150.00
138.58
160.00 147.82
150.00
138.58
150.00 138.58
140.00
129.34
150.00 138.58
l.50.00
138.58
150.00 138.58
140.00
129.34
130.00 120.10
:150.00
138.58
150.00 138.58
160.00
147.82
150.00 138.58
l70.00
157.06
150.00 138.58
150.00
138.58
150.00 138.58
150.00
138.58
160.00 147.82
140.00
129.34
160.00 147.82
160.00
147.82
140.00 129.34
l.50.00
138.58
1~40.00
129.34
l80.00
166.30
140.00 129.34
;160.00
147.82
160.00 147.82
1180.00
166.30
140.00 129.34
170.00
157.06
1150.00
138.58
3. 5 0 . 0 0
138.58
1:30.00
120.10
304
. .
(Table C4 continued)
160.00
147.82
150.00 138.58
160.00
147.82
150.00 138.58
180.00
166.30
140.00 129.34
160.00
147.82
150.00 138.58
160.00
147.82
130.00 120.10
160.00
147.82
160.00 147.82
170.00
157.06
150.00 138.58
180.00
166.30
160.00 147.82
160.00
147.82
160.00 147.82
160.00
147.82
130.00 120.10
160.00
147.82
120.00 110.87
160.00
147.92
120.00 110.87
150.00
138.58
140.00 129.34
160.00
147.82
150.00 138.58
180.00
166.30
130.00 120.10
170.00
157.06
130.00 120.10
160.00
147.82
140.00 129.34
170.00
157.06
150.00 138.58
160.00
147.82
150.00 138.51:
170.00
157.06
140.00 129.34
160.00
147.82
140.00 129.34
160.00
147.82
150.00 138.58
170.00
157.06
130.00 120.10
150.00
138.58
150.00 138.58
160.00
147.82
150.00 138.58
180.00
166.30
140.00 129.34
160.00
147.82
160.00 147.82
Avg (daN)
153.22
141.56
145.52 134.44
Std dev.
19.60
18.11
11.82
10.132
cv (%)
12.79
12.79
8.12
8.12
Max (daN)
200
184.78
170.00 157.06
Min (daN)
90
83.15
120.00 110.87
305
Table C!i:
"SINE 9 Block II @ SWC-range 8- 10 % g/g (Extract)
Implemerh
C u l t i v a t o r H S
W i d t h ( c m )
54 . 50
5 2 . 0 0
Depth (cm)
10 . ii! 5
9 . 0 0
T o t a l . tiine (s)
1 3 4 2 . 0 0
9 5 4 . 0 0
E f f e c : t i v e t i m e (s)
9 5 6 * 0 0
9 1 6 . 0 0
Effic:iency (8)
7 1 e S! 4
9 6 . 0 2
A n g l e o f p u l l
2 2! m 5 0
2 2 . 5 0
Y o k e ( c m )
90 . 0 0
1 2 0 . 0 0
D r a f t (CiaN)
P u l l
R e q .
P u l l
R e q .
1 4 0 * 0 0
1 2 9 . 3 4
100.00
92.39
1 3
1 2 0 . 1 0
8 0 . 0 0
7 3 . 9 1
? ? ? ? ? ? ?
1 5 0 . 0 0
1 3 8 . 5 8
1 0 0 . 0 0
9 2 . 3 9
1 4 0 . 0 0
3 2 9 . 3 4
1 1 0 . 0 0
7 3 . 9 1
3 4 0 . 0 0
1 2 9 . 3 4
1 0 0 . 0 0
9 2 . 3 9
1 4 0 . 0 0
129.34
9c. 00
8 3 . 1 5
1 5 0 . 0 0
1.38.58
1 0 0 . 0 0
8 3 . 1 5
1 4 0 . 0 0
1 2 9 . 3 4
1 0 0 . 0 0
9 2 . 3 9
1 5 0 . 0 0
1 3 8 . 5 8
1 1 0 . 0 0
7 3 . 9 1
14 0 . 00
1 2 9 . 3 4
1.0 0 . 0 0
9 2 . 3 9
14 0 . 0 0
l. 2 9 . 3 4
9 0 . 0 0
8 3 . 1 5
1. 4 0 . 0 0
l 2 9 . 3 4
3. 0 0 . 0 0
8 3 . 1 5
3. 3 0 . 0 0
l 2 0 . 1 0
3. 0 0 . 0 0
9 2 . 3 9
15 0 . 0 0
1 3 8 . 5 8
1 0 Cl . 0 0
9 2 . 3 9
3. 0 0 . 0 0
9 2 . 3 9
1. 0 0 . 0 0
9 2 . 3 9
14 0 . 0 0
l29.34
1. 0 0 . 0 0
9 2 . 3 9
3.3 0 . 0 0
1 2 0 . 1 0
1 0 0 . 0 0
9 2 . 3 9
1.0 0 . 0 0
9 2 . 3 9
3.2 C) . 0 0
8 3 . 1 5
1. 5 0 . 0 0
1 3 8 . 5 8
^ 0 0 . 0 0
9 2 . 3 9
l-2 cl . 0 0
1 1 0 . 8 7
:_ 10 . 0 0
7 3 . 9 1
:1 4 0 . 0 0
:129.34
:. 0 0 . 0 0
9 2 . 3 9
l4 0 . ‘0 0
1 2 9 . 3 4
9 0 . 0 0
8 3 . 1 5
:100.00
9 2 . 3 9
l 0 0 . 0 0
9 2 . 3 9
l2 10 . 0 0
1 1 0 . 8 7
l 1 0 . 0 0
7 3 . 9 1
1 2 0 . 0 0
1 1 0 . 8 7
l 0 0 . 0 0
9 2 . 3 9
: t o o . o o
9 2 . 3 9
:t 1 0 . 0 0
7 3 . 9 1
1 1 0 . 0 0
1 0 1 . 6 3
:t 0 0 . 0 0
9 2 . 3 9
: t o o . o o
9 2 . 3 9
8 0 . 0 0
7 3 . 9 1
: t o o . o o
9 2 . 3 9
:t 0 0 . 0 0
8 3 . 1 5
1 3 0 . 0 0
1 2 0 . 1 0
8 0 . 0 0
7 3 . 9 1
386
(Table C5 continued)
120.00
110.87
100.00
92.39
120.00
110.87
80.00
73.91
100.00
92.39
100.00
92.39
140.00
129.34
100.00
83.15
130.00
120.10
100.00
92.39
100,00
92.39
120.00
110.87
120.00
110.87
120.00
110.87
100.00
92.39
100.00
92.39
110.00
101.63
90.00
83.15
100.00
92.39
100.00
92.39
90.00
83.15
100.00
92.39
110.00
101.63
110.00
73.91
120.00
110.87
120.00
83.15
90.00
83.15
90.00
83.15
100.00
92.39
100.00
92.39
120.00
110.87
100.00
92.39
100.00
92.39
90.00
83.15
120.00
110.87
100.00
92.39
110.00
101.63
100.00
92 * 39
110.00
101.63
100.00
92.39
110.00
101.63
100.00
92.39
Avg (daN)
111.91
103.40
100.00
87.32
Std dev.
18.98
17.53
9.29
8.64
cv (SI
16.96
16.96
9.29
9.90
Max (daN)
150.00
138.58
120.00
110.87
Min (daN)
70.00
64.67
80.00
73.91
307
Table CE; : 'ARARA Black II @ SWC-range t3- 10 % g/g (Extra&)
1mplemen.t
Ridger -MaRA
Width (cm)
2 1 * 57
22.50
Depth (cm)
11.00
11.33
Total time (s)
1120 0, 00
91.0.00
Effective time (s)
624 (8 00
690.00
Efficiency (%)
55,.71
75.82
Angle of pull
'L! 2 . 0 0
2'2.00
Ycike (cm)
90, 00
120.00
Draft (daN)
Pull
Req.
Pull
Req.
100.00
92.72
160.00
148.35
80.00
74.17
150.00
139.08
90.00
83.45
170.00
157.62
100.00
92.72
150.00 139.08
100.00
92.72
180.00 166.89
90.00
83.45
200.00 185.44
120.00
111.26
200.00 185.44
90.00
83.45
130.00
120.53
lC~O.00
92.72
200.00 185.44
110.00 101.99
140.00 129.81
120.00
111.26
150.33
139.08
100.00
92.72
160.00
148.35
120.00 111.26
150.00
139.08
140.00 129.81
160.00 148.35
100.00
92.72
150.00 139.08
90.00
83.45
130.00
120.53
100.00
92.72
140.00
129.81
90.00
83.45
140.00
129.81
90.00
83.45
150.00
139.08
120.00 111.26
130.00
120.53
80.00
74.17
120.00 111.26
140.00 129.81
110.00 101.99
150.00 139.08
130.00
120.53
150.00 139.08
130.00 120.53
100.00
92.72
100.00
92.72
130.00 120.53
140.00
129.81
150.00 139.08
140.00 129.81
120.00 111.26
100.00
92.72
110.00 101.99
110.00 101.99
(Table CG continued)
130.00
120.53
120.00
111.26
100.00
92.72
130.00
120.53
140.00
129.81
100.00
92.72
130.00
120.53
110.00
101.99
120.00
111.26
130.00
120.53
1'20.00
111.26
140.00
129.81
120.00
111.26
130.00
120.53
110.00
101.99
140.00
129.81
120.00
111.26
140.00
129.81
120.00
111.26
150.00
139.08
140.00
129.81
110.00
101.99
130.00
120.53
120.00
111.26
130.00
120.53
120.00
111.26
100.00
92.72
150.00
1.39.08
110.00
101.99
120.00
111.26
130.00
120.53
140.00
129.81
140.00
129.81
140.00
129.81
120.00
111.26
130.00
120.53
130.00
120.53
150.00
139.08
130.00
120.53
120.00
111.26
140.00
129.81
120.00
111.26
140.00
129.81
120.00
111.26
130.00
120.53
160.00
148.35
100.00
92.72
140.00
129.81
120.00
111.26
120‘00
111.26
110.00
101.99
100.00
92.72
Avg [YdaN)
118.50
109.87
137.64
127.61
Std dev.
17.54
16.27
23.35
21.65
cv (81)
14.80
14.80
16.97
16.97
Max (daN)
150.00
139.08
200.00
185.44
Min (daN)
80.00
74.17
100.00
92.72
309
Table C7: 'UCF 10" I3loc:k 111 @ SWC-range 10 - 13 % g/g (Extra&)
i'mplement
Plow UCF
Width (cm)
22.75
22.50
Depth (cm)
Il.80
12.00
Total time (s)
2308.00
2472.00
Effect. time (s)
1.262.00
1634.00
Efficiency (8)
54.68
66.10
AngILe of pull
22.00
2.2.00
Yoke (cm)
90.00
120.00
Draft (ciaN)
Pull
Req.
Pull
Req.
:14 0 ,, 0 13
129.81
140.00 129.81
150 * 00
139.08
150.00 139.08
:t 5 0 0 0 0
139.08
150.00 139.08
l 6 0 0 0 0
148.35
160.00 148.35
:1 6 0 . . 0 0
148.35
150.00 139.08
1 8 0 ., 0 0
166.89
150.00 139.08
190.00
176.16
140.00 129.81
3. 6' 0 . 0 0
148.35
150.00 139.08
2 0 0 . 0 0
185.44
150.00 139.08
1. 5 0 . 00
Z-39.08
140.00 129.81
1 6 0 . 0 0
3.48.35
:i50.00 139.08
1. 7 0 . 0 0
157.62
I.40.00 129.81
160.00
1.48.35
l.50.00 139.08
160.00
l.48.35
3.50.00 139.08
15o.oc1
139.08
1.40.00 129.81
150.00
139.08
140.00 129.81
160.00
148.35
140.00 129.81
160.00
148.35
140.00 129.81
150.00
139.08
140.00 129.81
150.00
139.08
150.00 139.08
1.50.00
139.08
150.00 139.08
160.00
148.35
140.00 129.81
150.00
139.08
170.00 157.62
17~0.00
157.62
150.00 139.08
180.00
166.89
150.00 139.08
160.00
148.35
l50.00 139.08
200.00
185.44
150.00 139.08
150.00
139.08
170.00 157.62
160.00
148.35
160.00 148.35
170.00
157.62
140.00 129.81
310
(Table C7 continued)
160.00
148.35
140.00 129.81
170.00
157.62
150.00 139.08
160.00
148.35
140.00 129.81
160.00
148.35
150.00 139.08
120.00
111.26
150.00 139.08
130.00
120.53
140.00 129.81
150.00
139.08
150.00 139.08
160.00
148.35
130.00 120.53
170.00
157.62
140.00 129.81
150.00
139.08
150.00 139.08
170.00
157.62
150.00 139.08
150.00
139.08
140.00 129.81
170.00
157.62
150.00 139.08
150.00
139.08
150.00 139.08
200.00
185.44
140.00 129.81
160.00
148.35
150.00 139.08
150.00
139.08
160.00 148.35
150.00
139.08
160.00 148.35
140.00
129.81
150.00 139.08
200.00
185.44
150.00 139.08
180.00
166.89
160.00 148.35
160.00
148.35
160.00 148.35
160.00
148.35
160.00 148.35
160.00
148.35
150.00 139.08
160.00
148.35
170.00 157.62
160.00
148.35
160.00 148.35
180.00
166.89
150.00 139.08
Avg (daN)
154.72
143.45
146.67 135.99
Std dev.
15.99
14.82
8.46
7.84
CV(dec.)
9.93
9.93
5.68
5.68
Max (daN)
200.00
185.44
170.00 157.62
Min (daN)
120.00
111.26
130.00 120.53
311
Table C8 :: 'SINE 9 Block III @ SWC-range 10 - 13 % g/g {Extract)
Implement:
3-tine SINE 9
Width (cm)
52.50
52.00
Depth (cm)
10.00
10.00
Total time (s)
1480.00
1:332.00
Effect. time (s)
1116.00
986.00
Efficienc:y (8)
7 5 . ,4 1
74.02
Angle of pull
2 2 . 5 0
22 . 50
Coke (cm)
9
:L 2 0 0 0
.
???????
Draft :daN)
Pull
Req.
Pull
Req.
1. 3 0 . 0 0
120.10
80.00
73.91
1.2 0 . Cl 0
110.87
60.00
55.43
14 0 . 0 0
:129.34
70.00
64.67
1.2 0 . 0 0
110.87
80.00
73.91
1. 3 0 . 0 0
120.10
90.00
83.15
1. 4 0 . 00
129.34
80.00
73.91
14 0 . 0 0
129.34
90.00
83.15
1 5 0 . 0 0
138.58
9c.00
83.15
1 5 0 . 0 0
138.58
80.00
73.91
1 50 . 0 0
138.58
90.00
83.15
1 3 Q . 0 0
120.10
90.00
83.15
1 4 0 . 0 0
1L29.34
9Q.ûO
83.15
13 C) . 0 0
720.10
110.00
101.63
1 30 . 00
120.10
80.00
73.91
12 0 . 0 0
110.87
:100.00
92.39
14 0 . 0 0
129.34
120.00
110.87
1 50 . 0 0
138.58
:120.00 110.87
15 0 . 0 0
138.58
:100.00
92.39
14 0 . 0 0
129.34
l.10.00
101.63
14 0 . 0 0
1.29.34
120.00 110.87
14 0 . CIO
1.29.34
110.00
101.63
130.00
120.10
120.00 110.87
1 5 CI . CI 0
1.38.58
110.00 101.63
130.00
120.10
:too.oo
92.39
130.00
120.10
100.00
92.39
12 01 . 0 0
110.87
80.00
73.91
120.00
110.87
:1 0 0 . 0 0
92.39
120.00
110.87
:100.00
92.39
140.00
129.34
100.00
92.39
140.00
129.34
80.00
73.91
312
(Table C8 continued)
150.00
138.58
120.00
110.87
130.00
120.10
90.00
83.15
140.00
129.34
80.00
73.91
140.00
129.34
110.00
101.63
140.00
129.34
110.00
101.63
150.00
138.58
100.00
92.39
14'0.00
129.34
110.00
101.63
140.00
129.34
110.00
101.63
110.00
101.63
100.00
92.39
140.00
129.34
100.00
92.39
200.00
184.78
80.00
73.91
150.00
138.58
100.00
92.39
140.00
129.34
110.00
101.63
130.00
120.10
110.00
101.63
140.00
129.34
110.00
101.63
140.00
129.34
90.00
83.15
120.00
110.87
90.00
83.15
130.00
120.10
110.00
101.63
120.00
110.87
80.00
73.91
140.00
129.34
110.00
101.63
140.00
129.34
110.00
101.63
120.00
110.87
90.00
83.15
130.00
120.10
100.00
92.39
140.00
129.34
100.00
92.39
140.00
129.34
130.00
120.10
110.00
101.63
100.00
92.39
110.00
101.63
90.00
83.15
Avg (daN)
130.45
120.52
97.76
90.32
Std dev.
13.22
12.21
14.53
13.43
CV(%)
9.67
9.67
14.80
14.80
Max (daN)
200.00
184.78
130.00
120.10
Min (daN)
110.00
101.63
60.00
55.43
,313
Table C9: *AR.A,RA B~OCIC 11x @ SWC-range 10 - 13 % g/g W:xtract)
Implement
Ridger ?!iRARJi
Wi.dth (C:m)
23.00
23.00
Depth (cm)
10.33
1.0.00
Total time (s)
1520.00
1238.00
Effect. time (s)
fi 90 . 0 0
920.00
Efficiency (%)
El8 . 5 5
74.31
Angle of pull
22.00
22.00
Yoke (cm)
90.00
120.00
Draft (daN)
PL11 1
Req.
Pull
Req.
83.45
100.00
92.72
92.72
120..00
111.26
101.99
10 0 * 0 0
92.72
11.1.26
10 0 * 0 0
92.72
92.72
1 0 0 * 0 0
92.72
92.72
1 5 0 S 0 0
139.08
111.26
1 2 0 (I 0 0
111.26
l-1.26
10 0 * 0 0
92.72
139.08
9 0 ., 0 0
83.45
139.08
10 0 I) 0 0
92.72
139.08
10 0 <) 0 0
92.72
139.08
1 0 0 ., 0 0
92.72
92.72
1 0 0 I) 0 0
92.72
129.81
1 1 0 <) 0 0
101.99
120.53
100.00
92.72
111.26
!3 0 ., 0 0
83.45
101.99
1 0 0 <) 0 0
92.72
92.72
1 0 0 <) 0 0
92.72
1X1.26
1 0 0 <, 0 0
92.72
92.72
12 0 * 0 0
111.26
101.99
9 0 ,. 0 0
83.45
129.81
1 0 Q * 0 0
92.72
111.26
!3 0 , 0 0
83.45
111.26
140.00
129.81
i32.72
100 .oo
92.72
120.53
1:30.00
120.53
132.72
100.00
92.72
92.72
80.00
74.17
111.26
100.00
92.72
192.72
100.00
92.72
3 1 4
3.
.>w
(Table C9 continued)
130.00
120.53
100.00
92.72
.<,
100.00
92.72
90.00
83.45
100.00
92.72
90.00
83.45
100.00
92.72
90.00
83.45
,,
10.0.00
92.72
90.00
83.45
100,00
92.72
90.00
83.45
100.00
92.72
100.00
.1
92.72
110.00
101.99
100.00
92.72
100.00
92.72
100.00
92.72
100.00
," *
92.72
110.00
101.99
100.00
92.72
100.00
92.72
100.00
92.72
80.00
74.17
100.00
92.72
100.00
92.72
100.00
92.72
90.00
83.45
90.00
83.45
100.00
92.72
100.00
#~
92.72
90.00
83.45
90.00
83.45
100.00
92.72
120.00
111.26
90.00
83.45
*
100.00
92.72
130.00
120.53
100.00
92.72
100.00
92.72
100.00
92.72
90.00
83.45
110.00
101.99
90.00
83.45
120.00
111.26
100.00
92.72
100.00
92.72
100.00
92.72
100.00
92.72
90.00
83.45
100.00
92.72
90.00
83.45
110.00
101.99
100.00
92.72
Avg (daN)
109.88
101.98
100.00
9 2 . 7 2
Std dev.
16.16
14.99
13.19
12.23
C'V ( 8 )
14.60
14.60
13.10
13.10
Max (daN)
150.00
139.08
150.00
139.08
Min (daN)
90.00
83.45
80.00
74*17
Table. ClO: Output taxample of the CEEMAT data processing
system (WCF' plowr -Bugutub rice field)
Tl
T2
T3
T4
T5
#
hour:min Second Draft:
Radar
hr ::mn
S
daN
pulse
0103.
12:36.
0 0 0 0 .
108.5
53.00
0103.
12:36.
002.0
131.9
138.0
0103.
12:36.
004.0
102.7
162.0
0103.
12:36.
006.0
135.8
126.0
0103.
12:36.
008.0
132.0
122.0
0103.
12:36.
010.0
131.8
099.0
0103.
12:36.
012.0
158.3.
130.0
0103.
12:36.
014.0
122.3
115.0
0103.
12:36.
016.0
086.0
146.0
0103.
12:36.
018.0
152.6
147.0
0103.
12:36.
020.0
108.5
148.0
0103.
12:36.
022.0
094.4
140.0
0103.
12:36.
024.0
165.8
149.0
0103.
12:36.
026.0
117.2!
134.0
0103.
12:36.
028.0
26.84
28.00
0103.
12:36.
030.0
23.49
0.000
0103.
12:36.
032.0
1.195
0.000
0103.
12:36.
034.0
2.690
0.000
0103.
12:36.
036.0
17.38
18.00
13103.
12:36.
038.0
23.62
48.00
'3103.
12:36.
040.0
079.2
08.00
J3103.
12:36.
042.0
099.4
48.00
13103.
12:36.
044.0
148.Î
125.0
3103.
12:36.
046.0
1 0 1 . 0
096.0
0103.
12:36.
048.0
122.3
117.0
0103.
12: 36.
050.0
097.4
143.0
0103.
12:36.
052.0
099.8
196.0
0103.
12:36.
054.0
092.0
104.0
0103.
12:36.
056.0
088.5
082.0
0103.
12:36.
058.0
085.2
102.0
0103.
12:37.
0000.
104.1
127.0
0103.
12: 37.
002.0
131.8
134.0
0103.
12:3-l.
004.0
105.5
123.0
0103.
12:37.
006.0
136.'7
181.0
3103.
12:37.
008.0
103.,4
164.0
316
(Table Cl0 continued)
0103.
12:37,
010.0
132.7
139.0
0103.
12:37.
012.0
090.8
144.0
0103.
12:37.
014.0
117.8
143.0
0103.
12:37.
016.0
121.8
166.0
0103.
12:37.
018.0
096.8
138.0
0103.
12:37.
020.0
116.1
088.0
0103.
12:37.
022.0
158.8
113.0
0103.
12:37.
024.0
166.1
100.0
0103.
12:37.
026.0
169.7
40.00
0103.
12:37.
028.0
138.9
09.00
0103.
12:37.
030.0
108.3
094.0
0103.
12:37.
032.0
133.5
59.00
0103.
12:37.
034.0
146.6
144.0
0103.
12:37.
036.0
127.2
094.0
0103.
12:37.
038.0
180.8
110.0
0103.
12:37.
040.0
187.5
115.0
0103.
12:37.
042.0
142.8
092.0
0103.
12:37.
044.0
171.0
0.000
0103.
12:37.
046.0
165.7
0.000
0103.
12:37.
048.0
120.0
0.000
0103.
12:37.
050.0
130.1
0.000
0103.
12:37.
052.0
076.0
0.000
0103.
12:37.
054.0
103.1
08.00
0103.
12:37.
056.0
123.1
082.0
0103.
12:37.
058,o
115.4
073.0
0103.
12:38.
0000.
116.0
65.00
0103.
12:3a.
002.0
079.0
137.0
0103.
12:38.
004.0
127.4
072.0
0103.
12:38.
006.0
130.5
090.0
0103.
12:38.
008.0
136.4
089.0
0103.
12:38.
010.0
162.0
071.0
0103.
12:38.
012.0
170.8
086.0
0103.
12:38.
014.0
134.1
085.0
0103.
12:38.
016.0
136.9
63.00
0103.
12:38.
018.0
147.9
39.00
0103.
i2:38.
020.0
07.63
0.000
0103.
i2:38.
022.0
3.016
0.000
0103.
12:38.
024.0
0.872
0.000
0103.
12:38.
026.0
43.77
0.000
317
(Table Cl0 continued)
0103.
12:38.
028.0
091.7
53.00
0103.
12:38.
030.0
099.9
114.0
0103.
12:38.
032.0
69.18
120.0
0103.
12:38.
034.0
090.4
114.0
0103.
12:38.
036.0
121.9
100.0
0103.
12:38.
038.0
096.5
100.0
0103.
12:38.
040.0
093.8
081.0
0103.
12:38.
042.0
106.9
45.00
0103.
12:38.
044.0
089.3
092.0
0103.
12:3#8.
046.0
104.8
101.0
0103.
12:38.
048.0
2.494
5.000
0103.
12:38.
050.0
10.81
0.000
0103.
12:38.
052.0
0.934
33.00
0103.
12:38.
054.0
0.414
63.00
0103.
12:38.
056.0
32.81
3.000
0103.
12:38.
058.0
078.6
132.0
C103.
12:39.
0000.
113.5
176.0
0103.
12:39.
002.0
107.1.
182.0
0103.
12:39.
004.0
091.0
189.0
0103.
12:39.
006.0
075.9
183.0
0103.
12:39.
008.3
080.5
198.0
0103.
12:39.
010.0
088.51
184.0
0103.
12:39.
012.0
0 92 . 6
174.0
O103.
12:39.
014.0
110.8
167.0
0103.
12*3Q
*.. 2.
016.0
105.3
201.0
0103.
12.39
. . 2.
018.0
086.7
154.0
0103.
12:39.
020.0
114.5
174.0
0103.
12:39.
022.0
0 9 2 . 6
191.0
0103.
12.39
. . 4.
024.0
077.6
141.0
0103.
12:39.
026.0
080.8
144.0
0103.
12:39.
028.0
103.4
170.0
0103.
12::39.
030.0
109.3,
169.0
0103.
12::39.
032..0
092.3
108.0
0103.
12:39.
034..0
078.6
108.0
0103.
12::39.
036.0
096.2
154.0
0103.
12 : :3 9 .
038-O
089.9
121.0
0103.
12::39.
040,.0
090.2
127.0
0103.
12 : :3 9 .
042,.0
084.:3
176.0
3 1 8
(Table Cl0 continued)
0103.
12:39.
044.0
107.4
111.0
0103.
12:39.
046.0
67.62
130.0
0103.
12:39.
048.0
083.2
147.0
0103.
12:39.
050.0
071.3
140.0
0103.
12:39.
052.0
070.7
127.0
0103,
12:39.
054.0
60.34
120.0
0103.
12:39.
056.0
51.88
088.0
0103.
12:39.
058.0
65.32
084.0
0103.
12:40.
0000.
56.15
076.0
0103.
12:40.
002.0
071.2
101.0
0103.
12:40.
004.0
073.5
57.00
0103.
12:40.
006.0
30.47
31.00
0103.
12:40.
008.0
45.81
0.000
Storage
.L *..
c
--. s . . . . . ~
unit
P-
i
i
1
1P/i 1 :...:. : . . . . .- . ..-. : . . . . . . . . .7
R P M
Fuel cmmmption
Pull (3 pQiow
-T
--r-- 7-
A
Fgm EDLOG
Figure C3: CEEMAT data recording and processing _
APPENDIX D
APPENDJX D
FEEDLNG SYSTEMS .AVAILABLE TN M’EST-AFRICA
Table Dl: FU for msiintenance FUM of working cattle
-----
--=zzz.~----.-
11-----1---
Weight
Feeding Units (FU)
&e:)
y - y - - - - - - -
1 1 _ - - - - .-._----
- -
200
1.9’5
250
2.30
300
2.60
35’0
2.9:5
4010
3 . 2 5
450
3.55
500
3 . 8 5
--I
-=e-
..I-~-~~-
Source: CEEMA?‘, 1975
320
321
Table D2: Composition of some African feeds (As-Fed Basis)
-
Feeds
Dh
100 % DM bas,is
T-FU
-
-
-
-
n,
P
C a
Green grass pasture
Sudan (Andrononon gavanus)
18.0
6.2
0.02
0.35
0.66
Guinea (Panicum ma.ximum)
16.6
6.7
0.27
0.45
0.57
Elephant (Pennisetum purnureum)
18.5
6.9
0.26
0.60
0.58
Digitaria
15.3
9.4
0.23
0.49
0.60
Mixed fi-esh shoots
17.7
18.2
0.45
0.29
0.77
Mixed Young pasture
24.8
5.8
0.24
0.31
0.61
Mixed Young pasture + legumes
25.6
7.5
0.10
0.87
0.72
- _ .
hier pasture (standing hay)
suclan #1
29.7
4.2
0.12
0.34
0.57
Sudan#2
34.6
3.5
0.15
0.39
0.63
GUiJSfX
34.7
3.6
0.22
0.54
0.57
Digitaria
32.7
2.5
0.33
0.64
0.54
Mixed - average quality
30.8
3.2
0.18
0.30
0.38
Mixed - pour quality
39.6
2.0
0.10
0.48
0.35
Para gr&s (Br&hiaria)
28.6
0.08
0.22
0.58
-”
3rred hay - average quality
- -
64.0
2.6
0.09
0.69
0.46
-.-
;traw
SUdan
93.2
0.06
0.39
0.49
Guinea
95.7
0.08
0.48
0.49
Elephant
93.2
0.04
0.17
0.15
“Ivory Coast hay”
87.0
0.24
0.42
0.44
Brachiaria
94.1
0.06
0.62
0.37
Mixed &andina or tut)
96.4
0.10
0.40
.-
'0.30
?reen legumes
Stvlosanthes gracilis
21.0
12.4
0.23
1.65
0.73
Stylosanthes gracilis - older
44.0
5.9
0.18
1.22
0.72
Centrosema uubescens
21.8
!5.5
0 . 2 1
1.04
0.70
Centrosema pubescens - older
29.4
14.7
0 . 1 8
1.15
0.68
Egume hay
Stylosanthes
92.8
6.1
0 . 1 7
0.54
Centrosema
87.0
.2.5
0.27
0.54
Peanut hay (as-f4 basis = afb)
92.1
6.3
0.13
0.40
Bean hay (f)olichos 1ablab)afb
93.9
14.2
1.2
0.56
Cowpea hay (Vinnas sinensis)afb
89.0
7.6
0.26
0.61
Mai~ - whole Young plant
21.9
4.8
0.20
0.34
0.89
Maize - sarne + cob forming
48.2
0.25
0.69
0.75
Maize - dry Ieaves+stalk+husk
93.6
ii
w
m
0.53
Millet (Sorghum almum) - whole
27.4
Ii.9
0.54
0.62
0.70
Millet - whole plant - younger
15.0
4.7
0.35
0.30
0.41
Millet - same plant - older
30.9
2.7
0.28
0.23
0.42
Millet - ~IV. leaves + stalks
85.0
-
1.9
0.14
0.55
0.36
L_
-
-
ource: Watson, P.R. 198 1,
,322
Table D3: Composition of some African ffeeds (As-Fed Bah)
-SE--
-
WZ%J-=L
-6
Feetls
D M
100 % DM basis
1
-
-s-z
DP
Ca
----.
--_---
-
- -----
P
----- -. --,-
ksirls
Maize (Coastal West Africa)
87.0
6.61
0.34
0.03
1.05
Maize (Savanna West Af+ica)
92.6
7.73
0.33
0.02
1.08
Millet (Burkina Faso)
91.8
'1.3
0.37
0.40
0.97
Paddy rice
87.3
4.7
0.26
0.06
0.82
Sorghum
89.9
.5.9
0.29
0.02
0.92
Forlio
3.6
0.06
0.07
0.86
-~-
--~--------
88.4
-----.-
-._-_ -
-
--.-
Brans
Maize (traditional milling)
86.0
6.23
0.72
0.06
0.92
Millet
92.3
9.0
0.61
0.09
0.86
Rice
88.6
4,.16
0.41
0.09
0.42
Sorghum
- - - - - - - - - - - - - -
90.7
- - -
6.8
0.64
----- - - -
0.09
0.78
- - -
Cakes md me&
Peanut cake
92.7
47.3
0.65
0.11
1.93
Cottonseecl meal (industria])
-
- - - - -
- - - - - -
--.-
212
1.2
-.--_ -
-
0.15
0.56
-I_
seeds
Cottonseetl (whole)
94.4
!2.6
0.49
0.11
II .05
Peanuts
90.8
14.0
0.29
0.12
II. Il
Cowpeas (Vigna)
90.4
19.1
0.42
0.17
Il .06
Beans 1(Dolichos lablab)
- - - -
--_-_--II
89.6
----
19.9
0.29
-._-.-.
-
-
0.26
- - -
L.-caves
Banana (Tanganyika)
16.2
1.26
0.03
0.17
0.14
Manioc (Cassava)
27.3
9.4
0.51
0.92
0.64
Sahel acacia
60.8
10.1
1.90
3.05
0.46
Acacia Albida - dry leaves
92.8
7.2
0.15
0.23
0.69
Baobab - dry, West Africa
- - - -
---s--v
91.0
.----
5.7
0.40
-.---.
- - -
1.08
0.41
-_I
Miscellaneous
Brewcrs wet meal (local)
30.7
17.6
0.41
0.26
0.80
Brewers meal dried (local)
92.3
21.6
0.33
0.03
0.87
BreadfGt (Artocarpus m.)
30.0
0 0
0.04
0.02
0.3 1
Mango (greenlpulp) (munis)
14.5
0.16
0.01
0.02
0.18
Yam (fresh)
36.8
0 0
0.05
0.11
0.39
Yam (dlnecl)
89.6
5.04
0.12
0.19
OI.59
Cassava (manioc) - fresh
34.2
0.0
0.04
0.04
OI.34
Cassava - dried
88.0
0.0
0.08
0.09
Ot.98
Rice husks (&a.@
92.0
0.12
0.08
0.08
0.29
Ek3.n bulls (Vim sinensis)
39.3
5.60
0.44
0.85
Cocoa :puds
92.1
4.13
0.15
0.20
0.46
Maize cobs
88.3
0.0
0.03
0.01
0.40
Maize .. mature cob + grain chop
93.3
5'.4
0.01
0.22
0.84
Whole ‘banana
24.1
0..68
0.02
0 . 0 1
0.26
Banana pulp
24.1
0.55
0.02
0.02
0.25
Sugar cane moiasses
83.3
0.9
0.03
1.49
1.04
Nere powder (Mali)
-
1 .oo
Rice straw
0.35
-
-
-
-
c--x=
92.5
0.19
0.19
0.29
- - - _I=D - - =Es
Source: Watson, P.R. 198 1.
323
Table D4: Field operation execution time
1. Tillage
Manual: 28 a 35 man-dayska (8 hrs/day)
Animal traction: 6 a 8 ox-team-day/ha (4 a 5 hrs/day)
2. Manual seeding
Millet
20 man-day/ha
Maize
15 man-day/ha
Sorghum
3 5 man-day/ha
Groundnut
24 man-day/ha
3. Manual Weeding
Millet
20 man-day/ha
Maize
32 man-day/ha
Sorghum
17 man-day/ha
Groundnut
3 8 man-day/ha
4. Manual Harvesting
Millet
13 man-day/ha
Maize
13 man-day/ha
Sorghum
10 man-day/ha
Groundnut
18 man-dayi’ha
(Source: Le Moigne, 198 1)
“2
-.-x”..“‘-“mm~l.~--
_ - - _ . _ . - “ . , . - -
APPENDUIE
. -l
APPENDIX E
.
ONE COMPLETE SESSION OF THE EXPERT SYSTEM PROGRAM
EXPERT SYSTEM
MSU: Midigan Statt University
Figure El: Screen 1
324
3 2 5
Figure E2: Screen 2
3 2 6
Figure E3: Screen 3
--
I
~~.u--~u”NIIIIIIR(~
.--.
“ . _. .I_. _ -m..,-.-
-mm”.
327
-=SiBZZE-
mnj.m,-
II
iÏÏ’
..v
3
.:..
EXPERT SYSTEM FOR ANIMAL TRACTION EVALUATION
-,-v.
._L------p.---
The farmcr’s name is optional
Figure E4: Screen 4
328
Gewraphical Location and Fart Chractuistics
DB-1 9-t 997
I. Agro-eco-systcm Zone:
r-7
4
- Boulandor 1
1. Farm Sizc(ha]:
1 6.00 1
. Numbcr of Farm Workers:
r--iq
i. Main Cash trop:
Figure E5: Screen 5
329
Geographhal Location and Farm Charactuistics
08-l 3-l 997
.-
Types of Crops Grown
Type wf Soi1 @ the
@ the Fat-mi Level
Agro-ecosystem Level
Figure E6: Screen 6
330
Figure E7: Screen 7
331
- - , - -
I I - ~
r
UCF Moldbcaard Rw
AI
I component 1
Ratlng for field work:
Figure ES: Screen 8
3 3 2
.-
Figure E9: Screen 9
3 3 3
Figure ElO: Screen 10
c
_
._ .I_.--._.
..._.” ---.
.-...
“.*~..“,-“--.m”..-
334
Figure Ell: Screen 11
335
Farm Ekpipment Inwdment
-
-
II --- ~
Total Imglomcnt Cort
0
\\ 881df
Figure E12: Screen 1.2
.r
336
Figure E13: Screen 13
1.
-Il-.---I--P
“..
- - -
-
.s.‘-IIII,LIIP--
_- __.-_.____ “.1--.II__.-.- ..-- -““___. ..“.” -, l”,.- -1-1.-.--
r
338
P
F E
I R
EA
L 1
D 10
- _..--_ -. _
22.6
.---_._-_
17.6
-".-.-_ .-.
34.191
.---- -_ -i
339
_____ __..._,..-_.._._.. _ . . ..__ _ .._.. . . . . ._.
..__” _...-..,...
_ _ _ _ . . . . . _
. .._. __.__“.” . . ..-...<....-.....
_ . _ . . . -
. . _ . _ . _ _ . .
_ _ . . . . .
.
t
f
Figure E17: Screen i 7
342
Resuhd
FA- Fwd
- -I~F&
-
@‘U-j:
-_-
AvüikMe
.--.y
feed:
+.“....,.-.--
Figure E19: Screen 19
343
3 4 4
Figure E21: Screen 21
.
- .u-. .‘““.*I-“--lc
f
>.
REFERENCES
. “ .
R E F E R. E N C.: E S
AID, 1997. World water for agriculture (Climatc, Prccipitation Probabilities and
adequaçies for rainfed agriculture). Agency CI~ Intcr~mlional De\\~elopmcnt, \\j’ashingl:on
DC. L%A.
ATRIC: 1986 f’ield performance of maresha. Test report ‘Nos 4, 16 and 18. AIRIC:
Ethbpia.
Andersen, F.M. 1985. Draughf animal pou;er systems in %-lb-Saharan Afr-ica: their
production l:mpact and rescarch needs. In Copland, J W. (ed.) Draught Animal Pc~er li71
predrtction. Procee.dings of an internatiioral WCH k shop held at James Cook University,
Tow-îsville, Qld, Australia, 1 O- 16 July 1985. Canbei,ra AclAR Proceedings No. 10, 26-
31.
ARC. 1980. The nutrient lequirements of rurnbant li\\,estcKk Slough: Commonwealtla
Agi~uIfural Ri.ireaux.
Barwell and Ayre, 1982. The harnessing of dr-aught animais. Intermediate Technology
Publications, Ardington.
Retker J. and Klut&ac:.h H.D 1991. The role of donkeys in Agricul?ural mechanization in
Niger: Potsntial and limitations In Fielding ad Pcarson (eds), Donkeys, hiules, H~tses in
Tropic:al Aqicultural Dcvelopnxnt, CTV?vf Ed~nburgh: 1991.
Campbell, J 1930. Dipple s?ic.ks, dorkeys and diesels. IRRl, Los Banos, Philippines.
CFXMAT 1995. klanuel de cult~m avec tractjon arrimale: techniques rurales en .Urique.
Cadre d’Etudc et d’Experimentation du hfachinisme Agricole Tropical (CXEMAT) et
h%nktere de la Cooperation, Parc de Tout-voie, Antony, France. 336~.
CEEMAT 1974. Aide mcmeoire du moniteur de la culture atîelee. Centre d’Etude et
$‘Experimerttation du Machinisme Agricole TropicaE (CEEMAT) et Ministere de la
Cooye.ration, Parc de Tourvoie, Antor~y, France.
CEEWIAT/‘P’AC~.
1972. Tlsc ctnployment of draught animals in1 agriculture. Rome, Italy
346
347
C’IMMYT. 1984. Report of a networkshop on draught power and animal feeding. 4-6
October 1983, Ezulwini, Swaziland. Networking workshop Report 2. CIMMYT Eastern
and Southern Afnca Economies Programme: Mbabane, Swaziland.
Charreau, C. and Nicou, R. 1971. L’amelioration du profil cultural dans des sols sableux et
sabla-argileux de la zone Tropicale seche Ouest-Africaine et ses incidences Agronomiques
(d’apres des travaux des chercheurs de I’IRAT en Afrique de l’Ouest). Agronomie
Tropicale 26:903-978.
Conroy D. 1992. Advanced training techniques for oxen Tiller’s TechGuide. Tillers
International. Kalamazoo Michigan, USA.
Conroy D and Rice P. 1992. Selecting and training axer-. Tiller’s TechGuide. Tillers
International. Kalamazoo, Mjchigan, USA.
Crossley, P. and Kilgour, J. 1983. Small farm mechanization for developing countt’ies
John Wliley and Sons, Chichester, UK 2S3p.
RERBAC. 1992. Rapport d’activites. Campagne agricole 1991-l 992. Ziguinchor, Senegal.
Devnani, R.S. 1981, Design considerations for hamesses and yokes for draught animais
Technical bulletin no C%%E/8 1/30.
Dibbits, H.J. 1993. Human and draught anima.l power in trop production. Past experiznces
and outstanding problems. In Human and Draught Plnimal Power in Crop produc.tion.
Workshop proc.eedmgs, Harare, Zimbabwe, 18-22 January, 1993. Silsoe Research
Jnstitute-Agricultural Research Food Council. CEE-FAO, Rome 1993, Italy.
Ducreux, A. 1984. Caracterisation mecanique des sols sableux et sablo-argileux de la zone
tropicale seche de l’Afrique de l’Ouest. Etude d’un prototype d’outils permettant de les
travailler en periode seche. These de Docteur Ingenieur en Sciences Agronomiques-
Option Pedologie. Academie de Montpellier, Universite des Sciences et Techniques du
Languedoc. Montpellier, France.
Eicher, K.C. and Baker, D.C. 1982. Research on agricultural development in Sub-Saharan
Africa: A critical survey. MSU International Development Paper no 1.
Fall A. 1995. Factors affecting feed intake, energy expenditure and work output of oxen
and bulls used for draught purposes in Semi-Arid West Africa. Submitted to the University
of Edinburg in partial fulfiilment of the requirements for PhD degree.
Fall, A. 1985. Situation actuelle de l’environnement et de l’utilisation du parc de mater-iels
de culture attelee en Basse Casamance. Memoire de titularisation. Departement de
Recherches sur les Systemes de Production et Transfert de Technologie en Milieu Rural.
JSRA. (document interne).
--
.----s
-1
348
Fall, A. 1’390. Adoption et Principales cxmtraintes a la dif%rsictn des equipmcnts de traction
animale en Basse Casamance, Senegal. In. P. Starkey, M. Goe and A. l’aye (eds), Animal
traction for Agriculture Development. Wagcningen: Technical Center for Agricultural and
thF@JatiOIl (k%@-EEC hJJli2 ~OM3ltiOll).
FALL, A. 1990. Rolcs ac&.rels et potentiels des forgerons traditionnels dans la
mainfenance du materiel agricole en Basse Casamance 1 Senegal. In: P. Starkey, M. Goe
and A. Faye (eds), Animal traction for Agriculture Dçve!opment. Wageningen: Tcçhnical
Cent~- for Agricuhural and Cooperatiun (ACP-EEC Lorne Convention).
Fail, 13.N. 198 1. Etude sur la situation et les perspectives de l’industrie du machinisme
agricole au Senegal. Premiere consultation regionale pour le developperncnt de l’industrie
du machinisme agricole en Afrique. ON1[_!Dl-O‘CTA-CI:A-FIZO.
F.40. 1972. Manrral on the Emplo:yment ofdraught Animais in Agriculture. CEEMAT
FAO. 198Z!. Report of the FACI Expert Consultation w-r Appropriate use of animal energy
in Agriculture in Africa ami in Asia FAO. Rome.
Ffoulkes D. and Bamualim, 1989. Improving the nutrition level of draught animals usine,
availatrle feeds. In Hoffrnan D., Na.ri J. and Petheram .R.J. (eds), Draught Animais in Rural
Development. Proceedings of an Iternational Research Symposium, Cipansas, lndonesia.
ACIAR Proceedings no 27: 134-I 45.
Fath, D.H. 1990. Fundamentals of soi1 science. 8th edition. Wiley & Sons Publisher. New
York.
Gae, %I.R. 1983. Cu;rent status of research on animal ~traction. n’orld Animal Rer.;iew, 45,
2-17
Gae, M.R. 1987. Animal traction on smallholder farrns in Ethiopian highlands. University
Microfilms International: Ann Arbor, Michigan,
Goe, M.R. 1930. Overcoming constraints to animal traction through a collaborati\\.e
researçh network. In: P.H. Starkey, M.R. Goe and A. F’aye (eds). Proceedings ofthe third
WAATAN. 7-12 July 1988, Saly, Senegal. Animal Trac:tion for Developrnent. CTA.
Goe, MR., and ,McDowell, R.E. 1980. Animal Traction: Guidelines to utilization. Corne11
International Agriculture Mimeo. Ithaca, N. Y. Corne11 University.
Hagard:, :M. 1987. Le parc de mater:iels de culture attelee et des possibilites de sa
maintenance dans le depar-tement de Nioro: resultats d’enquetcs. Document de travail
1987:3. Departernent Systemes et Transfert, ISRA, Dakar, Senegal. 45~.
i
349
Havard M., Le Gal J.P. and Pirot R. 1994. Towards a decision support system integrating
different kinds of knowledge to choose and manage field equipment. U.R. Genie
Agronomique et Mecanisation, CIRAD-SAR,
Montpellier, France.
Havard M and Le Thiec G. 1995. Animal traction in Western Africa and Madagascar.
Influence of the environment on its develoyment and use. Communication in The
challenge of Animal Traction. ATNESA Network, 4-9 December 1995, Kenya. 1 1 p.
Hillel, D. 1982. Introduction to soi1 physics. Academi Press, Inc. New York, USA.
Hopfen, H.J. 1960. Farm implements for Arid and Tropical Regions. Agricultural
Development paper No 67. FAO,’ Rome, Italy.
Jaeger, KW. 1985. Agricultural mechanization: The economics of animal traction in
Burkina Faso. Doctor of Philosophy thesis submitted to IFPRI. Standford University
Jackson, P. 1990. Introduction to Expert Systems (2”d ed.). Addison-Wesley Pubhshing
Company, N.Y.
Jury W.A., Gardner W.R. and Gardner W.H. 1991. Soi1 physics. 5th Edition, Wiley &
Sons. New York.
Kartiarso, M.D. and Teleni, E. 1987. Nutritional study using a treadmill at James Cook
University. DAP Project Bulletin. 3 : 1- 15.
Keith M. 1992. Training Young steers. Tiller’s TechGuide. Tillers International.
Kalamazoo MI.
Kemp, D.C. 1987. Draught animal pow-er: some recent and current work. Worid Animal
Review. 63:7-14.
~LL.awrence, P.R. 1985. A review of the nutrient requirements of draught oxen. In Copland,
’
J.W. (ed.) Draught Animal Power for production: Proceedings of an inte.rnational
workshop held at James Cook University, Townsville, Qld, Australia, JO- 16 July 1!?85.
Canberra: ACIAR Proceedings No. 10, 59-63.
Lawrence, P.R. and Pearson, A. 1985. Factors affecting the measurements of draught
force, work output and power of oxen. J. agric. Sci. Camb. 105:703-714.
Lawrence, P.R. and Pearson A. 1989. hsfeasurement of Energy Expenditure in working
Animals: Methods for different c.onditions. In Hoffman D., Nari J. and Petheram R.J.
(eds), Draught Anima]s in Rural Development. Proceedings of an Itemational Research
Symposium, Cipansas, Indonesia. ACIAR Proceedings no 27: 155- 165.
Lawrence, P.R. and Pearson R.A. 1993, E’xperimental methods in draught animal science:
The need for standardization. In Human and Draught Animal Power in Crop production.
Workshop proce.edings, Harare, Zimbabwe, 18-22 January, 1993. Silsoe Research
Institute. Agricuftural Research Foo’d Council CEE-FAO, Rome 1993, Italy.
Lawrence, P.R. and Stibbards R.J. 1990. The encrgy c.osts of walking, carrying and
pulling Ioad:; on flat surfaces by Brahman cattle and swamp buffalo. Animal Produl:tion,
50: 29-39.
+( Lawrenov P R. and Zerbini E3. 1993. Recent trends in research on draught animal nutrition,
1 JR Huma:n and Braught A~limal qower in Crop production. Workshop proceedings,
Harare, Zimbabwe, 18-22 Januat-y, li993. Silsoe Research Tnstitute-14gricultural Research
Food Council. CEE-FAO, R.ome 1993, ‘Italy.
I,ce J. V. d., Udo W. M. J. and Brouwer B. 0. 1993. Design and validation of an Animal
Traction Module for a smallholder L,ivesteck Systems Simulation Model. Agricultural
Spstems 43 ( 1993) 199-227.
Le Moigne, M. 198 1. Contralintes posees par l’insertion de !!a mecanisation dans fes unites
de production agricole en zone sahelienne. Etude methodologique. CEEMAT. Paris,
France.
Martin, D., and Teleni, E. 1989. Fatigue in buffaloes QII diflèrent work loads. DAP Project
BufPetin, ‘3, 2-6.
Le Moigne, M. 1984. Missio:n d’appui au programme Mijchiisl7le Agricole en Casamance.
CEEMAT. 33p.
Le Thiec, G. 1991. Les hatmachememts pour bovins - “‘De l’Europe a l’Afrique”. CIRAD-
SAR, CEIEMAT, Montpellier, France.
Le ‘IXe,, G. and Havard M. 1995. Quels jougs pour les bovins de trait en Afrique Sub-
Saharienne? Communication in The challenge of AnimaU Traction. ATNESA Network, 4-9
Deçember 1995.
Le Thiec G a:nd Havard M. 1995. Developing suitable ylokes for draft oxen in Sub-Saharan
Africa. Communication in Meeting the challenges of Animal Traction. ATNESA Network,
4-9 December 1995. I Op.
M.A..T. 1,985. Special animal traction. No 91. Juillet-Septembre.
M.A.T. 1988 Activites CEEMAT 1986-87. CEEMAT. Parc de Tourvoie, Antony,
France.
351
Matthews, M.D.P. and Pullen. 1975. Cultivation trials with ox-draw-n irnplernents using
Xdarna cattle in the Gambia. Tropical Agriculturaf Engineering Information, Report series;
Overseas Department. NIAE, Silsoe, Bedford, England.
Mbengue E%M. 1986. LA mecanisation de la transformation des cereales au Senegal:
Aspects techniques et nutritionnels, I§RA CNRA Bambey. DRSAEA. Dakar, Senegal
MDR. 1986. Etude du Secteur Agricole: Plan Cerealier. Ministere du Developpement
Rural, Senegai.
Monnier, J. 1983. Pour un nouveau developpement de la culture attelee en region. Lecons
des experiences acquises au Senegal, Madagascar, Cote d’ivoire. Les roles de la
R.echerche et de la Vulgarisation. Montpellier, France.
Morris J. 1983. The ec,onomics of small farm mechanization in Crossley P. and Kilgour J.
(eds). Wiley and Sons. NY.
Montoroi J.P. 199 1. Etude du comportement hydrique des sols rouge et beige du bassin
versant de Djiguinoum (Basse Casamance - Campagne de mesures 1990). ORSTOM,
Programme CEE DG XII, Contrat no TS2A-0216-M (CD).
Niane A. B. 1884. Etudes cartographiques et agro-pedologiques des sols de p1atea.u de
Basse Casamance. Memoire de confirmation. Cra Djibelor, ISRA. Il0 pp.
Ndiame, F. 1986. Aspects economiques de l’utilisation de la traction animale bovine et de
sa promotion par le biais du Credit Special du PIDAC pour le materiel agricole. ISRA.
DRSAEA. Memoire de titularisation. Dakar, Senegal.
Ndiame, F. 1990. Animal traction in the Lower Casamance: Technical aspects and socio-
economic implications. In: Paul Starkey and Adama Faye (eds), Animal traction for
Agriculture Development. Wageningen: Technical Center for Agricultural and
Cooperation (ACP-EEC Lame Convention).
Ndiame F. and Fall A. 1989. Par quelle cereale remplacer le riz ? Quelques problemes de
la culture du maïs en Basse Casamance. In Economie des Filières en Regions Chaudes.
Actes du X- seminaire d’economie et de sociologie. 1 l- 15 Septembre 1989, Montpellier,
France. Une publication du CIRAD.
Nicou, R. 1975. Le probleme de la prise en masse a la dessication des sols sableux et
sabla-argileux de la zone tropicale seche. In Agronomie Tropicale XxX-4: 325-343.
Nolle, J. 1975. Le tracteur animal. M. A. T. 5 1:46-56
OVeill, D.H., Howell P.J.L., Paice, M.E.R. and Kemp, D.C. 1987. An i.nstr-umentation
system to measure the performance of draught animais at work. In: Proc. Natl. Seminar
Animal Energy Utilization. 53-72. Central Institute of Agricultural Engineering: Bhopal,
India.
352
O’Neili Dl, Hayton S. and Sirns B. 1989. Measurement lofdraught animal performance. In
HolTtnan D., Nari J. and Petherarn R.J. (eds), Draught Animais in Rural Development.
Proceedings of an lternational Research Symposium, Cipansas, lndonesia ACIAR
Proc.eedings no 27:264-27 1.
‘y, Pearson R.A. and Fall A. 1993. R.esearch on the nutrition of draught animals. In IJuman
and draught animal poser in trop production. Workshop proceedings 18-22 Ja.nua\\y,
1993, Ha.rare Zimbabwe. Silsoe Research Institute and Agricultural Research and Food
Council, CEE-FAO, Rome, Italy.
Pelissier, .P. 1966. Les paysans du Senegal. L,es civilisations agraires du Cayor a Ii~
Casan-mce. Saint Yrieux, :France. (rd. Fabreque.
l?eret S., Pirot R. and Razaka B. 1.994. In-fiefd nleasurement of tillage irnplements
pelfi3rl~~~~nces:
Practical applications in soil-tillage relations studies. Paper AG-ENG 94.
Ping.all P., Bigot Y and Binswanger 13. 1987. Agricultural nle.chanization and the e\\olution
of f&ning systems in Sub-Saharan Africa. A World Rank Publication, John Hopkims
Press, Baltimore, MD. USA.
Pirot, R. 1993. Automation of in-feld data recording: Recent research work at CIRAD-
SAR. CIRAD-SAR., Montpellier, :Fra,nce.
Posner, J.L. 1985. Contribution a la connaissance agronorrrique de la Basse Casamance
(synr hese bibliographique). Document de tra\\:aii 1985:3. Deyartement SystenÎes et
Transfert, ISRA, Dakar, Senegal.
Posner J.L. and al. 1988. Les systemes de production en Basse Casamance et les strategies
paysannes face au deficit pluviometriquc. lSRA/?vlSU International Developrnent Papers.
Raman I-l., Mohan S., and Rangacharya N.C.V. 1992. Decigion Support for Crop Planning
during Droughts. In Journal of Itigation and Drainage Engineering, vol 118 no 2 ,229-
241.
RatlifL.F, Ritchie J.T., and Cassel D.K. 1983. Field-meaasured Lirnits of Soi1 Water
availnbility as Related to Laboratory-Measured Properties. In Journal of SSSA, vol 47 no.
4: 770-775.
Rehak, D.R. 1983. Expert systerns for drought nlanagen~ent. 113 proceeding of Conference
on Ernerg,ing Computer Techniques in Storrn. Water and Flood hlanagement W Ja:n-tes
(ed), ASCE. New York, NY, 333-387.
R.itchie, J.T. 1972. Mode1 for predicting evapotranspiration fr#otn a row trop with
incornplete caver. In Journal of Water resources Research October 1972, 8: 1204- 12 13
353
Roose E. 1967. Dix annees de mesure de l’erosion et du ruissellement au Senegal.
Agronomie Tropicale 1967 2”: 123-l 52.
Roosenberg R. 1987. Updating animal-powered mechanization by building on history.
ASAE proceedings: December 15 - 18, 1987 Winter meeting in Chicago IL.
Roosenberg R. 1992. Neck yoke design and fit: Ideas from dropped hitch point traditions.
Tiller’s TechGuide. Tillers International. Kalamazoo MI.
Sebillotte M. 1990. Les processus de decision des agriculteurs. Deuxieme partie:
consequences pour les demarches d’aide a la decision. In modelisation systemique et
systeme agraire. Decision et organisation, J. Brossier, B. Vissac et J.L. Le Moigne, eds ,
Paris, France, INRA: 10% 117.
Sivakumar M.V.K. 1988. Predicting rainy season potential from the onset of rains in the
Sahelian and Sudanian cfimatic zones of West Af?ica. Agricultural and Forest
hdeteorology 42: 295-305
Smith A. J. 1990. Using science to understand the biological constraints that limit work
animal productivity. In: P. Starkey, M. Goe and A. Faye (eds), Animal traction for
Agriculture
Development. Wageningen: Technical Center for Agricultural and
Cooperation (ACP-EEC Lome Convention).
Smith A. J. 198 1. Draught animal Research: A neglected subject. World Anin~al review,
40, 43-48
Sonko, M L. 1990. Disponibilite des animaux de trait et contraintes structurelles en Basse
Casam;ance. In: P. Starkey, M. Goe and A. Faye (eds), Animal traction for 14griculture
Development. Wageningen: Technical Center for Agricultural and Cooperation
(ACP-EEC Lome Convention).
Sonko, M.L. and Fall, A. 1992. Budgets de culture: Recueil de donnees technico-
economiques de production en Basse Casamance. FAO contrat d’auteur no 7 CMT 37503.
Srivastava, N.S.L. and Ojha, T.P. (eds). 1987. Utilization and economics of draught
animal power. Central Institute of Agricultural Engineering: Bhopal, India.
Stapleton, J.H. 1995. Linear Statistical Models. Wiley and Sons, New York. USA.
Starkey P. 1986. Draught animal power in Africa: priorities for development, research and
training. Networking paper 14. FSSP, University of Florida: Gainsville, Florida.
Starkey, P. 1989. Harnessing for cattle and buffaloes: Options and research. ‘In Hoi%nan
D., Nari J. and Petheram R.J. (eds), Draught Animals in Rural Development. Proceedings
of an Iternational Research Symposium, Cipansas, Indonesia. ACIAR Proceedings no
27:280-291.
Starkey P. 1989 Hamessing and Trnplemen~s for Animal traction - AR animal traction
resource book for Africa. GATEIGTZ, Vieweg and s~r‘ts, IWwschweig, Wiesbaden.
Germny.
Surr!ner GI. 1’388. Precipitation: Process and analysis John Wiley and Sons. New ‘York,
USA.
Turban E. and Meredith J.R. 1994. Fundamentals of Marrag,ement Science. 6th Edition
Irwin, Bcxton Ma. USA.
Teleni E ;and Hogan J.P. 1989. Nutrition of draught animais. In Hoffnlan D., Nari J. and
Pet heram R. J. (eds), Draught Animal s in Rural Developmet:~t. Proceedings of an
TteInafkxal Rese.arch Symposium~, Ci!pansas, Jndonesia. .4CIAR Proccedings no 2’7: 1 18-
133.
Tourte, R 1971. Themes legers - Themes lourds: Systc;nes intensifs; Voies difYercnres
ouwtec, a.u Developpement Agriwk du Scnegal, (,Simple technical innovation,
Sophisticated technical innovation, 3ntensive Systerns; Thret- dif??rent Approaches to
Agrkulture Developpement in Senegal). Agronomie Tropiçale, No 26 (5).
‘%\\Il Traverse~ S. 1974. Le boeuf Ndama et la riziculture en Basse Çasamance. Station Rizicole
“x!e Djibelor. ISRA, Dakar, Senegal.
L7padhya’y. R.C. 1989. Performance limiting Gctol s in draught animals: Can they he
manipuhited to improve ouIput? In Hoffma,n D., Kari J. ond Petheram R.J. (eds), Draught
Animal~ in Rural Development. Proceedings of an Iternational Research Symposium,
Cipansas., lndonesia. ACIAR .Proceedings no 27, I66- 173.
IJzure.au, C. 1974. Animal draR in West Afi-ka. \\~‘orlci Crops. 26: 112-l 14
Wanapat,, :M. 1989. Supplementaq fe:eds: Their ttfiects on pwformance of draught animals
and their acceptability to Sout.heast Aslan krmers. In Woffnxx~ Il., Nari J. and Pethcram
R.J. (eds), Draught Animais in Rural Development Proccedings of an Iternational
Research Symposium, Cipansas, Indonesia. ACIAR Proceedings no 27: 146-l 53.
Willocks, T.J. 1979. Tillage energ>/ requirements for semi-arid trop production in
Botswana. In R. Lai (ed). Soi1 tillage and Cq production. Proceedings series NO.~. 30%
323. IITA: Rxdan, Nigeria.