Previous Page Table of Contents Next Page

Maintaining crossbred populations of dairy cattle in the tropics


Under nearly all conditions of climate and management throughout the world, except the very worst, the milk yield of a local cattle breed can be increased by crossing it with an improved dairy or dual-purpose (milkmeat) breed. Thus, as countries with a temperate climate develop their dairy industries they inevitably replace their local breeds by such improved breeds, either by grading up (displacement crossing) or by importing both cows and bulls. The same applies to developed countries in subtropical regions, such as Israel, South Africa, southern Australia and the southern United States.

As long as the European dairy cattle are well fed, protected from disease, shielded from direct sun and given a chance to cool down at night and in the winter, they do not appear to be seriously inconvenienced by high environmental temperatures. It may be difficult to get cows in calf during the hot wet season and there may be a seasonal decline in yield but in general a European dairy breed is the most economic type for these regions. The same applies to those tropical regions in which the climate is moderated by altitude, as in many parts of east and south-central Africa, or by proximity to the sea, as in the islands of the Caribbean and the South Pacific.

However, there comes a point as we approach the deep tropics where purebred European cattle require more elaborate management than is possible under the existing economic and technological conditions. Many trials in east Africa, India, southeast Asia and Central America have shown that crossbreds between local breeds and European dairy breeds not only produce more milk than the local breed but also more than the pure European (see Table 1 in this article, and Tables 2 and 3 in Meyn and Wilkins in this issue, pages 25 and 27). Naturally, if the first cross is half way between the two breeds and the backcross to the European breed is half way between halfbred and purebred then the genetic situation is purely additive and grading up to the European breed is indicated. This article is concerned with the case where optimum performance is given by some intermediate between the two breeds, and will discuss breeding schemes which can be used to maintain this hybrid vigour.

Maintaining a crossbred population

There are three ways of mating the F1 females after the initial cross. They can be mated to sires of the local breed or of the improved breed or they can be mated to F1 sires. On an experimental basis, all three matings should be made. This will make possible a comparison between the F1 and F2 and between halfbreds and three-quarterbreds. Hopefully, purebreds will also be present for comparison.

If the halfbred or threequarter is clearly superior then inter se breeding within this level of exotic blood should begin at once. There is no point in trying further to refine the exact proportions which may be optimal. There will be so much variation within grades that only an approximation is necessary and selection can at once be based on performance rather than on proportion of blood. Because the well-known Santa Gertrudis beef breed was based on 3/8 zebu and 5/8 British, there has been a tendency to extol this as the ideal proportion and it has been followed in several other new breeds; e.g., the Bonsmara of South Africa and the Jersind and Brown Sind of Allahabad. This is pure formalism. In practice there are three principal ways of maintaining a crossbred population, whether it starts from an F1 or a backcross:

  1. The crosses may be bred inter se to produce a new breed which will have, at least initially, either 50 percent or 75 percent of the blood of the exotic breed.

  2. The crossbred bulls may be used to mate with, and eventually to grade up, the local cows. The proportion of exotic blood will thus rise gradually to reach the limiting values of 50 percent, or 75 percent when the displacement is complete and a new breed is formed.

  3. The F1 cows can be bred to bulls of the two pure breeds in alternate generations in a crisscrossing system. When equilibrium is reached the cows will alternate in successive generations between 66.6 and 33.3 percent exotic blood.

I.L. Mason is Animal Breeding Officer, Animal Production and Health Division, FAO, Rome.

In systems (1) and (2) 50 percent of the hybrid vigour of the F1 is maintained in the new breed. System (3) retains two thirds of the hybrid vigour. On theoretical grounds this would be the system to use if a large decline in performance from F1 to F2 indicated the presence of hybrid vigour (see Tables 2 and 3).

However, important criteria are: which system can be operated in practice and which offers the best scope for a selection programme to follow the initial crosses? Effective genetic improvement in milk yield depends firstly on an efficient milk-recording system, and secondly on a bull progeny testing scheme based on the results of this recording.

If such a system is available it can be applied at once to the new breed. The breed formation will have to be made on a scale sufficient to provide opportunity for progency testing several bulls each year. This cannot be restricted to a single small government herd. Within a single experimental herd milk recording is easy enough but cows may be insufficient in number to allow an adequate progeny test of several bulls and a choice between them.

If no milk-recording system is available then it may be better to exploit the improvement programme in another country (or possibly in a more developed part of the same country) by importing improved bulls or their semen and using a crisscrossing or rotational breeding system.

Formation of new breeds

The activities of breeders and breed societies in western Europe and North America during the last century have taught us to think of breeds as closed population of nearly identical animals. From the present point of view the important characteristic of a breed is that it is self-contained. It generates its own supply of bulls and does not have to rely on an outside source of bulls or semen.

Because of the aura from the past and the emphasis on uniformity there has been a reluctance to take the step of attempting to form a new breed. People are afraid that breeding from crossbreds will lead to excessive variation, and therefore that formation of a new breed is only possible if immense resources of animals are available. In fact, the increase in variation on breeding an F2 generation has been much exaggerated. To begin with, it is only the genetic variation which is increased, and with characters of low heritability this is only a small part of the total variation. Secondly, the obvious effects are due to segregation of colour, pattern and morphological characters (e.g., horns) which are controlled by single genes. The increase in variation between F1 and F2 for such quantitative characters as growth rate and milk yield will be very small, and in view of other sources of variation may not be detectable.

Table 1. Performance of European x zebu crossbred cows at military dairy farms in India during 1934–61

Grade of cowFirst lactationFemale calves, birth to first calving
  Kg Percent
Sahiwal1181 7806128
1/4-European241 5703732
3/8-      "962 1207941
1/2-      "1172 560244
5/8-      "722 350822
3/4-      "3072 33019418
7/8-      "2182 24023730
15/16-   "952 11015436
31/32-   "101 8407345

Source: Amble and Jain, 1967.


1. Taylor cow, Patna, India.

Table 2. Performance of succeeding generations of Ayrshire X Red Sindhi crosses in Madras

GenerationNumberAge at first calvingAverage lactation yield
Red Sindhi (Selected)15-1 558
Ayrshire × Sindhi   
2731.52 280
2336.51 496
1138.01 646

Source: Littlewood, 1973.

Table 3. Performance of Sinhala cattle and their crosses with Jersey and Friesian in Sri Lanka1

 NumberMilk yieldAge at first calvingCalving interval
Jersey Sinhala    
2141 21436.6370
Friesian × Sinhala    
1571 57137.4393

Source: Wijeratne, 1970.
1 Yield corrected for age and season.

Certainly there is a danger that if the initial crossing is not done on a large enough scale then problems of inbreeding may be encountered. This is particularly the case if the breed is based on a single herd or on only one or two imported bulls. Inbreeding leads to decline in fertility, viability and growth rate to a greater or lesser extent. This has often not been realized and inbreeding has been deliberately employed in order to concentrate the blood of outstanding sires or to produce uniformity in a new breed. But uniformity is an elusive goal, a will-o' -the-wisp. It must not be sought at the expense of productivity. Indeed, it is essential at the outset to have as much variation as possible in order to allow scope for selection. The aim of crossbreeding is to combine the high yield of the European breed with the resistance of the local breed, and intense selection is needed to find the few animals combining both characters.

However, one should not be put off by takes about the formation of the Santa Gertrudis breed, which is said to have been based on a herd of 25 000 cattle on a ranch of 400 000 hectares (a million acres). In fact, only a few hundred of these animals were used for breed formation and the rest of the herd was merely graded up. The lesson from the Santa Gertrudis (as from the Shorthorn) is the danger of concentrating too much on single bulls. Both breeds were linebred to a single bull — Monkey for the Santa Gertrudis and Shakespeare for the Shorthorn — and it is probably no coincidence that both breeds have at times run into severe infertility problems.

If a breed is to be formed in a single experimental herd it is therefore essential to use a large number of bulls (at least 10-15) in the formative stages. As early as possible the programme should be extended to other herds and bulls should be interchanged between herds. A large cow population will be needed for progeny testing the bulls. Above all, animals should be selected not according to colour, conformation or other fancy points but for important economic characters which define total dairy merit, i.e., milk yield, viability and thrift of the calf, fertility of the cow, ease of milk let-down, udder and teat characteristics (including durability and suitability for machine milking, if appropriate), docility, and possibly also milk composition. Further details of a breeding scheme are available in the paper by Amble and Jain (1965).

At the present time crossbreeding between zebus and dairy breeds is taking place at several centres in the tropics. It would be most desirable to coordinate the activities within each crossing breed (Friesian, Jersey, Brown Swiss and Danish Red) and encourage exchange of breeding stock between countries to ensure a wide genetic base for each of the new breeds.

New breeds by introducing European bulls into a zebu population

There have been several attempts in the past to form new breeds by introducing European dairy genes into the zebu cattle of India and Africa, but none has come to fruition.

In Tanzania, for instance, Hutchison (Mahadevan and Hutchison, 1964) started a crossbred herd of European dairy cattle x east African zebus in 1946, but it was disbanded in 1966. Earlier the new strain had been officially baptized “Taurindicus”. It is clearly easier to name a new breed than to maintain it, and the reports from India contain many attractive and descriptive names such as Jersind, Brown Sind and Karan Swiss which presumably will all be eventually absorbed in the Kamaduk (see Figure 2). In tropical South America one has read of the Ocampo (Venezuela), Jerdi (Brazil), Tropical, Tropicana and Suisbu (Argentina). It would be interesting to have more particulars of these breeds and any others of European x zebu origin. While there are still no good examples of new dairy breeds whose genesis has been planned and executed on the basis of a European male X zebu female cross, there are two interesting examples of populations developed haphazardly from this cross in the past which have remained as distinct entities because of their economic value. These are the Taylor cattle of India and the Hatton cattle of Sri Lanka.

2. Diagram of breeding programme for evolving a new breed of dairy cow in India — the Kamaduk (Bhat, 1972).

The Taylor cattle of Patna (Bihar state) originate from four Shorthorn and Channel Island bulls introduced by Commissioner Taylor in 1856 for crossing onto the local zebus. According to Sinha (1951) there were in his time about 2 500 Taylor cows, and their milk yield averaged 6–8 kg daily.

The Hatton (or Cape) cattle are said to derive originally from European imports which the Dutch brought to Ceylon via the Cape of Good Hope between 1765 and 1815. Until the more recent imports of European dairy breeds they were the principal milk producers on the island (Mahadevan, personal communication).

Neither of these populations has had the benefit of either a scientific breeding programme or a society for breed promotion. That they should have survived so long indicates that they must have considerable intrinsic merit. They deserve to be called “breeds” and merit a suitable improvement plan. Indeed, a programme of milk recording, progeny testing and breed promotion might even pay quicker dividends than starting now the laborious business of forming a new breed from scratch. Such a programme could, of course, include the introduction of outstanding halfbred bulls in order to benefit from the genes of today's improved dairy breeds.

New breeds by introducing zebu bulls into a European breed population

The first cattle introduced into the New World (i.e., America and Australia) came from Europe. When it was observed that in the tropical and subtropical areas of these continents they performed less well than in their homelands it was thought that the explanation might be lack of heat tolerance. The remedy suggested was the introduction of zebu blood.

For this purpose the dairy zebus from Pakistan, the Red Sindhi and the Sahiwal, have been the most popular. Red Sindhis were imported into the southern states of the United States for experimental crossing with Jersey, Holstein-Friesian and Brown Swiss. The result of the cooperative experiments carried out in Georgia, Louisiana, Maryland and Texas (Branton et al., 1966) showed that as the proportion of zebu blood increased, the milk yield fell (see Figure 5). It is true that heat tolerance increased, but this did not have any economic advantage. The differences in fertility, if significant, were in favour of the pure European breed.

These experiments have been given up, and the milk yield of European breeds in the North American subtropics is being increased by attention to feeding and management combined with selection for productivity within the hot environment.

Nevertheless, the conclusions of Branton et al. (1966) include the sentence: “If the hypothesis of lower nutritive requirements is acceptable, then under conditions of adverse climatic conditions and poor nutrition, some proportion of zebu breeding may prove desirable.” There are two areas where this has proved to be the case and two new breeds, the Jamaica Hope and the Australian Milking Zebu (AMZ), have been developed.

3 and 4. Two examples of the Jamaica Hope breed — “Basil's Rosie” and “Blossom's Basil Bertie”.


5. Milk yield of Jersey x Red Sindhi crosses in the United States (figure from Branton et al., 1966).

At the government farm (Hope Agricultural Station) in Jamaica the beneficial effect on milk yield of using halfbred zebu bulls on Jersey cows was recognized before 1920, and in that year a single Sahiwal bull was imported to continue the good work (Lecky, 1949). Fortunately he turned out to be an outstanding animal, and the Jamaica Hope breed owes most of its zebu genes to him. The new breed contains about 20 percent zebu blood and 70–75 percente Jersey, the rest being from Friesian and other breeds. The bulls produced in the government herd were used to grade up the general Jersey-zebu stock on the island and now most of the dairy cows belong to this breed.

Mahadevan (1966) gives an excellent short account of the origin of this breed and Mahadevan et al. (1970) show that its further development is being limited by the restriction of progeny testing of bulls to the government herd. This means that the number of daughters per bull is too low for an accurate assessment, or else that too few bulls can be tested to given an adequate intensity of selection. All farmers should be involved in testing, by using young bulls extensively through artificial insemination and milk-recording their daughters.

In Australia the approach to forming a new breed based on Jersey and zebu has been much more systematic, and is described in the article by R.H. Hayman in this issue of World Animal Review. In fact, the first experimental results were similar to those in the United States — in both cases the F1 had a much lower milk yield than the Jersey. At this point the Americans gave up but the Australians were convinced that zebu blood had something to offer in the warmer, tick-infested parts of northern New South Wales and in Queensland. The results which Hayman gives in his Table 4 indicate that they were right. The breeding system used has been crossbreeding and selection up to the F3 generation, followed by grading up the local population of Jerseys to the halfbred bulls and at the same time progeny testing these bulls. The question arises, why should zebu blood be an advantage in New South Wales (and in the West Indies) but not in the similar climate of the southern United States? Several differences can be noted which may be significant. For instance, cattle ticks and tick-borne fevers are present in northern New South Wales (and in Jamaica) but are not a problem in the United States. Feeding and management levels are clearly different in Australia and the United States, judging by the mean yield of 2 989 kg for Jerseys given by Branton et al. (1966) compared with 1 944 kg in Hayman's Table 1. In Australia there is more reliance on pasture. The F1 cows also have more difficulty in establishing a lactation in the absence of the calf. In the United States the F2; yielded less than the F1 (see Figure 2); in Australia (see Hayman's Table 1) the F2 and F3 yielded more than the F1. This increase was presumably achieved by intense selection among their sires.

6. Jamaica Hope cows, Bodles, Jamaica.

New breeds of multiple racial origin

The above discussion of new breed formation is concerned with a foundation based primarily on two breeds only — one improved temperate dairy breed and one zebu. (The additional European blood in the Taylor, Hatton and Jamaica Hope breeds and the use of two zebu breeds in the formation of the AMZ, do not really affect the issue.) In a planned breed formation three or more breeds could be used as a matter of deliberate policy. This will increase the genetic variance and hence the scope for selection, but it will also make the programme more complicated and delay the final birth of the breed. It may also overload the foundation with genes from inferior breeds — note how Hayman had to get rid of the inferior Sindhi crosses.

7. F1 Sindhi × Jersey cross at Jeanerette, La., United States.

The Indian Council of Agricultural Research has launched an ambitious programme to create a new breed — the All-India Coordinated Research Project on dairy cattle breeding (Bhat, 1972). It is based on the current crossbreeding between Jersey, Friesian and Brown Swiss on the one hand and various local zebu breeds on the other (Hariana, Sahiwal, Gir, Ongole, Hallikar, Tharparkar) at eight different centres. The breeding scheme is shown in Figure 2. The first crosses will be backcrossed to a second European breed and the backcrosses combined in pairs. Thus it will be three generations (say 15 years) before the foundation stock is ready. Then will start the important phase of progeny testing bulls on a large enough scale so that those rare animals can be picked out which combine (in some measure at least) the genes for high milk and butterfat yield and growth rate and fertility of the European breeds with those for disease resistance and heat tolerance of the Indian breeds.

8. Sahiwal × Friesian cross at Morula Farm, Naivasha, Kenya.

This is a very ambitious programme. It might be quicker and easier to think in terms of two new breeds — one based on the Jersey for early maturity and butterfat, the other on the Friesian for milk and beef. Table 4 shows the preliminary results from the crossbreeding work at Haringhata, West Bengal, which was started under the auspices of an FAO project and has now been incorporated into the All-India Coordinated Research Project. If these are confirmed by later results they would appear to indicate the advisability of concentrating on Jersey and Friesian rather than Brown Swiss.

9. Backcrosses of jersey × Hariana to Jersey (left) and Hariana (right) at Haringhta, India.

Crisscrossing between local and improved breeds

The examples given above show that while it is easy to increase the milk yield of a zebu breed by crossing with a European dairy breed the difficulty is to maintain the advantage in later generations. It needs an extensive system of milk recording and progeny testing, preferably based on an artificial insemination service so that each bull can have many daughters distributed through several herds. If no such facilities are available it is better to use a crisscross system of breeding.

When equilibrium is reached, the cows in alternate generations have one-third improved or one-third local blood. These two types should be kept in separate herds, the first run with a European bull and the second with a local bull. If this is not possible, the presence or absence of hump (assuming that the local breed is a zebu) will indicate which breed of bull should be used on each cow. If artificial insemination is used the separation into herds is not necessary; the humped cows will be inseminated with European semen and the humpless with zebu.

The total resources of the farm can be used for keeping cows and breeding their replacements. All male calves will be slaughtered and bulls will be brought in as necessary. The continuing genetic improvement stems from these bulls. They will come from herds in which a rigid breedimprovement system is in operation. If artificial insemination is available then semen can also be obtained from outside the country, but if the exotic semen comes from a temperate country the question of genotype x environment interaction must be borne in mind — the best bulls as tested under temperate conditions may not be the best in the tropics.

Naturally a source of improved local bulls is also necessary, and this highlights the importance of a selection programme such as that described by Meyn and Wilkins for the Kenya Sahiwal in this issue (see page 24). Replacement cows will be chosen from the best cows of the two types. The two-thirds exotic may well give more milk than the one-third exotic, but the temptation to choose more two-third than one-third exotics as replacements should be avoided. This would mean a preponderance of onethird exotics in the subsequent generation. It would therefore be desirable to start the crisscrossing by backcrossing half the F1 cows to the local and half to the exotic bull. This would enable the two types to be kept in equilibrium from the start.

Table 4. Fertility and milk yield of purebred and crossbred heifers at Haringhata. West Bengal, in 1972/73

 HarianaCrossbred heifers
Foundation cowsControl heifersJersey crossFriesian crossBrown Swiss cross
Body weight at 18 months (kg) 1931(27)233266250
Conception rate (%)     
to first service
in 1971/72
15.522.3 49.7 
Mortality to 18 months, all years (%)  12.511.919.0
Age at first service (months) 28.517.818.720.0
Age at first calving (months) 41.428.028.530.5
Lactation milk yield (kg)266432651 898 (15)2 162 (13)1 821 (7)
Lactation length (days)298127304 (15)345 (13)349 (7)
Milk fat (%)
Milk SNF (%)

Source: Guha, 1973.
1 Figures in brackets give number in mean for less than 50 observations.
2 Calf present.
3 Calf absent.

Crisscrossing has very definite genetic advantages. By breeding always from crossbred cows it exploits any hybrid vigour in fertility as well as in yield. It can be modified by introducing a third breed into the rotation, e.g., a second European breed, if it appears desirable to maintain a higher proportion of European blood. Table 5 shows some results of the beginning of a rotational crossbreeding scheme using three breeds at Turrialba, Costa Rica. However, since the last cross was by the breed with the highest milk yield it is not clear how much of the advantage of the three-breed cross is due to additive gene action and how much to heterotic effects. A comparable system has been in operation among beef-cattle breeders in Africa for many years on an ad hoc basis. The defensive, almost guilty way they describe it is entirely unnecessary. The system is genetically sound and should be exploited for both beef and dairy cattle.

In their article in this issue, Meyn and Wilkins present evidence that in the semiarid and coastal areas of Kenya the Sahiwal-Ayrshire cross gives more milk than the purebred Ayrshire, and the Sahiwal-Friesian cross more calves than the purebred Friesian. They also demonstrate that Kenya has a source of improved Sahiwals and of locally adapted European dairy breeds. They therefore recommend a crisscrossing programme of the type outlined above, particularly for use in dairy ranching.

Likewise Mahadevan (1970) presents evidence of the advantages of a rotational crossbreeding system for the lowlands of Sri Lanka.

Such a scheme has the further advantage of flexibility. The breeds used for crossing can be changed immediately according to changes in environment or demand (milk or meat). The two types of animals produced may have different roles to play, e.g., the two-thirds zebu steer may be more suitable for draught, or in a dairy ranching enterprise the two-thirds zebu cows may be used for suckling.

If beef production is also important, then the crisscross cows not used for breeding replacements can be put to a “terminal” beef sire and all their progeny raised for beef.

The tradition of the uniform purebred should not be allowed to obscure the many advantages of this relatively novel method of breeding dairy cattle for the tropics.

Table 5. Production averages by breed groups in the Turrialba station dairy herd, Costa Rica

 Breed or crossbred group1
CriolloJerseyJ × C
Ayrshire × JC
Age at first calving (months)36313230
Adult weight (kg)406331375-
Services per pregnancy1.
First lactation    
305-day milk yield (kg)
11421 8001 7321 911
butterfat (%)
305-day milk yield (kg)
1 9452 1512493-
butterfat (%)

Source: Anon. 1973.
1 Numbers per group vary from 12 to 28. These averages are not completely contemporary, not free from the effects of culling or selection. The Criollo yields are probably overestimates.


Amble, V.N. & Jain, J.P. 1965. A plan for evolving a new breed of dairy cattle by crossing indigenous and exotic breeds. J. Genetics, 59: 91–109.

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.

Anon. 1973. Breeding to improve milk production in the tropics. Activities in Turrialba, July 1973, p. 6–7.

Bhat, P.N. 1972. Kamaduk: the cow of plenty. Indian Farming, October 1972.

Branton., C., McDowell, R.E. & Brown, M.A. 1966. Zebu-European crossbreeding as a basis of dairy cattle improvement in the U.S.A. Southern Cooperative Series. Bull. 114 (40 p.)

Guha, H. 1973. Annual technical report of the UNDP Project since merged with the All India Co-ordinated Research Project on cattle at Haringhata for the year 1972-73. West Bengal, Directorate of Animal Husbandry, 16 p. (Mimeograph)

Hayman, R.H. 1974. The development of the Australian Milking Zebu. Wld Anim. Rev. (FAO), 11.

Lecky, T.P. 1949. The Hope Jerseys. A Study of the breeding of Jersey cattle at Hope Agricultural Station, Jamaica. Jamaica, Department of Agriculture. Bulletin No. 42.62 p.

Littlewood, R.W. 1933. Crossbreeding for milk. Indian J. Vet. Sci., 7: 263–272.

Mahadevan, P. 1966. Breeding for milk production in tropical cattle. Tech. Commun. Commonw. Bur. Anim. Breed., Genet., Edinburgh. Farnham Royal, Bucks., England, Commonwealth Agricultural Bureaux. 154 p.

Mahadevan, P. 1970. A note on the maintenance of heterosis in crossbred cattle in the coconut triangle of Ceylon. Ceylon Cocon. Quart., 21: 92–93.

Mahadevan, P. & Hutchison, H.G. 1964. The performance of crosses of Bos taurus and Bos indicus cattle for milk production in the coastal region of Tanganyika. Anim. Prod., 6: 331–336

Mahadevan, P., Wellington, K.E. & Roache, K.L. 1970. An evaluation of Jamaica Hope bulls. J. agric. Sci Camb., 74: 473–476.

Meyn, K. & Wilkins, J. 1974. Breeding for milk in Kenya, with particular reference to the Sahiwal stud. Wld Anim. Rev. (FAO), 11.

Sinha, B.N. 1951. Taylor cows of Patna. Indian vet. J., 27: 272–276.

Wijeratne, W.V.S. 1970. Crossbreeding Sinhala cattle with Jersey and Friesian in Ceylon. Anim. Prod., 12: 473–483.

Previous Page Top of Page Next Page