8. The formation of composite breeds using Sahiwal cattle

8.1 The objectives of crossbreeding
8.2 The formation of composite breeds as an alternative to rotational crossbreeding systems
8.3 The use of composite breeds with a Sahiwal component in tropical Africa

8.1 The objectives of crossbreeding

The basic objectives of cattle crossbreeding systems are to take advantage of breed differences in additive genetic merit for specific characters so that performance characteristics are synchronized with adaptability to a given environment and with market requirements, while at the same time making the best use possible of non-additive heterosis effects. In addition, crossbreeding systems may achieve complementarity in part of a self-contained herd through the use of terminal sire breeds that have greater additive genetic merit for growth than the breed of the. cow herd.

Major differences among breeds of Bos taurus and Bos indicus cattle have been demonstrated for most of the characters that contribute to production efficiency (Gregory et al, 1978a, 1978b, 1978c, 1978d, 1978e, 1979a, 1979b; Koch and Dikeman, 1977; Koch et al, 1976; Laster et al, 1976, 1979; Notter et al, 1978a, 1978b; Smith et al, 1976a, 1976b; Young et at, 1978a, 1978b). These breed differences indicate that there are probably major differences in average gene frequencies relating to economically important characters.

Breeds of Bos indicus cattle differ, in response capability for both milk and meat production and also in their adaptability to climate and associated factors, including tolerance of different diseases and parasites. Breeds of Bos taurus cattle differ substantially in performance characters associated with milk and meat production and probably in terms of adaptability to a variety of environmental conditions. Certain populations of Bos taurus cattle in Africa and South America have become adapted to adverse conditions through natural selection, but, similar to Bos indicus breeds, they have a relatively low response capability for milk and meat production when environmental conditions are improved. However, some of these populations may potentially make important contributions to composite breeds for greater milk and meat production under adverse environments.

Results based on both experimentation and computer simulation indicate that differences in additive genetic merit of breeds for specific characters can be used in crossbreeding systems to synchronize genetic resources with other production resources and to provide complementarity through the use of terminal sire breeds (Cartwright et al, 1975; Fitzhugh et al, 1975; Morris and Wilton, 1976; Notter et al, 1979; Smith, 1976; Wilton and Morris, 1976). Experimental results evaluating rotational crossbreeding systems indicate that high levels of heterosis are sustained in successive generations and that the relationship between loss of heterosis and loss of heterozygosity approaches linearity (Boston et al, 1976a, 1976b; Cartwright et al, 1964; Chapman et al, 1970, 1971; Crockett et al, 1978a, 1978b; Franke, 1978, 1979a, 1979b; Gregory and Cundiff, 1980; Koger et al, 1975; Peacock and Koger, 1979; Touchberry, 1970). These results suggest that heterosis in cattle may be primarily the result of the dominance effects of genes.

It has been shown that the cumulative effects of heterosis on characters contributing to weight of calf weaned per cow exposed to breeding are about 23% for crosses among Bos taurus breeds (Gregory et al, 1965; Wiltbank et al, 1967; Cundiff et al, 1974a, 1974b) and 50% or more for crosses between Bos taurus and Bos indicus breeds (Cartwright et al, 1964; Koger et al, 1975). Further, these results show that 60% or more of the observed cumulative heterosis is the result of heterosis effects on maternal traits. Thus, crossbreeding systems that are organized to include a high percentage of the reproducing female population as crossbreds are generally favoured.

The basic outlines for two- and three-breed rotational crossbreeding systems are shown in Tables 26 and 27. The estimate of 8.5% heterosis for individual traits is based on results presented by Gregory et at (1965) and Wiltbank et at (1967) from the first phase of a comprehensive crossbreeding experiment involving the Hereford, Angus and Shorthorn breeds. Their estimate was the result of increases in survival and preweaning growth rate of crossbred calves compared with their straightbred half sibs when all calves were produced by similar straightbred dams. In the second phase of this experiment, crossbred and straightbred cows produced crossbred calves that were equal in individual heterozygosity and additive genetic composition. In this case, the crossbred cows produced 14.8% more calf weight per cow exposed to breeding as a result of both increased reproductive performance and improved maternal ability (Cundiff et at, 1974a, 1974b).

As shown by Dickerson (1969, 1973), heterozygosity stabilizes after seven generations for two- and threebreed rotation crossbreeding systems (Tables 26 and 27). The heterozygosity at equilibrium as a percentage of the maximum (F₁) for n breeds in a rotation is equal to (2ⁿ - 2)/(2ⁿ - 1). After seven generations in a two-breed rotation system, the fluctuation in additive genetic composition reaches equilibrium with two-thirds of the breed of the sire and one-third of the breed of the maternal grandsire (Table 26). After seven generations in a three-breed rotation system, the additive genetic composition reaches equilibrium at four sevenths of the sire breed, two-sevenths of the breed of the maternal grandsire and one-seventh of the breed of the maternal great-grandsire, which is the breed to which the females are mated (Table 27). Because of the wide fluctuation between generations in additive genetic composition in breed rotation crossbreeding systems, the breeds used in a breed rotation system should be reasonably comparable in characters such as birth weight to avoid major calving problems (Laster et at, 1973) and should be compatible in performance characteristics such as size and lactation potential to facilitate common management of all breed-of-sire groups (Cundiff, 1977).

Table 26. Genetic composition and heterosis expected in a two-breed rotation (%)

		Additive genetic composition				Heterozygosity % relative to F1		Estimated % increase in weight weaned per cow exposeda
		Dam		Calf
Generation	Sire breed	A	B	A	B	Dam	Calf
1	A		100	50	50	0	100	8.5
2	B	50	50	25	75	100	50	19.0
3	A	25	75	63	37	50	75	13.8
4	B	63	37	31	69	75	63	16.4
5	A	31	69	66	34	63	69	15.2
6	B	66	34	33	67	69	66	15.8
7	A	33	67	67	33	66	67	15.5
8	B	67	33	33	67	67	67	15.5

^a Based on heterosis effects of 8.5% for individual traits and 14.8% for maternal traits when loss of heterosis is proportional to loss of heterozygosity.

Table 27. Genetic composition and heterosis expected in a three-breed rotation (%)

		Additive genetic composition						Heterozygosity % relative to F1		Estimated % increase in weight weaned per cow exposeda
		Dam			Calf
Generation	Sire breed	A	B	C	A	B	C	Dam	Calf
1	A			100	50	0	50	0	100	8.5
2	B	50	0	50	25	50	25	100	100	23.3
3	C	25	50	25	12	25	62	100	75	21.2
4	A	12	25	62	56	12	31	75	88	18.6
5	B	56	12	31	28	56	16	88	88	20.5
6	C	28	56	16	14	28	58	88	84	20.2
7	A	14	28	58	57	14	29	84	86	19.7
8	B	57	14	29	29	57	14	86	86	20.0

^a Based on heterosis effects of 8.5% for individual traits and 14.8% for maternal traits when loss of heterosis is proportional to loss of heterozygosity.

The experiment from which Gregory et al (1965) and Wiltbank et al (1967) estimated heterosis for individual traits and Cundiff et al (1974a, 1974b) estimated heterosis for maternal traits was continued into a third phase. A preliminary analysis of the results from this experiment carried through two generations of troth two- and three-breed rotational crossbreeding showed retention of heterosis equal to, and in some cases greater than, expectations relative to percentage of F₁ heterozygosity retained (Gregory and Cundiff, 1980).

Existing breeds of each species are mildly inbred lines, and to the extent that heterosis is due to the dominance effects of genes, it represents the recovery of accumulated inbreeding depression (Dickerson, 1973; Cundiff, 1977), Deviation. of heterosis from linear association with heterozygosity results from the epistatic effects of genes. For loss of favourable epistatic combinations that may either have become fixed or are maintained by selection in parental breeds, loss of heterosis is greater than loss of heterozygosity; however, for loss of unfavourable epistatic combinations that may have become fixed through chance, loss of heterosis is less than loss of heterozygosity. Both genetic situations may exist, but favourable epistatic combinations are more likely to exist in parental breeds than unfavourable. Also, heterosis may be greater than heterozygosity if a threshold effect for heterozygosity should exist.

Appropriate investigations have not been conducted to show the importance of favourable epistatic combinations in cattle. There is little likelihood that fixed favourable epistatic combinations are important, however, other than for characters such as fitness for which there has been natural or automatic selection, because the selection goals that have generally characterized cattle breeding are continuously changing.

The higher level of heterosis achieved between crosses of Bos indicus and Bos taurus breeds than between crosses of Bos taurus breeds alone is well documented for beef production characters (Cartwright et al, 1964; Koger et al, 1975; Peacock and Koger, 1979; Crockett et al, 1978a, 1978b). Much of this higher level of heterosis may he the result of a low level of climatic adaptability, and thus of poor purebred performance, of the Bos taurus breeds included in these studies and perhaps a low production response capability of the Bos indicus breeds studied in a moderately intensive beef production environment. The crossbred is a genetic intermediate that is most likely hefter suited than the parental pure breeds for both adapting to the climatic environment and meeting beef production requirements in moderately intensive production systems (Cundiff, 1970). Major differences in level of heterosis probably do not exist between crosses of Bos taurus breeds which are reasonably well adapted to a given production environment (Gregory et al, 1978a, 1978b, 1978d, 1978e). Breeds that show generally higher mean heterosis in crosses with other breeds are probably relatively highly inbred: the higher average heterosis may he the result of greater accumulated inbreeding depression reflected in their contribution to the purebred mean.

8.2 The formation of composite breeds as an alternative to rotational crossbreeding systems

The utilization of heterosis through organized breed rotation crossbreeding systems is restricted in many situations due to two factors. For one thing, the fluctuation between generations in additive genetic composition may severely limit the achievement of optimum contributions by different breeds. For instance, in a two-breed rotation crossbreeding system, the genes contributed by the breeds of the sire and the maternal grandsire fluctuate between one-third and two-thirds between generations.

Such wide fluctuations between generations make it difficult to synchronize climatic adaptability and performance characteristics appropriate to a given management level and natural environment. If Bos indicus breeds should contribute one-fourth or one-half of total germ-plasm to achieve the best climatic adaptability, then rotational crossbreeding systems using pure Bos indicus and Bos taurus sires will not provide the optimum Bos indicus input consistently (Tables 26 and 27). Also, because of the requirement to use compatible breeds in a rotation, rotational crossbreeding systems cannot exploit complementarily (Cartwright, 1970) unless they are combined with terminal-sire systems by breeding mature cows to terminal sires after replacements are produced in a breed-rotation system with younger cows (Notter et al, 1979).

The second major limitation on the use of breed rotation crossbreeding systems is the fact that a substantial proportion of the world's cattle population is kept in management units which are too small to use crossbreeding systems on a self-contained basis. An alternative or supplement to crossbreeding systems is needed to utilize heterosis and to achieve the most favourable additive genetic composition so that a high production response capability for milk and beef production in specific climatic, nutritional, disease and parasite environments is obtained. The formation of composite breeds based on a multibreed foundation is a potentially attractive alternative or supplement to continued crossbreeding. Once a composite breed is formed on the basis of initial crossbreeding, the animals are allowed to breed freely inter se. Thus, the herd is managed as a straightbred population and the management problems that are associated with small herd size in a rotational crossbreeding system are avoided.

A further potential advantage of composite breeds is that their response to selection should be greater than that of parental breeds because of increased genetic variation. Greater selection intensity is also possible as a result of a higher reproduction rate due to heterosis (Dickerson, 1973; Cundiff, 1977).

Table 28. Heterozygosity of different mating types and estimated increase in performance as a result of heterosis

Mating type	Heterozygosity % relative to F1	Estimated % increase in weight weaned per cow exposeda
Pure breeds	0	0
Two-breed rotation	66.7	15.5
Three-breed rotation	85.7	20.0
Four-breed rotation	93.3	21.7
Two-breed composite:
F3: ½A, ½B	50.0	11.6
F3: 5/8A, 3/8B	46.9	10.9
F3: ¾A, ¼B	37.5	8.7
Three-breed composite:
F3: ½A, ¼B, ¼C	62.5	14.6
F3: 3/8A, 3/8B, ¼C	65.6	15.3
Four-breed composite:
F3: ¼A, ¼B, ¼C, ¼D	75.0	17.5
F3:3/8A, 3/8B, 1/8C, 1/8D	68.8	16.0
F3 :½A, ¼B, 1/8C, 1/8D	65.6	15.3
Five-breed composite:
F3: ¼A, ¼B, ¼C, 1/8D, 1/8E	78.1	18.2
F3: ½A, 1/8B, 1/8C, 1/8D, 1/8E	68.8	16.0
Six-breed composite:
F3: ¼A, ¼B, 1/8C, 1/8D, 1/8E, 1/8F	81.3	18.9
Seven-breed composite:
F3: 3/16A, 3/16B, 1/8C, 1/8D, 1/8E, 1/8F, 1/8G	85.2	19.8
Eight-breed composite:
F3: 1/8A, 1/8B, 1/8C, 1/8D, 1/8E, 1/8F, 1/8G, 1/8H	87.5	20.4

^a Based on heterosis effects of 8.5% for individual traits and 14.8% for maternal traits, assuming that loss of heterosis is proportional to loss of heterozygosity. This assumption has not been validated for composite breeds.

Wright (1922) showed that retention of initial heterozygosity following crossing and subsequent random mating within the crosses is proportional to (n-1)/n where n is the number of breeds included in the cross, assuming equal contribution by each breed. The loss in heterozygosity occurs between the F₁ and F₂ generations, Table 28 provides information on the level of heterozygosity relative to the F₁ which is retained after equilibrium is reached for two-, three- and four-breed rotation crossbreeding systems and is presented for comparative purposes for two-, three-, four-, five-, six-, seven and eight-breed composites, with breeds contributing in different proportions in several instances. Heterozygosity relative to the F₁ generation has been adjusted in the- composite populations in which contribution by parenta1 breeds is unequal. Percentage of mean F₁ heterozygosity retained is proportional to 1 - S _iⁿ P_i², where P_i is the fraction of each of n breeds used in the pedigree of a composite breed; i.e. the heterozygosity retained in a three-breed composite formed from ³/₈ breed A, ³/₈ breed B and ¼ breed C is equal to 1 - [(³/₈)² + (³/₈)² + (¼)⁴ ] or 65.6. Obviously, as many breeds as possible should be used which can contribute to a favourable additive genetic composition for the production and marketing situation because retention of heterozygosity is a function of the number of breeds included in the foundation (Wright, 1922; Dickerson, 1969, 1973). Based on the assumption that loss of heterosis is approximately proportional to loss of heterozygosity, estimates of increase in weight produced per cow exposed to breeding for beef production systems are presented in Table 28 for each mating type. The assumption that loss of heterosis is proportional to loss of heterozygosity has not been validated for composite breeds of beef cattle, however, or for any other species. The heterosis levels on which the estimates in Table 28 are based are appropriate for crosses of Bos taurus breeds. As mentioned earlier, heterosis has been shown to be greater in crosses of Bos taurus with Bos indicus breeds.

Dickerson (1969) was the first to point out that composite breeds may have performance advantages from retained heterosis approaching those obtainable by more complex crossbreeding systems. He discusses the possible importance of loss in composite breeds of favourable epistatic combinations that may either have become fixed or are maintained by selection in the parental breeds and presents documentation of recombination loss of epistatic superiority of parental populations in Drosophila, in maize and in poultry. Livestock breeding experiments have not been designed to determine the importance of recombination loss of epistatic purebred superiority, however, because of the problem of separating heterosis retention in composite breeds from the renewed inbreeding and selection which are generally practiced immediately after formation of composite populations. Dickerson (1969) emphasizes the importance of maintaining sufficiently large populations so that the initial advantage of increased heterozygosity is not dissipated by early re-inbreeding of composite breeds. Perhaps the failure of some early efforts at composite cattle breed formation was due to early re-inbreeding and the use of a small number of inadequately characterized parental breeds.

Although composite breed formation would not provide directly and immediately the advantage of mating sires with high breeding value for growth rate to dams of smaller mature size, it does make it possible, through between-breed selection, to increase general adaptability to the climatic and/or nutritive environment at a rapid rate. Further, if specific crossbreeding systems are indicated through the use of breeds that excel in either paternal or maternal characters, composite breed formation provides the opportunity to optimize additive genetic composition rapidly for specific role(s) through the choice of foundation breeds that may he specialized for either paternal or maternal use.

To make composite breed formation a predictable procedure, however, candidate foundation breeds need to he characterized in a range of environments and the extent to which the loss of heterosis is linearly associated with loss of heterozygosity needs to he determined. Information is already available characterizing a number of Bos taurus and Bos indicus breeds in different ecological zones. Although much of this information may he observational or subjective in nature, it is probably adequate to identify with reasonable precision candidate foundation breeds for the establishment of promising composite breeds in different environments. The simultaneous characterization of environments is also necessary to extend the relevance of breed characterization data to the broadest possible range of production situations.

At the same time, experiments are in progress at the U S Meat Animal Research Center in Clay Center, Nebraska, to determine the linearity of association of loss of heterosis and loss of heterozygosity in composite cattle populations and to determine selection response in composite populations compared to the pure foundation breeds. Even if the loss of heterosis in composite breeds should prove to be greater than the loss of heterozygosity, however, the formation of composite breeds may still he the most promising approach to increasing genetic response for meat and milk production in a variety of environments.

The inclusion of the maximum number of breeds of demonstrated potential in the composite will result in the retention of the highest proportion of original (F₁) heterosis. The population of a new composite breed should also be sufficiently large so that the initial advantage of increased heterozygosity is not dissipated by early re-inbreeding. It is suggested that not less than 12 sires be used per generation: this number will result in a rate of increased inbreeding of only about 1% per generation. Once the composite breed is formed in a given production environment, further selection within the population should focus on criteria that contribute to increased meat and milk production, as these factors will also reflect adaptability.

8.3 The use of composite breeds with a Sahiwal component in tropical Africa

Trail (1980) classified Africa south of the Sahara from the standpoint of cattle adaptability considerations into five ecological zones based primarily on the criteria of rainfall and elevation. The five ecological zones are listed in Table 29, along with their estimated cattle populations and the percentage of the total regional cattle population ascribed to each zone (FAO, 1977).

It is recognized that variations in rainfall distribution, elevation, soil fertility, solar radiation, temperature and relative humidity within each of these five ecological zones influence the climate and the nutritional, disease and parasite environment and thus the type of cattle that are adapted and their response capability for milk and beef production. Trail (1980) expressed the view that the environment of the arid to semi-arid ecological zone generally favours indigenous and improved indigenous cattle, while the environment of the semi-arid to humid ecological zone can support exotic-indigenous crosses. He further concluded that parts of the temperate highlands can support pure exotic Bos taurus cattle from the standpoint of climatic considerations as long as the nutritive environment is sufficient to support their high milk and meat production response capability.

Table 29. Cattle populations in sub-Saharan Africa by ecological zone

Ecological zone	Annual rainfall	Estimated cattle population	Percentage of total for sub-Saharan Africa
Very arid	< 400 mm	9000000	6%
Arid to semi-arid	400-600 mm	33000000	22%
Semi-arid to humid	> 600 mm	66000000	44%
Temperate highlands	> 600 mm	33000000	22%
Humid, tsetse-infested	> 600 mm	9000000	6%
Total		150000000	100

Source: FAO (1977).

It is appropriate here to consider the development of composite breeds involving a contribution by the Sahiwal for meat and milk production in the semi-arid to subhumid ecological zones where 44% of the cattle in Africa south of the Sahara are located. Consideration is also given to the use of Sahiwals in the lower-potential areas of the temperate highlands and the higher-potential areas of the semiarid highlands. The herds that contributed records to this study are located as shown in Table 30.

From the standpoint of level of nutritive environment, the five herds are ranked from highest to lowest, in the order listed in Table 30, with perhaps minimal difference between Deloraine Estates and Cedarvale Farms. For rank on nutritive environment from the standpoint of output, El Karama Ranch would perhaps be ranked equal to Kilifi Plantations because most of the output at El Karama is beef. Even though both Deloraine and Cedarvale are classified in the temperate highlands, the nutritive environment provided, as determined by economic feasibility, is not considered adequate to support pure Bos taurus cattle successfully, though Deloraine maintains some straightbred Ayrshire cows on a preferential treatment basis.

Table 30. Ecological zones of study sites in Kenya

Herd	Ecological zone
Kilifi Plantations	subhumid
Deloraine Estates	temperate highlands (lower- to middle-potential area of zone)
Cedarvale Farms	temperate highlands (lower-potential area of zone)
El Karama Ranch	semi-arid highlands (higher-potential area of zone)
Ilkerin Project	semi-arid highlands (higher-potential area of zone)

The Sahiwal breed has the greatest potential for milk production among the Bos indicus breeds generally available in Africa. It is also assumed that the Sahiwal breed is approximately equal to indigenous breeds of Bos indicus cattle in terms of adaptability in the ecological zones of Africa that are the focus of this study. Results from the analysis of records from the five herds in Kenya provide the basis for this assumption. Based on the records of herds composed of Sahiwals in various combinations with other Bos indicus and Bos taurus breeds, an appropriate strategy for maximizing milk and beef production in a given environment appears to be the formation of a specific composite breed for each location, including different proportions of the Sahiwal breed. In the ecological zones where the five herds are located, a minimum Sahiwal contribution of one-fourth is needed to provide adaptability to the climatic, nutritive, disease and parasite environment, while a Sahiwal contribution of more than three-fourths would probably not be favourable in terms of maximizing milk and beef production.

Table 31 gives the breed composition and the percentage of heterozygosity retained relative to the F₁ generation for the reciprocal crosses of 11 different breeds and breeding groups represented in the five herds under analysis The values given in this table for the percentage of heterozygosity retained emphasize the desirability of including three breeds, rather than two, in the foundation of a composite breed, and the importance of having approximately equal contributions by each breed if this can be done while at the same time achieving optimum additive genetic composition. In a composite breed based on equal contributions by three breeds, the proportion of mean F₁ heterozygosity would be

1 - [(¹/₃)² + (¹/₃)² + (¹/₃)² ] = ²/₃, or 0.67.

A composite breed based on equal contributions by four breeds would retain an even higher proportion of mean F₁ heterozygosity, or

1 - [(¼)² + (¼)² + (¼)² + (¼)² ] = ¾, or 0.75.

The percentage of mean F₁ heterozygosity retained by these two composite breeds should be compared with the level of 0.67 retained by a two-breed rotation crossbreeding system and 0.86 retained by a rotation crossbreeding system with three breeds.

The selection of foundation breeds for inclusion in a composite breed needs to he based on performance under environmental conditions similar to those in which the composite breed will be maintained. For instance, a comprehensive experiment involving Hereford cattle (Pahnish et al, 1979; Butts et al, 1971) showed the importance of the interaction of genotype with climate and associated forage resources. In an environmental situation which favours a contribution by Bos indicus breeds greater than one-fourth, a three-breed composite including the Boran breed, as well as the Sahiwal and a Bos taurus breed, would be desirable. The milk and beef production potential of the Boran breed is approximately equal to that of the Sahiwal over a range of African environments.

Table 31. Breed composition and heterozygosity of progeny from different mating types

		Sire Breed Groupa
Breed Dam Groupa		1 Sahiwal	2 Ayrshire	3 Friesian	4 Brown Swiss	5 ¾S -¼A
1. Sahiwal		-	½S -½A	½S -½F	½S ½B	-
	heterozygosity % relative to F1	-	50.0	50.0	50.0
2. Ayrshire		½S- ½A	-	-	-	3/8S - 5/8A
	heterozygosity relative to F1	50.0				46.9
3. Friesian		½S-½F	-	-	-	3/8S-1/8A-½F
	heterozygosity % relative to F1	50.0				59.4
4. Brown Swiss		½S-½B	-	-	-	3/8S -1/8A -½B
	heterozygosity % relative to F1	50.0				59.4
5. ¾S - ¼A		-	3/8S - 5/8A	3/8S - 1/8A - ½F	3/8S - 1/8A - ½B	¾S - ¼A
	heterozygosity % relative to F1		46.9	59.4	59.4	37.5
6. ½S- ½A		¾S - ¼A	¼S- ¾A	¼S - ¼A - ½F	¼S - ¼A - ½B	5/8S - 3/8A
	heterozygosity % relative to F1	37.5	37.5	62.5	62.5	49.9
7. ¼S - ¾A		5/8S - 3/8A	-	-	-	½S-½A
	heterozygosity % relative to F1	46.9				50.0
8. 2/3S- 1/3A		-	1/3S - 2/3A	1/3S - 1/6A - ½F	1/3S.- 1/6A - ½B	17/24S - 7/24A
	heterozygosity % relative to F1		44.5	61.1	61.1	41.3
9. 1/3S- 2/3A		2/3S- 1/3A	-	-	-	13/24S - 11/24A
	heterozygosity % relative to F1	44.5				49.6
10. ½S - ½F		¾S - ¼F	¼S- ¼F - ½A	¼S- ¾F	¼S - ¼F - ½B	5/8S - 1/8A - ¼F
	heterozygosity % relative to F1	37.5	62.5	37.5	62.5	53.1
11. ½S - ½B		¾S - ¼B	¼S - ¼B - ½A	¼S - ¼B - ½F	¼S - ¾B	5/8S - 1/8A - ¼B
	heterozygosity % relative to F1	37.5	62.5	62.5	37.5	53.1

^a S = Sahiwal, A = Ayrshire, F = Friesian and B = Brown Swiss.

Table 31, cont.

		Sire Breed Groupa
Breed Dam Groupa		6 ½S - ½A	7 ¼S - ¾A	8 2/3S - 1/3A	9 1/3S - 2/3A	10 ½S -½F	11 ½S - ½B
1. Sahiwal		¾S -¼A	5/8S - 3/8A	-	2/3S - 1/3A	¾S - ¼F	¾S - ¼B
	heterozygosity % relative to F1	37.5	46.9		44.5	37.5	37.5
2. Ayrshire		¼S - ¾A	-	1/3S - 2/3A	-	¼S - ½A - ¼F	¼S - ½A - ¼B
	heterozygosity % relative to F1	37.5		44.5		62.5	62.5
3. Friesian		¼S - ¼A - ½F	-	1/3S - 1/6A - ½F	-	¼S - ¾F	¼S - ½F - ¼B
	heterozygosity % relative to F1	62.5		61.1		37.5	62.5
4. Brown Swiss		¼S - ¼A - ½B	-	1/3S - 1/6A - ½B	-	¼S - ½B - ¼F	¼S - ¾B
	heterozygosity % relative to F1	62.5		61.1		62.5	37.5
5. ¾S - ¼A		5/8S - 3/8A	½S - ½A	17/24S - 7/24A	13/24S - 11/24A	5/8S - 1/8A - ¼F	5/8S - 1/8A - ¼B
	heterozygosity % relative to F1	46.9	50.0	41.3	49.6	53.1	53.1
6. ½S - ½A		½S - ½A	3/8S - 5/8A	7/12S - 5/12A	5/12S - 7/12A	½S - ¼A - ¼F	½F - ¼A - ¼B
	heterozygosity % relative to F1	50.0	46.9	48.6	48.6	62.5	62.5
7. ¼S - ¾A		3/8S - 5/8A	¼S - ¾A	11/24S - 13/24A	7/24S - 17/24A	3/8S - 3/8A - ¼F	3/8S - 3/8A - ¼B
	heterozygosity % relative to F1	46.9	37.5	49.6	41.3	65.6	65.6
8. 2/3S - 1/3A		7/12S - 5/12A	11/24S - 13/24A	2/3S - 1/3A	½S - ½A	7/12S - 1/6A - ¼F	7/12S - 1/6A - ¼B
	heterozygosity % relative to F1	48.8	49.6	44.5	50.0	56.9	56.9
9. 1/3S - 2/3A		5/12S - 7/12A	7/24S - 17/24A	½S - ½A	1/3S - 2/3A	5/12S - 1/3A - ¼F	5/12S- 1/3A - ¼B
	heterozygosity % relative to F1	48.6	41.3	50.0	44.5	65.3	65.3
10. ½S - ½F		½S - ¼A - ¼F	3/8S - 3/8A - ¼F	7/12S - 1/6A - ¼F	5/12S - 1/3A - ¼F	½S - ½F	½S - ¼F - ¼B
	heterozygosity % relative to F1	62.5	65.6	56.9	65.3	50.0	62.5
11. ½S - ½B		½S - ¼A - ¼B	3/8S - 3/8A - ¼B	7/12S - 1/6A - ¼B	5/12S - 1/3A - ¼B	½S - ¼B - ¼F	½S - ½B
	heterozygosity % relative to F1	62.5	65.6	56.9	65.3	62.5	50.0

^a S = Sahiwal, A = Ayrshire, F = Friesian and B = Brown Swiss.

The Bos taurus stock used in the crossbreeding programmes described in this study were originally imported from Europe or North America. There they were selected according to their performance levels in a temperate climate, which is probably not an appropriate indicator of performance in the tropics. Since it was first introduced, however, the Ayrshire breed has been used extensively in all the production environments covered in this study and the adaptability and response capability for milk and beef production of Ayrshire crosses have been well documented. In addition, limited information available from Cedarvale Farms indicates that Sahiwal crosses with the Friesian breed are comparable with equivalent Sahiwal × Ayrshire crosses in terms of several production characters At Kilifi Plantations, limited use has also been made of the Brown Swiss breed in a crossbreeding programme with Sahiwals and Ayrshires, and initial results are promising. Although limited use has been made of Simmental sires at Deloraine Estates, sufficient data on their progeny were not available to assess their potential contribution to composite breeds in similar environments.

A four-breed composite based on equal contributions by the Sahiwal, Ayrshire, Friesian and Brown Swiss breeds could be created by crossing Friesian-Brown Swiss sires with Sahiwal-Ayrshire cows. Such a breed should be well adapted to the climatic environment represented by Deloraine Estates. However, the nutritive requirements for such a composite breed would likely be in the range of the level provided by Kilifi Plantations. Such a composite may be on the borderline in terms of climatic adaptability at Kilifi, with only one-fourth Sahiwal contribution, but this might be adequate given the favourable nutritive environment.

The results from the analysis of records contributed by the five herds provide a basis for suggesting specific composite breeds for each environmental situation. All possible reciprocal crosses of the 11 breeds and breeding groups represented in the five herds under study are shown in Table 31. The level of initial heterozygosity retained in the various potential composite breeds depicted in the table ranges from 0.375 to 0.656.

Detailed suggestions for composite breed formation in specific situations must be based on considerations of the natural environment, the level of management which is economically feasible and the relative importance of milk and beef as production goals. More specifically, consideration needs to be given to average temperatures, rainfall amount and distribution, solar radiation, relative humidity, soil fertility and the quality of natural forages available, along with the degree of environmental modification which is economically justified and technically feasible. Input costs and output values need to be projected for the medium and longer term, because the development of composite breeds has both intermediate and longer-term implications. The five situations described in this study vary considerably in terms of climatic factors, potential nutritive environment and production goals.

For environmental situations represented by Ilkerin, which lies in a higher potential area of the arid to semi-arid ecological zone, the optimum combination for a composite breed may be in the range of five-eighths to three-fourths Sahiwal with the Ayrshire breed contributing the remaining one-fourth to three-eighths. These combinations, however, have the disadvantage of retaining relatively low proportions of initial heterozygosity. A ⁵/₈ Sahiwal - ³/₈ Ayrshire combination would retain 0.469 of F₁ heterozygosity, while a ¾ Sahiwal - ¼ Ayrshire combination would retain only 0.375 (Table 31).

El Karama Ranch also lies in a higher-potential area of the arid to semiarid ecological zone. In this environment the Sahiwal breed should probably contribute from three-eighths to five-eighths to a composite breed with the Ayrshire breed a strong candidate to contribute the remainder. Such a composite breed would retain a minimum of 0.469 of F₁ heterozygosity, or 0.50 if the Sahiwal and Ayrshire breeds each contribute equally.

Cedarvale Farms lies in a lower-potential area of the temperate highlands. Suitable composite breeds for this situation, as shown in Table 31, might be reciprocal cross combinations 10-2 (¼S - ½A - ¼F) or 11-2 (¼S - ½A ¼B). An Ayrshire contribution of one-half is suggested, with no more than one-fourth of either Friesian or Brown Swiss, because Ayrshires are probably more suited to this fairly restricted feed environment because of their smaller size and lower milk production potential. The two composite breeds suggested would retain 0.625 of the mean heterozygosity,

Deloraine Estates lies in a lower- to middle-potential area of the temperate highlands, with temperatures similar to those at Cedarvale but with higher average rainfall. For such environmental situations, with some improvement in the natural nutritive environment, a number of composite breeds appear promising. These are shown in Table 31 as reciprocal cross combinations 10-2 (¼S - ½A ¼F), 10-4 (¼S - ½B - ¼F), 11-2 (¼S - ½A - ¼B) and 11-3 (¼S - ½F ¼B). All of these composite breeds would retain 0.625 of the mean F₁ heterozygosity.

Kilifi Plantations lies in the subhumid zone, with a relatively favourable nutritive environment but a rather stressful climate. The nutritive environment will support a fairly high response capability, so that a number of composite breeds may be considered, as shown in Table 31. These would include reciprocal cross combinations 11-6 (½S - ¼A - ¼B), 11-7 (³/₈S ³/₈A - ¼B), 11-8 (⁷/₁₂S - ¹/₆A - ¼B), 11-9 (⁵/₁₂S- ¹/₃A - ¼B), 11-10 (½S - ¼F ¼B), or possibly reciprocal cross combinations 10-6 (½S - ¼A - ¼F), 10-7 (³/₈S- ³/₈A - ¼F), 10-8 (⁷/₁₂S - ¹/₆A - ¼F) or 10-9 (⁵/₁₂S - ¹/₃A ¼F). A Brown Swiss contribution might be favoured over a similar Friesian contribution because a composite breed with one-half Brown Swiss and the remainder Sahiwal and Ayrshire has shown from initial trials to be adapted to this climatic and feed environment. A composite breed of reciprocal cross combinations 10-11 (½ Sahiwal - ¼ Friesian - ¼ Brown Swiss) should also be we'll adapted to this environment with a relatively high response capability for both milk and beef production. A contribution by the Sahiwal breed in the range of three-eighths to seven-twelfths should favour climatic adaptability. The range in percentage of mean F₁ heterozygosity retained in the composite breeds suggested is from 0.569 to 0.656, which approaches the level achieved in a two-breed rotation crossbreeding system (Table 26).