Nutrition of draught oxen in semi-arid west Africa. 2....
Nutrition of draught oxen in semi-arid west Africa. 2. Effect of work
on intake, apparent digestibility and rate of passage of food through
the gastro-intestinal tract in draught oxen given trop residues
A. Fall’, R. A. Pearson’, P. R. Lawrence’t and S. Fernhndez-Rivera?
Abstract
Tzclo experimenfs were co~~dwtcd to inzmfigate thc wlationships betzueen zoork mil irztnkc md digestion @foorl b!y
draught oxei7 gicen millet stozw. 117 thefirsf cxpcrinmt, intakc of millet stozw, wnfcr infakc, liac u~eighf, plasnla
concerltrafiorzs of triiocloflzyronine, thyroxine nnd urcn-nifrogerz wcre measurcd in 18 animais thnt workcd 0, 2 OI
4 h/day in scqzlence during fhree 3-zocek experimental yeriods. Digestibilify alld rate of passage ~ffood residws
through the digesfizw tract zuere mcasured in a secoud experimcnf 017 12 alzimals zuorkin,q either 0, 2.5 or 5 h/day irl
sequcnce during tkrce 2-zueek experimcntal periods. Feediflg behaviour zoas monitored OJZ six nnimnls zoorkirlg
cifher 0, 2.5 or 5 h/day. York did not affect intake of millet stover, apparent digcstibilities ami fhe rat? cfpnssagc of
digesta fhrougk the yastro-intestinal tract. This suggests fhnf fhe 17ut~ie~t wpply fmm irztake of uoqhngcs lny
workirlg oxetl is wzlikely to be sufficient to compensate for the extra energy rxpended during zoork. Food intakc was
aficted by the quality of the millet sfover r~ffercd. T/7c lezle of infake of millet stover was proportiotlal to the a1m1rnf
of leaves in the stover. Food infake incrcased nlso as work progressed. Hozoever,
animais mobilized their bodly
reserzles to perforrn work. Animais colzsumed more zuater OII workiny dnys fhan 011 days they zoere af-rcsf ii1 shade.
The heat stress that working arlirnals zoere subjcctcd to did nef 17ppm~ to interfcre mith flzeiy digcstizv fllrlctiou.
Keywords: digestibility, draughf anirnals, food infake, millet sfovcr, zoork.
Introduction
to the depletion of circulating energy substrates.
Ideally, draught oxen must consume sufficient food
With
sustained
exercise,
muscles
draw
before and during the cropping season SO they cari
energy-yielding substrates from body reserves. Work
start work with a reasonable live weight and perform
therefore imposes a higher energ? demand, which
work. However the scarcity and poor quality of
would be expected to stimulate mtake to supply
foods available before and during the early part of
energy to muscle and to replenish depleted body
the cropping season in semi-arid areas often limit
nutrients (Weston, 1985). The occurrence of fatigue is
their nutrient intake. Food intake cari be influenced
a natural result of sustained muscular activity. The
positively or negatively by work through direct or
desire to eat and ruminate may be suppressed by
indirect mechanisms. Direct effects of work on food
fatigue (Pearson and Lawrence, 1992). Physiological
intake occur through physiological changes resulting
changes in working animals also include increased
from exercise. Muscular activity induces a higher
body temperature due to heat gained from solar
metabolic rate in working animals as compared with
radiation and increased metabolism during work.
animals at-rest (Preston and Leng, 1987). This leads
The resulting heat stress could depress food intake in
working animals (Collier and Beede, 1985).
t Present address: Centre for Agriculture in the Tropics
One indirect effect of work on intake stems from the
and Subtropics. Institute of Animal Production,
reduced time animals have access to food. Limited
Department of Animal Nutrition and Aquaculture.
time available to eat and ruminate is a major
Universitat Hohenheim (480), D-7093, Stuttgart, Germany.
constraint to increased food intake in working
2 1 7
.
218
Fall, Pearson, Lawrence and Fernandez-Rivera
ruminants (Pearson and Lawrence, 1992). Time of
feeding also affects food intake. Bakrie and Teleni
(1991) reported reduced food intake by animais
given roughages before work as compared w i t h
animals given food aftcr work.
Millet stover
Concentrate
Experiment
1
2
1
2
Work also has the potential to affect digestibility of
foods by oxen directly and indirectly through
Dry mattcr (DM) (g/kg)
902
910
905
935
changes 111 a range of factors including increases in
Crude protein (CI’)
33
36
293
177
body temperature, food particle residence time in the
Grass energy (C;E)
gastro-intestinal
tract
and
effectiveness of
(MJ/kg DM)
17.5
18.0
18.1
14.;
mastication on particle breakdown (Weston, 1985).
Organic matter (OM)
964
973
898
756
Positive effects of work on food digestibility may
Neutra-dctergent
s t e m f r o m t h e e n h a n c e m e n t o f m i c r o b i a l
fibre (NDF)
789
7 8 1
293
197
Acid-detergent fibre (ADF) 539
519
131 72
fermentation though greater mixing of rumen
Hemicellulose (HEM)
274
261
162
1 2 5
contents due to exercise (Matthewman and Dijkman,
1993) and higher but moderate body temperatures
resulting from work. Detrimental effects of work on
food digestibility may result from the shift of blood
16.00 and 18.00 h. Food troughs and buckets were
flow from the gut to muscles and peripheral tissues,
withdrawn from a11 the pens when oxen on exercise
reduction in meal frequencies (Matthewman and
schedule were working. Therefore a11 animals had
Dijkman, 1993), and the less thorough mastication of
equal time of access to food, but assuming work
food because of limited time to ruminate (Pearson
restricted rumination, they did not have similar time
and Smith, 1994). There is need for a clear
available to ruminate. Orts were regularly removed
understanding of the relation between work and
from troughs. Oxen at rest were tethered out of the
digestive physiology for the feeding management of
yens and in direct sunlight when other teams were
draught oxen
to be improved. This study
working.
investigated the relationships between food intake
and the efficiency of utilization of foods and work
£xpeui~n~~tnl des@. A Latin-square cross-over design
performance.
with repeated measures was adopted for this
experiment. Treatment consisted of the number of
hours worked per day: 0, 2 and 4 h/day pulling a
Material and methods
loaded sledge along a flat circuit or performing
common field operations (cultivation). Oxen were
Ayrirnnls ancl fce&rlg. This experiment was conducted
allotted according to their initial body weight in
from July to September 1993 at the International
three groups with average weights of 245, 273, and
Livestock Research Institute (ILRI), International
390 kg for groups 1, 2 and 3, respectively. Oxen in
Crop Research Institute for the Semi-Arid Tropics
groups 1,2 and 3 were allotted to squares 1, 2 and 3,
(ICRISAT) Sahelian Centre, Sadore, Niger. Eighteen
respectively SO that each square was formed with
local zebu oxen, aged 4 to 8 years, average live
animals of similar liste weight. Rows of each square
weight 302 (s.e. 78) kg, were used. Oxen were housed
w e r e formed b y o x
teams whereas columns
in individual concrete-floored pens roofed with zinc
represented experimental periods. The experiment
sheets. Each pen was fitted with a halved empty oil
lasted 9 weeks, which was divided into three 3-week
drum as a food trough and a graduated metal bucket
periods. Treatments were applied in sequence to
for water. Individual pens were separated with
teams during experimental pcriods. During each
wooden planks to prevent mixing of food and food
period, three teams were idle, three teams were
spillages.
working 2 h/day and three teams were working
4 h/day. Each team worked 3 days/week. Teams
Oxen were trained to pull common farm implements
working for 4 h/day worked 2 h in the morning and
and 55-kg metal sledges. They were given chopped
2 h in the afternoon.
millet stover nd libitum except during the working
periods. Millet stover was supplemented with a
n/Ieb?surerneizts. Work output, distance travelled and
concentrate mix made Ltp (g/kg) of wheat bran (600),
elapsed working time were continuously measured
groundnut cake (300) and bone meal(lO0) at a rate of
using an ergometer (Lawrence and Pearson, 1985).
21.3 g/kg M’1’Ts pcr day (Table 1). The concentrate
W e e k l y b l o o d s a m p l e s w e r e t a k e n f o r t h e
was offered when animals returned from work in the
determination of plasma urea-N (PLJN), thyroxine
morning at about 11.00 h. Millet stover was given
(TJ) and triiodothyronine (T3). Body weight was
when oxen finished eating the concentrate, and at
measured every week. Food offered and refusals
Nutrition of draught cattle - 2
219
were weighed every day. Refusals on the floor and
Eqwinwrzt 2
food left in troughs were collected separately
A~~imnls nnd feeding. This experiment was conducted
because of contamination of floor spillage by urine
from December 1994 to February 1995 at the
and water.
ICRISAT Sahelian Centre in Niger. Twelve oxen,
aged 4 to 7 years, average weight 288 (s.e. 11) kg, at
Lnborntory nnnlysis. Daily food samples were pooled
the start of the experiment, were used. They were
each week. A sample was taken and dried in a forced
housed as in experiment 1. Al1 oxen were given
air-oven to constant weight at 55°C and ground to
chopped millet stover ncE libitum except during the
pass a l-mm screen. The following determinations
working hours. The stover was chopped by hand to
were made on the weekly pooled samples of foods:
lengths of about 12 to 20 cm. The millet stover was
acid-detergent fibre (ADF), neutral-detergent fibre
supplemented with a concentrate mix made up (g/
(NDF), nitrogen (N) gross energy (GE), ash and
kg) of wheat bran (400), groundnut cake (300), rock
organic matter (OM) according to the Association of
phosphate (lOO), crushed bone (100) and common
Officiai Analytical Chemists (1990).
salt (100) (Table 1). The concentrate was given at a
daily rate of 10 g dry matter per M0.7s at 12.00 h after
Plasma T, was analysed using the fluorescence
the morning working session. Daily food allowance
polarization immunoassay technique with an Abbot
was adjusted SO that refusais were at least equal to a
TDx Analyser (Abbot Laboratories, USA). The
proportion of 0.50 of food offered.
analysis for plasma T, used the IMx total T, assay
based on the microparticle enzyme immunoassay
Treatments. Treatments consisted of levels of work
technique (Abbot Laboratories, USA). PUN was
performed: 0, 2.5 and 5 h/day achieved by walking
assayed by an enzymatic method using a Bayer
0, 6 and 12 km/day, respectively. Each team in an
Diagnostic RA-2000 random access chemistry
exercise treatment pulled a metal sledge loaded with
analyser (Bayer Diagnostics, Basingstoke, UK).
weights SO that the draught force exerted was
equivalent to proportionately 0.10 of the team live
Dntn niznlysis. The following statistical mode1 was
weight. Work was performed continuously, 7 days/
used to analyse food and water intake, weight
week, pulling the sledge around a flat circuit. Work
change, plasma thyroid hormones and PUN
stopped when the set distance or set time was
concentrations:
completed or when one of the oxen in the team was
unwilling to continue.
Expevimentnl des@ A Latin-square crossover design
was used. Twelve oxen were assigned to the three
where: Y = dependent variable (food intake, water
treatment groups, two teams in each group. The
intake, M change, plasma thyroid hormones
rows of the squares represented individual oxen,
concentration, urea-nitrogen concentration); u =
whereas columns were experimental periods. The
overall mean; Si = effect of ith square, i = 1, 2, 3;
experiment lasted 10 weeks divided into five 2-week
T(S),,, = effect of the jth team nested within ith
periods. Observations were repeated every week in
square, j = 1,2, 3 ; P(s),i]k = effect of the kth
periods 1, 3 and 5. No treatment was applied during
experimental period nested within ith square, k =
periods 2 and 4 to dissipate carry-over effects from
1,2,3; Aci, = effect of the Ith work level, 1 = 1: 0 h/
previous periods. Each square included oxen of
day, 1 = 2: 2 h/day and 1 = 3: 4 h/day; W,,, = effect of
similar live weight. Treatments were applied in
the rrrth week, llz = 1,2,3; W X Piirk = interaction
sequence during experimental periods SO that during
between the nrth week and the kth period; W X A,,,, =
each period four oxen were idle, four oxen were
interaction between the rnth week and the lth work
working 2.5 h/day and four oxen were working 5 h/
level; W X T(S),,,,, = interaction between the mth
day.
week and the jth team in the ith square; E,,,,,, = effect
peculiar to the jth team in the ith square subjected to
Mensurements.
Sampling,
measurements
a n d
the lth level of work in 172th week of the kth period.
laboratory analyses were as in experiment 1 with the
following amendments and additions.
The term T(S),,,, was used as the error term to test the
effect of work. The sums of squares for treatment
Work: an ergometer was used only during the
and week were further partitioned into single
preparation phase of this experiment, to measure
degrees of freedom using polynomial contrasts (i.e.
work performed, distance travelled and elapsed
A,). Weekly live-weight changes were estimated by
working time for different known work loads. A
regression analysis and were further subjected to
regression analysis of force on work load was
analysis of variante using generalized linear models
derived and used to determine the load required for
(Statistical Analysis Systems Institute (SAS), 1985).
each team SO that the draught force exerted was
on intake, apparent digestibility and rate of passage of food through
the gastro-intestinal tract in draught oxen given trop residues
A. Fall’, R. A. Pearson’, P. R. Lawrence’t and S. Fernhndez-Rivera?
Abstract
Tzclo experimenfs were co~~dwtcd to inzmfigate thc wlationships betzueen zoork mil irztnkc md digestion @foorl b!y
draught oxei7 gicen millet stozw. 117 thefirsf cxpcrinmt, intakc of millet stozw, wnfcr infakc, liac u~eighf, plasnla
concerltrafiorzs of triiocloflzyronine, thyroxine nnd urcn-nifrogerz wcre measurcd in 18 animais thnt workcd 0, 2 OI
4 h/day in scqzlence during fhree 3-zocek experimental yeriods. Digestibilify alld rate of passage ~ffood residws
through the digesfizw tract zuere mcasured in a secoud experimcnf 017 12 alzimals zuorkin,q either 0, 2.5 or 5 h/day irl
sequcnce during tkrce 2-zueek experimcntal periods. Feediflg behaviour zoas monitored OJZ six nnimnls zoorkirlg
cifher 0, 2.5 or 5 h/day. York did not affect intake of millet stover, apparent digcstibilities ami fhe rat? cfpnssagc of
digesta fhrougk the yastro-intestinal tract. This suggests fhnf fhe 17ut~ie~t wpply fmm irztake of uoqhngcs lny
workirlg oxetl is wzlikely to be sufficient to compensate for the extra energy rxpended during zoork. Food intakc was
aficted by the quality of the millet sfover r~ffercd. T/7c lezle of infake of millet stover was proportiotlal to the a1m1rnf
of leaves in the stover. Food infake incrcased nlso as work progressed. Hozoever,
animais mobilized their bodly
reserzles to perforrn work. Animais colzsumed more zuater OII workiny dnys fhan 011 days they zoere af-rcsf ii1 shade.
The heat stress that working arlirnals zoere subjcctcd to did nef 17ppm~ to interfcre mith flzeiy digcstizv fllrlctiou.
Keywords: digestibility, draughf anirnals, food infake, millet sfovcr, zoork.
Introduction
to the depletion of circulating energy substrates.
Ideally, draught oxen must consume sufficient food
With
sustained
exercise,
muscles
draw
before and during the cropping season SO they cari
energy-yielding substrates from body reserves. Work
start work with a reasonable live weight and perform
therefore imposes a higher energ? demand, which
work. However the scarcity and poor quality of
would be expected to stimulate mtake to supply
foods available before and during the early part of
energy to muscle and to replenish depleted body
the cropping season in semi-arid areas often limit
nutrients (Weston, 1985). The occurrence of fatigue is
their nutrient intake. Food intake cari be influenced
a natural result of sustained muscular activity. The
positively or negatively by work through direct or
desire to eat and ruminate may be suppressed by
indirect mechanisms. Direct effects of work on food
fatigue (Pearson and Lawrence, 1992). Physiological
intake occur through physiological changes resulting
changes in working animals also include increased
from exercise. Muscular activity induces a higher
body temperature due to heat gained from solar
metabolic rate in working animals as compared with
radiation and increased metabolism during work.
animals at-rest (Preston and Leng, 1987). This leads
The resulting heat stress could depress food intake in
working animals (Collier and Beede, 1985).
t Present address: Centre for Agriculture in the Tropics
One indirect effect of work on intake stems from the
and Subtropics. Institute of Animal Production,
reduced time animals have access to food. Limited
Department of Animal Nutrition and Aquaculture.
time available to eat and ruminate is a major
Universitat Hohenheim (480), D-7093, Stuttgart, Germany.
constraint to increased food intake in working
2 1 7
.
218
Fall, Pearson, Lawrence and Fernandez-Rivera
ruminants (Pearson and Lawrence, 1992). Time of
feeding also affects food intake. Bakrie and Teleni
(1991) reported reduced food intake by animais
given roughages before work as compared w i t h
animals given food aftcr work.
Millet stover
Concentrate
Experiment
1
2
1
2
Work also has the potential to affect digestibility of
foods by oxen directly and indirectly through
Dry mattcr (DM) (g/kg)
902
910
905
935
changes 111 a range of factors including increases in
Crude protein (CI’)
33
36
293
177
body temperature, food particle residence time in the
Grass energy (C;E)
gastro-intestinal
tract
and
effectiveness of
(MJ/kg DM)
17.5
18.0
18.1
14.;
mastication on particle breakdown (Weston, 1985).
Organic matter (OM)
964
973
898
756
Positive effects of work on food digestibility may
Neutra-dctergent
s t e m f r o m t h e e n h a n c e m e n t o f m i c r o b i a l
fibre (NDF)
789
7 8 1
293
197
Acid-detergent fibre (ADF) 539
519
131 72
fermentation though greater mixing of rumen
Hemicellulose (HEM)
274
261
162
1 2 5
contents due to exercise (Matthewman and Dijkman,
1993) and higher but moderate body temperatures
resulting from work. Detrimental effects of work on
food digestibility may result from the shift of blood
16.00 and 18.00 h. Food troughs and buckets were
flow from the gut to muscles and peripheral tissues,
withdrawn from a11 the pens when oxen on exercise
reduction in meal frequencies (Matthewman and
schedule were working. Therefore a11 animals had
Dijkman, 1993), and the less thorough mastication of
equal time of access to food, but assuming work
food because of limited time to ruminate (Pearson
restricted rumination, they did not have similar time
and Smith, 1994). There is need for a clear
available to ruminate. Orts were regularly removed
understanding of the relation between work and
from troughs. Oxen at rest were tethered out of the
digestive physiology for the feeding management of
yens and in direct sunlight when other teams were
draught oxen
to be improved. This study
working.
investigated the relationships between food intake
and the efficiency of utilization of foods and work
£xpeui~n~~tnl des@. A Latin-square cross-over design
performance.
with repeated measures was adopted for this
experiment. Treatment consisted of the number of
hours worked per day: 0, 2 and 4 h/day pulling a
Material and methods
loaded sledge along a flat circuit or performing
common field operations (cultivation). Oxen were
Ayrirnnls ancl fce&rlg. This experiment was conducted
allotted according to their initial body weight in
from July to September 1993 at the International
three groups with average weights of 245, 273, and
Livestock Research Institute (ILRI), International
390 kg for groups 1, 2 and 3, respectively. Oxen in
Crop Research Institute for the Semi-Arid Tropics
groups 1,2 and 3 were allotted to squares 1, 2 and 3,
(ICRISAT) Sahelian Centre, Sadore, Niger. Eighteen
respectively SO that each square was formed with
local zebu oxen, aged 4 to 8 years, average live
animals of similar liste weight. Rows of each square
weight 302 (s.e. 78) kg, were used. Oxen were housed
w e r e formed b y o x
teams whereas columns
in individual concrete-floored pens roofed with zinc
represented experimental periods. The experiment
sheets. Each pen was fitted with a halved empty oil
lasted 9 weeks, which was divided into three 3-week
drum as a food trough and a graduated metal bucket
periods. Treatments were applied in sequence to
for water. Individual pens were separated with
teams during experimental pcriods. During each
wooden planks to prevent mixing of food and food
period, three teams were idle, three teams were
spillages.
working 2 h/day and three teams were working
4 h/day. Each team worked 3 days/week. Teams
Oxen were trained to pull common farm implements
working for 4 h/day worked 2 h in the morning and
and 55-kg metal sledges. They were given chopped
2 h in the afternoon.
millet stover nd libitum except during the working
periods. Millet stover was supplemented with a
n/Ieb?surerneizts. Work output, distance travelled and
concentrate mix made Ltp (g/kg) of wheat bran (600),
elapsed working time were continuously measured
groundnut cake (300) and bone meal(lO0) at a rate of
using an ergometer (Lawrence and Pearson, 1985).
21.3 g/kg M’1’Ts pcr day (Table 1). The concentrate
W e e k l y b l o o d s a m p l e s w e r e t a k e n f o r t h e
was offered when animals returned from work in the
determination of plasma urea-N (PLJN), thyroxine
morning at about 11.00 h. Millet stover was given
(TJ) and triiodothyronine (T3). Body weight was
when oxen finished eating the concentrate, and at
measured every week. Food offered and refusals
Nutrition of draught cattle - 2
219
were weighed every day. Refusals on the floor and
Eqwinwrzt 2
food left in troughs were collected separately
A~~imnls nnd feeding. This experiment was conducted
because of contamination of floor spillage by urine
from December 1994 to February 1995 at the
and water.
ICRISAT Sahelian Centre in Niger. Twelve oxen,
aged 4 to 7 years, average weight 288 (s.e. 11) kg, at
Lnborntory nnnlysis. Daily food samples were pooled
the start of the experiment, were used. They were
each week. A sample was taken and dried in a forced
housed as in experiment 1. Al1 oxen were given
air-oven to constant weight at 55°C and ground to
chopped millet stover ncE libitum except during the
pass a l-mm screen. The following determinations
working hours. The stover was chopped by hand to
were made on the weekly pooled samples of foods:
lengths of about 12 to 20 cm. The millet stover was
acid-detergent fibre (ADF), neutral-detergent fibre
supplemented with a concentrate mix made up (g/
(NDF), nitrogen (N) gross energy (GE), ash and
kg) of wheat bran (400), groundnut cake (300), rock
organic matter (OM) according to the Association of
phosphate (lOO), crushed bone (100) and common
Officiai Analytical Chemists (1990).
salt (100) (Table 1). The concentrate was given at a
daily rate of 10 g dry matter per M0.7s at 12.00 h after
Plasma T, was analysed using the fluorescence
the morning working session. Daily food allowance
polarization immunoassay technique with an Abbot
was adjusted SO that refusais were at least equal to a
TDx Analyser (Abbot Laboratories, USA). The
proportion of 0.50 of food offered.
analysis for plasma T, used the IMx total T, assay
based on the microparticle enzyme immunoassay
Treatments. Treatments consisted of levels of work
technique (Abbot Laboratories, USA). PUN was
performed: 0, 2.5 and 5 h/day achieved by walking
assayed by an enzymatic method using a Bayer
0, 6 and 12 km/day, respectively. Each team in an
Diagnostic RA-2000 random access chemistry
exercise treatment pulled a metal sledge loaded with
analyser (Bayer Diagnostics, Basingstoke, UK).
weights SO that the draught force exerted was
equivalent to proportionately 0.10 of the team live
Dntn niznlysis. The following statistical mode1 was
weight. Work was performed continuously, 7 days/
used to analyse food and water intake, weight
week, pulling the sledge around a flat circuit. Work
change, plasma thyroid hormones and PUN
stopped when the set distance or set time was
concentrations:
completed or when one of the oxen in the team was
unwilling to continue.
Expevimentnl des@ A Latin-square crossover design
was used. Twelve oxen were assigned to the three
where: Y = dependent variable (food intake, water
treatment groups, two teams in each group. The
intake, M change, plasma thyroid hormones
rows of the squares represented individual oxen,
concentration, urea-nitrogen concentration); u =
whereas columns were experimental periods. The
overall mean; Si = effect of ith square, i = 1, 2, 3;
experiment lasted 10 weeks divided into five 2-week
T(S),,, = effect of the jth team nested within ith
periods. Observations were repeated every week in
square, j = 1,2, 3 ; P(s),i]k = effect of the kth
periods 1, 3 and 5. No treatment was applied during
experimental period nested within ith square, k =
periods 2 and 4 to dissipate carry-over effects from
1,2,3; Aci, = effect of the Ith work level, 1 = 1: 0 h/
previous periods. Each square included oxen of
day, 1 = 2: 2 h/day and 1 = 3: 4 h/day; W,,, = effect of
similar live weight. Treatments were applied in
the rrrth week, llz = 1,2,3; W X Piirk = interaction
sequence during experimental periods SO that during
between the nrth week and the kth period; W X A,,,, =
each period four oxen were idle, four oxen were
interaction between the rnth week and the lth work
working 2.5 h/day and four oxen were working 5 h/
level; W X T(S),,,,, = interaction between the mth
day.
week and the jth team in the ith square; E,,,,,, = effect
peculiar to the jth team in the ith square subjected to
Mensurements.
Sampling,
measurements
a n d
the lth level of work in 172th week of the kth period.
laboratory analyses were as in experiment 1 with the
following amendments and additions.
The term T(S),,,, was used as the error term to test the
effect of work. The sums of squares for treatment
Work: an ergometer was used only during the
and week were further partitioned into single
preparation phase of this experiment, to measure
degrees of freedom using polynomial contrasts (i.e.
work performed, distance travelled and elapsed
A,). Weekly live-weight changes were estimated by
working time for different known work loads. A
regression analysis and were further subjected to
regression analysis of force on work load was
analysis of variante using generalized linear models
derived and used to determine the load required for
(Statistical Analysis Systems Institute (SAS), 1985).
each team SO that the draught force exerted was