Crossbreeding for dairy production in the lowland tropics of Kenya

Livestock Production Science 63 (2000) 55–63 www.elsevier.com / locate / livprodsci

Crossbreeding for dairy production in the lowland tropics of Kenya II. Prediction of performance of alternative crossbreeding strategies a, b c a A.K. Kahi *, G. Nitter , W. Thorpe , C.F. Gall a

Institute of Animal Production in the Tropics and Subtropics, Hohenheim University, 70593 Stuttgart, Germany b Institute of Animal Breeding and Husbandry, Hohenheim University, 70593 Stuttgart, Germany c International Livestock Research Institute, P.O. Box 30709, Nairobi, Kenya Received 13 November 1998; received in revised form 21 June 1999; accepted 1 July 1999

Abstract The predicted performance of nine crossbreeding strategies for milk production and reproductive traits and cow live weights (LW) were compared. The strategies were selected because of their superiority to many others. The data were from crosses of Ayrshire (A), Brown Swiss (B), Friesian (F) and Sahiwal (S) cattle. Performances were predicted from the results of a genetic model based on additive breed differences, dominance and additive 3 additive interaction effects for the breeds involved. The crossbreeding strategies were: first cross (F 3 S), two-breed rotation (AS) Rot , three-breed rotation (BFS) Rot and two- (F and S), three- (B, F and S) and four-breed synthetic breeds based on equal and unequal contribution of the foundation breeds. The mean lactation milk yield (MY) of production systems based on F 3 S, (AS) Rot or (BFS) Rot cows and in which replacements are raised from within, was also predicted under varying number of calvings (NL) and reproductive performance (RP). At the individual cow level, differences in the predicted MY between the F 3 S cross, three-breed rotation and the synthetic breeds were small. While the F 3 S cross was superior to the two-breed rotation for predicted MY, its performance was similar for MY expressed per unit of metabolic weight. Among the synthetic breeds, differences in MY expressed per unit metabolic weight were small. At the production system level, it was predicted that MY for production systems based on F 3 S cows become superior to those based on (AS) Rot cows only at a NL higher than 4 and were inferior to those based on (BFS) Rot . This study shows that F 3 S cows were closely rivalled by the (BFS) Rot and the synthetics. Its inferiority was particularly shown at the production system level. It is concluded that the first cross is not generally the best suited for dairy production systems in the tropics. There is the need to promote greater awareness of the potential of synthetic breeds and to formulate strategies for developing and exploiting them.  2000 Elsevier Science B.V. All rights reserved. Keywords: Crossbreeding; Crossbreeding strategy; Dairy cattle; Synthetic breeds; Tropics

*Corresponding author. Tel.: 149-711-4593294; fax: 149-711-4593290. E-mail address: [email protected] (A.K. Kahi) 0301-6226 / 00 / $ – see front matter  2000 Elsevier Science B.V. All rights reserved. PII: S0301-6226( 99 )00132-3

56

A.K. Kahi et al. / Livestock Production Science 63 (2000) 55 – 63

1. Introduction There is a substantial amount of improvement in production from crossing Bos indicus to Bos. taurus dairy breeds for milk production in the tropical countries. Crossbreeding optimises the additive genetic and nonadditive (heterotic) breed effects of B. taurus and B. indicus cattle (Gregory and Cundiff, 1980). Cunningham and Syrstad (1987) reviewed reports on dairy cattle crossbreeding in 25 countries across the tropics and showed consistent improvement of lactation milk yield (MY) and calving interval (CI) with increasing European gene fraction up to 50% and thereafter an increase in CI but with no clear trend in MY. This result has been confirmed elsewhere in the tropics (Madalena et al., 1990; Thorpe et al., 1993; Rege et al., 1994; Talbott, 1994). In addition to the additive difference between the two breeds, their first crosses show an improvement in performance due to the fact that they possess 100% heterozygosity with respect to breed of origin. Strategies for utilisation of different exotic breeds for milk production in the tropics vary and depend on the production environment. Madalena et al. (1990) recommended the utilisation of the F 1 for milk production in Brazil. Syrstad (1996) suggested rotational systems and the synthetic breed strategy for large and small farms, respectively. Results summarised by Rege (1998) showed that at the same level of indigenous genes, crosses of different exotic breeds differed in their performance indicating that no one breed, crossbreed or crossbreeding strategy will have superior aggregate performance in all production environments. The challenge therefore is to match the genotype to the production environment, of which the feed resource availability is the major factor distinguishing amongst the environments. In most of the tropics, the smallholder dairying sector contributes a great deal to the total national milk output (Walshe et al., 1991). In Kenya, for example, over 80% of the national milk supply is produced in medium to high potential areas in the highlands by smallholder farmers keeping various grades of exotic dairy cattle. As such, any genetic improvement programme must necessarily be targeted at the smallholder dairy sector. Currently, most smallholders practise systems of upgrading

indigenous breeds to higher exotic grades without following a defined breeding programme. This process has resulted in problems, which can be ascribed to lack of adaptation to tropical stresses of poor nutrition, disease challenge and heat stress. Therefore there is the need to estimate the comparative performances of crossbreeding strategies involving various dairy breeds in relation to their implementation in the smallholder sector in the tropics. In this paper, performance of crossbreeding strategies are predicted using the results of a genetic model based on additive breed differences, dominance and additive 3 additive interaction effects (Kahi et al., 2000a).

2. Material and methods A detailed description of the data used in this study has been presented in Part I of this report (Kahi et al., 2000a). Data were from crosses of Ayrshire (A), Brown Swiss (B), Friesian (F) and Sahiwal (S) cattle. Milk production and reproductive traits and cow live weights were predicted and compared for the following crossbreeding strategies: first cross (based on F and S), two-breed rotation (A and S), three-breed rotation (B, F and S) and two- (F and S), three- (B, F and S) and four-breed synthetic breeds based on equal and unequal contribution of the foundation breeds. Abbreviations and definitions of the crossbreeding strategies and the gene contribution of each breed to the respective strategy are shown in Table 1. Milk production and reproductive traits were: lactation milk yield (MY; kg); lactation length (LL; days); calving interval (CI; days); annual milk yield (AMY; kg); and daily milk yield (DMY; kg). The AMY was calculated as (MY/ CI) 3 365, while DMY was defined as MY/ LL. Apart from the (AS) Rot , the crossbreeding strategies were chosen because preliminary analysis indicated that they gave the highest MY in their respective groups. Among the two-breed rotations, the (AS) Rot was included because it has received more attention in recent years (Trail and Gregory, 1981; Thorpe et al., 1994). However more importantly, the crossbreeding strategies and breeds involved were chosen from among those for which data were available from Kilifi Plantations. Kahi et al. (2000a) showed that the use of the

A.K. Kahi et al. / Livestock Production Science 63 (2000) 55 – 63

57

Table 1 Definition of the crossbreeding strategies Abbreviation

Crossbreeding strategy

Breed contribution (%)a A

F3S (AS) Rot (BFS) Rot (FS) Syn (3 / 4F 1 / 4S) Syn (BFS) Syn (3 / 8B 3 / 8F 1 / 4S) Syn (ABFS) Syn (1 / 8A 1 / 4B 1 / 2F 1 / 8S) Syn a

First cross Two-breed rotation Three-breed rotation Two-breed synthetic (equal contribution) Two-breed synthetic (unequal contribution) Three-breed synthetic (equal contribution) Three-breed synthetic (unequal contribution) Four-breed synthetic (equal contribution) Four-breed synthetic (unequal contribution)

B

F

S

50 33.3 50

50 50 33.3 50

75

25

33.3

33.3

33.3

37.5

37.5

25

25

25

25

25

12.5

25

50

12.5

50 33.3

A, Ayrshire; B, Brown Swiss; F, Friesian; S, Sahiwal.

Kinghorn model based on hypothesis X (Kinghorn, 1980) did not improve the accuracy of estimated crossbreeding parameters compared to those estimated by the dominance model. Dillard et al. (1980) and Robison et al. (1981) indicated that omission of epistasis effects in the genetic model did not significantly affect the accuracy of prediction of genotypes in B. taurus 3 B. taurus crosses. This is due to the fact that there is a high co-linearity of dominance and epistasis coefficients. When one of them is fitted the other may have little additional effect on prediction of performance of untested genotypes (Kinghorn and Vercoe, 1989). This also appears true for B. taurus 3 B. indicus cattle. A model used by Grosshans et al. (1994) which included additive breed effect, dominance and additive 3 additive interaction effects, was used to predict performance of various crossbreeding strategies (Model D in Kahi et al., 2000a). Model D gives a detailed view of epistatic interactions which leads to clear results for the epistasis effects and the heterotic effects caused by dominance gene effects. The results for the parameter estimates of this model were used here but these parameters can linearly be transformed to those of the Dickerson model (Dickerson, 1973). Predicted performances were computed for the traits of interest as follows:

M 5 m 1 f( gB )gB 1 f( gF )gF 1 f( gS )gS 1 c(dAB )dAB 1 c(dAF )dAF 1 c(dAS )dAS 1 c(d BF )d BF 1 c(d BS )d BS 1 c(d FS )d FS 1 c(aaAB )aaAB 1 c(aaAF )aaAF 1 c(aaAS )aaAS 1 c(aa BF )aa BF 1 c(aa BS )aa BS 1 c(aa FS )aa FS where M is the mean of the trait of interest predicted from a particular crossbreeding strategy (this was done for all strategies); m is general mean (represents the mean performance of breed A); f( gi ) is gene fraction of breed i; gi is additive breed effect for breed i; d ij and aa ij are the dominance and additive3 additive interaction effects between i and j, respectively; and c(d ij ) and c(aa ij ) were their coefficients, which were calculated as p fi p mj 1p mi p fj and ( f fp p fj 1 p im p jm 1 c(d ij ) ) / 2, respectively (Akbas et al., 1993; Grosshans et al., 1994), where p fi and p mi denote the gene proportion of breed i in the sire and dam, respectively. Since the number of crossbred generations containing F genes was insufficient to allow estimation of effects that are due to crosses with breed F (Kahi et al., 2000a), it was assumed that the dominance and additive3additive interaction effects that are due to A3F and B3F crossing were similar to those due to A3B while those due to F3S were

A.K. Kahi et al. / Livestock Production Science 63 (2000) 55 – 63

58

similar to those due to B3S crossing. A recent summary of 80 reports in the literature on crossbreeding in the tropics indicated that F crosses were not any better than crosses involving Jersey, A, B or Red Dane (Rege, 1998). Table 2 shows the gene fraction of breeds (breed contribution) other than A and coefficients of dominance and additive3 additive interaction effects for different crossbreeding strategies. The mean MY of production systems that are based on F3S, (AS) Rot or (BFS) Rot cows were predicted under the assumption of partial self-replacement by varying parameters of NL (a proxy of length of productive life) and reproduction. Only three of the nine strategies were examined because these are strategies with ‘by-products’ in terms of pure-bred cows, which are needed to produce bulls. The purebred cows are required even when these three strategies are fully established. The synthetic breed strategies only require pure-bred cows during their development and when fully established are managed as pure-bred population with no ‘by-products’. It was assumed that semen from any B. taurus breed can be purchased from outside the production system so that the burden of keeping pure-bred cows of these breeds is ignored. However, the replacement of F3S cows requires pure-bred S cows, which have to be inseminated with purchased F semen. Furthermore, the rotations (AS) Rot and (BFS) Rot require S cows to produce S bulls needed as sires to run the rotation. These cows are considered as part of the

production system. Therefore, the mean performance of a production system (MPS) based on F3S or rotational cows can be represented by: MPS F3S 5 vMS 1 (1 2 v)MF3S ,

(1)

MPS (AS) Rot 5 wMS 1 (1 2 w)M(AS ) Rot

(2)

and MPS ( BFS) Rot 5 yMS 1 (1 2 y)M( BFS) Rot

(3)

where MS , MF3S , M(AS ) Rot and M(BFS ) Rot are the mean predicted performance of pure-bred S, F3S, twoand three-breed rotational cows, respectively while v, w and y are the proportion of S cows to be kept in each of the production systems. These proportions are dependent on the number of calvings (NL), calves reared per parturition (CRP) and culling rate of cows (CRC), i.e. proportion of females not suitable for breeding. Therefore, assuming a sex ratio of 0.5, v can be calculated as: 2 v 5 ]]]]]]] NL 3 CRP 3 (1 2 CRC) 2 v 5 ]]] NL 3 RR

or

or

2 v5] RP

(4)

where RR5CRP3(12CRC) is the net reproduction rate while RP5NL3RR is the lifetime reproductive performance, i.e. number of calves reared and suitable for breeding. The proportions w and y are further dependent on the sire / dam ratio (SDR, i.e.

Table 2 Breed contribution and coefficients of dominance and additive3additive interaction effects used in prediction of performance of different crossbreeding strategies Crossbreeding strategy a

F3S (AS) Rot (BFS) Rot (FS) Syn (3/4F 1/4S) Syn (BFS) Syn (3/8B 3/8F 1/4S) Syn (ABFS) Syn (1/8A 1/4B 1/2F 1/8S) Syn a

See Table 1.

Breed contribution

Coefficients of dominance and additive3additive interaction effects

B

F

S

dAB

dAF

dAS

d BF

d BS

d FS

aaAB

aaAF

aaAS

aa BF

aa BS

aa FS

0.000 0.000 0.333 0.000 0.000 0.333 0.375 0.250 0.250

0.500 0.000 0.333 0.500 0.750 0.333 0.375 0.250 0.500

0.500 0.500 0.333 0.500 0.250 0.333 0.250 0.250 0.125

0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.125 0.063

0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.125 0.125

0.000 0.667 0.000 0.000 0.000 0.000 0.000 0.125 0.031

0.000 0.000 0.286 0.000 0.000 0.222 0.281 0.125 0.250

0.000 0.000 0.286 0.000 0.000 0.222 0.188 0.125 0.063

1.000 0.000 0.286 0.500 0.375 0.222 0.188 0.125 0.125

0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.125 0.063

0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.125 0.125

0.000 0.445 0.000 0.000 0.000 0.000 0.000 0.125 0.031

0.000 0.000 0.190 0.000 0.000 0.222 0.281 0.125 0.250

0.000 0.000 0.190 0.000 0.000 0.222 0.188 0.125 0.063

0.500 0.000 0.190 0.500 0.375 0.222 0.188 0.125 0.125

A.K. Kahi et al. / Livestock Production Science 63 (2000) 55 – 63 Table 3 Parameters used in the prediction of the mean performance of production systems based on F3S, two- and three-breed rotational cows Parameter

Range of values examined

Number of calvings (NL) Calves reared per parturition (CRP) Culling rate of cows, i.e. proportion not suitable for breeding (CRC) Culling rate of sires (CRS) Sire / dam ratio (SDR)

4, 6, 8 0.70, 0.80, 0.90, 1.00 0.10 0.90 1:50, 1:200

number of cows mated to one sire) and on the culling rate of sires due to unsuitability for breeding (CRS): 2 3 1/2 w 5 ]]]]]]] RP 3 SDR 3 (1 2 CRS)

(5)

The factor 1 / 2 is included because half of the rotational cows are mated by sires of breed A. For y this factor is 1 / 3 as two-thirds of rotational cows are mated by B and F sires: 2 3 1/3 y 5 ]]]]]]] RP 3 SDR 3 (1 2 CRS)

(6)

Table 3 shows the parameter values (Peeler and Omore, 1997) used to calculate the proportion of S cows to be kept in each of the systems.

59

3. Results Table 1 shows the abbreviations and definitions of all nine crossbreeding strategies while Table 4 shows the predicted trait means under these strategies for milk production and reproductive traits and LW on cow level per lactation. The need to maintain the parent pure-bred populations required to produce the crosses is ignored. The difference in predicted MY between the F3S cross, the three-breed rotation and the synthetic breeds was small. F3S cross produced 745 kg more than (AS) Rot . Among the synthetic breeds, the (3 / 4F 1 / 4S) Syn and (1 / 8A 1 / 4B 1 / 2F 1 / 8S) Syn had higher predicted MY than the others. Predicted MY in the (BFS) Syn , (3 / 8B 3 / 8F 1 / 4S) Syn and (ABFS) Syn was similar. The (AS) Rot had the longest and (FS) Syn the shortest predicted CI. The differences between synthetic breeds in predicted CI, DMY, LL and LW were small. The F 1 had the heaviest predicted LW but were not different from those of (BFS) Rot and two- and three-breed synthetics based on both equal and unequal contribution of the foundation breeds. The (AS) Rot were 74 kg lighter than the F3S cross. Although the F3S cross was superior to (AS) Rot for predicted MY, their performance was similar for predicted MY expressed per unit of MW. However, the F3S cross produced 3.03 kg or about 9% more predicted AMY expressed per unit MW than the (AS) Rot . Among the synthetic breeds, the (3 / 4F 1 / 4S) Syn and (1 / 8A 1 / 4B 1 / 2F 1 / 8S) Syn were superior in predicted MY and AMY

Table 4 Predicted average performance per lactation under alternative crossbreeding strategies for milk production and reproductive traits and cow live weight Crossbreeding strategy a

F3S (AS) Rot (BFS) Rot (FS) Syn (3 / 4F 1 / 4S) Syn (BFS) Syn (3 / 8B 3 / 8F 1 / 4S) Syn (ABFS) Syn (1 / 8A 1 / 4B 1 / 2F 1 / 8S) Syn a

Traits b

Per unit metabolic weight

MY

AMY

CI

LL

DMY

LW

MY

AMY

3922 3177 3772 3726 4151 3694 3784 3626 3907

3672 2919 3509 3619 3977 3527 3586 3398 3658

388 398 389 376 377 378 379 382 380

320 316 322 308 323 316 321 325 331

10.84 10.12 10.66 10.65 11.87 10.47 10.79 10.58 11.42

485 411 477 473 471 479 478 463 469

40.47 40.71 39.43 37.12 41.96 37.77 39.19 38.95 41.77

37.04 34.01 36.14 35.44 39.48 35.47 36.50 35.36 38.10

See Table 1. AMY, annual milk yield; CI, calving interval; DMY, daily milk yield; LL, lactation length; LW, cow live weight; MY, lactation milk yield. b

A.K. Kahi et al. / Livestock Production Science 63 (2000) 55 – 63

60

Table 5 Predicted average MY of production systems based on F3S and rotational cows under varying number of calvings and reproduction rates Parameter a NL

RR

Production system RP

F3S v

b

(AS) Rot MY

c

SDR

(BFS) Rot w

b

MY

50 4

6

8

0.63 0.72 0.81 0.90 0.63 0.72 0.81 0.90 0.63 0.72 0.81 0.90

2.52 2.88 3.24 3.60 3.78 4.32 4.86 5.40 5.04 5.76 6.48 7.20

0.79 0.69 0.62 0.56 0.53 0.46 0.41 0.37 0.40 0.35 0.31 0.28

2972 3091 3183 3257 3289 3368 3429 3479 3447 3507 3553 3590

c

SDR

w

b

MY

200 0.08 0.07 0.06 0.06 0.05 0.05 0.04 0.04 0.04 0.03 0.03 0.03

3141 3146 3149 3152 3153 3156 3158 3160 3159 3161 3163 3164

SDRc

yb

MY

0.05 0.05 0.04 0.04 0.03 0.03 0.03 0.02 0.03 0.02 0.02 0.02

3717 3724 3729 3734 3735 3740 3744 3746 3745 3748 3751 3753

50 0.02 0.02 0.02 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01

3168 3169 3170 3170 3171 3171 3172 3172 3173 3173 3174 3174

SDRc

yb

MY

0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.00

3758 3760 3761 3762 3763 3764 3765 3766 3765 3766 3767 3767

200

a NL, number of calvings; RR, reproduction rate; RP, overall reproduction performance, i.e. number of calves reared and suitable for breeding per lifetime (5NL3RR). b v, w and y are the proportion of pure-bred S cows per production system and are calculated as described in the text. c SDR, sire / dam ratio.

expressed per unit metabolic weight while the differences in other synthetic breeds were small. Table 5 shows the predicted means of MY for production systems in which replacements are raised from within the system and are based on F3S or on rotational cows under varying alternatives of NL, RP and SDR. It was predicted that production systems based on F3S cows become superior to those based on (AS) Rot cows only at a NL higher than 4 and were inferior to those based on (BFS) Rot . The inferiority of the production system based on F3S is due to keeping pure-bred S dams that have a low genetic potential. In this production system, the proportion of pure-bred S dams required for the replacement of the F3S cows was substantially higher than that required for the replacement of the rotational cows. As expected, there was a linear relationship between the RP and the mean MY in both the production systems. However, the gain in MY with increasing RP was higher in production systems based on F3S cows than in those based on rotational cows. In production systems with rotational cows, varying the SDR from 50 to 200 resulted in small changes in MY especially when the RP was high.

4. Discussion The aim of this study was to predict and compare the performance of cows under nine crossbreeding strategies and the performance of the production systems applying the strategies in which there is the necessity to keep dams that have low genetic potential. Prediction was based on previous results on additive breed difference, dominance and additive3 additive epistatic effects (Kahi et al., 2000a). The best crossbreeding system (first cross) should be that which optimises heterozygosity jointly with additive breed effects (Cunningham and Syrstad, 1987). However, the results indicate that the F3S cross is rivalled by the synthetics in almost all traits. The high levels of management (nutrition, disease control) at Kilifi Plantations, which are not typical conditions in the tropics, might have influenced this outcome. Cunningham (1981) suggested that production in a good environment is influenced heavily by additive breed effects and small heterotic effects and in a poor environment by heterosis. Therefore, it is expected that in a poor environment heterosis would have benefited the F 1 (exemplified by the

A.K. Kahi et al. / Livestock Production Science 63 (2000) 55 – 63

F3S cross in the present study) in the comparisons. However, logistical aspects for the use of the F 1 strategy in dairy production systems (especially the smallholder dairy sector) in the tropics remain problematic. Based on results from Brazil, Madalena (1993) presented an F 1 continuous replacement scheme to capitalise on the superiority of the F 1 hybrid. This system might not be practical because the number of pure-bred females required is too large to be kept in a nucleus or to be found concentrated in a few herds (Kosgey et al., 1998). This is also the case when a production system is considered as shown in the present study. To solve the problem of having to maintain pure-bred B. indicus cows, Rutledge (1996) suggested the use of in vitro fertilisation to continuously produce F 1 embryos in a central laboratory. Such a system is not sustainable because of the technology and the costs involved in the production of the embryos. Also efficient dissemination of embryos would require a developed infrastructure, which is partially or completely lacking in most developing countries. Furthermore, such a system would be too sophisticated to suit the poor conditions in small farms and villages. While this system can be implemented in dairy ranches supplying F 1 heifers, it may have some drawbacks on a regional scale because of health controls and transport costs (Madalena, 1981) and the initial cost of replacement heifers. Therefore the real challenge is to establish breeding programmes that allow for on-farm raising of replacement heifers, because replacement costs are important in determining profit from any dairy enterprise (Van Arendonk and Brascamp, 1990). Onfarm raising of replacement F 1 heifers is possible. However as shown in the present paper, this will require keeping a high proportion of purebred dams that have low genetic potential and might result in depressed overall productivity and hence sustainability. The result that the F3S cross and (AS) Rot were similar in MY expressed per unit of metabolic weight indicated that higher yields are partly due to higher LW and thus have little economic advantage. In production systems where stratification for the supply of dairy replacements is not possible, rotational crossbreeding systems are likely to be optimal

61

breeding systems for dairy production (Thorpe et al., 1993). Rotational crossing would lead to better dairy performance mainly because of larger heterosis (Syrstad, 1996). It is only at a NL higher than 4 that production systems based on F3S cows were superior to those based on (AS) Rot cows. The NL is highly influenced by the overall mean performance of the cow, the management levels and the culling policies and rarely exceeds the value of 4 (Amble and Jain, 1967; Silva et al., 1986; Thorpe et al., 1994; Lemos et al., 1996; Kahi et al., 2000b). A comparison of results in Tables 4 and 5 indicates that the ranking of the three strategies at the cow level is not identical to the ranking of strategies at the production system level. The (BFS) Rot ranked inferior to F3S at the cow level but superior at the production system level at all alternatives of NL and RP. However, the utilisation of heterosis through organised breed rotation is constrained due to the fact that a high percentage of cattle are kept in management units that are too small (Trail and Gregory, 1981). Rotation systems would be the strategy of choice when organisational and management problems can be overcome as in large and well organised farms (Syrstad, 1996). However, the wide fluctuation in breed composition of the cows between generations make it difficult to synchronise climatic adaptability and performance characteristics with a given management level and natural environment (Trail and Gregory, 1981). Greater awareness of the potential of synthetic breeds for dairying in the tropics should be promoted. A herd of a synthetic breed is managed as a pure-bred population and the management problems associated with small herd sizes in a rotational crossbreeding system are avoided. Gregory et al. (1982) have discussed the utilisation of synthetic breeds for dairy production in the tropics. Although synthetic populations show maximum recombination loss (Dickerson, 1973), they are a useful alternative to rotational crossing for retaining heterosis. Furthermore theoretically, their response to selection can be greater than that of parental breeds because of the increased genetic variation (Dickerson, 1973). At least six countries have invested large resources to develop synthetic breeds from crossbred foundations (McDowell et al., 1996). Examples of some syn-

62

A.K. Kahi et al. / Livestock Production Science 63 (2000) 55 – 63

thetic breeds include the Jamaica Hope which is 3 / 4 Jersey, the Cuban Mambi which is 3 / 4 Holstein, the Cuban Sibovey which is 5 / 8 Holstein, the Indian Karan Fries which is 3 / 4 B. taurus (i.e. Brown Swiss and Holstein), the Australian Milking Zebu which is 3 / 4 Jersey and the Brazilian Pitanqueiras which is 5 / 8 Red Poll (McDowell et al., 1996). Because of its organisational simplicity, the synthetic breed strategy is the most realistic approach to utilising the advantage of crossbreeding in small-scale dairying (Syrstad, 1996). A problem to be addressed is the need for an efficient genetic improvement scheme in the synthetic breed comparable to those applied in exotic B. taurus breeds. This may be difficult as a sufficient size of the breeding population is necessary. Large private sector herds, such as Kilifi Plantations, which are sources of breeding stock to the surrounding smallholder farmers, have the resources to run a selection scheme. Additional improvement can be achieved by selecting the best indigenous B. indicus cows to set up synthetic populations. Syrstad and Ruane (1998) have discussed how this can be done using a small number of animals. Cunningham (1980), Smith (1988) and Bondoc and Smith (1993) have described the methodology of open nucleus breeding schemes applicable for the specific situation in developing countries. With such a system, the population size will be large enough for genetic improvement through within-breed selection. This is particularly appropriate for situations such as in Kenya where it is difficult to establish a large-scale field-recording scheme.

have been used, without due regard to the overall system. The results presented here suggest that there is the need to promote greater awareness of the potential of synthetic breeds and to formulate strategies for developing and exploiting them. Because of the diversity in tropical dairy production systems, choice of breeds and crossbreeding strategy should be determined by the ecological and socio-economic characteristics of its target production system. To determine whether the genetic differences among strategies and breeds lead to greater economic benefits, economic evaluations using appropriate economic evaluation criteria (Kahi et al., 1998) are needed. For the breeds and strategies examined in the present paper, this will be discussed in a later paper (Kahi et al., 2000b).

Acknowledgements Valuable suggestions and comments of two anonymous reviewers are gratefully acknowledged. We thank C.D. Wilson of Kilifi Plantations, for permission to use the records from the herd, the German Academic Exchange Service (DAAD, Bonn, Germany) and Hohenheim University (Stuttgart, Germany) for funding the work, and the International Livestock Research Institute (ILRI, Nairobi, Kenya) for provision of facilities. This paper was written when the first author was on leave from Egerton University (Njoro, Kenya).

References 5. Conclusions This study shows that F3S cows were closely rivalled by the (BFS) Rot and the synthetics. Their inferiority is particularly marked in a production system requiring replacements to be raised from within the system. This conclusion is contrary to the prevalent recommendation that the first cross is best suited for dairy production in the tropics, a recommendation based on comparisons of different blood levels in static crossbreeding systems. Even when data from rotational systems have been used, breed composition of individual animals (at a point in time)

Amble, V.N., Jain, J.P., 1967. Comparative performance of different grades of crossbred cows on military farms in India. J. Dairy Sci. 50, 1695–1702. Akbas, Y., Brotherstone, S., Hill, W.G., 1993. Animal model estimation of nonadditive genetic parameters in dairy cattle, and their effect on heritability estimation and breeding value prediction. J. Anim. Breed. Genet. 110, 105–113. Bondoc, O.L., Smith, C., 1993. Deterministic genetic analysis of open nucleus breeding schemes for dairy cattle in developing countries. J. Anim. Breed. Genet. 110, 194–208. Cunningham, E.P., 1980. Methods for recording, evaluation and selection in adverse environments. FAO / UNEP Technical Consultation on Animal Genetic Resources, Conservation and Management, Rome, 12 pp. Cunningham, E.P., 1981. Selection and crossbreeding strategies in

A.K. Kahi et al. / Livestock Production Science 63 (2000) 55 – 63 adverse environments. In: Animal Genetic Resources Conservation and Management. FAO Anim. Prod. and Health Paper No. 24, pp. 279–288. Cunningham, E.P., Syrstad, O., 1987. Crossbreeding Bos indicus and Bos taurus for milk production in the tropics. FAO Anim. Prod. and Health Paper No. 68. Dickerson, G.E., 1973. Inbreeding and heterosis in animals. In: Proc. Animal Breeding Symposium in Honour of Jay Lush, Am. Soc. Anim. Sci. and Am. Dairy Sci. Association, Champaign, IL, pp. 54–77. Dillard, E.U., Rodriquez, O., Robison, O.W., 1980. Estimation of additive and nonadditive direct and maternal genetic effects from crossbreeding beef cattle. J. Anim. Sci. 50, 653–663. Gregory, K.E., Cundiff, L.V., 1980. Crossbreeding in beef cattle: evaluation of systems. J. Anim. Sci. 51, 1224–1242. Gregory, K.E., Trail, J.C.M., Koch, R.M., Cundiff, L.V., 1982. Heterosis, crossbreeding and composite breed utilisation in the tropics. Proc. 2nd World Congr. Genet. Appl. Livest. Prod. VI, 279–282. Grosshans, T., Distl, O., Seeland, G., Wolf, J., 1994. Estimation of individual crossbreeding effects on milk production traits of the German Black Pied dairy cattle using different genetic models. J. Anim. Breed. Genet. 111, 472–492. Kahi, A.K., Kosgey, I.S., Cardoso, V.L., van Arendonk, J.A.M., 1998. Influence of production circumstance and economic evaluation criteria on economic comparison of breeds and breed crosses. J. Dairy Sci. 81, 2271–2279. Kahi, A.K., Thorpe, W., Nitter, G., Baker, R.L., 2000a. Crossbreeding for dairy production in the lowland tropics of Kenya. I. Estimation of individual crossbreeding effects on milk production traits and cow live weight. Livest. Prod. Sci. 63, 39–54. Kahi, A.K., Thorpe, W., Nitter, G., van Arendonk, J.A.M., Gall, C.F., 2000b. Economic evaluation of crossbreeding for dairy production in a pasture based production system in Kenya. Livest. Prod. Sci. (in press). Kinghorn, B., 1980. The expression of ‘recombination loss’ in ¨ ¨ quantitative traits. Z. Tierzuchtg. Zuchtgbiol. 97, 138–143. Kinghorn, B.P., Vercoe, P.E., 1989. The effect of using the wrong genetic model to predict the merit of crossbred genotypes. Anim. Prod. 49, 209–216. Kosgey, I.S., van Arendonk, J.A.M., Kahi, A.K., 1998. Potential benefit and implementation of nucleus breeding scheme for milk production from crossbred cattle in Kenya. In: Conference held on 27–30 January 1998, Nairobi, Kenya, Food, Lands and Livelihood. Setting Research Agendas for Animal Science, pp. 97–98. Lemos, A.M., Teodoro, R.L., Madalena, F.E., 1996. Comparative performance of six Holstein-Friesian3Guzera grades in Brazil. 9. Stayability, herd life and reasons for disposal. Brazilian J. Genet. 19, 259–264. Madalena, F.E., 1981. Crossbreeding strategies for dairy cattle in Brazil. World Anim. Rev. 38, 23–30. Madalena, F.E., 1993. A simple scheme to utilise heterosis in tropical dairy cattle. World Anim. Rev. 74 / 75, 17–25. Madalena, F.E., Teodoro, R.L., Lemos, A.M., Monteiro, J.B.N.,

63

Barbosa, R.T., 1990. Evaluation of strategies for crossbreeding dairy cattle in Brazil. J. Dairy Sci. 73, 1887–1901. McDowell, R.E., Wilk, J.C., Talbott, C.W., 1996. Economic viability of crosses of Bos taurus and Bos indicus for dairying in warm climates. J. Dairy Sci. 79, 1292–1303. Peeler, E.J., Omore, A.O., 1997. Manual of Livestock Production Systems in Kenya. KARI / DFID National Agricultural Research Project II. Rege, J.E.O., 1998. Utilization of exotic germplasm for milk production in the tropics. Proc. 6th World Cong. on Genetics Applied to Livest. Prod. 25, 193–200. Rege, J.E.O., Aboagye, G.S., Akah, S., Ahunu, B.K., 1994. Crossbreeding Jersey with Ghana Shorthorn and Sokoto Gudali cattle in a tropical environment: additive and heterotic effects for milk production, reproduction and calf growth traits. Anim. Prod. 59, 21–29. Robison, O.W., McDaniel, B.T., Rincon, E.J., 1981. Estimation of direct and maternal additive and heterotic effects from crossbreeding experiments in animals. J. Anim. Sci. 52, 44–51. Rutledge, J.J., 1996. Cattle production systems based on in vitro embryo production. J. Reprod. Dev. 42, 18–22. Silva, H.C.M., Wilcox, C.J., Spurlock, A.H., Martin, F.G., Becker, R.B., 1986. Factors affecting age at first parturition, life span, and vital statistics of Florida dairy cows. J. Dairy Sci. 69, 470–476. Smith, C., 1988. Genetic improvement of livestock, using nucleus breeding units. World Anim. Rev. 65, 2–10. Syrstad, O., 1996. Dairy cattle crossbreeding in the tropics: choice of crossbreeding strategy. Trop. Anim. Health Prod. 28, 223– 229. Syrstad, O., Ruane, J., 1998. Prospects and strategies for genetic improvement of the dairy potential of tropical cattle by selection. Trop. Anim. Health Prod. 30, 257–268. Thorpe, W., Kang’ethe, P., Rege, J.E.O., Mosi, R.O., Mwandotto, B.A.J., Njuguna, P., 1993. Crossbreeding Ayrshire, Friesian and Sahiwal cattle for milk yield and preweaning traits of progeny in the semiarid tropics of Kenya. J. Dairy Sci. 76, 2001–2012. Thorpe, W., Morris, C.A., Kang’ethe, P., 1994. Crossbreeding of Ayrshire, Brown Swiss, and Sahiwal cattle for annual and lifetime milk yield in the lowland tropics of Kenya. J. Dairy Sci. 77, 2415–2427. Talbott, C.W., 1994. Potential to increase milk yield efficiency in tropical countries. Ph.D. Thesis, North Carolina State University, Raleigh, NC. Trail, J.C.M., Gregory, K.E., 1981. Sahiwal cattle: An evaluation of their potential contribution to milk and beef production in Africa. Int. Livest. Ctr. Africa. Monogr. No. 3. Int. Livest. Ctr. Africa. Addis Ababa. Ethiopia. Van Arendonk, J.A.M., Brascamp, E.W., 1990. Economic considerations in dairy cattle breeding. Proc. 4th World Congr. Genet. Appl. Livest Prod. XVI, 78–85. Walshe, M.J., Grindle, J., Nell, A., Bachman, M., 1991. Dairy development in Sub-Saharan Africa: a study of issues and options. World bank technical paper No. 135. Africa Technical Department series.