Previous Page Table of Contents Next Page

A simple scheme to utilize heterosis in tropical dairy cattle

The need for cross-breeding
The Brazilian cross-breeding trial
Superiority of the F1 cross
Cross-breeding strategies
The F1 replacement scheme (F1R)

F.E. Madalena

At the time of writing, the author was Senior Officer, Animal Breeding and Genetic Resources, Animal Production and Health Division, FAO, Rome, Italy His present address is EPAMIG, Escola de Veterinaria-UFMG, Cx P. 567, 30000 Belo Horizonte-MG, Brazil.

The need for cross-breeding

Tropical dairy production systems are characterized by low animal performance. This is not only the result of the direct effects of climate, but also, and more important, it is caused by poor roughage quality, low concentrate feed inputs and the high incidence of disease and parasites. Poor husbandry practices are generally a reflection of underlying socio-economic factors such as the low level of farmers' education and the lack of supporting research, extension and animal health services. Low labour costs, capital scarcity and limited infrastructure, including roads, communications, electricity and marketing, also influence the feasibility of adopting advanced technologies.

To be economically efficient, dairy cattle must survive, reproduce, milk and grow in the environment determined by the ecological and socio-economic factors described. Local tropical breeds are adapted to the environmental stresses but have very low milk yield, while the specialized temperate breeds, despite their high genetic potential to milk in the intensive systems in which they were selected, cannot withstand the harsh tropical conditions, to the point of not even being able to sustain their numbers (de Vaccaro, 1990). Major changes would be required, both at the farmer and industry levels, to improve management and technical services extended to farmers in order to make intensive systems based on temperate breeds profitable. In spite of the Israeli example of rapid intensification (Volcani, 1973), it would appear that under most circumstances levels of cattle nutrition, health and reproduction may not be substantially increased in the short term because of economic and organizational constraints. Under these conditions, cross-breeding of high-yielding temperate breeds with the adapted local tropical breeds results in higher productivity than the breeding of pure-bred cattle (McDowell, 1972; Cunningham and Syrstad, 1987).

The purpose of this article is to present a simple, practical cross-breeding scheme based on the results of research conducted in Brazil, which are briefly summarized below.

The Brazilian cross-breeding trial

A trial was initiated in 1975 by the National Dairy Cattle Research Centre in Brazil with assistance from FAO and the United Nations Development Programme (UNDP) to establish a breeding strategy for the southeast region, which includes the states of Minas Gerais, São Paulo, Rio de Janeiro and Espírito Santo. At the time, the region's 7.5 million cows produced 5.3 tonnes of milk, and the average annual milk production per cow was 666 kg.

Husbandry practices were extremely varied, from intensive systems based on specialized breeds to- dairy ranching of Zebus. Nonetheless, European (mainly Holstein-Friesian) x Zebu (mainly Gir and Guzera) crosses predominated in the more specialized dairying areas. Gir and Guzera are Brazilian breeds developed from Bos indicus imported from India. The large cross-bred cattle population had been kept intermediate between both breed types for the last 60 years or so, but, since there was no firm experimental evidence on which to base breeding recommendations and considering the economic and social importance of the industry, a trial was set up to compare the following cross-breeding strategies:

grading up to Holstein-Friesian (HF);
forming a new breed from an HF x Zebu foundation;
criss-crossing HF x Zebu, i.e. alternating sires of each breeding every generation;
modified criss-crossing, i.e. repeating the HF - sire breed for two generations followed by one generation of Zebu sires.

This last procedure was expected to be more suitable for the more intensive systems as it would maintain the herd at a higher HF fraction than criss-crossing while retaining heterozygosity (Madalena, 1981).

To evaluate these strategies, batches of contemporary heifers of six red and white HF x Guzera (Gu) cross-bred groups were generated and distributed to two experimental stations and 65 cooperating farms, where their performance was recorded under each farm's management system. Farms of varying management levels were chosen, and for analytical purposes they were grouped into two classes: "high" and "low" management. The "high" class included mostly experimental station results. On all farms, the local practice of milking with the calf at foot was followed.

The six cross-bred groups were:

First back-cross to Gu (1/4 HF fraction)
F1 or first cross of HF x Gu (1/2 HF fraction)
5/8 HF sires and dams mated inter se
First back-cross to HF (7/8 HF fraction)
Second back-cross to HF (7/8 HF fraction)
Fifth or higher back-cross to HF (>31/32 HF fraction)

The sixth group represented the HF absorption alternative, the third group represented the new breed alternative and the other groups represented the crisscrossing schemes.

A total of 527 heifers were distributed. The experiment is still ongoing, since total lifetime performance is being measured. However, an analysis was made of accumulated performance based on data available for the first eight years. Madalena (1989) gave a fuller description of the background of the experiment and its procedures.

Superiority of the F1 cross

The main results of the experiment were presented in detail elsewhere (Madalena et al., 1990a and 1990b) and will be summarized here in graphical form only. Figure 1 shows results for milk fat and protein yield per day of calving interval. Heterosis had a very important effect on performance, as reflected in the superiority of the F1 cross. Performance declined in back-crosses as the HF gene fraction departed from 1/2 for grades both higher and lower than the F1. The inter se 5/8 cross also had very poor milking performance.

It should be mentioned that all lactations were included in the analyses, irrespective of their being short or "abnormal", as both conditions were influenced by the cross-bred group, and the deletion of those lactations resulted in biased comparisons (Figure 2). Thus, the F1 superiority for milk yield and related traits was lower in studies where short lactations were deleted or yield was adjusted for lactation length. Milking with or without the calf at foot did not appear to influence the trends in crossbred group effects in the Brazilian results for Guzera and Gir crosses; however, short lactation problems were reported in F1 crosses in some studies in other countries (Madalena, 1986). A word of warning should be expressed, then, before these results are applied generally to all other Zebu breeds or samples.

The importance of heterosis, not only for milk yield, but also for other economically important factors such as lactation length, herd life, mortality rate, price of cull cows, temperament, age and weight at puberty and tick resistance, was substantiated.

Some of the findings on resistance to parasites are presented in Figure 3. The differences between the F1 and the 1/4 HF groups (3/4 Zebu) were small in practical terms. Small differences in heat tolerance and other adaptation traits between pure-bred Zebus and F1 in crosses with British beef breeds were also reported by Frisch (1987). Differences between cross-bred groups were very important for culling and mortality in the "low" management class in Brazil, again favouring the F1, as shown in Figure 4. The mortality rate was low in the "high" management class and there were no differences between the cross-bred groups.

Profit per day of herd life was calculated for each group based on the sale of milk, calves, cull cows and heifers and the expenses incurred for concentrate feeds, animal deaths, milking time, interest and other inputs. Profits calculated from the prevailing prices during the experimental period from 1980 to 1985 are shown in Figure 5. The F1 was the more profitable group under all circumstances on both management levels.

Since heterosis was important for most traits, the F1 superiority increased as more traits were considered, particularly under "low" management. This may be seen by comparing the slopes in Figure 5 with those of Figure 1.

Cross-breeding strategies

The expected outcomes of the four cross-breeding strategies described above were based on the performance of the six experimental groups, to which genetic models were applied (Dickerson, 1973; Eisen, 1989), as illustrated in Figure 1. Details were given by Madalena et al. (1989).

1 Milk fat and protein yield per day of first calving interval of Holstein-Friesian x Guzera crosses - Rendement journalier en matière grasse et en protéines du lait du premier intervalle de vêlage des croisements Holstein-Frisonne x Guzera - Producción de grasa de leche y proteína por día de intervalo entre partos de cruces de Frisón x Guzerá

2 Milk yields based on 42 to 57 first lactations in the "low" management class - Rendement laitier fondé sur 42 à 57 premières lactations au niveau d'exploitation «bas» - Rendimientos de leche basados en las primeras 42 a 57 lactancias en el manejo «bajo»

3 Parasite burdens in six Holstein-Friesian x Guzera grades - Charges parasitaires de six animaux issus de croisement Holstein-Frisonne x Guzera - Cargas parasitarias de seis cruces de Frisón x Guzerá

Holstein-Friesian in the "low" management classes - Holstein-Frisonne dans les niveaux d'exploitation «bas» - Vacas Frisonas en los niveles de manejo «bajo»

Holstein-Friesian in the "high" management classes - Holstein-Frisonne dans les niveaux d'exploitation «haut» - Vacas Frisonas en los niveles de manejo «alto»

4 Culling and mortality rates up to mean age of 6.6 years in the "low" management class based on 37 to 57 females per cross-bred group - Taux de réforme et de mortalité jusqu'à un âge moyen de 6,6 ans au niveau d'exploitation «bas» fondés sur 37 à 57 femelles par groupe issu de croisement - Tasas de descarte y mortalidad hasta la edad media de 6,6 años en el nivel de manejo «bajo» basadas en 37 a 57 hembras por grupo de cruces

5 Expected profit per day from first calving up to the end of recording for different cross-bred groups - Profit prévu par jour du premier vêlage jusqu'à la fin de la consignation des résultats pour différents groupes issus de croisements - Beneficio diario esperado desde el primer parto hasta el final del registro para diferentes cruces

6 The F1 replacement scheme - Schéma de renouvellement des F1 - Esquema de reemplazas con F1

The superiority of the F1 was maintained under a wide range of economic circumstances simulated by varying the cost-price structure. Some results are shown in Table 1.

The results of the trial may be considered in two scenarios.

Prevailing present situation, represented by "low" management:

- The second-best alternative - criss-crossing or HF-Zebu rotation - was well behind the F1 in terms of profit. Criss-crossing requires the selection of at least one bull from each breed, recording the breed of the sire of females retained for breeding so that mating groups may be established for the next generation, and the use of controlled mating groups. This scheme is not suitable for small farms, therefore, unless they are well organized and use artificial insemination or unless the careful and ongoing exchange of breeding material is organized across a group of small farms.

- Developing a new breed (5/8 HF: 3/8 Zebu) also proved inferior to the F1 strategy in terms of profit. This inferiority could be overcome by strong selection, but it would take some 20 to 40 years to achieve. Of course, such a new breed would be easy to use once developed as no special matings would be required.

- Upgrading to HF is not advisable, as has now been demonstrated repeatedly.

Improved management, represented by "high" management:

- Under a regime of low price for fat and no payment for protein, the second-best and equal alternatives were upgrading to HF and HF-HF-Zebu rotation, respectively. The latter was preferable under higher milk component prices.

- The new breed had very poor performance.

Under the present conditions and without any changes in management, therefore, profits from dairying may be increased considerably by using F1 females. This, of course, does not mean that husbandry practices should not be improved, but rather that both the genetic and environmental components of the production system should be matched so that the resources available are used adequately.

The F1 females responded well to improved management, as shown by their superiority in the "high" level farms. This is a very important element in encouraging farmers to improve the management of their farms, as was verified both in Israel and Brazil (M. Soller and E.A. Alves, personal communication). Should management continue to improve, it would be easy to upgrade to HF by back-crossing. Parallel evidence obtained in Brazil showed that pure-bred HF did not differ from their 3/4 and 7/8 crosses with Gir at the very reasonable production level of 10 kg of milk per day of calving interval, which is five times higher than the national average. This indicates that Zebu genes would not be a hindrance until husbandry practices are improved substantially (Madalena, 1989).

The F1 replacement scheme (F1R)

To capitalize on the superiority of the F1 hybrid, a scheme of continuous replacement of F1 heifers needs to be devised. In the proposed scheme, some farms would specialize in providing female replacements to the dairy farms, as in the poultry industry. A focal agency (say, a dairy cooperative or a milk plant) would organize the production of the required F1 heifer calves, for example, by contract mating with a Zebu ranch. Part of the Zebu herd would be bred to Zebu to maintain the pure-bred herd, but most would be bred to Holstein-Friesian to obtain the F1 heifers. Dairy farmers would buy these replacement calves from the cooperative and then dispose of their own (Figure 6).

For the dairy farmer, the increased profit gained from the F1 female should make purchase attractive, provided that costs or the health risk are not unduly increased. The costs of rearing heifers on ranches should not be higher than those on dairy farms.

For the Zebu rancher, the sale of dairy females should be an added value. The proportion of the Zebu herd that could be used for crossing would depend on the reproduction and survival rates obtained. Under good management, 60 percent of the herd could be crossed with HF and 40 percent would have to be mated with Zebu bulls in order to perpetuate the Zebu herd (see Appendix); however, under normal management only 30 percent would be crossed. Consequently, the scheme would only be attractive to ranches with efficient management and where artificial insemination is used. Unfortunately these are not common, but not many would be required. Six hundred Zebu females could support a population of 1000 F1 dairy cows. Restricting the production of heifers to only a few units would facilitate health monitoring by the organizing agency. Under reasonable management, the growth and feed conversion of F1 males would be superior to that of Zebus, providing added benefits to the ranch (Paiva et al., 1992).

From a national perspective, benefits would accrue at all levels. There is a risk that crossing could extend beyond the proportions required to sustain the Zebu herd, however, and adequate precautions should be taken to avoid this danger.


The above results are in line with the general interaction between production systems and genotypes found in other species (Moav, Soller and Hulata, 1976).

Since tropical dairy production should not be based on either pure-bred Zebus or temperate breeds, a feasible cross-breeding programme is required. In Brazil, small farmers typically keep their herds intermediate by periodically switching bull breeds, which often results in high proportions of low-producing animals with too much or too little Zebu breeding. Rotational crossing is not a practical option for small farmers. However, although developing a new breed is not a short-term proposition, it should not be discarded. Yet another scheme has been proposed in which F1 bulls would be produced for natural mating using imported semen. This scheme, however, has not yet been tested. At least half the heterosis would be lost using this system, and there are indications that the selection for high yield may decrease reproductive efficiency in temperate countries and the herd life of female offspring under tropical conditions (Abubakar, McDowell and van Vleck, 1987).

1/4 Holstein-Friesian x 3/4 Guzera in the "high" and "low" management classes - 1/4 Holstein-Frisonne x 3/4 Guzera dans les niveaux d'exploitation «haut» et «bas» - 1/4 Frisón x 3/4 Guzerá en clases de manejo «alta» y «baja» (A)

1/4 Holstein-Friesian x 3/4 Guzera in the "high" and "low" management classes - 1/4 Holstein-Frisonne x 3/4 Guzera dans les niveaux d'exploitation «haut» et «bas» - 1/4 Frisón x 3/4 Guzerá en clases de manejo «alta» y «baja» (B)

5/8 Holstein-Friesian x 3/8 Guzera inter se under "high" management - 5/8 Holstein-Frisonne x 3/8 Guzera inter se dans le niveau d'exploitation «haut» - Inter se 5/8 Frisón x 3/8 Guzerá en manejo «alto»

1 Profit per cow per day of herd life under alternative strategies of cross-breeding of Holstein-Friesian x Zebu - Profit par vache et par jour de vie dans le troupeau selon diverses stratégies de croisement d'Holstein-Frisonne x zébu - Ganacia por vaca, por día de vida en el rebaño, bajo varias alternativas de cruzamiento entre Holstein-Frisón x Cebú

Management level



Characteristics of management levels

Mean first lactation milk yield1 (kg)



Mean first lactation length (days)



Mean first calving interval (days)



Concentrates fed (kg/cow/days)



Cost-price situation









F1 profit (kg/d)2









F1 performance (percentage)


HF-HF-Z rotation









HF-Z rotation









Upgrading to HF









New breed (5/8 HF: 3/8 Z)









A: 1980 to 1985 prices (protein not paid for).
B: Fat differential tripled and protein paid for at the same rate.
C: Cost of concentrates halved.
D: Beef value of animals doubled.
HF = Holstein-Friesian.
Z = Zebu.
1 All milked twice with calf suckling stimulus.
2 Profit expressed per kg of milk (3.3% fat) = US$0.16.
Source: Madalena et al., 1989.

F1 Holstein-Friesian x Guzera under "high" management - F1 Holstein-Frisonne x Guzera dans les niveaux d'exploitation «haut» - Cruces F1 de Frisón x Guzerá en manejos «alto»

F1 Holstein-Friesian x Guzera under "low" management - F1 Holstein-Frisonne x Guzera dans les niveaux d'exploitation «bas» - Cruces F1 de Frisón x Guzerá en manejos «bajo»

Considering the difficulties of the alternative crossbreeding schemes, the F1R scheme appears to be a feasible option, not only because of its known high profit, but also because it would be simple to apply. In fact, it is already being applied in Brazil on a limited scale. McDowell (1985) suggested that breeding programmes be established to retain the merits of the first cross.

The F1R scheme was viewed in the past with some misgivings (e.g. Madalena, 1981), but several new developments give cause to reconsider its merits: the proportion of Zebu cows needed is smaller than what was estimated earlier because of the high reproduction rate that is possible on well-managed ranches and because of the long herd life of both the Zebu and F1 cows; it is possible to monitor health in the few ranches needed to produce replacements; and the transport costs are apparently low (US$10 per heifer for typical distances in Brazil). F1 superiority appears to be greater than previously thought.

The available experimental results consistently justify the F1R scheme, and its practical aspects should be documented in pilot projects on a commercial scale. It is true that the proposed scheme does require some organizational effort, but this is well within a realistic range for any development action.


The proportion of the Zebu herd that can be used for crossing is calculated as follows:

H = number of Zebu cows bred to Holstein-Friesian

Z = number of Zebu cows bred to Zebu

h = number of first calving heifers per cow in the herd, given by the product of calving rate x sex ratio x survival rate x 1 - culling rate

r = replacement rate or proportion of heifers needed annually to replace each cow removed from the herd, = 1/(average herd life).

In order to maintain the Zebu herd numbers, there is a yearly requirement of (H+Z)r Zebu heifers, which are obtained from the Z Zebu cows. Therefore, (H+Z)r = Zh, and the proportion of Zebu cows in the herd is Z/(H+Z) = r/h. Assuming h = 0.38 and herd life is 6.8 years (Oliveira Filho, Abreu and Bianchini Sobrinho, 1991), r/h = 0.3868 of the herd must be mated with Zebu and the remaining 0.6132 may be used for crossing.

The herd life of the F1 is estimated to be 7.2 years (A.M. Lemos, personal communication); therefore, its replacement rate is 1/7.2 = 0.1389. A population of 1000 F1 cows then would require 139 replacement heifers per annum, out of H = 139/0.38 = 366 Zebu cows. Consequently, Z = 366 x 0.3868/0.6132 = 231 Zebu cows are to be bred to Zebu sires to obtain Zh = 88 Zebu replacement heifers per annum. The total Zebu herd would have 597 = 366 + 231 cows (Figure 6).


Abubakar, B.Y., McDowell, R.E., & van Vleck, L.D. 1987. Interaction of genotype and environment for breeding efficiency and milk production of Holsteins in Mexico and Colombia. Trop. Agric. (Trinidad), 64: 17-22.

Cunningham, E.P. & Syrstad, O. 1987. Crossbreeding Bos indicus and Box taurus for milk production in the tropics. FAO Animal Production and Health Paper No. 68. Rome, FAO.

Dickerson, G.E. 1973. Inbreeding and heterosis in animals. In Proceedings of Animal Breeding and Genetics Symposium in honour of Dr J.L. Lush, p. 54-77. Champaign, IL, USA, American Society of Animal Science and American Dairy Science Association.

Eisen, E.J. 1989. Genetic models to predict crossbred performance: review. Revista Brasileira de Genetica, 12 (Suppl.): 13-26.

Frisch, J.E. 1987. Physiological reasons for heterosis in growth of Bos indicus x Bos taurus. J. Agric. Sci. (Camb.), 109: 213230.

Lemos, A.M., Teodoro, R.L., Oliveira, G.P. & Madalena, F.E. 1985. Comparative performance of six Holstein-Friesian x Guzera grades in Brazil. 3. Burdens of Boophilus microplus under field conditions. Anim. Prod, 41: 187-191.

Madalena, F.E. 1981. Crossbreeding strategies for dairy cattle in Brazil. World Anim Rev, 38: 23-30.

Madalena, F.E. 1986. Economic evaluation of breeding objectives for milk and beef production in tropical environments. In Proceedings of 3rd World Congress on Genetics Applied to Livestock Production, 9: 33-43.

Madalena, F.E. 1989. Cattle breed resource utilization for dairy production in Brazil. Revista Brasileira de Genetica, 12 (Suppl.): 183-220.

Madalena, F.E., Lemos, A.M., Teodoro, R.L. & Barbosa, R.T. 1989. Effect of cost-price structure on the economic evaluation of alternative dairy cattle crossbreeding schemes Revista Brasileira de Genetica, 12: 887-893.

Madalena, F.E., Lemos, A.M., Teodoro, R.L., Barbosa, R.T. & Monteiro, J.B.N. 1990a. Dairy production and reproduction in Holstein-Friesian and Guzera crosses. J. Dairy Sci, 73: 1872- 1886.

Madalena, F.E., Teodoro, R.L., Lemos, A.M., Monteiro, J.B.N. & Barbosa, R.T. 1990b. Evaluation of strategies for crossbreeding of dairy cattle in Brazil. J. Dairy Sci, 73: 1887-1901.

Madalena, F.E., Lemos, A.M. & Teodoro, R.L. 1991. Effects of reducing the variation in lactation length on the evaluation of dairy cattle breeds and crosses. Revista Brasileira de Genetica.

McDowell, R.E. 1972. Improvement of livestock production in warm climates. San Francisco, CA, USA, Freeman.

McDowell, R.E. 1985. Crossbreeding in tropical areas with emphasis on milk, health and fitness. J. Dairy Sci., 68: 24182435.

Moav, R., Soller, M. & Hulata, G. 1976. Genetic aspects of the transition from traditional to modem fish farming. Theor. Appl Genet., 47: 285-290

Oliveira, G.P. 1986. Avaliação da susceptibilidade ao berne (Dermatobia hominis L. Jr. 1781) em novilhas HVB:Guzera, em condições naturais. Anais do XIII Congresso Brasileiro de Zoologia, p. 90. (Abstr.).

Oliveira Filho, E.B., Abreu, U.G.P. & Bianchini Sobrinho, E. 1991. Avaliação genética quantitative do intervalo entre partos em um rebanho Nelore, variedade pele rose. Anais da 28a. Reunião Anual da Sociedade Brasileira de Zootecnia.

Paloschi, C.G. 1981. Carga média, relações fêmeas -machos e número de ovos por fêmeas em populaçôs de Cooperia sp., em bezerros Holandes e mestiços de Holandes x Guzera. Universidade Federal de Rio do Janeiro. 91 pp. (M.Sc. thesis)

Paiva, J.A.J., Madalena, F.E., Teodoro, R.L. & Campos, A.T. 1992. Food conversion efficiency in six groups of Holstein-Friesian/Zebu crosses. Livest. Prod Sci, 30: 213222.

Vaccaro, L.P. 1990. Survival of European dairy breeds and their crosses with Zebus in the tropics. Anim. Breed. Abstr., 58: 475-494.

Volcani, R. 1973. Alta producción lechera en condiciones naturales desfavorables. Revista Mundial de Zootecnia, 8: 28-33.

Previous Page Top of Page Next Page