Small Ruminant Research, 9 ( 1993 ) 353-366
© 1993 Elsevier Science Publishers B.V. All rights reserved. 0921-4488/93/$06.00
Comparative performance of several breeds of Caribbean hair sheep in purebreeding and crossbreeding R.K. Rastogi a, M.J.
K e e n s - D u m a s b a n d F.B. L a u c k n e r c
aDepartment of Livestock Science, The University of West Indies, Trinidad, bDivision of Agriculture, Tobago House of Assembly, Tobago and cCaribbean Agricultural Research and Development Institute, Trinidad, West Indies (Accepted 21 February 1992 )
ABSTRACT Rastogi, R.K., Keens-Dumas, M.J. and Lauckner, F.B., 1993. Comparative performance of several breeds of Caribbean hair sheep in purebreeding and crossbreeding. Small Rumin. Res., 9: 353-366. The relative merit of the Barbados Blackbelly ( BB ), West African (WA), Blackhead Persian ( BH P ), Blenheim Grade (BLG) and Virgin Island White (VIW) breeds of hair sheep was examined as straightbreds and in 2-, 3- and 4-breed combinations. Lamb performance traits studied were birth weight, preweaning average daily gain, weaning weight and survival to weaning at 8 weeks. Purebred WA, BB and BLG lambs were similar in growth, but excelled the other two breeds. Straightbred and 2-breed cross lambs were similar in growth performance; however, 3- and 4-breed crosses were significantly superior in average daily gain and weaning weight. Differences in general combining ability of breeds and reciprocal effects were significant for birth weight only. Differences in maternal ability of BB, WA and BHP ewes were significant for average daily gain and weaning weight, with WA ewes excelling. Three-breed cross lambs sired by WA rams also excelled in growth. BHP lambs or crossbred lambs born to BHP ewes exhibited highest survival to weaning. With respect to ewe performance traits, BB ewes ranked highest in frequency of multiple births, prolificacy, number of lambs weaned and total weight of lamb(s) weaned per ewe lambing, followed by WA, BLG and BHP ewes. Performance of Virgin Island White lambs and ewes was generally inferior to other breeds; however, the data were insufficient to draw definitive conclusions. Crossbreeding did not significantly improve number of lambs weaned or total weight of lamb(s) weaned.
Sheep have been raised by small farmers in the Commonwealth Caribbean since the days of colonization. They are traditionally kept for subsistence and as a means to raise quick cash in times of need (Devendra, 1977 ). Caribbean people have a preference for meat from sheep. However, local Correspondence to: Dr. R.K. Rastogi, Department of Livestock Science, The University of West Indies, Trinidad, West Indies.
R.K. RASTOGI ET AL.
supply fulfills only about 15% of the market demand. Consequently, during the last 15 years several Caribbean countries have accorded priority to small ruminant production and development programmes. The Government of the Republic of Trinidad and Tobago in 1971 established a 'Sheep Research and Multiplication Unit' at Blenheim Estate in Tobago. Sheep are not indigenous to the Caribbean. The colonists introduced European wool breeds and African slaves brought hair sheep with them. The present day hair sheep of the Caribbean are the result of centuries of crossing and natural selection (Combs, 1983 ). Several breeds of hair sheep are found throughout the Caribbean. Reports exist comparing some of these breeds in pure breeding (Bodisco et al., 1973; Rastogi et al., 1983 ), but none exist comparing these breeds in crossbreeding. The Blenheim Sheep Station was reorganised in 1981 and its objectives were redefined to provide contemporary comparison of five breeds in purebreeding and crossbreeding, and to determine the possibility of developing a new synthetic breed for commercial use throughout the Caribbean. This study reports results from analyses of lamb and ewe performance. EXPERIMENTAL PROTOCOL
Blenheim sheep station The station comprises 40 ha of land on the windward side of Tobago. The topography varies from hilly to mildly undulating and is considered rather unsuitable for cultivation. Tobago's climate is humid-subtropical with relative humidity between 80-85% and daily temperature between 23-32°C. There is a major dry season from January to May and a short dry season in October. The mean annual rainfall is 1,700 mm. Pastures are mainly in Pangola grass (Digitaria decumbens) with some Guinea grass ( Pannicum maximum). Elephant grass (Pennisetum purpureum) was established in less steep areas and was used as fresh-chop forage during the night; the excess in the wet season was ensiled to provide silage for the dry season. Pasture was fertilised at the rate of 25 kg N / h a / y r . Pasture productivity was poor, and an analysis of composite forage samples in 1982 indicated low mineral levels particularly of P, Na, Cu, Se and Zn (Conrad, 1984). Gliricidia (Gliricidia maculata) was used as live fence, and cuttings from it were routinely harvested and fed to sheep. Ewes were let out to pasture at 08 : 00 h and brought in at 15 : 00 h. If available, commercial dairy concentrate ( 16% crude protein) was fed to ewes at the rate of 0.5 k g / h e a d / d prior to mating and lambing, and after lambing for a period of 3 to 4 wk. In order to overcome mineral deficiencies in the forage, sheep were offered mixed mineral supplement so that by 1984 mineral profiles of lamb's blood were found to be normal (Youssef, 1986 ).
PERFORMANCEOF CARIBBEANHAIR SHEEP
Ewes and lambs were confined for a week following lambing and then let out to pasture daily. Beginning in 1985, lambs were creep fed up to weaning. To take advantage of year-round fertility of hair sheep, starting in 1983 all non-pregnant ewes were exposed to rams for 1 month at 4 month intervals. Thus, most lambings occurred in January/February, May/June, and September/October. Coccidiosis and helminth parasitism constituted the major causes of debilitation and mortality. All sheep were routinely dewormed using available broad spectrum anthelmintics and coccidiostats. However, reinfestation occurred easily. A foot bath of copper sulphate solution was used to prevent foot rot, particularly in the rainy season. Isolated cases of dermatophilosis were also observed (Keens-Dumus and Mannasmith, 1984). A n i m a l resources
As of September 1982, the station was stocked with 246 breeding ewes of four breeds. Barbados BlackbeUy (BB) numbered 72 (29.3%), West African TABLE 1
Experimental design and frequency of different birth types for 1982-85 data (phase I ) Lamb breed/mating type
Purebreeding Barbados Blackbelly (BB) West African (WA) Blackhead Persian (BHP) Blenheim Grade (BLG) Virgin Island White (VIW) Sub-total/mean 2-breed crosses/matings Ram X Ewe BBXWA WAXBB BB×BHP BHPXBB WA X BHP BHPXWA VIWXBB VIWXWA
Sub-total/mean Total/overall mean
% ewes producing S
296 260 134 382 63
150 157 103 245 51
29 50 74 52 78
48 42 23 39 20
22 8 3 9 2
137 143 78 154 58 95 119 53
82 82 60 79 51 54
39 48 72 25 86 44
52 35 25 58 14 43
9 17 3 17 13
S = singles; Tw = twins; Tr + = triplets and higher.
R.K. RASTOGIET AL.
TABLE 2 Experimental design and frequency of different birth types for 1986-88 data (phase II ) Lamb breed/mating type
% ewes producing S
Purebreeding Barbados Blackbelly (BB) West African (WA) Blackhead Persian ( B H P ) Virgin Island White (VIW)
83 52 58 81
43 33 46 65
30 52 78 77
51 42 20 22
19 6 2 1
67 53 76 67 43 51 16 26 16 8 12 20
38 32 38 37 25 33 15 20 13 8 7 2
34 44 32 41 44 52 93 70 77 100 43 100
55 47 42 35 44 42 7 30 23
11 9 26 24 12 6
3-breed matings BBXWA.BHP BB X WA.VIW BBXBHP.BB BB X BHP.WA BB × VIW.WA BBXVIW.BHP WA X BB.BHP WA X BHP.BB WA×V1W.BB WA × VIW.BHP BHPXBB.WA BHPXWA.BB BHPXVIW.BB BHPX V1W.WA VIWXBB.WA V1WXWA.BB VIW×BB.BHP VIWXBHP.BB VIW × BHP.WA V1WXWA.BHP
25 2 2 29 24 5 32 55 25 5 63 26 28 7 43 32 8 35 14 6
19 2 1 19 21 4 17 38 21 5 39 20 22 6 23 21 4 26 9 3
53 100 100 53 86 75 35 69 81 100 49 70 73 83 30 52
42 14 25 47 18 19
Sub-total/mean 2-breed matings Ram × Ewe WA×BB BHPXBB VIWXBB BB×WA BHPXWA VIWXWA BBXBHP WAX BHP VIWXBHP BB X VIW WA×VIW BHPXVIW Sub-total/mean
73 56 100
41 30 27 17 57 48 100 23 33
PERFORMANCE OF CARIBBEAN HAIR SHEEP
TABLE 2 Con 't. Lamb breed/mating type
% ewes producing S
4-breed matings WA × BHP. (VIW.BB) WA × VIW. (BHP.BB) BHP×BB.(WA.VIW) BHP × BB. (VIW.WA) BHP × WA.(VIW.BB) BHP × VIW. (BB.WA) BHP × VIW. (WA.BB) VIW × BB. (WA.BHP) VIW X BB. (BHP.WA) VIW × WA. (BB.BHP) VIW×WA.(BHP.BB) VIW × BHP. (BB.WA) VIW X BHP. (WA.BB) Sub-total/mean Total/overall mean
30 20 4 29 44 16 29 26 79 36 113 35 44
18 21 3 26 32 12 26 19 59 20 76 29 24
67 81 67 88 66 67 88 63 71 45 54 48 58
33 14 33 12 31 33 12 37 27 40 43 48 42
2 15 3 4
S = singles; Tw=twins; T r + =triplets and higher.
(WA) 57 (23.2%), Blackhead Persian (BHP) 50 (20.3%), and various crossbred ewes named Blenheim Grade (BLG) 67 (27.2%). Virgin Island White (VIW) sheep were introduced to the project in late 1983 when 21 ewes and nine rams were imported from St. Croix, US Virgin Islands. Twenty ewes and ten rams of the Blackbelly breed were also imported from Barbados in 1983. Offspring of some ewes later entered the breeding flock.
Mating plan The mating plan was carried out in two phases. In phase I, purebred and 2breed matings were made. The experimental design for phase I ( 1982-85 ) is shown in Table I. Numbers and kinds of matings made in phase I were designed to produce the various kinds and numbers of progeny required for breeding in phase II. Limited pen facilities and managerial requirements made multiple sire breeding groups a necessity. Generally, three rams were used in each lambing period for BB, WA, BHP, and BLG, while two rams were used for VIW. A new random assignment of ewes to rams was made for each mating period. The experimental design for phase II ( 1986-88 ) is shown in Table 2; purebred, 2-, 3- and 4-breed matings were made and only purebred rams were used.
R.K. RASTOGIET AL.
TABLE 3 Least squares constants for lamb performance traits through weaning for 1982-85 data (phase I) Effects
Overall mean Error SD Lambing season-year (LS) Sep-Nov, 82 Apr-Jul, 83 Jan-Jun, 84 Sep-Nov, 84 Jan-Jun, 85 Sep-Nov, 85 Sex Male Female Type of birth Single Twin Triplet and higher Crossbreds vs. purebreds Purebreds (PB) Crossbreds (CB) Purebreds Barbados Blackbelly
Birth vet. (kg)
Preweaning ADG (g)
Weaning wt. (kg)
0.15 0.04 -0.15 0.11 -0.02 -0.17
151 139 218 97 270 94
-11.7 7.6 8.7 -24.4 10.6 9.1
-0.61 0.48 0.50 - 1.37 0.51 0.48
316 487 166 484 485
24.4 - 8.0 - 16.4 NS 0.6 -0.6
1.34 - 0.44 -0.90 NS 0.02 -0.02
9.6 - 11.4
0.50 - 0.64
** 715 208 294 135 373 130
** 672 643
0.07 - 0.07 ** 0.51 -0.01 -0.50 NS -0.02 0.02 ** 0.16
0.14 - 0.30
396 645 274 670 645
West African (WA) Blackhead Persian
(BHP) General combining ability Barbados Blackbelly West African Blackhead Persian Maternal ability Barbados Blackbelly Western African Blackhead Persian Reciprocals BB.WA, WA.BHP, BHP.BB WA.BB, BHP.WA, BB.BHP LS X PB LS × CB Regr. on weaning age (d) Regr. on birth wt. R2
- 0.10 NS - 0.03 0.01 0.02 ** -0.07
1.6 -5.7 4.1
0.12 -0.33 0.21
-0.3 10.7 -10.4 NS -0.5
-0.04 0.58 -0.54 NS -0.2
NS NS - 0.44**
*P< 0.05; **P < 0.01 ; NS = not significant ( P > 0.05 ).
NS NS 0.13** 1.49**
PERFORMANCE OF CARIBBEAN HAIR SHEEP
Least squares breed means for lamb performance traits through weaning for 1982-85 data ( p h a s e l ) Breed
Birth wt. (kg)
W e a n i n g wt. (kg)
A D G (g)
Purebreds Barbados Blackbelly ( B B ) West African ( W A ) Blackhead Persian ( B H P ) Blenheim Grade ( B L G ) a Virgin Island White
295 243 132 375 63
** 2.85 2.83 2.38 2.76 2.61
143 209 132 282 47
** 152 160 139 156 140
** 11.17 I 1.54 10.39 11.73 10.12
135 142 75 145 58 90 115 51
** 2.77 2.87 2.77 2.58 2.63 2.76 2.59 2.69
101 105 60 106 47 66 72 35
** 156 143 141 155 139 160 152 159
** 11.39 10.62 10.59 11.32 10.48 11.55 10.84 ll.19
2-breed crosses BB×WA WA × BB BBXBHP BHPXBB WA × B H P BHPXWA VIW×BB VIW×WA **P<0.01.
aLambs born during the first three lambing seasons only. bLambs born during the last four lambing seasons only.
Data collection Data were collected on lamb and ewe performance from 1982 to 1988. Lambs were weighed within 12 h of birth and again at weaning. Lambs were weaned at an average age of 8 wk with a range of 5-12 wk. Lambs weaned before 6 wk or after 11 wk of age and weighing < 6 kg or > 19 kg at weaning were excluded from this study. Measures of ewe reproductive performance were frequency of different types of birth and litter size. Measures of litter size for each ewe lambing were total number born, number born alive and number weaned. Measures of lamb survival were percent lambs born alive of the total number born (prenatal), and percent lambs weaned of the number born alive (postnatal). Individual lamb weights at weaning were adjusted to 8-wk age basis prior to being added to arrive at total weight of lamb (s) weaned per ewe lambing. Ewes that gave birth to dead lamb (s) were included with weaning weight being zero.
Statistical analyses Data were analysed by analysis of variance (ANOVA) using least-squares procedures (Harvey, 1960). Use of ANOVA assumes normality of the data,
R.K. RASTOGI ET AL.
and litter size and lamb survival data apparently are not normally distributed. However, this method was still used as ANOVA permitted quantification of the effects of more variables, and presentation of the results from the different traits in the same format. Several models with fixed effects were used to analyse the different traits. The results of analyses presented in Table 3 for lamb performance traits were based on total data except BLG and VIW lambs. Thus means for BLG and VIW breeds in Table 4 were derived from separate analyses with similar models but without combining ability, maternal ability, and reciprocal effects; this comment also applies to means for these two breeds presented in Table 5. The effect of ewe breed in crossbreeding in Table 5 was also calculated from a separate analysis of data on ewes that were used for 2-breed matings. Each recorded lambing was considered an independent observation. Covariances among repeated observations on the same ewe and genetic relationships among ewes were not considered in analyses. RESULTS AND DISCUSSION
Phase I - lamb growth
Effects of lambing season-year, sex and type of birth were highly significant (Table 3) for measures of lamb growth and were according to expectations from the literature. Average differences between purebred and crossbred lambs were not significant. Heterosis may not be very important for these traits, particularly when lambs are weaned at an early age of 8 weeks as done in this study. Differences between purebreds were highly significant. Body weights and growth of BB, WA and BLG lambs were similar and markedly superior to BHP and VIW lambs (Table 4). Differences between crossbred lambs were also highly significant. Overall, WA ewes weaned the heaviest crossbred lambs, followed by BB ewes. Lamb growth performance in this study was markedly superior to that reported for a much smaller set of lambs of the same four pure breeds born during October 1974 at the same station (Rastogi et al., 1979). Thus, with the exception of BHP lambs whose growth performance remained similar to that in 1974, the average daily gain improved by 13-25% and weaning weight by 8-18%. Harricharan et al. (1984) reported slightly superior performance for BHP lambs born and raised at the Lowland Estate in Tobago which supplied the original BHP sheep to the Blenheim station. Lallo et al. ( 1988 ) studied growth performance of BB grade lambs raised intensively on sugarcane and agro-by-products; birth weight ranged between 2.7-3.4 kg, 8 to 9 week weaning weight between 10.9-14.7 kg, and preweaning average daily gain between 140-225 g. The performance of BB lambs was comparable with that reported from Barbados (Patterson, 1983 and 1984). Lamb performance in
PERFORMANCE OF CARIBBEAN HAIR SHEEP
our study was superior to that reported from Venezuela for the same pure breeds (Bodisco et al., 1973; Combellas et al., 1980). Differences in general combining ability of breeds and reciprocal effects were significant only for birth weight. Breeds differed significantly in maternal ability for average daily gain and weaning weight (Table 3). WA ewes exhibited superior maternal ability. No comparable reports for hair sheep were found in the literature.
Reproductiveperformance Barbados Blackbelly (BB) ewes had a higher percentage of multiple births (70%) and included 11 litters of quadruplets and four litters of quintuplets (Table 1 ). WA ewes were next with 50% multiple births and included two litters of quadruplets and one litter of quintuplets. Ewes of the other three breeds produced no quadruplets and dropped triplets with a frequency of less than 9%. The wide variability for litter size in BB and to a lesser extent in WA ewes may be suggestive of the existence of a major gene controlling litter size and needs to be investigated. Lambing season-year had a significant effect only on number weaned (Table 5 ). No significant differences were detected between purebred and 2-breed matings. Differences among purebreds were highly significant. The mean number of lambs born total for BB ewes was the highest (1.93). This was appreciably greater than for WA (1.60) and BLG ( 1.56 ) ewes, which in turn was markedly higher than BHP ( 1.27 ) and VIW (1.19) ewes. Ranking of breeds was the same for number weaned and also when ewes were used in 2breed matings (Table 5 ). Fitzhugh and Bradford ( 1983 ) summarized reported averages for litter size in BB, WA, BHP and VIW ewes. They found BB ewes to have the highest mean litter size, followed by VIW, WA and lastly, BHP. Their averages for BB, WA and BHP were all lower than observed in this study, while their figure for mean litter size for VIW was superior. Mason (1980) provided a subjective assessment of the performance of VIW ewes in the US and British Virgin Islands and quoted higher litter size figures. Mason ( 1980 ), however, stated that sheep production in St. Croix was on the decline and that these sheep looked smaller in size, less fertile and mixed with wool breeds. It should be recalled that VIW sheep in this study were imported from St. Croix. The observation by Mason, in addition to sampling errors, may help explain the rather poor prolificacy of VIW ewes in this study. Our litter size figures for BB ewes are comparable to those reported from Barbados by Patterson (1983 and 1984). Litter size figures for BHP ewes are similar to those reported by Harricharan et al. (1984) for the Lowland Estate flock in Tobago. Our results were superior to the report from Venezuela for BB and WA ewes (Mazzarri et al., 1976). Litter size in this study was appreciably greater than the previous report for ewes of the same breeds which lambed during 1976 and 1977
R.K. RASTOGIET AL.
TABLE 5 Least squares means for litter size, lamb survival and ewe productivity for 1982-85 data (phase I) Effects
No. ewes lambing
% Lamb survival
Total lamb born total born alive No. wnd. prenatal postnatal wt. wnd. (kg)
Overall mean Error SD
Lambing season-year (LS) Sep.-Nov, 82 Apr.-Jul, 83 Jan.-Jun, 84 Sep.-Nov, 84 Jan.-Jun, 85 Sep.-Nov, 85
114 132 184 90 225 84
NS 1.64 1.73 1.65 1.50 1.66 1.49
NS 1.61 1.65 1.55 1.46 1.60 1.42
** 1.46 1.51 1.29 1.19 1.28 1.19
NS 99.0 97.3 96.6 98.6 97.6 97.2
** 93.1 92.2 82.3 82.8 82.6 84.8
** 15.58 17.38 14.67 11.86 15.30 13.16
Mating system Purebred matings 2-breed matings
NS 1.60 1.62
NS 1.56 1.54
NS 1.29 1.35
NS 98.4 97.0
* 84.0 88.6
NS 14.37 14.93
Purebred matings ( P M ) Barbados Blackbelly (BB) West African (WA) Blackhead Persian (BHP) Blenheim grade (BLG) a Virgin Island White (VIW)b
161 157 103 245 51
** 1.93 1.60 1.27 1.56 1.19
** 1.88 1.56 1.24 1.53 1.16
** 1.50 1.30 1.07 1.31 0.96
NS 97.5 97.6 98.0 98.4 97.5
NS 83.2 86.8 88.9 86.8 84.9
** 16.34 15.18 11.59 15.03 10.77
82 82 60 79 50 55
. . 1.75 1.82 1.29 2.02 1.13 1.71
1.41 1.49 1.06 1.64 1.10 1.40
NS 96.3 99.0 97.0 96.1 100.0 95.4
NS 83.2 84.2 82.4 86.8 95.5 85.8
** 16.00 15.29 12.49 16.96 12.92 15.92
161 137 110
** 1.86 1.69 1.21
** 1.51 1.39 1.06
NS 97.6 96.3 99.1
NS 85.8 85.0 88.4
** 15.66 15.84 12.32
2-breed matings (2-BM) BB×WA WA×BB BBXBHP BHP×BB WAXBHP BHP×WA Ewe breed in crossbreeding BB WA BHP
. 1.65 1.77 1.23 1.87 1.14 1.57
** 1.78 1.59 1.17
*P< 0.05; **P < 0.01; NS = not significant ( P > 0.05 ). aEwes lambed during first three lambing seasons only. bEwes lambed during last four lambing seasons only. N.B.: First order interactions between LS on the one hand, and PM and 2-BM on the other, were not significant. R 2 for the complete model was less than 16% for all traits.
PERFORMANCE OF CARIBBEAN HAIR SHEEP
at the same station (Rastogi et al., 1983 ). Thus, litter size improved for all breeds ranging from 15% increase for BHP ewes to 45% for BB ewes. Lamb survival Overall, 98% of the total lambs born were alive at birth and 86% of those alive at birth were present at weaning (Table 5 ). The postnatal lamb survival was significantly superior in crossbreds. As expected, survival of lambs born to BHP ewes was generally better. Lamb survival figures in this study represented a significant improvement over those reported in the literature (Bodisco et al., 1973; Rastogi et al., 1983). Ewe productivity All effects except mating system were highly significant sources of variation for total lamb weight weaned per ewe lambing (Table 5 ). BHP ewes weaned the lightest litters, followed by VIW and BLG. BB ewes weaned the heaviest litters in purebreeding but WA ewes excelled in 2-breed matings. Comparable figures were not found in the literature. Phase H - lamb growth ANOVA indicated that lambing season-year,sex, and type of birth were all highly significant sources of variation (data not shown ), as in phase I. Significant differences due to system of mating (purebred vs. 2- vs. 3- vs. 4-breed crosses) were observed for preweaning average daily gain and weaning weight at 8 wk of age. Four-breed crosses weighed heaviest at weaning ( 11.3 kg), while purebreds were the lightest ( 10.6 kg). Significant differences were noted among the four purebreds for birth weight only. VIW lambs had the highest birth weight (2.84 kg). As in phase I, WA lambs were slightly heavier at weaning ( 11.5 kg) than the rest of purebreds. Breed of sire did not cause significant variation in growth performance of 2- and 4-breed cross lambs; however, it was a significant source of variation in growth performance of 3-breed cross lambs. Three-breed cross lambs sired by WA rams exhibited superior growth performance. Phase I results also indicated that WA ewes exhibited better maternal ability for growth in 2-breed cross lambs (Table 3). Reproductive performance Overall frequency of multiple births was lower compared to phase I (Table 2 ), perhaps due to a younger group of ewes, over-crowding, decline in general flock management, nutrition and health. The complete model accounted for less than 12% of total variation in litter size indicating large environmental effects. The system of mating had a highly significant effect on litter size at birth but not at weaning. Differences among purebreds had a significant effect on litter size at birth and weaning, with BB
R.K. RASTOGI ET AL.
ewes excelling. Effect of ewe breed in 2-breed matings was significant for litter size at birth but not at weaning. Effect of type of crossbred ewe in 3-breed matings was not significant but it was highly significant in 4-breed matings for litter size at birth. Overall, reproductive performance of ewes was markedly inferior compared to phase I due to above-mentioned reasons. Generally, it can be concluded that the reproductive performance of crossbred ewes was inferior to purebred BB and WA ewes. This can be explained on the basis of excellent prolificacy of these ewes and better maternal ability of WA ewes. Some 2-breed cross ewes (e.g., BHP.BB and BHP.WA), showed promise in terms of total productivity but the advantage was not enough to warrant their use as crossbred foundation for developing a new breed.
Ewe productivity Differences due to all effects studied were not significant except for seasonyear of lambing. WA ewes weaned markedly heavier litters compared to all other types of ewes. In phase I, BB ewes weaned the heaviest litters, followed closely by WA ewes. Reports of crossbreeding among breeds of hair sheep are not available in the literature and our results could not be compared. CONCLUSIONS
( 1 ) Breed differences do exist but appear to be small, particularly between BB, WA and BLG. (2) Under an improved management system, where lamb survival can be optimised, BB should be the breed of choice. (3) Under the small holder system of production, which is prevalent in the Caribbean, WA may out-perform or do as well as BB. (4) BLG held its own in lamb growth and was comparable to BB and WA in total productivity. Because of its heterogenity and the scope it offers for selection, BLG should not be excluded from consideration. In fact, BLG sheep may prove themselves quite well under the small holder system of production. (5) Performance of this small sample of VIW sheep was quite inferior to that reported from the Virgin Islands itself. Given the possibility of sampling variation, further evaluation of VIW sheep is suggested. (6) BHP sheep of South African origin certainly appear to be inferior in performance to their more prolific counterparts in the Caribbean which are of possible West African origin. (7) Relative ranking of these breeds should be investigated under small holder system together with intervention at the level of animal health management so as to minimize the incidence of internal parasites. (8) The possibility of a major gene controlling prolificacy in BB (and perhaps WA) sheep needs to be researched. (9) The results indicated that crossbreeding among Caribbean hair sheep and develo p m e n t of a new synthetic breed was not of practical significance.
The Blenheim Sheep Project was a cooperative effort between the Ministry of Food Production and Marine Exploitation of Trinidad and Tobago, the Caribbean Agricultural Research and Development Institute, The University of the West Indies, and Winrock International Institute for Agricultural Development. Many people contributed to the implementation of this project and its programme; to them our sincere thanks.
REFERENCES Bodisco, V., Duque, C.M. and Valles, A.S., 1973. Compartamiento productivo de ovinos tropicales en el periodo 1968-1972 (Productive performance of tropical sheep during the 19681972 period). Agronomia Tropical, 23: 517-540. Combellas, J. de, Martinez, N. and Gonzalez, E., 1980. A study of factors which influence birth and weaning weight in lambs. Trop. Anim. Prod., 5:261-265. Combs, W., 1983. A history of the Barbados Blackbelly Sheep. In: H.A. Fitzhugh and G.E. Bradford (Editors), Hair Sheep of Western Africa and the Americas: A Genetic Resource for the Tropics, Westview Press, Boulder, pp. 179-200. Conrad, J.H., McDoweil, L.R. and Ellis, G.L., 1984. Mineral deficiencies of grazing ruminants in the Caribbean basin. In: Proc. V Regional Livestock Meeting, The University of the West Indies Library, Trinidad (W.I.), 10 pp. Devendra, C., 1977. Sheep of the West Indies. World Rev. Anim. Prod., 13: 31-38. Fitzhugh, H.A. and Bradford, G.E., 1983. Productivity of hair sheep and opportunities for improvement. In: Hair Sheep of Western Africa and the Americas: A Genetic Resource for the Tropics, Westview Press, Boulder, pp. 23-54. Harricharan, H., Mannasmith, C., Keens-Dumas, M.J., Ramlal, H. and Nobriga, A. de., 1984. The reproductive performance of Blackhead Persian and other ewes and growth of the lambs reared on the Lowland Estate in Tobago. In: Proc. V Regional Livestock Meeting, The University of the West Indies Library, Trinidad (W.I.), 16 pp. Harvey, W.R., 1960. Least squares analysis of data with unequal subclass numbers. ARS 20-8, USDA, Washington, DC, pp. 141-157. Keens-Dumas, M.J. and Mannasmith, C.H., 1984. Sheep production in Tobago. In: Proc. V. Regional Livestock Meeting, The University of the West Indies Library, Trinidad (W.I.), 17 PP. Lallo, C.H.O., Neckles, F.A. and Garcia, G.W., 1988. A system for intensive hair sheep production under zero-grazing conditions utilizing sugarcane and byproduct feedstuffs. In: Proc. 2nd Annu. Seminar on Agric. Res., The University of the West Indies Library, Trinidad (W.I.), pp. 135-149. Mason, I.L., 1980. White Virgin Island Sheep. In: Prolific Tropical Sheep, FAO Anita. Prod. Hlth. Paper 17, Rome, pp. 29-32. Mazzarri, G., Fuenmayor, C.E., Reveron, A.E. and Arriojas, M., 1976. Sincronizacion y fertilidad en ovejas tropicales en diferentes epocas del ano e intervalos entre parto y servicio (Synchronisation and fertility in tropical ewes at different times of year and intervals between lambing and mating). Instituto Investigaciones Zootecnical, Maracay, Venezuela, Mimeo. 7p. Patterson, H.C., 1983. Barbados Blackbelly and crossbred sheep performance in an experimental flock in Barbados. In: H.A. Fitzhugh and G.E. Bradford (Editors), Hair Sheep of Western
R.K. RASTOGI ET AL.
Africa and the Americas: A Genetic Resource for the Tropics, Westview Press, Boulder, pp. 151-162. Patterson, H.C., 1984. Performance of Barbados Blackbelly sheep in two experimental flocks in Barbados. In: Proc. V Regional Livestock Meeting, The University of the West Indies Library, Trinidad (W.I.), 19 pp. Rastogi, R.K., Youssef, F.G., Keens-Dumas, M.J. and Davis, D., 1979. Note on early growth rates of lambs of some tropical breeds. Trop. Agric. (T'dad), 56:259-261. Rastogi, R.K., Archibald, K.A.E. and Keens-Dumas, M.J., 1983. Sheep production in Togabo with special reference to Blenheim Sheep Station. In: H.A. Fitzhugh and G.E. Bradford (Editors), Hair Sheep of Western Africa and the Americas: A Genetic Resource for the Tropics, Westview Press, Boulder, pp. 141-150. Youssef, F.G., 1986. The mineral status of forages and grazing ruminants in Trinidad and Tobago. In: Proc. Short Course/Seminar on Small Ruminant and Rabbit Health and Production, The University of the West Indies Library, Trinidad (W.I.), 22 pp. RI~SUMI2 Rastogi, R.K., Keens-Dumas, M.J. and Lauckner, F.B., 1993. Le rendement comparatif en croisement et en race pure de plusieurs race de moutons ~ poiles de la Caraibe. Small Rumin. Res., 9: 353-366. Les m4rites relatifs de plusieurs races de moutons h poils, plus exactement: Barbados Blackbelly (BB), West African (WA), Blackhead Persian (BHP), Blenheim Grade (BLG) et Virgin Island White (VIW) ont 6t6 examin6 e n t a n t que race pure et 6galement en croisement de 2-, 3et 4-. Les traits de rendement des agneaux 6tudi6s 6taient: le poids ~ la naissance, le gain quotidien de la naissance au sevrage, le poids au sevrage aussi bien que le taux de viabilit6 des agneaux au sevrage/~ 8 semaines. La croissance des agneaux de race pure de WA, BB et BLG 6tait semblable, mais sup6rieure aux autres races. La croissance des agneaux de race pure et de double croisement 6tait semblable; n6ammoins, les croisements de 3- et 4-6taient sup6rieurs d'une mani6re significative, en gain quotidien et en poids au sevrages. Les diff6rences dans la 'general combining ability' des races et des effets r6ciproques 6taient significatives uniquement pour le poids ~ ia naissance. Les diff6rences en 'maternal ability' des brebis des races BB, WA, et BHP 6taient significatives pour le gain quotidien et le poids au sevrage; les brebis de race WA montrant leur sup6riorit6. Les agneaux crois6s des trois races, issus de b61iers WA, ont aussi bien d6montr6 une croissance sup6rieure. Les agneaux BHP ou les agneaux crois6s n6s de brebis BHP 6taient plus viables au sevrage. En ce qui concerne le rendement des brebis, les brebis BB comptaient parmi les meiUeures dans la fr6quence de naissances multiples, prolificit6, nombre d'agneaux sevr6s et la productivit6 num6rique au sevrage, suivies par les brebis WA, BLG et BHP. En g6n6ral, le rendement des agneaux et des brebis de race VIW 6tait inf6rieur aux autres races, n6ammoins, les donn6es 6taient insuffisantes pour tirer des conclusions d6finitives. Le croisement n'a pas augment6 d'une mani6re significative, le nombre d'agneaux sevr6s ou la productivit6 num6rique au sevrage.