Genetic improvement: selection and crossing

Contents - Previous - Next

Selection-crossing schemes developed for intensive breeding on a commercial scale in temperate climates are not necessarily the most appropriate for backyard rabbitries, having been designed for a quite different purpose.

A small unit may have from 10 to 60 breeding does. Local populations and populations bred locally from imported sires and dams of different breeds should be used to upgrade the stock.

Effective genetic improvement in this context should be a group effort with scientific and technical support from the country's research and development organizations. The improvement programme could focus on a village or preferably a group of villages, or on all the rabbitries in a province or the whole country.

Genetic improvement demands technical specialization. There should therefore be breeder-selectors with larger units of 100 to 200 does who work together to coordinate the selection methodology with the support of research and development services. The selectors should be excellent breeders themselves, using production systems that are well adapted to local conditions and resources. Sophisticated equipment should not be used. The point is that selective breeding should be conducted in conditions in line with those of the best commercial rabbitries in the country. Health care and sanitation, in particular, must be exemplary.

A selection unit must be effective on 2 levels: breeding and production. The extra costs entailed by the technical side of the selection work should be borne by the group of breeders benefiting from the genetic improvement. The cost of research devoted to a genetic improvement programme for the whole country should be shared by a larger group.

For developing countries interested in developing rabbit production, there are two ways to organize selection. It can be done by government organizations or by breeders using a programme devised by government organizations.

The first system is used in Mexico. Selection is carried out by the Dirección General de Avicultura y Especies Menores, which has cooperated with INRA in France since 1976. This method is not incompatible with the second which can be promoted at the same time if interested breeder-selectors can be found.

Group research agencies working on selection should first determine whether the selection methods being used are truly effective, and create new genetic material to upgrade national production. They should then devise strategies for the best use of local and imported animals, and make comparative studies of the different breeds and crossbreeding experiments.

The object of selection is to upgrade performance by enhancing an animal's genetic value where breeding and feeding techniques permit expression of genetic value. In fact, breeding and feeding techniques must be improved at the same time as the genetic value. Selection and crossbreeding should increase the annual output per doe and speed the growth rate for earlier slaughter and better carcass and meat quality.

Crossing is a supplementary benefit to intrapopulation selection. But genetic progress from crossbreeding is not cumulative from one generation to the next as is progress from selection, except where selection is used to improve crossing.


A major objective of selection is to improve annual numerical productivity per doe. This global character depends on the breeder, the animal and the environment. The breeder establishes the theoretical reproduction rate of the does. For backyard rabbitries, it is assumed that weaning takes place at 42 days, servicing at 24 days after kindling, and the average conception rate is 70 percent. This gives an average of 6 litters per doe per year.

A culled doe is immediately replaced by a young doe ready for mating. If the stock renewal rate is 100 percent per year, the annual number of litters per doe will be roughly 5.5. If an average 6 young per litter are weaned and 5.5 reach slaughter or reproduction age, the objective is then 30 rabbits per doe annually.

This modest goal is realistic for backyard rabbitries not based exclusively on pelleted feed. If necessary the weaning age can be extended by delaying presentation of the doe for servicing beyond day 24. The theoretical reproduction rate can be stepped up if the goal is too easily achieved or too modest in terms of the potential of the stock and environment. The doe could be brought for servicing at day 17 after kindling with weaning at 35 or 42 days. This would give an additional litter per doe, raising the annual goal to 35 rabbits per doe. A more intensive breeding objective could produce 40-50. For many countries, however, this would not be a realistic goal.

Whatever reproduction rate is adopted, it is important to have fertile does which accept the buck and can produce many large litters with good kindling-to-weaning survival rates. This implies a whole range of characters: acceptance of the buck, gestation, fertility, viability of young, milk production and longevity. These characters and performances can be summed up by the selection criterion: average number of weaned per litter from the first 3 litters obtained within a predetermined period. There is a close correlation between performance during the first 3 litters and the doe's total output. In practice, the following principle could be followed:

As weaning age is variable, the number of rabbits weaned can be calculated on litter size at 28 days so the doe's genetic value can be estimated more rapidly.

Chapter 9 describes an even simpler system of choosing breeding stock. which can he done directly in the rabbitry.

The other group of characters for selection have to do with weight gain. One selection criterion is average daily weight gain from weaning to slaughter age, say at day 70. The difference between individual weight at day 70 and individual weight at weaning is divided by the number of days elapsed between these 2 dates. The idea is to speed up post weaning growth. There is no need to measure the quantity of feed consumed, except for experimental purposes or to compare genetic types for selection for feed utilization.

It is not easy to measure the quantity of feed or dry matter eaten by the animals. and when they are given different feeds and local forage feed conversion efficiency is difficult to calculate. Speeding up post weaning growth indirectly reduces the amount of dry matter needed for every kg of liveweight gain.

Slaughter yield, carcass quality (meat/bone ratio, fat) and organoleptic qualities of meat arc complex characters to select for because they can only be measured in carefully controlled slaughter conditions. Direct intrapopulation selection for these characters would he unrealistic. Breeders can check sample figures for these characters for the population they arc using and if improvement is necessary crosses can be made with trucks from good meat breeds (described earlier in this chapter).

With stock being used for selection it is necessary to:

All rabbits are identified at weaning when separated from the dam by a numbered ear tag or a number tattooed in the ear. This might be the date of birth plus a day-of-year identity number. Depending on the size of the group an individual number might have 4 or 5 digits (up to 999 or 9 999 births a year) or even 6 if necessary. Another number indicating genetic type (breed or cross) could be added to the animal's cage card. The following characters should be measured:

On doe and litter cards:

These data are recorded on a doe card, listing each service and litter. Where individual weights are recorded (needed in selection for growth rate) another doe-litter card will be filled out for each litter.

On buck cards.

This card will show the number of gestations and prolificacy of does serviced by each buck.

On genealogical cards.

The number of the breeding animal (buck or doe) and the numbers of its sire and dam. These cards will make up the genealogical records.

Four kinds of card can be used for the management of the stock. The number of the breeding animal's cage can be marked on the doe, buck and litter cards for more effective follow-up. These cards are designed for manual or computer processing and for setting up data banks. The cards are used as monthly breeding balance sheets and in selection.

Having defined the selection objectives, the choice of a method is based on:

Selection is made by truncation: in keeping for reproduction the animals with the best additive genetic values (or selection index). The genetic progress expected per unit of time is expressed as:

where t is the generation interval and i is selection intensity. This is the selection differential expressed as standard deviation. The selection differential over the index is the difference between average index values in the selected group and in the population before selection. In a normal distribution, i is expressed as a function of the selected percentage p by

where Z is the ordinate of the normal curve reduced to the truncation point. Numerical tables give values of i for different values of p.

RGG is the correlation between the additive genetic value of the animal and the G estimated by the index.
RGG measures the precision of the index which depends on the heritability of the character and on the selection method.
G is the additive genetic standard deviation of the character selected, which is a datum of the population.

The choice of system aims at maximizing E ?G, by decreasing the generation interval, the percentage selected (ie, by increasing i) by increasing RGG. The objective is an optimum in terms of the breeding characteristics of the species.

Two selection methods for increasing litter size can be compared. The most practical method is mass selection in which does are chosen by an index based on performance. The progeny of these does will produce the bucks and does for herd renewal. The important criteria are numbers born alive and weaned per kindling.

Allowing for the weak heritability of this character, selection should be based on the average of k successive kindlings for each doe. starting with k = 3. The rate of selection can be p = 25 percent for the dams of daughters kept for breeding. An average of 4 daughters are kept from each doe selected. The p might be 10 percent for dams of sons kept for breeding. This means one male offspring from each doe selected is kept. At this rate a generation interval of 10-12 months is practicable.

The index value of an individual animal varies with the number k of litters checked; as k increases the figure becomes more exact. The servicing acceptance of the does and their gestation rates must also be considered, selecting on the basis of an index which takes account of the interval of time E DG between the first and k kindling.

The index is more exact when the group is depicted as families of females. showing the performance of the doe, her full sisters and her paternal half sisters. A more complete index will include repeated performances over successive litters of the doe, her full and half sisters and her dam. The selection index is the multiple linear regression equation of additive genetic value to he estimated on the basis of predictive phenotypes.

At INRA's Animal Genetic Improvement Station in Toulouse a combined selection index is used based on the number of rabbits weaned per litter. The predictive phenotypes are the litter sizes of the doe proposed for selection and of her contemporary sisters and dam. The genetic progress expected from selection on the number weaned per litter assumes that certain parameters of the selected population are known (Table 43).

In these conditions the genetic progress E (DG) expected in litter size from one generation to the next is:

It may he concluded that mass selection that takes into account repeated performances of the does can significantly improve the number of young rabbits weaned per doe ( + 1.5 young over 10 generations). The method is recommended to small-scale breeders.

In theory, progress by mass selection on post-weaning growth rate is swifter than progress on litter size, because the character is more heritable and can be measured in troth sexes. Young males and females can both he selected. Does can he selected on the basis of litter size index and the progeny of these does selected on the basis of post-weaning growth rate. This requires individual weighing of the young and is more difficult to carry out in practice than is selection based on litter size. Only specialized rabbitries can engage in this kind of selection. The breeder who renews his herd on the basis of the best does by litter size and fertility can make a second choice of which young to keep based on their health and weight at a given age.

Table 43.-Genetic parameters of the character

Average number weaned per litter
Phenotypic variance
Phenotypic standard deviation
Additive genetic variance
Additive genetic standard deviation
Correlation between phenotypes of full sisters t = 0.10
Selection rate of mothers to daughters 25%
Selection rate of mothers to sons 10%
Selection intensity of mothers to daughters if = 1.27
Selection intensity of mothers to sons im = 1.75

Here there are different cases to consider: ( 1 ) a rabbitry practicing combined selection based on litter size: (2) mass selection based on the same character for a sizeable number of breeding does in the strain (200); (3) smaller groups.

Case 1: Selection of a strain on the basis of litter size at weaning (INRA, Toulouse). Combined selection, separate generations.

The theoretical plan calls for raising the stock in separate breeding groups, each group constituting a generation. In each generation 196 does are bred with a batch of 42 males. Twenty-five percent of these does are selected according to the results of the first 3 litters, the theoretical reproduction rate permitting a generation interval of 10 months. Each doe selected produces an average of 4 replacement female offspring, so the group is made up of families of full sisters and paternal half sisters.


The mating programme is implemented in accordance with the composition of the breeding groups. Table 44 shows that the females of each of the 14 families are distributed among 14 breeding groups with 3 males ( 1 and 2 alternates) and 14 females. One breeding female is chosen at random per family from among the 196 does.

This mating programme means the genetic value of each doe can be figured according to her performance and those of related females (family average). The plan can also be implemented with fewer than 14 families and 14 breeding groups (eg, 10, or a total of 100 breeding females). The breeding groups system offers the practical advantage of matching a production layout in which families are represented by mother cages. Here, the 14 females' cages and the 3 males' cages are arranged side by side in rows in the rabbitry.

Case 2: Selection of a strain for litter size at weaning at the Irapuato National Rabbit Breeding Centre in Mexico. Mass selection, overlapping generations.

The first system (case 1) using separate generations can be recommended for assessing genetic progress and compiling more exact selection indices, the animals compared being contemporaries. In small rabbitries the numerical productivity of the does is lower and a simultaneous large number of female offspring (future breeders) cannot be supplied by the does selected. In this case the best does supply female offspring for stock renewal over a phased period. The first system, which uses separate breeding groups, does offer the advantage of a sanitary strip between each group and good conditions for experimental studies, but does not make the best use of the mother-cages which are not fully occupied.

For these reasons the Irapuato centre adopted a management system using 15 breeding groups and 15 families with overlapping generations and mass selection. The advantage of this system is that it is easy for breeders to use and does not demand computer processing. The criterion is the number of rabbits weaned per doe and per month of production. Chapter 9 gives a practical guide for stock renewal.

Case 3: Selection in smaller groups.

The breeding groups system can be used with low stock numbers. If there is little or no introduction of outside breeding animals the groups are made up of 2 males and 4-5 females. The families then consist of the young breeding stock produced by these groups. For renewal the young females from one group are distributed among the other groups. The breeding bucks from one group are renewed by does and their progeny selected from their own group. Mother-son matings are avoided in overlapping generations and brother-sister matings in separate generations.

Rouvier (INRA, Toulouse) selected a strain of 60 and then 40 breeding does over 10 generations, using separate generations. At the Irapuato centre in Mexico good meat strains are selected in smaller groups than strains selected for maternal performance, and this is the method used.

Once the breeding animals for herd renewal have been chosen' a mating programme is the next step. Selection may be random, but mating between close relatives such as full brothers-sisters, half-brothers-half-sisters, mother-son or father-daughter must be avoided. In organizing a servicing calendar a practical method is to assign breeding animals to cages according to breeding groups, taking family origin into account. The family is the original breeding group. A breeding group consists of 2-3 cages of sires and l 0-14 cages of dams, or proportionately fewer if the colony is numerically small.


Crossbreeding to take advantage of heterosis and nicking ability requires the selection of several pure populations. The first step is to select local populations and strains derived from imported animals. They are crossbred to test the potential improvement of the local populations.

Another aspect of crossing to be considered is the establishment of synthetic strains.

1. Simple or 2-breed crossing. Females of a local population, or breed A. will be crossed with males of breed C to improve the growth and muscular development of young meat rabbits and for a heterotic effect on the numerical productivity of does. Using this system the breeder can cross the pure breed A with part of his stock (perhaps 20 percent) for self-renewal of the female stock. The other females will be terminally crossed with C males. which can be obtained from another breeder. All the progeny of this cross are destined for the butcher.

2. Two-stage or 3-breed crossing. Breeding animals of 2 populations (A and B) will be crossed to get an AB crossed female for terminal crossing with males of breed C. The first crossing might be between B males of a good breed for size. fertility and maternal performance with females of a local A population. Using this system the breeder must rely on breeders or multipliers for female AB breeding animals and C sires which demands careful timing and organization.

The system can be elaborated by the use of C sires which have themselves been crossed according to a system widely used in poultry breeding.

3. Rotating and alternative crossbreeding. Using several breeds and local populations for improvement, such as A . B and C, the breeder can apply the following system:

Rotating and alternative crossbreeding.

The advantage of this system is that it offers both heterosis and nicking ability, and the breeder can himself produce his female replacement stock; only the male breeding animals need to he acquired elsewhere. When this system is used with only 2 breeds it is called alternative crossbreeding.

Systems 1 and 3, in which the breeder acquires male breeding animals for stock improvement but can select females from his own rabbitry, are well adapted to small-scale production.

The A B crossing produces an F. generation, which is crossed F. x F. to produce F2, followed by F. and so on. A synthetic strain can thus be constituted which enjoys the nicking ability of the two A and B populations and half the initial heterosis. A synthetic strain can also be established from more than 2 breeds. This system is recommended for improving local populations.


Genetic improvement and the implementation of selection plans should be worked out at a regional or national level. Questions of genetic progress and its extension to breeders, and of scientific technical training, should be studied at these levels.

The genetic improvement plan for rabbits in France, for instance, covers the following aspects:

Guidelines and technical training in selection for collective action (performance checks, establishment of a database and computer processing of data, testing of strains, etc) come under the Livestock Division of the Ministry of Agriculture.

The back-up research and development programme for scientific and technical support and the study of production economics is conducted by INRA and ITAVI (Institut technique de l'aviculture et des animaux de basse-cour) working with producers' associations. These associations are assigned one technician for every 5 000 breeding does. ITAVI and l'Enseignement agricole hold training courses for rabbit breeders and experts.

A country wishing to develop backyard rabbitries could begin by:

1. Setting up a system to record technical and economic data in private rabbitries-production costs, sale prices, climate, local natural resources, consumer patterns, handcrafts (utilization of skins), motivation of breeders for raising rabbits.
2. For experimental purposes, making an analysis of breeding techniques, feeding, specific and nonspecific diseases, and a study of breeding behaviour of local populations and imported breeds.
3. Setting up a system to train breeders and provide them with technical back-up through trained extension agents.


Domestic rabbits are not as widespread as the other species of domestic mammals that are traditionally used to provide meat, milk, wool and skins. But rabbits are genetically very flexible, which makes them adaptable and productive in a wide range of production systems.

FIGURE 17.-Genetic progress (selection, performance checks, testing) and extension (multiplication, production) in France (Rouvier, 1981)

FIGURE 1. New Zealand White rabbit

FIGURE 2. Bouscat Giant White rabbit

FIGURE 3. French Lop rabbit

FIGURE 4. Californian rabbit

FIGURE 5. Dutch belted rabbit

FIGURE 6. French Giant Papillon rabbit

FIGURE 7. Vienna Blue rabbit

FIGURE 8. Flemish Giant rabbit

FIGURE 9. Creole (Guadaloupe) rabbits

FIGURE 10. A "family package" of bucks and breeding does supplied by a Mexican programme

FIGURE 11. Wooden hutches with mesh floors arranged in a 2-storey system (Guadaloupe)

FIGURE 12. Open drinkers supplied semi-automatically from a fitted bucket (Guadaloupe)

FIGURE 13. Fattening cages built entirely in wire mesh, placed outdoors in superimposed rows (France)

FIGURE 14. Cages arranged in a plastic greenhouse, protected with a reed lattice (France)

FIGURE 15. Exterior view of the same greenhouse, photographed in winter

FIGURE 16. Italian system for arrangement of fattening cages (Italy)

FIGURE 17. Mesh cages arranged in the Californian system (France)

FIGURE 18. Cages for the collection and transport of rabbits to the abattoir (Hungary)

FIGURE 19. Faeces from rabbits receiving feed with a normal proportion of roughage, slightly deficient in roughage, and deficient but without diarrhoea

Research on rabbit breeding behaviour and production development started only recently-less than 20 years ago-though formal genetic research has a longer history. For production and selection this can be both an advantage and a drawback. The advantage is that there is less temptation for countries to import ready-made solutions without examining their own specific problems. Genetic variability is also equal to the challenge. A relative drawback is that the newness of this field of research makes it necessary to establish the appropriate genetic improvement pattern for each region or country. The major constraint here is dependence on the environment, and a careful study of the production environment is essential.

The great genetic flexibility of the species and its short life cycle are definite assets. This flexibility is a function of a genetic variability which can be traced to the recent domestication of the rabbit and the lack of intensive artificial selection. This has made possible the rapid emergence of breeds varying greatly in adult size and muscle development (adult weight varies at a ratio of 1 to 8). Doe prolificacy depends basically on the breed. In breeds of comparable mature size, average prolificacy is relatively independent of the environment. This character can be turned to good account in deciding how to use local populations.

Various breeds and populations are available for developing or modernizing production. A minimum of environmental factors must be mastered first; for the rest the rabbit will simply adjust to the physical and human constraints of its habitat.

In most developing countries, highly intensive production with complete control over technical and environmental detail is precisely the opposite of what is needed. What is needed are careful country profiles of the production environment (technical studies of local feed resources and genetic and sociological studies) and training for rabbit breeders. For genetic improvement, the first step could be a study of the breeding performance of pure and crossed local populations. Local populations are usually smaller in size and less prolific and appropriate crosses with local and imported breeds for better productivity could be worked out. Existing local populations should always be conserved and selected on an intrapopulation basis so they can be used to improve production in their own environment.

Contents - Previous - Next