Nutrition of draught oxen in semi-arid west Africa ...
Nutrition of draught oxen in semi-arid west Africa
3. Effect of body condition prior to work and weight losses during
work on food intake and work output
A. Fall’, R. A. Pearson’ and S. Fernandez-Rivera”
‘Centrcfor Tropical Veterinwy Medicirw, Utiiversity 4 Edinbur~~h, Enster Bush, Roslirl, Midlotltiarl EH25 9RG
~lnter~~ational
Lizlesfock Rescarch hstitute, lnternaf~or~al
Crop Rcwarch hstitute for the Semi-Arid Troyics, Sahcliarz
Centre, BP 22404, Niamey, Nigo
Abstract
Ei$ltcen oxen zwe nlloffcd fo thw ft’f.?fltmtIt &voupi; nccordir~g f o fheir body corldifiou: yow, mcrfiiw ad gond.
Work oufput, spced, livc weighf arzd body conditiotz zuere measurrd dwiug 7 wecks ZL~CII n~~imals suorkcd 3 days/
teck, 4 k/day, pullitzg loads equivahlt fo 12.5 k~~/IOO kg 1Wr weight. The auimals were gizw millet stozw acl
libitum dwing kours they did net zoork plus IOgjkg M of a co?lcenfratr rnix. Work did flot i~~flmwcc iizfake of
millet sfover. Hou~euer, fbod infakc improvcd as work yvopsrrd and animais if! hi condition afr more millet
stozler fhan animais in good body condition. Work performance was affecfed by lizw zwi~yl~t but Ilof bod!/ cwdition.
Live-wcight losses did net havc 17 defrime~ztal effcct ou work yerformarlce. Poz~er oufput improvcd dztrilzg the cowse
of tkc cxpcriment whilc animals zwe losing zueight Animais it7 011 treatmen t groups ht body ;cu@t during th 7
weeks of zuork but zc~eighf losses were more pronounced in oxerz in good tkan ii1 poor l~ohy cofldifion. At thc end of
fhe working period, animais zuert! put ou nafural postures z~GNiouf supplemwzfatiot1. lt took 4 wccksfur ariimals in
poor and medium bodly condition and 6 weeks for animais in good bodl/ comlition to rcach fhir prc-icork linc
zueigh f .
i
Keywords: bod!/ condition, drau,ght animais, fond intake, live wcight, millet sfozw, iaorlk
1
Introduction
than 4 weeks). Although the cropping season is short
Draught oxen in good condition and of suitable live
in semi-arid areas, many oxen are hired out or
weight are required to ensure timeliness in soi1
loaned to other farmers and used for transport.
preparation and in planting, crucial for successful
Hence those oxen that are available may be used
cropping in semi-arid areas of west Africa. This is
over extended periods.
hecause work output is a function of body size and
working animals preferably use long-chain fatty
Supplementary food is often recommended as a
acids from fat reserves to fuel muscular activity
means of producing draught oxen in good condition
during sustained exercise (Preston and Leng, 1987;
and live weight to optimize power available.
Pethick, 1993). Unfortunately, draught oxen often
However, food supplementation is expensive.
lose weight during the dry season (Wilson, 1987).
Furthermore, investigations of the effect of dry
Therefore they have minimum live weight and body
season supplementation on draught oxen have
reserves at the start of the cultivation period when
generally failed to show any significant benefit of
farm power demand is highest. This is also a time
committing scarce food resources to work output
when food resources are scarce and do not match the
and consequently to better trop production (Dick0
nutrient
requirements
of draught oxen for
and Sangaré, 1984; Astatke ~>t al.. 1984; Khibe and
maintenance, let alone work. While body condition
Barthotomew, 1993). Feeding strategies for draught
and weight losses during work may not constrain
oxen cari be better planned if the minimum working
performance when animals are only used for short
weight for cultivation is known and if the losses in
periods (3 weeks), they take on greater significance
weight and body condition that animais cari tolerate
where animals are used for longer periods (more
before work output is affected are known. Hence in
,
227
228
Fall, Pearson and Fernhdez-Rivera
this study the relationship between body weight,
thermometer. Live weight was measured for 3 days
body condition, weight loss during work and work
consecutively
at the beginning of each week.
output of draught oxen were investigated.
Body condition was assessed each week as defined
by Nicholson and Butterworth (1986). Each day,
individual food allowances and refusals were
Material and methods
weighed and a food sample taken, dried and ground
for laboratory analysis (Fall et III., 1997%).
This experiment was conducted from July to
September 1994 at the International Crop Research
At the end of ithe 7 weeks of work, 10 animais were
Institute for the Semi-Arid Tropics (ICRISAT)
monitored for 2 months to investigate the rate of
Sahelian Centre in Niger. Eighteen oxen, aged 4 to 7
weight gain after work. The animals grazed rainy
years, with an average live weight of 326 kg, were
season natural pastures from 08.30 to 17.00 h without
used. They were given chopped millet stover nd
supplementation. They were kept in stables after
liEitum, supplemented daily with 10 g dry tnatter
grazing and had access to water nd libitm. M was
(DM) per kg M’)‘7i of a concentrate mix made up (g/
measured for a 2-monthgeriod every 2nd day in the
kg) of wheat bran (500), groundnut cake (350), rock
morning before grazing.
phosphate (50), crushed bone (50) and common salt
(50). Animals at rest were allowed to eat when other
Dntn omlysis
ox teams were working.
During the course of the experiment two oxen, one in
team 3 (‘poor’ IBC) and one in team 6 (‘medium’
IBC), were impaired by joint disorders. They were
Treatments consisted of three levels of body
consequently
allowed to rest from the 5th week of
condition before work (IBC). The oxen were given
the experiment. Sound oxen in these pairs were
enough food during the 3 months before the
teamed LIP SO that they could continue work for the
experiment in order to reach contrasting body
rest of the experiment. Therefore, different sets of
condition scores as defined by Nicholson and
data were used to analyse parameters of interest in
Butterworth (1986) on a scale from 1 (emaciated) to 9
this study. ‘The data set used to analyse daily intake
(obese). Three pairs of oxen were assigned to each of
of millet stover and daily weight changes included
the three treatment groups with average body
a11 weeks and a11 oxen except ox 17 and 25. Teams 3
condition scores of 2.33, 3.67 and 5.67 for groups 1, 2
and 6 wcre excluded from the analysis of speed,
and 3, respectively. Average M of teams in groups 1,
power and work output.
2 and 3 were 615,650 and 692 kg, respectively.
Statistical models used to analyse food intake,
The experiment lastcd 7 weeks. Teams worked for 4
wcight change, body condition, speed, power and
days each week. Work consisted of pulling a loaded
work output using Statistical Analysis Systems
sledge around a flat circuit. Each day, the teams
Institute (1985) are given below:
worked for 4 h to complete 10 laps of the circuit.
Work stopped when the set distance or the set time
(1) intake of millet stover (g DM per day per kg M
was complete or when oxen were unwilling to
and g DM per day per kg Mn”i)
continue or when it was judged that the oxen were
too tired to continue working. During the
Yljh, = p -t C, + T,,,i + A, + W, + C X A, + C X W,,
preparation phase an ergometer (Lawrence and
+ W X T,,,,/ + ei,k/
Pearson,
1985) was used to measure work
performed, distance travelled and elapsed working
(2) change in live weight (g/day) and body condition
time for different known loads. A regression analysis
(points per week)
of force on sledge load was performed and used to
determine the load required for each team SO that the
Yii, = p + C, + Tciil + W, + C X W,, + W X T,i,,, + c,;,
draught force exerted was equivalent to 12.5 kgf per
100 kg M. The following equation was used to
(3) speed (m/s) and power ( W and W / 100 kg M)
determine work loads: load (kg) = 0.201 X force (N) -
744.
Yi,k,ri = ~ + Ci +
+ Wk + R,,, + C X W,~ + C X R,,,,
T,;,j
+ W X TclJik + eijhrir
Animals were allowed to stop for 3 to 4 min after
each lap. Respiration rate and rectal temperature
where: Y := one observation of daily food intake,
were then recorded. Respiration rate was assessed by
dailv weight (change, weekly body condition score,
counting the number of flank movements for 30 s.
force, distance, speed, power or work; u = mean; Ci =
Rectal temperature was tneasured with a clinical
ith TBC score, i = 1,2 and 3 (1 = ‘poor’, 2 = ‘medium’,
Nutrition of draught cattle - 3
229
3 = ‘good’ IBC); T,,,; = jth oxen team nested within
Differences
(P < 0.01) in daily weight gain due to IBC
the ith IBC group, j = 1,2 and 3; A, = kth activity, k =
were observed. Al1 oxen lost weight during the
0: rest, k = 1: work; W, = Ith experimental week, 1 = 1,
experiment but weight losses were highest in oxen in
2. . .7; C X Aik = interaction between the ith IBC and
‘good’ IBC. Daily weight losses were 456 (se. 103.3),
the kth activity; C X W, = interaction between the ith
308 (s.e. lO3.3) and 719 (s.e 89.5) g/day for oxen in
IBC and Ith week; W X T,,,j, = interaction between the
‘pool’,
‘medium’ and ‘good’ IBC, respectively.
ith week and the jth team in the ith IBC group; R,,, =
Weight losses averaged 21.9 kg for oxen in the ‘poor’
rnth lap of the circuit travelled; C X R,,,, = interaction
IBC group, 14.8 kg for oxen in the ‘medium’ IBC
between the ith initial body condition and the fnth
group and 34.5 kg for oxen in the ‘good’ TBC group
lap travelled; r = random error.
over 7 weeks. These weight losses were equivalent to
proportionately 0.074,0.047 and 0.099 of the initial M
The effect of IBC on daily intake, weight change,
for oxen in ‘poor’, ‘medium’ and ‘good’ IBC,
speed, power and work was tested using team
respectively.
within condition (T,,,,) as the error term. The
interaction between wéek and team within condition
Weight losses estimated from polynomial regressions
(W X T,,,,,) served as the error term to test the effects
are illustrated in Table 1. The pattern of live-weight
of week and the interaction between week and other
changes was the same irrespective of IBC. Daily
factors
included in
t h e
model.
Orthogonal
weight losses were highest during the 1st week of the
polynomial regressions were fitted for variables such
experiment and decreased from week 1 to week 4.
as week, lap, and their interaction with IBC to
There was a steady increase in weight losses from
investigate the trend in food intake, weight change,
week 5 to week 7. The regression of daily weight
speed, power and work over time.
losses on intake of millet stover showed no
association between these two measurements.
Results
Body condition scores of a11 oxen declined over time.
Linear effect of IBC (P < 0.05) and the linear and
The regression of body condition on time showed
quadratic effect of week (P < 0.05) on daily intake of
that the better the IBC the more severe was its
millet stover (g/day per kg M; g/day per kg M”.75)
deterioration Body condition score declined at a rate
were seen (Table 1). The poorer the IBC the higher
of 0.006, 0.107 and 0.235 points per week, for oxen in
was the intake of millet stover. Intake of millet stover
‘poor’, ‘medium’ and ‘good’ IBC, respectively.
increased steadily over time and reached a plateau
by the 4th week. Food intakes on working and
In the 10 oxen monitored on good quality natural
non-working days were similar.
pastures after work, rapid weight gains were
Daily weight losses (g/day)
estimated by
Daily food
Daily food intake
polynomial regression
intake g/ kg M
g/ kg M”“”
IBC
Sources of variation
Mean
se.
Mean
se.
Significance ‘Poor’
‘Medium’ ‘Good’
Initial body condition
Linear”
‘Poor’
18.1
0.26
75.1
1.08
‘Medium’
17.2
0.26
72.9
1.07
‘Good’
15.2
0.22
64.8
0.93
Activity
R e s t
17.2
0.20
72.2
0.86
Work
16.6
0.19
69.7
0.79
Week
Linear**+
1
13.8
0.22
58.1
0.94
Quad.***
206
409
1193
2
15.4
0.24
64.8
1.00
4 7 7
306
650
3
16.3
0.23
68.4
0.95
4 2 1
2 1 1
449
4
17.4
0.23
72.9
0.95
272
185
478
5
17.0
0.22
71.2
0.93
263
288
628
6
17.7
0.22
73.6
0.94
625
583
788
7
17.2
0.24
71.6
1.03
1592
1129
847
-.~---.
.-.- .-...---
~..
-
230
Fall, Pearson and Fernhndez-Rivera
observed as illustrated by the following regression
equations of M (kg) on time (60 days [D] after work):
M = 257 (se. 12) + 0.825 (se. 0.054) X D for oxen in
‘poor’ IBC (at the start of the experiment), M = 302
(s.e. 12) + 0.967 (se. 0.041) X D for oxen in ‘medium’
Power
IBC (at the start of the experiment), M = 303 (se. 11)
SPWd
P«wer
(W per
Source of \\~ariation
(m/s)
W)
100 kg M)
+ 0.870 (se. 0.037) X D for oxen in ‘good’ IBC (at the
start of the experiment).
Initial body condition
‘Poor’
0.94
637
117
The overall rate of change of M was similar
‘Medium
0%
637
107
irrespective of the IBC score. However oxen in ‘poor’
‘C;ood’
0.96
780
1 1 8
and ‘medium’ IBC had higher M gains (3.20 g/day
s.e.
om5
4
0.6
per kg M) than oxen in ‘good’ IBC (2.87 g/day per kg
La\\7
M). Oxen in ‘poor’ and ‘medium’ IBC were able to
1
0.95
706
1 1 7
reach their initial live weight 4 weeks, on average,
2
0.97
722
130
3
0.97
724
1 2 0
after work stopped. It took 6 weeks after the
4
0.95
706
117
cessation of work for oxen in ‘good’ IBC to reach
5
0.92
690
115
their pre-work M.
6
0.92
68s
1 1 4
7
0.89
666
1 1 1
IBC did not affect speed of work of teams. Power (W)
8
048
651
109
developed by teams in ‘poor’ and ‘medium’ IBC was
9
02v
651
1 0 8
similar but significantly lower than power output
10
0.87
(746
1 0 7
(W) of oxen in ‘good’ IBC (P < 0.01). However, when
s.e.
048
6
1
power was expressed relative to live weight (W per
Week
1
0.78
,590
9 5
100 kg M), the effect of IBC was no longer significant.
2
044
631
103
Three oxen teams with approximately similar M and
3
0438
653
107
in different body condition (team 1: ‘poor’ IBC,
1
0.98
725
1 2 0
719 kg, team 2: ‘medium’ IBC, 721 kg; team 3: ‘good’
s
1 .oo
744
124
IBC, 739 kg) developed similar power output (team
6
0.96
713
120
1: 775 W; team 2: 697 W; team 3: 741 W).
7
0.99
739
1 2 6
se.
0.07
5
1
Differences
in speed and power output due to week
of work were significant (P < 0.01, Table 2). Speed
Early in lactation increases in DM intake were seen
and power output increased steadily over weeks for
in cows with a low body condition as compared with
a11 teams irrespective of their IBC. Even though oxen
those in a better body condition (Jones and
lost body weight throughout the experiment, there
Garnsworthy, 1989). As suggested by Faverdin ct nl.
was a significant weekly increase of 0.035 m/s and
(1995), undernutrition induces an increase in food
25.1 W in speed and power, respectively.
intake when food unavailability is no longer a
limiting factor.
Lap number had a significant effect (P < 0.01) on
speed and power output. Two contrasting phases
The steady increase in food intake and power output
were observed in the pattern of power output each
over 7 weeks observed in this study suggests that
day. During the first three laps of work (1st h) there
oxen undcrwent an adaptation to work during the
was an increase of 0.01 tn/s and 9 W in speed and
first days of work. Thcy became more adapted as
power each lap. In the second phase starting from
work went on and they were therefore able to
the fourth lap a steady decline in speed and work
increase their intake. Bartholomew cf d. (1994) also
output of 0.014 m/s and 10.3 W were observed for
attributed increases in speed of working teams over
each lap completed.
time to adaptation to work. Therefore, working
during the dry season would bave the adïantage
Discussion
that oxen are fit when cultivation starts and do not
As in previous studies (Fall ct nl., 1997) work did not
have to undergo this adaptation period in the
affect intake of millet stover. These results are
cropping season.
consistent with most other studies that indicated no
differences
in intake in working animals as
During the preparation phase of the experiment oxen
compared with animals at-rest (reviewed by Pearson
were given food to reach the targeted body condition
and Dijkman, 1993). Poor body condition before
and live weight. This was however difficult to
work was conducive to higher intake of millet stover
achieve because live-weight changes were associated
than good body condition before work over 7 weeks.
with changes in body condition. At the start of the
”
.
Nutrition of draught cattle - 3
231
experiment heavier animals tended to have better
820 N (Khibe and Bartholomew, 1993). They could
body condition than lighter animais. In order to
however pull heavily loaded carts without undue
minimize the confounding effect of M and IBC on the
stress.
rate of work (W) in different treatment groups,
power output was expressed relative to M (W per
Jt was expected that weight losses would adversely
100 kg M). The use of power relative to M to
affect work output. In this experiment power output
investigate the effect of IBC on work performance
improved over weeks while oxen were undergoing
was based on the assumption that oxen in ‘good’ IBC
weight losses. The same trends in live-weight change
had a higher fat content per kg M than oxen in ‘poor’
and power output were reported by Bartholomew et
or ‘medium’ IBC.
RI. (1993). Therefore the weight losses oxen cari
tolerate before work output is affected are difficult to
Power output is a function of speed and draught
estimate. Continuous and severe weight losses cari
force. The latter was set in this experiment to be
compromise the health of the animal, or even its life.
proportional to M. Therefore any advantage of oxen
However in this experiment, oxen were able to
in good condition over those in poor condition
regain li\\,e weight at a rapid rate when they had
would bave been expressed as a higher average
access to good quality pasture during the rainy
walking speed. As compared with oxen in poor
season. It took them 4 to 6 weeks to reach their
condition, oxen in good condition would perform
pre-work live weight.
work at a higher rate or for a longer period of time
because they would have more body reserves to
At the start of the cropping season, various body
draw on to fuel muscle activity. However walking
conditions are found in oxen because of differences
speed was not significantly affected by body
in food resources, management practices and the
condition in this experiment. Furthermore, the effect
disease situation in farming systems in the semi-arid
of IBC on power output was no longer significant
zone. Observations from this study suggest that a
when power output was expressed in relation to M.
bodv condition score of between 2 and 3 as defined
The similarity of power output relative to M for a11
by fiicholson and Butterworth (1986) would be a low
oxen suggested that animais with same body mass,
critical score below which work may irreversibly
irrespective of fat content, generated the same power
damage the oxen’s health. The ideal body condition
output This suggests that power output is more
score would range between 4 and 6. These animals
dependent on body mass than body condition. Oxen
are not too fat nor too lean and cari perform well if
in good condition did not out-perform oxen with
they are in good health. Oxen with a body condition
equal M but in poor body condition. These results
score of over 6 may be too fat to move comfortably
are consistent with those reported by Bartholomew cf
and are more susceptible to heat stress than leaner
al. (1994). These authors evaluated the relative
oxen. Moreover, the feeding level required to reach a
importance of body weight and body condition on
body condition score over 6 is unlikely to be
work performance by applying the same load of
profitable as far as feeding for work performance is
367 N to groups of oxen weighing 33.0 and 360 kg
concerned.
and in good and poor condition. Light oxen in good
condition could not sustain the work level applied.
Oxen should be supplemented during the cropping
They concluded that live weight rather than body
season when work is performed for more than 6
condition is the single most important determinant
weeks. If the working period is short, weight losses
of work output. Therefore, it seems that farmers
could be tolerated as animais Will regain their live
should be encouraged
to Select large-framed animals
weight rapidly. Supplementary feeding is however
for draught purposes. This may be constrained by
worthwhile for animals scheduled to be sold for
the fact that more Young animals are being used for
meat after work, even for short working periods, SO
draught purposes with a rapid on-farm turn-over
that work does not adversely affect their market
rate of these animals, apparently driven by an
value. In
semi-arid
w e s t
Africa
food
attractive meat market. There is therefore a genuine
supplementation under these circumstances may be
need to investigate feeding and management
based on whole cotton seed, groundnut,
sesame or
practices that Will optimize power and meat output
cotton cakes if available. These foods are rich in
in these farming systems.
protein and energy and provide substrates
(long-chain fatty acids, glycogenic compounds) that
Good IBC may allow work for longer periods of time
cari be directly used for work.
in the cropping season. In the present study, oxen in
‘poor’ IBC sustained average draught forces of
Acknowledgements
682 N. These animals might not be able to perform
The
authors
thank
thc
Overseas
Devclopmcnt
ploughing or ridging for extended periods because
Administration (ODA) of the United Kingdom and the
these two activities require draught forces of about
International Livestock Rese,wch Institute for funding this
-
--------~-~~
232
Fall, Pearson and Fernhdez-Rivera
study and E. A. Hunter for statistical advice during
Jones, G. P. and Garnsworthy, P. C. 1989. The effects of
preparation of the manuscript.
dictary energy content on thc rcsponse of dairy cows to
body condition at calving. Auirml Produifiorl 49: 183-191.
Ref erences
Khibe, T. and Bartholomew, P. W. 1993. Recherches aux
possibilités de conditionnement des boeufs de labour à base
Astatke, A., Reed, J. D. and Butterworth, M. H. lY8h. Effect
de paille de brousse enrichie 3 l’ut+ et .@valuation de son
of diet restriction on work performance and weight loss of
impact sur leur capacité de travail. In Rcmrrh .for
local Zebu and Friesian-Boran crossbred oxen. II.CR BI&+~JZ
de7~Gmr1~t o f ,mimnl trncth ~II kV& Afiico (ed. P. R .
23: 11-14.
Lawr&ce, K.‘Lawrence, J. T. Dijkman and P. H; Starkey).
Bartholomew, P. W., Khibe, T. and Little, D. A. 199-I. Effect
Prwmfi~zgs o,f thc jurrfh workslmp c$ t/w h’csf @kn A~~imz/
of liveweight and body condition on work output from
Trncfh Akt7uork, Kmo, Nigrrin FI-13 ]~/y 1990, pp. 175-179.
draft oxen. Troyic.l7/ A??ima/ Hcnltlr md Prod~~ctim 26: 58-64.
Wcst Africa Animal Traction Network and International
Bartholomew, P. W., Khibe, T., Little, D. A. and Ba, S.
Livestock Centre for Africa (ILCA), Addis Ababa, Ethiopia.
19Y3. Effect of change in body weight and condition during
Lawrence, P. R. and Pearson, R. A. 1985. Factors affecting
the dry season on capacity for work of draft oxcn. Tropical
the measurement of draught force, work output and power
A~~inml Htalth md Prod~rt-lio~i 25: 50-58.
o f oxen. \\ournn/ o f A,~ric~rltwnl
Scicwcr, Cnf~lbrirfxc 1 0 5 :
Dicko, M. S. and Sangaré, M. 1984. DP I’cwplai dc In trnction
703-714.
n1rirmle rrl iOlll’
Sulrlienru~: de d e
/a srrn.‘l~flrc,ntation
Nicholson M. J. and Butterworth, M. 19%. A pide to
a/ir/rrrltnirr.
D o c u m e n t d e p r o g r a m m e n o . A%N 3 .
cowditior~ scoring of~ebu cuttk. International Livestock Centre
International Livestock Centre for Africa, Addis Ababa,
for Africa (ILCA), Addis Ababa, Ethiopin.
Ethiopia.
Pearson, R. A, and Dijkman, J. T. 1994. Nutritional
Fall, A., Pearson R. A. and Lawrence, P. R. 1997. Nutrition
implications of work in draught animais. Pmwdiqs c!f tlw
of draught oxen in semi-arid west Africa. 1. Energy
NlltritiOlz Socr’r~ty 53: 169-179.
cxpenditure by oxen working on soils of different
Pethick, D. W. 1993. Carbohydrate and lipid oxidation
consistencies. Animnl Scictw 64: 209-215.
during excrcise. Alrstrulim @ml c?f Apklrltwol Rrwwlr
Fall, A., Pearson, R. A., Lawrence, P. R., Fernandez-Rivera,
44: 431-141.
S. 1997b. Nutrition of draught oxen in semi-arid west
Preston, T. R. ,and Leng, R. A. 1987. Mntshing rrmimulf
Africa. 2. Effect of work on intakc, digestibility and rate of
prod~rctiorl systmrs 7dth nvnilnblr rfwwccs in flfr frq7ics nfirf
passage of food through the gastro-intenstinal tract in
sub-tropics. Penarnbul Books, Armidale, Australia.
draught oxen given trop residues. A~~imnl Srimce 64:
217-225.
Statistical Analysis Systems *Instittite. 1985. SAS IISCT’S
pidP: sfntistics. Statistical Analysis Systems Institute Inc.,
Faverdin, P., Baumont, R. and Ingvartsen, K. L. 1995.
Cary, NC.
Control and prediction of feed intake in ruminants. In
Rccmt dcvrlopmts irl thc rlrrtritim of herhinoree (ed. M.
Wilson, R. T. 1987. Livestock production in central Mali:
factors influencing growth and liveweight in agro-pastoral
Journet, E . Genet, M.-H. Farce, M. Theriez and C .
Demarquilly). Procwdhp of’ tk~ fuurfl~ ir~lermfionnl
cattle. ‘hyicn/ Animal Hmlfh md Production 19: 103-114.
syrnposiwn on fhe rn~tritim of hrrbivorrs, pp. 95-120. INRA
Editions, Paris
(Rrcriwd 29 ]~~r~my 1996-Accytrd 13 Srptrmber 1996)
3. Effect of body condition prior to work and weight losses during
work on food intake and work output
A. Fall’, R. A. Pearson’ and S. Fernandez-Rivera”
‘Centrcfor Tropical Veterinwy Medicirw, Utiiversity 4 Edinbur~~h, Enster Bush, Roslirl, Midlotltiarl EH25 9RG
~lnter~~ational
Lizlesfock Rescarch hstitute, lnternaf~or~al
Crop Rcwarch hstitute for the Semi-Arid Troyics, Sahcliarz
Centre, BP 22404, Niamey, Nigo
Abstract
Ei$ltcen oxen zwe nlloffcd fo thw ft’f.?fltmtIt &voupi; nccordir~g f o fheir body corldifiou: yow, mcrfiiw ad gond.
Work oufput, spced, livc weighf arzd body conditiotz zuere measurrd dwiug 7 wecks ZL~CII n~~imals suorkcd 3 days/
teck, 4 k/day, pullitzg loads equivahlt fo 12.5 k~~/IOO kg 1Wr weight. The auimals were gizw millet stozw acl
libitum dwing kours they did net zoork plus IOgjkg M of a co?lcenfratr rnix. Work did flot i~~flmwcc iizfake of
millet sfover. Hou~euer, fbod infakc improvcd as work yvopsrrd and animais if! hi condition afr more millet
stozler fhan animais in good body condition. Work performance was affecfed by lizw zwi~yl~t but Ilof bod!/ cwdition.
Live-wcight losses did net havc 17 defrime~ztal effcct ou work yerformarlce. Poz~er oufput improvcd dztrilzg the cowse
of tkc cxpcriment whilc animals zwe losing zueight Animais it7 011 treatmen t groups ht body ;cu@t during th 7
weeks of zuork but zc~eighf losses were more pronounced in oxerz in good tkan ii1 poor l~ohy cofldifion. At thc end of
fhe working period, animais zuert! put ou nafural postures z~GNiouf supplemwzfatiot1. lt took 4 wccksfur ariimals in
poor and medium bodly condition and 6 weeks for animais in good bodl/ comlition to rcach fhir prc-icork linc
zueigh f .
i
Keywords: bod!/ condition, drau,ght animais, fond intake, live wcight, millet sfozw, iaorlk
1
Introduction
than 4 weeks). Although the cropping season is short
Draught oxen in good condition and of suitable live
in semi-arid areas, many oxen are hired out or
weight are required to ensure timeliness in soi1
loaned to other farmers and used for transport.
preparation and in planting, crucial for successful
Hence those oxen that are available may be used
cropping in semi-arid areas of west Africa. This is
over extended periods.
hecause work output is a function of body size and
working animals preferably use long-chain fatty
Supplementary food is often recommended as a
acids from fat reserves to fuel muscular activity
means of producing draught oxen in good condition
during sustained exercise (Preston and Leng, 1987;
and live weight to optimize power available.
Pethick, 1993). Unfortunately, draught oxen often
However, food supplementation is expensive.
lose weight during the dry season (Wilson, 1987).
Furthermore, investigations of the effect of dry
Therefore they have minimum live weight and body
season supplementation on draught oxen have
reserves at the start of the cultivation period when
generally failed to show any significant benefit of
farm power demand is highest. This is also a time
committing scarce food resources to work output
when food resources are scarce and do not match the
and consequently to better trop production (Dick0
nutrient
requirements
of draught oxen for
and Sangaré, 1984; Astatke ~>t al.. 1984; Khibe and
maintenance, let alone work. While body condition
Barthotomew, 1993). Feeding strategies for draught
and weight losses during work may not constrain
oxen cari be better planned if the minimum working
performance when animals are only used for short
weight for cultivation is known and if the losses in
periods (3 weeks), they take on greater significance
weight and body condition that animais cari tolerate
where animals are used for longer periods (more
before work output is affected are known. Hence in
,
227
228
Fall, Pearson and Fernhdez-Rivera
this study the relationship between body weight,
thermometer. Live weight was measured for 3 days
body condition, weight loss during work and work
consecutively
at the beginning of each week.
output of draught oxen were investigated.
Body condition was assessed each week as defined
by Nicholson and Butterworth (1986). Each day,
individual food allowances and refusals were
Material and methods
weighed and a food sample taken, dried and ground
for laboratory analysis (Fall et III., 1997%).
This experiment was conducted from July to
September 1994 at the International Crop Research
At the end of ithe 7 weeks of work, 10 animais were
Institute for the Semi-Arid Tropics (ICRISAT)
monitored for 2 months to investigate the rate of
Sahelian Centre in Niger. Eighteen oxen, aged 4 to 7
weight gain after work. The animals grazed rainy
years, with an average live weight of 326 kg, were
season natural pastures from 08.30 to 17.00 h without
used. They were given chopped millet stover nd
supplementation. They were kept in stables after
liEitum, supplemented daily with 10 g dry tnatter
grazing and had access to water nd libitm. M was
(DM) per kg M’)‘7i of a concentrate mix made up (g/
measured for a 2-monthgeriod every 2nd day in the
kg) of wheat bran (500), groundnut cake (350), rock
morning before grazing.
phosphate (50), crushed bone (50) and common salt
(50). Animals at rest were allowed to eat when other
Dntn omlysis
ox teams were working.
During the course of the experiment two oxen, one in
team 3 (‘poor’ IBC) and one in team 6 (‘medium’
IBC), were impaired by joint disorders. They were
Treatments consisted of three levels of body
consequently
allowed to rest from the 5th week of
condition before work (IBC). The oxen were given
the experiment. Sound oxen in these pairs were
enough food during the 3 months before the
teamed LIP SO that they could continue work for the
experiment in order to reach contrasting body
rest of the experiment. Therefore, different sets of
condition scores as defined by Nicholson and
data were used to analyse parameters of interest in
Butterworth (1986) on a scale from 1 (emaciated) to 9
this study. ‘The data set used to analyse daily intake
(obese). Three pairs of oxen were assigned to each of
of millet stover and daily weight changes included
the three treatment groups with average body
a11 weeks and a11 oxen except ox 17 and 25. Teams 3
condition scores of 2.33, 3.67 and 5.67 for groups 1, 2
and 6 wcre excluded from the analysis of speed,
and 3, respectively. Average M of teams in groups 1,
power and work output.
2 and 3 were 615,650 and 692 kg, respectively.
Statistical models used to analyse food intake,
The experiment lastcd 7 weeks. Teams worked for 4
wcight change, body condition, speed, power and
days each week. Work consisted of pulling a loaded
work output using Statistical Analysis Systems
sledge around a flat circuit. Each day, the teams
Institute (1985) are given below:
worked for 4 h to complete 10 laps of the circuit.
Work stopped when the set distance or the set time
(1) intake of millet stover (g DM per day per kg M
was complete or when oxen were unwilling to
and g DM per day per kg Mn”i)
continue or when it was judged that the oxen were
too tired to continue working. During the
Yljh, = p -t C, + T,,,i + A, + W, + C X A, + C X W,,
preparation phase an ergometer (Lawrence and
+ W X T,,,,/ + ei,k/
Pearson,
1985) was used to measure work
performed, distance travelled and elapsed working
(2) change in live weight (g/day) and body condition
time for different known loads. A regression analysis
(points per week)
of force on sledge load was performed and used to
determine the load required for each team SO that the
Yii, = p + C, + Tciil + W, + C X W,, + W X T,i,,, + c,;,
draught force exerted was equivalent to 12.5 kgf per
100 kg M. The following equation was used to
(3) speed (m/s) and power ( W and W / 100 kg M)
determine work loads: load (kg) = 0.201 X force (N) -
744.
Yi,k,ri = ~ + Ci +
+ Wk + R,,, + C X W,~ + C X R,,,,
T,;,j
+ W X TclJik + eijhrir
Animals were allowed to stop for 3 to 4 min after
each lap. Respiration rate and rectal temperature
where: Y := one observation of daily food intake,
were then recorded. Respiration rate was assessed by
dailv weight (change, weekly body condition score,
counting the number of flank movements for 30 s.
force, distance, speed, power or work; u = mean; Ci =
Rectal temperature was tneasured with a clinical
ith TBC score, i = 1,2 and 3 (1 = ‘poor’, 2 = ‘medium’,
Nutrition of draught cattle - 3
229
3 = ‘good’ IBC); T,,,; = jth oxen team nested within
Differences
(P < 0.01) in daily weight gain due to IBC
the ith IBC group, j = 1,2 and 3; A, = kth activity, k =
were observed. Al1 oxen lost weight during the
0: rest, k = 1: work; W, = Ith experimental week, 1 = 1,
experiment but weight losses were highest in oxen in
2. . .7; C X Aik = interaction between the ith IBC and
‘good’ IBC. Daily weight losses were 456 (se. 103.3),
the kth activity; C X W, = interaction between the ith
308 (s.e. lO3.3) and 719 (s.e 89.5) g/day for oxen in
IBC and Ith week; W X T,,,j, = interaction between the
‘pool’,
‘medium’ and ‘good’ IBC, respectively.
ith week and the jth team in the ith IBC group; R,,, =
Weight losses averaged 21.9 kg for oxen in the ‘poor’
rnth lap of the circuit travelled; C X R,,,, = interaction
IBC group, 14.8 kg for oxen in the ‘medium’ IBC
between the ith initial body condition and the fnth
group and 34.5 kg for oxen in the ‘good’ TBC group
lap travelled; r = random error.
over 7 weeks. These weight losses were equivalent to
proportionately 0.074,0.047 and 0.099 of the initial M
The effect of IBC on daily intake, weight change,
for oxen in ‘poor’, ‘medium’ and ‘good’ IBC,
speed, power and work was tested using team
respectively.
within condition (T,,,,) as the error term. The
interaction between wéek and team within condition
Weight losses estimated from polynomial regressions
(W X T,,,,,) served as the error term to test the effects
are illustrated in Table 1. The pattern of live-weight
of week and the interaction between week and other
changes was the same irrespective of IBC. Daily
factors
included in
t h e
model.
Orthogonal
weight losses were highest during the 1st week of the
polynomial regressions were fitted for variables such
experiment and decreased from week 1 to week 4.
as week, lap, and their interaction with IBC to
There was a steady increase in weight losses from
investigate the trend in food intake, weight change,
week 5 to week 7. The regression of daily weight
speed, power and work over time.
losses on intake of millet stover showed no
association between these two measurements.
Results
Body condition scores of a11 oxen declined over time.
Linear effect of IBC (P < 0.05) and the linear and
The regression of body condition on time showed
quadratic effect of week (P < 0.05) on daily intake of
that the better the IBC the more severe was its
millet stover (g/day per kg M; g/day per kg M”.75)
deterioration Body condition score declined at a rate
were seen (Table 1). The poorer the IBC the higher
of 0.006, 0.107 and 0.235 points per week, for oxen in
was the intake of millet stover. Intake of millet stover
‘poor’, ‘medium’ and ‘good’ IBC, respectively.
increased steadily over time and reached a plateau
by the 4th week. Food intakes on working and
In the 10 oxen monitored on good quality natural
non-working days were similar.
pastures after work, rapid weight gains were
Daily weight losses (g/day)
estimated by
Daily food
Daily food intake
polynomial regression
intake g/ kg M
g/ kg M”“”
IBC
Sources of variation
Mean
se.
Mean
se.
Significance ‘Poor’
‘Medium’ ‘Good’
Initial body condition
Linear”
‘Poor’
18.1
0.26
75.1
1.08
‘Medium’
17.2
0.26
72.9
1.07
‘Good’
15.2
0.22
64.8
0.93
Activity
R e s t
17.2
0.20
72.2
0.86
Work
16.6
0.19
69.7
0.79
Week
Linear**+
1
13.8
0.22
58.1
0.94
Quad.***
206
409
1193
2
15.4
0.24
64.8
1.00
4 7 7
306
650
3
16.3
0.23
68.4
0.95
4 2 1
2 1 1
449
4
17.4
0.23
72.9
0.95
272
185
478
5
17.0
0.22
71.2
0.93
263
288
628
6
17.7
0.22
73.6
0.94
625
583
788
7
17.2
0.24
71.6
1.03
1592
1129
847
-.~---.
.-.- .-...---
~..
-
230
Fall, Pearson and Fernhndez-Rivera
observed as illustrated by the following regression
equations of M (kg) on time (60 days [D] after work):
M = 257 (se. 12) + 0.825 (se. 0.054) X D for oxen in
‘poor’ IBC (at the start of the experiment), M = 302
(s.e. 12) + 0.967 (se. 0.041) X D for oxen in ‘medium’
Power
IBC (at the start of the experiment), M = 303 (se. 11)
SPWd
P«wer
(W per
Source of \\~ariation
(m/s)
W)
100 kg M)
+ 0.870 (se. 0.037) X D for oxen in ‘good’ IBC (at the
start of the experiment).
Initial body condition
‘Poor’
0.94
637
117
The overall rate of change of M was similar
‘Medium
0%
637
107
irrespective of the IBC score. However oxen in ‘poor’
‘C;ood’
0.96
780
1 1 8
and ‘medium’ IBC had higher M gains (3.20 g/day
s.e.
om5
4
0.6
per kg M) than oxen in ‘good’ IBC (2.87 g/day per kg
La\\7
M). Oxen in ‘poor’ and ‘medium’ IBC were able to
1
0.95
706
1 1 7
reach their initial live weight 4 weeks, on average,
2
0.97
722
130
3
0.97
724
1 2 0
after work stopped. It took 6 weeks after the
4
0.95
706
117
cessation of work for oxen in ‘good’ IBC to reach
5
0.92
690
115
their pre-work M.
6
0.92
68s
1 1 4
7
0.89
666
1 1 1
IBC did not affect speed of work of teams. Power (W)
8
048
651
109
developed by teams in ‘poor’ and ‘medium’ IBC was
9
02v
651
1 0 8
similar but significantly lower than power output
10
0.87
(746
1 0 7
(W) of oxen in ‘good’ IBC (P < 0.01). However, when
s.e.
048
6
1
power was expressed relative to live weight (W per
Week
1
0.78
,590
9 5
100 kg M), the effect of IBC was no longer significant.
2
044
631
103
Three oxen teams with approximately similar M and
3
0438
653
107
in different body condition (team 1: ‘poor’ IBC,
1
0.98
725
1 2 0
719 kg, team 2: ‘medium’ IBC, 721 kg; team 3: ‘good’
s
1 .oo
744
124
IBC, 739 kg) developed similar power output (team
6
0.96
713
120
1: 775 W; team 2: 697 W; team 3: 741 W).
7
0.99
739
1 2 6
se.
0.07
5
1
Differences
in speed and power output due to week
of work were significant (P < 0.01, Table 2). Speed
Early in lactation increases in DM intake were seen
and power output increased steadily over weeks for
in cows with a low body condition as compared with
a11 teams irrespective of their IBC. Even though oxen
those in a better body condition (Jones and
lost body weight throughout the experiment, there
Garnsworthy, 1989). As suggested by Faverdin ct nl.
was a significant weekly increase of 0.035 m/s and
(1995), undernutrition induces an increase in food
25.1 W in speed and power, respectively.
intake when food unavailability is no longer a
limiting factor.
Lap number had a significant effect (P < 0.01) on
speed and power output. Two contrasting phases
The steady increase in food intake and power output
were observed in the pattern of power output each
over 7 weeks observed in this study suggests that
day. During the first three laps of work (1st h) there
oxen undcrwent an adaptation to work during the
was an increase of 0.01 tn/s and 9 W in speed and
first days of work. Thcy became more adapted as
power each lap. In the second phase starting from
work went on and they were therefore able to
the fourth lap a steady decline in speed and work
increase their intake. Bartholomew cf d. (1994) also
output of 0.014 m/s and 10.3 W were observed for
attributed increases in speed of working teams over
each lap completed.
time to adaptation to work. Therefore, working
during the dry season would bave the adïantage
Discussion
that oxen are fit when cultivation starts and do not
As in previous studies (Fall ct nl., 1997) work did not
have to undergo this adaptation period in the
affect intake of millet stover. These results are
cropping season.
consistent with most other studies that indicated no
differences
in intake in working animals as
During the preparation phase of the experiment oxen
compared with animals at-rest (reviewed by Pearson
were given food to reach the targeted body condition
and Dijkman, 1993). Poor body condition before
and live weight. This was however difficult to
work was conducive to higher intake of millet stover
achieve because live-weight changes were associated
than good body condition before work over 7 weeks.
with changes in body condition. At the start of the
”
.
Nutrition of draught cattle - 3
231
experiment heavier animals tended to have better
820 N (Khibe and Bartholomew, 1993). They could
body condition than lighter animais. In order to
however pull heavily loaded carts without undue
minimize the confounding effect of M and IBC on the
stress.
rate of work (W) in different treatment groups,
power output was expressed relative to M (W per
Jt was expected that weight losses would adversely
100 kg M). The use of power relative to M to
affect work output. In this experiment power output
investigate the effect of IBC on work performance
improved over weeks while oxen were undergoing
was based on the assumption that oxen in ‘good’ IBC
weight losses. The same trends in live-weight change
had a higher fat content per kg M than oxen in ‘poor’
and power output were reported by Bartholomew et
or ‘medium’ IBC.
RI. (1993). Therefore the weight losses oxen cari
tolerate before work output is affected are difficult to
Power output is a function of speed and draught
estimate. Continuous and severe weight losses cari
force. The latter was set in this experiment to be
compromise the health of the animal, or even its life.
proportional to M. Therefore any advantage of oxen
However in this experiment, oxen were able to
in good condition over those in poor condition
regain li\\,e weight at a rapid rate when they had
would bave been expressed as a higher average
access to good quality pasture during the rainy
walking speed. As compared with oxen in poor
season. It took them 4 to 6 weeks to reach their
condition, oxen in good condition would perform
pre-work live weight.
work at a higher rate or for a longer period of time
because they would have more body reserves to
At the start of the cropping season, various body
draw on to fuel muscle activity. However walking
conditions are found in oxen because of differences
speed was not significantly affected by body
in food resources, management practices and the
condition in this experiment. Furthermore, the effect
disease situation in farming systems in the semi-arid
of IBC on power output was no longer significant
zone. Observations from this study suggest that a
when power output was expressed in relation to M.
bodv condition score of between 2 and 3 as defined
The similarity of power output relative to M for a11
by fiicholson and Butterworth (1986) would be a low
oxen suggested that animais with same body mass,
critical score below which work may irreversibly
irrespective of fat content, generated the same power
damage the oxen’s health. The ideal body condition
output This suggests that power output is more
score would range between 4 and 6. These animals
dependent on body mass than body condition. Oxen
are not too fat nor too lean and cari perform well if
in good condition did not out-perform oxen with
they are in good health. Oxen with a body condition
equal M but in poor body condition. These results
score of over 6 may be too fat to move comfortably
are consistent with those reported by Bartholomew cf
and are more susceptible to heat stress than leaner
al. (1994). These authors evaluated the relative
oxen. Moreover, the feeding level required to reach a
importance of body weight and body condition on
body condition score over 6 is unlikely to be
work performance by applying the same load of
profitable as far as feeding for work performance is
367 N to groups of oxen weighing 33.0 and 360 kg
concerned.
and in good and poor condition. Light oxen in good
condition could not sustain the work level applied.
Oxen should be supplemented during the cropping
They concluded that live weight rather than body
season when work is performed for more than 6
condition is the single most important determinant
weeks. If the working period is short, weight losses
of work output. Therefore, it seems that farmers
could be tolerated as animais Will regain their live
should be encouraged
to Select large-framed animals
weight rapidly. Supplementary feeding is however
for draught purposes. This may be constrained by
worthwhile for animals scheduled to be sold for
the fact that more Young animals are being used for
meat after work, even for short working periods, SO
draught purposes with a rapid on-farm turn-over
that work does not adversely affect their market
rate of these animals, apparently driven by an
value. In
semi-arid
w e s t
Africa
food
attractive meat market. There is therefore a genuine
supplementation under these circumstances may be
need to investigate feeding and management
based on whole cotton seed, groundnut,
sesame or
practices that Will optimize power and meat output
cotton cakes if available. These foods are rich in
in these farming systems.
protein and energy and provide substrates
(long-chain fatty acids, glycogenic compounds) that
Good IBC may allow work for longer periods of time
cari be directly used for work.
in the cropping season. In the present study, oxen in
‘poor’ IBC sustained average draught forces of
Acknowledgements
682 N. These animals might not be able to perform
The
authors
thank
thc
Overseas
Devclopmcnt
ploughing or ridging for extended periods because
Administration (ODA) of the United Kingdom and the
these two activities require draught forces of about
International Livestock Rese,wch Institute for funding this
-
--------~-~~
232
Fall, Pearson and Fernhdez-Rivera
study and E. A. Hunter for statistical advice during
Jones, G. P. and Garnsworthy, P. C. 1989. The effects of
preparation of the manuscript.
dictary energy content on thc rcsponse of dairy cows to
body condition at calving. Auirml Produifiorl 49: 183-191.
Ref erences
Khibe, T. and Bartholomew, P. W. 1993. Recherches aux
possibilités de conditionnement des boeufs de labour à base
Astatke, A., Reed, J. D. and Butterworth, M. H. lY8h. Effect
de paille de brousse enrichie 3 l’ut+ et .@valuation de son
of diet restriction on work performance and weight loss of
impact sur leur capacité de travail. In Rcmrrh .for
local Zebu and Friesian-Boran crossbred oxen. II.CR BI&+~JZ
de7~Gmr1~t o f ,mimnl trncth ~II kV& Afiico (ed. P. R .
23: 11-14.
Lawr&ce, K.‘Lawrence, J. T. Dijkman and P. H; Starkey).
Bartholomew, P. W., Khibe, T. and Little, D. A. 199-I. Effect
Prwmfi~zgs o,f thc jurrfh workslmp c$ t/w h’csf @kn A~~imz/
of liveweight and body condition on work output from
Trncfh Akt7uork, Kmo, Nigrrin FI-13 ]~/y 1990, pp. 175-179.
draft oxen. Troyic.l7/ A??ima/ Hcnltlr md Prod~~ctim 26: 58-64.
Wcst Africa Animal Traction Network and International
Bartholomew, P. W., Khibe, T., Little, D. A. and Ba, S.
Livestock Centre for Africa (ILCA), Addis Ababa, Ethiopia.
19Y3. Effect of change in body weight and condition during
Lawrence, P. R. and Pearson, R. A. 1985. Factors affecting
the dry season on capacity for work of draft oxcn. Tropical
the measurement of draught force, work output and power
A~~inml Htalth md Prod~rt-lio~i 25: 50-58.
o f oxen. \\ournn/ o f A,~ric~rltwnl
Scicwcr, Cnf~lbrirfxc 1 0 5 :
Dicko, M. S. and Sangaré, M. 1984. DP I’cwplai dc In trnction
703-714.
n1rirmle rrl iOlll’
Sulrlienru~: de d e
/a srrn.‘l~flrc,ntation
Nicholson M. J. and Butterworth, M. 19%. A pide to
a/ir/rrrltnirr.
D o c u m e n t d e p r o g r a m m e n o . A%N 3 .
cowditior~ scoring of~ebu cuttk. International Livestock Centre
International Livestock Centre for Africa, Addis Ababa,
for Africa (ILCA), Addis Ababa, Ethiopin.
Ethiopia.
Pearson, R. A, and Dijkman, J. T. 1994. Nutritional
Fall, A., Pearson R. A. and Lawrence, P. R. 1997. Nutrition
implications of work in draught animais. Pmwdiqs c!f tlw
of draught oxen in semi-arid west Africa. 1. Energy
NlltritiOlz Socr’r~ty 53: 169-179.
cxpenditure by oxen working on soils of different
Pethick, D. W. 1993. Carbohydrate and lipid oxidation
consistencies. Animnl Scictw 64: 209-215.
during excrcise. Alrstrulim @ml c?f Apklrltwol Rrwwlr
Fall, A., Pearson, R. A., Lawrence, P. R., Fernandez-Rivera,
44: 431-141.
S. 1997b. Nutrition of draught oxen in semi-arid west
Preston, T. R. ,and Leng, R. A. 1987. Mntshing rrmimulf
Africa. 2. Effect of work on intakc, digestibility and rate of
prod~rctiorl systmrs 7dth nvnilnblr rfwwccs in flfr frq7ics nfirf
passage of food through the gastro-intenstinal tract in
sub-tropics. Penarnbul Books, Armidale, Australia.
draught oxen given trop residues. A~~imnl Srimce 64:
217-225.
Statistical Analysis Systems *Instittite. 1985. SAS IISCT’S
pidP: sfntistics. Statistical Analysis Systems Institute Inc.,
Faverdin, P., Baumont, R. and Ingvartsen, K. L. 1995.
Cary, NC.
Control and prediction of feed intake in ruminants. In
Rccmt dcvrlopmts irl thc rlrrtritim of herhinoree (ed. M.
Wilson, R. T. 1987. Livestock production in central Mali:
factors influencing growth and liveweight in agro-pastoral
Journet, E . Genet, M.-H. Farce, M. Theriez and C .
Demarquilly). Procwdhp of’ tk~ fuurfl~ ir~lermfionnl
cattle. ‘hyicn/ Animal Hmlfh md Production 19: 103-114.
syrnposiwn on fhe rn~tritim of hrrbivorrs, pp. 95-120. INRA
Editions, Paris
(Rrcriwd 29 ]~~r~my 1996-Accytrd 13 Srptrmber 1996)