Small Ruminant
Research
Small Ruminant Research 20 ( 1996) 23-29Nutritional consequences among ingredients of
free-choice feeding Awassi lambs
M. Gargiilii”, H.R. Kutlu”,*, E. Demirb, 0. ozti.irkcan”, J.M. ForbesC
“Cukurova University. Agricultural Faculty, Departmentof
Animal Science, 01330 Adana, TurkeybGaziosmanpaga University, Agricultural Faculty, Department of Animal Science, Tokat, Turkey ‘Leeds University, Department of Animal Physiology and Nutrition, LS2 9JT, Leeds, UK
Accepted 1 June 1995
Abstract
The present study was designed ( 1) to test whether growing lambs can select an adequate diet when offered an unrestricted choice among feed ingredients, and (2) to determine whether choice feeding compared with single-feeding would alter the fattening performance of lambs reared under normal husbandry conditions. Thirty, 3-month-old male Awassi lambs were used. Each dietary treatment had three replicate groups, comprising five lambs each. Ground barley, wheat bran, cottonseed meal and
alfalfa hay were used as the main feed ingredients. One of the dietary treatments (choice-fed) was offering the main feed
ingredients as choices, while the other (single-fed) was feeding with a control diet (containing 2.4 Meal ME kg- ’ and 17.0% CP) prepared using the same feed ingredients for a 56 day fattening period. Lambs offered choices among the feed ingredients could select diets to meet their nutrient requirements according to age throughout the fattening period. For the entire experimental period the choice-fed group attained higher daily live-weight gain (346 g vs. 299 g; P < 0.01) and better feed conversion efficiency (feed/gain; 5.1 vs. 5.7, P < 0.05) than the group given access to the single diet. No difference (P > 0.05) between daily feed intake of the choice-fed and single-fed groups ( 1776 g vs. 1704 g) was observed. The lambs consistently selected diets high in CP relative to ME when allowed to choose feed ingredients. It is concluded that lambs, when given choices among feed ingredients (ground barley, cotton seed meal, wheat bran and alfalfa straw), can select diets that meet their nutrient requirements according to stage of maturity and they attain better growth performance than lambs fed with a single diet.
Keywords: Lamb; Diet selection; Feed ingredients; Fattening performance -
1. Introduction
In animal nutrition diet selection techniques have
elicited a great deal of interest. It has been thought that
dietary self-selection studies could be of value in under-
standing nutritional requirements of animals, especially
during changes in nutrient requirements as affected by
physiological or environmental factors. Diet selection
* Corresponding author. Tel: + 90-322-3386822; Fax: + 90-322- 3386576; E-mail: hrk@cubim2.cc.cu.edu.tr
0921~4488/96/$15.00 0 1996 Elsevier Science B.V. All rights reserved SSDIO921-4488(95)00777-6
techniques may have also the potential to avoid the
need to formulate and manufacture complete diets.
Because of the potential advantages of a choice-feeding system, much research has been carried out to assess the abilities of farm animals to choose their own nutri- ents successfully.
Choice-feeding studies have mostly been focused on
monogastric animals and conducted to test the ability
of chickens or pigs to select a diet that meets their
24 M. GBrgiilii et al. /Small Ruminant Research 20 (1996) 23-29
However, there have been few studies on ruminant animals that test whether these animals can select a diet
of adequate composition between a suitable pair of
feeds or among feedstuffs. Kyriazakis and Oldham
( 1993) attributed the shortage of self-selection studies
in ruminant nutrition to lack of understanding of the
main factors that control feed intake of ruminant
animals.
The literature concerning diet selection in ruminant
animals shows that sheep can select a diet that meets
their protein requirements when they are given a choice
between two feeds varying in protein content, one in excess, the other deficient if given alone (Cooper and Kyriazakis, 1993; Kyriazakis and Oldham, 1993). Kyr- iazakis and Oldham (1993) also reported that sheep
offered feeds containing low and high protein as a
choice avoid excess protein intake and grow at genetic
potential. As far as costs of diet formulation and prep-
aration are concerned, the cheapest, simplest and most attractive option would be where sheep could select an
adequate diet from basic raw feed ingredients offered
separately. However, there is no study in the literature on feed ingredient selection in sheep to allow them to
make their own diet to meet main nutrient (energy,
protein) requirements. Therefore, the present study
was proposed (1) to test whether growing lambs can select an adequate diet when offered an unrestricted
choice among feed ingredients, and (2) to determine
whether choice feeding compared with single-feeding
would alter the fattening performance of lambs reared
under normal husbandry conditions.
2. Materials and methods
Thirty, 3-month-old fat-tailed male Awassi lambs
weaned at 7680 days old were obtained from Ceylan- pinar State Farm and transported to the Experimental
Unit of Cukurova University, Agricultural Faculty,
Department of Animal Science, Adana, Turkey. Lambs
were adapted to the experimental unit and fed a diet
containing 2.4 Meal ME kg-’ and 17.0% CP for 1
week. At the beginning of the experiment lambs were
numbered with ear tags and divided into nine diet treat- ment groups of similar live weight. Each treatment group had three replicates, comprising five lambs each.
The experimental unit was semi-open and naturally
ventilated with two rows, each of three pens, measuring
Table 1
Composition of the control diet given to the single-fed group of Awassi male lambs
Ingredients %
Ground barley 47.92
Wheat bran 18.06
Cotton seed meal 21.82
Alfalfa straw 10.00 Feed additives 2.20 Limestone I .I5 Mineral mixture” 0.10 Vitamin mixtureb 0.10 lonophore (Na-lasalocid) 0.25
Calculated composition (per kg fresh diet)
ME (Meal) 2.4 DM (%) 89.70 Crude ash (%) 5.14 CP (%) 17.00 Crude fibre (%) 8.13 Crude oil (%) 2:06 Ca (%) 0.87 P (%) 0.29 Na (%) 0.03
“Mineral mixture contains 10.000 mg Mn, 10.000 mg Fe, 10.000 mg Zn, 5.000 mg Cu, 100 mg Co, 100 mg I, 100 mg Se, 369.650 mg CaCO, per kg.
‘Vitamin mixture contains 15 000 000 1U vitamin A, 3 000 000 IU vitamin D, 30 000 mg vitamin E, 100 000 mg niacin per kg.
400 X 350 cm’ each, to which replicate groups were
randomly allocated. Each pen contained one metal
trough (50 cm wide, 250 cm length, 30 cm deep), which was divided into four for the choice treatments and a metal bucket for fresh water.
Ground barley, wheat bran, cotton seed meal and
alfalfa straw were used as the feed ingredients. One of
the dietary treatment groups (choice-fed) was offered
the main feed ingredients as choices, while the other
treatment group (single-fed) was fed a control diet
prepared from the same feed ingredients (Table 1) with nutrient content slightly higher than that recommended by the NRC (1985).
Each of the main feed ingredients given to choice
feeding was mixed with limestone, mineral mixture,
vitamin mixture and ionophore (Na-lasalocid) at the
levels used in the control diet in order to ensure the
micronutrient intakes of the choice-fed group and also
to prevent any possible effect of micronutrients on feed
M. GBrgiilii et al. /Small Ruminant Research 20 (1996) 23-29 25
Table 2
Initial live weight and the performance biweekly and the entire periods of the experiment of Awassi lambs offered a single diet or feed ingredients as choices
Period (days) Item Single-fed Choice-fed
O-14 ILW (kg)
LWG (g day-‘) FI (g day-‘) ME1 ( Meal day _ ’ ) CPI (g day-‘) FCE (FI/LWG) 31.15kO.64 31.44*0.51 323.80+35.10 376.70 f 3 I .40 1523.OOk95.0 1456.30 +45.0 3.66 If: 0.23 3.34kO.11 258.91 f 16.27 236.97 + 6.17 4.77 f 0.35 3.91 k 0.25 15-28 2942 43-56 LWG (g day-‘) FI (gday-‘) MEI ( Meal day ) CPI (gday-‘1 FCE (FI/LWG) 280.50 f 37.60 259.10 + 24.50 1561.30* 13.00 1775.70 *42.50** 3.75 + 0.23 4.25 + 0.08** 265.43 f 2.28 264.60 i 5.68 5.74kO.64 6.95 * 0.53 LWG (g day-‘) 312.4Ok35.00 380.50 f 24.00* FI (gday-I) 1784.3Ok 143.0 1807.70 + 57.00
MEI ( Meal day - I ) 4.28 + 0.34 4.46kO.13
CPI (g day-‘) 303.34 f 24.35 263.30 * 10.59
FCE (FULWG) 5.77 f0.35 4.80 k 0.32
LWG (g day-‘) FI (g day-‘) ME1 ( Meal day _ ’ ) CPI (g day-‘) FCE (FULWG) 28 I .OO + 29.60 369.50 f 16.90** 1947.00+87.0 2060.30 f 70.0 4.67 k 0.21 4.97kO.15 33 I .05 * 14.82 286.74~11.11 7.04 f 0.59 5.65 +0.53 O-56 LWG (gday-‘) FI (g day-‘) MEI (Meal day ’ ) CPI (gday-‘) FCE (FI/LWG)
*Row values are different at P < 0.05. **Row values are different at P < 0.0 1,
299.40 f 6.80 346.4Ok 9.90**
1704.00_+ 17.0 1776.70+49.0
4.09 + 0.04 4.26kO.10
289.68 f 2.99 263.19+7.54*
5.70*0.13 5.13*0.12*
ILW, initial live weight; LWG. live-weight gain; FI, feed intake (on fresh basis): MEI, metabolizable energy intake; CPI, crude protein intake; FCE, feed conversion efficiency.
saltlick in the troughs and all sections of the troughs. The experiment lasted for 56 days. At the beginning the lambs for feed choices were given the opportunity to experience the ingredients which were subsequently
to be given as choices. During this training period
alfalfa straw was at all times available and the other
ingredients were offered alternately every 3 days for a
period of 6 days. This training period was adapted by
employing relatively short periods but more choices
from the method described by Kyriazakis et al. ( 1990). A training period has been found to be an important
part of the choice-feeding trial both for monogastric,
e.g. pigs (Kyriazakis et al., 1991), and ruminant ani-
mals, e.g. sheep (Kyriazakis and Oldham, 1993). In
the choice-fed group the main feed ingredients were
given always in the same position in the trough through- out the experiment.
Lambs were weighed individually every 2 weeks
after 4 h fasting. Feed or feed ingredient intake and
feed conversion efficiency were recorded biweekly per replicate for each treatment group. Lambs received nat-
ural day light (mean, 13.5 h day-‘) and the ambient
temperature ranged from 10 to 25°C throughout (from 21 March to 15 May 1994) the experiment.
Data obtained in the experiment with respect to per-
26 M. G~rgiilii et al. /Small Ruminant Research 20 (1996) 23-29
(Bek and Efe, 1988). Results obtained in this study are feed ingredients increased daily live-weight gain
presented as means per lamb with ( + ) standard errors
(P
<
0.01 ), daily feed, ME intakes and feed conversionof the mean (SEM) efficiency, while reducing
(P
<
0.05) CP intake.3. Results
Results with respect to daily live-weight gains, daily feed, ME, CP intakes and feed conversion efficiencies of the lambs fed the control diet and lambs offered feed
ingredients during the experiment are shown in Table
2. Results showed that when the lambs were given choices among ground barley, wheat bran, cotton seed meal and alfalfa straw, they successfully regulated their nutrient (energy and protein) intakes as indicated by their attainment of better growth performance than the lambs fed the single diet.
The average composition of the diet selected by the
choice-fed group differed from the composition of the control diet over the entire study (Table 3). When a
comparison is made between the composition of the
diet made by the choice-fed group and the control diet, the proportions of ground barley and alfalfa straw were similar in both diets, while the diet selected by the
choice-fed group contained approximately 10% less
(P-CO.05)
cotton seed meal and approximately 10%more
(P<O.O5)
wheat bran than the control diet(Table 3). Therefore, the diet of the choice-fed group was similar to the control diet in ME content, while
lower
(P
<
0.05) in CP content and CPf ME ratio.During the first 4 weeks of the experimental period
the choice-fed group showed numerically greater live-
weight gain, feed, ME intakes and numerically lower
CP intake than the single-fed group, while no differ- ences in feed conversion efficiency were observed dur- ing the same period. In the second 4 weeks the
choice-fed group achieved greater (P < 0.01) live-
weight gain, numerically higher feed conversion effi-
ciency, feed and ME intakes and significantly lower
(P
< 0.05)
CP intake than the single-fed group. Whenthe data were examined for the entire experimental
period, it was determined that giving choices between
Compositions of the diets selected by the lambs var-
ied throughout the experiment (Table 3). Diets in the
first 4 weeks of the experimental period contained 2.35
Meal ME and 15.5% CP with a ratio of CP/ME of 6.6 while the same lambs in the second 4 weeks made a
diet containing 2.44 Meal ME and 14.9% CP with a
CP/ME ratio of 5.8. Differences in the composition of the diets made by choice-fed groups in the first and
second 4 weeks of the experiment were significant
(PCO.05).
Results with respect to feed intake in the choice-fed group also showed that roughage (alfalfa hay) intake varied throughout the fattening period; roughage intake Table 3
Compositions of the diets selected by the lambs offered feed ingredients as choices throughout the experiment
Period (days) Weigh periods Early Late Overall
o-28 29-56 o-56
o-14 IS-28 29-43 44-56
Ingredients (%) Ground barley Wheat bran Cotton seed meal Alfalfa straw Feed additives” 35.14* 1.59 47.23 + 2.06 30.56 f 0.73 29.73k2.13 17.03 f 1.83 11.25+0.57 15.07*0.45 9.59 f 0.53 2.20+0.00 2.20 * 0.00
Calculated composition (per kg fresh diet)
ME (Meal) 2.30+0.01 2.39 f 0.02 CP (%) 16.29kO.40 14.90 f 0.08 CPIME 7.09*0.19 6.23 +0.03 5s.o0+ 1.44 21.24*0.69 10.63 f 1.33 7.93 f I .oo 2.20+0.00 2.47 + 0.01 14.56 f 0.29 5.91 f0.12 52.15* 1.99 27.50 f I .I1 6.88k 1.17 11.27&0.43 2.20*0.00 2.41 kO.01 13.92 + 0.26 5.77kO.13 41.79+ l.59h 30.10+ l.5lb 13.84f0.64b 12.07f0.15b 2.20*0.00 2.35 fO.Olb 15.52f0.14b 6.6 1 f 0.07h 54.86+ 1.53 24.60 f 0.96 8.64& 1.17 9.70 f 0.57 2.20 f 0.00 2.44*0.01 14.22 f 0.25 5.s4*0.1 I 48.92+ 1.36 27.10? 1.19’ 11.00+0.41~ 10.78 + 0.37 2.20 + 0.00 2.40 f 0.01 14.81*0.08’ 6.18+0.04’
“Contain limestone, mineral and vitamin mixtures and Na-lasalocid at the same level used in the control diet ‘Compositions of the diets made in the first and second 4 week periods are different (P < 0.05).
M. GGrglilir et al. / Smnll Ruminant Research 20 (1996) 23-29 27
of the lambs in the first period was higher (P < 0.05) than the intake in the second period, as the choice-fed group consumed 12% alfalfa straw in total intake in the first 4 weeks, while the same lambs consumed alfalfa straw less than 10% in the second 4 week period.
4. Discussion
Lambs offered choices among ground barley, cotton seed meal, wheat bran and alfalfa straw successfully
regulated their nutrient intakes and attained better
growth performance than those given access to the sin-
gle diet prepared with the same ingredients, which was
in excess of calculated NRC requirements (1985).
Observed protein intake regulation in this study corre- sponds with the findings of Cooper and Kyriazakis ( 1993) and Kyriazakis and Oldham ( 1993). Nutrient
selection of the lambs varied according to stage of
maturity, as energy intake gradually increased, while
protein intakes gradually decreased throughout the
experiment. The energy requirement of 3-month-old
male Awassi lambs appear to be lower during the first stage ( 14 weeks) than during the second stage (5-8 weeks) of an 8 week fattening period, while the protein
requirement of the lambs were higher in the first stage
than in the second stage of the fattening period.
Changes in nutrient intakes of the lambs during the fattening period may be a reflection of changes in com- position of live-weight gain, as the proportion of fat is increased, while the proportion of protein is decreased
in gain with age (Theriez et al., 1982; Orskov, 1987; Webster and Povey, 1990). Protein to energy ratios required in the diet decrease when the animal becomes older or heavier, agreeing with Orskov ( 1987), who noted that the protein to energy ratio required in feed
decreases with increasing stage of maturity, since the
protein to energy ratio in carcasses changes as the lambs increase in weight. Our results with respect to energy and protein intakes of choice-fed group also showed that lambs offered a choice among feed ingredients consumed greater energy and protein than their actual
energy and protein requirements as published by the
NRC ( 1985), but not in the first 2 week period for energy (Figs. 1 and 2). This may indicate that to grow
at genetic potential fat-tailed male Awassi lambs
require higher energy (except during the first 2 week period while the body weights of lambs increased from 31 kg to 35 kg) and protein than the NRC ( 1985) recommendation.
Although the choice-fed group consumed less pro-
tein and similar feed and energy in comparison with
the single-fed group, they attained about 16% higher live-weight gain and about 11% better feed conversion efficiency than the single-fed group. As no significant (P > 0.05) difference between the feed intakes of the
single-fed and choice-fed groups was observed, the
increase in live-weight gain of the choice-fed group is
simply a reflection of the improvement in feed conver-
sion efficiency. The difference between feed conver- sion efficiencies of the choice-fed and single-fed groups
could be partly attributed to the difference between
5.5 - NRC (1985) 5.0 3
3
2
4.5
-j 4.0 !s 3.5 3.0 - - Single-fed (0 .eC-. __--- Choice-fed I 1 35 40 Live Body Weight (kg)28 M. Giirgiilii et al. /.%a11 Ruminant Research 20 (1996) 23-29
320
6 240
220
30 35 40
Live Body Weight (kg)
45
Fig. 2. Dietary protein intake of Awassi lambs in relationship to the NRC ( 1985) recommendation during the study period.
compositions of diets consumed by the groups. Choice-
fed lambs selected their nutrient intakes according to
their physiological status and environmental condi-
tions, while the single-fed lambs had to consume the
control diet, which may be imbalanced due to a high
level of CP compared with the diet selected by the
choice-fed group, particularly as the lambs got older.
There is a metabolic cost to deaminating excess amino
acids (Kyriazakis and Oldham, 1993), which may
result in lowered feed conversion efficiency. In order
to assess the influence of choice feeding on feed con- version efficiency a series of single diets differing in
nutrient compositions (protein or protein to energy
ratio) and choice feeding should be compared with a dose response curve.
Improvement in feed conversion efficiency by
choice feeding may also be attributed to an improved C/N ratio for rumen bacteria and a possible synchro- nization of nutrient supply and its positive effect on microbial growth and microbial protein production in the rumen. Differences between feeds in rates of deg-
radation of protein or energy substrates may cause
short-term imbalances between nitrogen and energy
supply to ruminal microorganisms. Theoretically, such
asynchronous nutrient supply can impair microbial
growth and duodenal protein supply (Newbold and
Rust, 1992; Russell et al., 1992). Studies on the syn-
chronization of nutrient supply suggest that microbial
protein production, efficiency and also performance in
ruminant animals can be increased by feeding diets,
which are formulated to provide synchronized energy
and nitrogen release (Clark et al., 1990; Newbold and Rust, 1992; Russell et al., 1992; Sinclair et al., 1992).
However, the application of synchronous nutrient sup-
ply is difficult in practical feeding, as it requires a spe- cific diet (Sinclair et al., 1992; Henning et al., 1993) or alternate feeding with two diets differing in protein to energy ratio throughout a day (Gijrgiilii et al., 1995). By offering energy and protein sources and roughage
as choices, lambs can have an opportunity to synchro-
nize their nutrient intakes in order to develop a syn-
chronous nutrient supply for rumen microorganism.
The result could be optimal usage of end-product in the
rumen, resulting in an improvement in feed conversion
efficiency, and live-weight gain. As no diurnal feeding pattern (feeding order, frequency and interval of meals
from each feed ingredient) of the lambs were moni-
tored in this study, whether the choice-fed lambs syn- chronized their nutrient intakes is not known. This area, therefore, requires new studies in order to assess the
possible role of choice feeding in synchronization of
nutrient supply in ruminant animals by monitoring
feeding patterns of lambs when they are offered protein and energy sources and also roughages as choices.
Changes in roughage intake of the lambs with age may be associated with changes in energy requirement. In order to increase energy intake, the lambs preferred consuming high energy sources rather than low energy sources. This can be seen with the changes in the pro- portion of ground barley and alfalfa straw according to age in this experiment.
M. Gtirgiilii et al. /Small Ruminant Research 20 (1996) 23-29 29
Growing male Awassi lambs, when given choices
among feed ingredients (ground barley, cotton seed
meal, wheat bran and alfalfa straw), can select diets
that meet their nutrient requirements according to stage
of maturity. They attain better growth performance than
lambs fed the single diet (containing 2.4 Meal ME
kg-’ and 17.0% CP) prepared with the same ingredi- ents, which was in excess of calculated requirements. Feeding ingredients as choices has the potential for ( 1)
simple methodology to ascertain nutrient requirements
of lambs, (2) diet formulation and preparation, (3)
better growth performance in ruminant animals by
increasing feed efficiency, which may be a result of
synchronized nutrient supply to the animal.
References
Bek, Y. and Efe, E., 1988. Experimental Design I. Textbook. Uni- versity of Cukurova, Faculty of Agriculture, Adana, Turkey, pp. 395.
Clark, J.H., Klusmeyer, T.H. and Cameron, M.R., 1990. Synchro- nous energy and protein supply for optimization of rumen fer- mentation. J. Anim. Sci., 68 (Suppl. 1): 736A.
Cooper, S.D.B. and Kyriazakis, I., 1993. The diet selection of lambs offered food choices of different nutrient density. Anim. Prod., 56: 469 (abstract)
Gorgtlltl, M., Kutlu, H.R., Demir, E. and ijzttirkcan, 0.. 1995. The effect of alternate day feeding with the diets differing in energy and protein contents on fattening performance of male Awassi lambs. J. Faculty of Agriculture, Cukurova University, in press. Henning, P.H., Steyn, D.G. and Meissner, H.H., 1993. Effect of
synchronization of energy and nitrogen supply on ruminal char- acteristics and microbial growth. J. Anim. Sci., 71: 2516-2528.
Kyriazakis, I. and Oldham, J.D., 1993. Diet selection in sheep: the ability of growing lambs to select a diet that meets their crude protein (nitrogenX6.25) requirements. Br. J. Nutr., 69: 617- 629.
Kyriazakis, I., Emmans, G.C. and Whitemore, CT., 1990. Diet selec- tion in pigs: choices made by growing pigs given foods of dif- ferent protein concentrations. Anim. Prod., 51: 189-199. Kyriazakis, I., Emmans, G.C. and Whitemore, CT., 1991, The
effects of feed protein content on the performance of pigs pre- viously given feeds with low or moderate protein content. Anim. Prod., 52: 165-173.
Newbold, J.R. and Rust, S.R., 1992. Effect of asynchronous nitrogen and energy supply on growth of ruminal bacteria in batch culture. J. Anim. Sci., 70: 538-546.
NRC, 1985. Nutrient Requirements of Sheep. Sixth Revised Edition, National Academy Press, NRC, Washington, DC, pp. 77. Orskov, O.R., 1987. Early weaning and fattening of lambs. In: I.
Fayez, M. Marai and J.B. Owen (Editors), New Techniques in Sheep Production. Butterworths, London, pp. 189-195. Rose, S.P. and Kyriazakis, I., 199 1. Diet selection of pigs and poultry.
Proc. Nutr. Sot., 50: 87-98.
Russell, J.B., O’Connor,C.D.,Fox, D.C.,Van Soest, P.J. andsniffen, C.J., 1992. A net carbohydrate and protein system for evaluating cattle diets. I. Ruminal fermentation. J. Anim. Sci., 70: 351-361. Sinclair, L.A., Garnsworthy, P.C., Newbold, J.R and Buttery, P.J.,
1992. The effect of dietary nitrogen and energy synchronization on rumen fermentation and microbial protein synthesis in sheep. Br. Sot. Anim. Prod. Winter Meeting, 23-25 March 1992. Pro- gram and Summaries. Paper No. 5, British Society of Animal Production, Edinburgh, pp. 56.
Theriez, M., Villette, Y. and Castrillo, C., 1982. Influence of meta- bolizable energy content of the diet and of feeding level on lamb performances. I. Growth and body composition. Livestock Prod. Sci., 9: 471-485.
Webster, G.M. and Povey, G.M., 1990. Nutrition of finishing lamb. In: C.F.R. Slade and T.L.J. Lawrence (Editors), New Develop- ment in Sheep Production. Occasional Publication of Br. Sot. Anim. Prod., No. 14. British Society of Animal Production, Edin- burgh, pp. 71-82.