PAPERS PRESENTED AT THE EXPERT CONSULTATION (Contd.)

B. MONITORING ANIMAL GENETIC RESOURCES AND CRITERIA FOR PRIORITY ORDER OF ENDANGERED BREEDS

MONITORING ANIMAL GENETIC RESOURCES AND CRITERIA FOR PRIORITIZATION OF BREEDS

K. Maijala¹

1 Introduction

1.1 Need for information and documentation

The importance of collecting information on animal breeds has been stressed in many meetings and papers concerned with the conservation of animal genetic variation (e.g. 3, 7, 8, 9–11, 13, 18, 22). A record of these meetings and recommendations is given by Hodges in an early paper in these proceedings, entitled “Review of past and present activities and prospects for the future”. Details of some information already assembled at the FAO Animal Genetic Data Bank is given in the paper by Ruane in these proceedings; similarly a review of the data in the EAAP/FAO Animal Genetic Data Bank may be found in Simon (25).

1.2 Purposes and kinds of information to be collected

The information should serve decision making for both conservation and for improved use in different production environments. It should be in standard and comparable form making it possible to classify breeds and strains on the basis of the urgency of their conservation, their genetic and phenotypic traits and according to their suitability to given production systems and environments. The Global Animal Genetic Data Banks should serve member countries of FAO by the publication of a World Watch List as an early warning system, which is regularly undated. A critical feature of this list would be the indication of declining or small population sizes.

In earlier discussions because of the large number of breeds and the cost of conservation, the evaluation of breeds was stressed as a first step in the programme. More recently it has been realized, that acquiring sufficient information is so difficult, time-consuming and expensive that it is not possible to evaluate each breed properly before it is lost. The emphasis has shifted to identifying the populations which are threatened and therefore in need of urgent conservation activities. Collections programmes should therefore be restricted initially to the most important and urgent information to permit rapid decisions on the need for conservation.

The different types of information needed for different types of decisions were discussed and evaluated by Maijala (17). There are today several types of questionnaires. The type of information needs further study and discussion which is now addressed in this paper.

¹ Haapatie 13.D., 00780 Helsinki, Finland.

2 Different types of questionnaire and their suitability

The following types of questionnaire have been developed:

2.1 SABRAO

The Society for the Advancement of Breeding Research in Asia and Oceania (SABRAO) developed forms for documentation of performance of the breeds of 11 species in Asia and Australia. Eight of these are summarized in Appendix 1. In addition there were forms for guinea fowls, geese and quails.

The forms are divided into two parts: schedule 1 for a general description of each breed, and schedule 2 for actual performance data collected on a particular flock or herd, or from a particular survey or evaluation study. Hence, the forms are rather extensive and not very suited for preliminary and fast mapping of threatened populations.

2.2 FAO Descriptors

The FAO descriptors were developed in trials organized by FAO and UNEP in Africa, Asia and Latin America in 1983–85 (9–11). They are summarized in Appendix 2. They are intended for extracting data from various published or unpublished source documents and entering into a computer system. These lists, too, are divided into two parts: master records about the physical characteristics of each breed/strain, and slave records about the performance characteristics and environments. Since the forms cover all possible traits of interest and occurrence, they are massive and difficult in use by untrained people for identifying threatened populations in need of swift conservation actions.

2.3 EAAP/FAO Questionnaires

The EAAP/FAO-forms were developed gradually by a working party of the European Association for Animal Production (EAAP), for collecting information of European breeds for which conservation might be needed (18, 20, 25). The forms are summarized in Appendix 3.

In planning the questionnaires, the EAAP working party partly utilized the SABRA forms as models; partly its own experiences from the European surveys in 1982 and 1985; those of the Scandinavian Animal Genetic Resources Working Party and those of FAO/UNEP. It tried to standardize the forms for different species and sought an optimum compromise between many-sided, reliable information and on the other hand maintaining the interest of those supplying the information.

Originally, there were five forms (A-E), but when the Hannover Animal Genetic Data Bank was expanded in cooperation with FAO to provide global coverage forms F and G were added for additional information (25, 14) Buffaloes and chicken were added to the species covered. In spite of the attempts to standardize and simplify the forms, they have continued to cause problems of completion, even in Europe. Further simplification may be needed for identifying the breeds needing urgent conservation actions (17). The task of completing the forms can be eased by pre-filling them with the information from previous surveys. In this way only checking is needed for the part of the form.

Care has been taken to provide clear instructions for completing the forms. The use of scientific or popular terms is clarified and those completing the forms are urged to provide only objective information. The language problem has been solved by using only English. This is important for the using the information, although those completing the forms would sometimes prefer their own language for speed. In some parts of the world it may be necessary to accept some other important language simply to get the information.

2.4 Shortened FAO-form

The shortened FAO form (23), summarized in Appendix 3, was developed for collecting central information of non-European and Russian breeds and of all buffalo breeds, listed by Mason (21). It ia intended for fast identification of breeds needing immediate conservation action. The intention is to send them out pre-filled with the basic information of breeds names etc. from Mason's World Dictionary of Livestock Breeds (21).

2.5 Selected Parts of EAAP/FAO Questionnaires

An alternative for the shortened FAO forms is to use forms A and B of he EAAP/FAO questionnaires. The former could be pre-filled from Mason's book. Answers should be sought only for paragraph 4 (breeding population numbers). In addition, opportunity should be found for the supply brief information on the outstanding performance traits and special environmental adaptations (Form D, paragraph 4) qualifications.

2.6 Recommendations

The SABRA forms and the FAO descriptors are obviously too massive for getting enough basic information rapidly on a large number of breeds for assessment of threatened status, especially in developing countries. Even the EAAP/FAO forms are too complicated and difficult to fill in for this purpose; but some parts of them (Form A, Form B, para.4, Form D, para.4) could be used, as well as the shortened FAO-forms (23). Experts involved in these matters operationally will have to discussion and decide which to use or whether a new alternative should be developed for the preliminary listing.

In the second round, the complete EAAP/FAO questionnaires might be appropriate in many cases and even the FAO descriptors or SABRA forms could be used for collecting information from different sources and evaluation studies. In the later rounds of surveys, the proportion of populations for which these forms are appropriate will be increased.

3 Suggestions as to the frequency of survey

3.1 Frequency for the EAAP/FAO questionnaire

According to the European experiences, attempts to collect information simultaneously on population sizes and various other characteristics impairs the frequency, completeness and speed of replies. Hence, there is reason to consider separating regular censuses of population sizes from surveys on characteristics of the breeds.

3.1.1 Shortened Forms for Census Data

In view of the urgency of getting basic information for possible conservation actions, it is well-founded strategy to carry out surveys of a shortened form as soon as possible in all developing countries and in those industrial countries in which no survey has been made before. Some information about performance and the existence of specially interesting genetic traits could also be requested. The part of the form concerning animal numbers and the recent trends of these and crossbreeding should be repeated at 3-year intervals. Where numerically minor breeds or breeds being extensively crossed with males of other breeds exist, even more frequent surveys would be desirable. However, advice should be taken from local experts on whether this is realistic. In the follow-up surveys, attempts should be made to use the more complete forms.

3.1.2 Complete EAAP/FAO Questionnaires

On the subject of repetition of information the complete questionnaires can be divided into two classes:

• The origin of the breed, breed description, qualification of breed, management conditions, environment and socio-management system need not be repeated frequently. On average 10-year intervals suffice, but the contact persons should be encouraged to check and up-date any item, in which essential changes have taken place.

• Some of the performance records and additional information deserve to be checked and up-dated as soon as essential new information becomes available, but at least every fifth year. It would be sensible to repeat the surveys at shorter intervals for species with high reproduction rate and short generation interval. However, maintaining the cooperation of people and organizations responsible for replying to the questionnaires has to be considered. Hence, the question deserves to be discussed in an expert group or to be solved in different ways in different regions and for different species.

3.2 Frequency for Different Regions

It is not possible to give fixed repetition schedules for the frequency of different surveys for different regions or country types, since the situations vary. The situation of many breeds in many developing countries would speak in favour of more frequently surveys than in developed countries, but the difficulty of getting the data may make this unrealistic. On the other hand, the efficient application of A.I. and other techniques in the developed countries can accelerate breed changes increasing the need for frequent up-dating. Thus, it would be desirable to have a competent contact person or organization in each country for assessing the needs of up-dating the surveys for different species and breeds. In this case, relevant information of essential changes could be requested annually from each country. As an example of such a need the change in the breeding policy of the Chinese pig breed Taihu may be cited. Extensive crossbreeding has begun recently and could make it a threatened breed, in spite of large numbers.

Possibilities of getting data on breed performance in developing countries are probably less than in developed countries. However, it may be possible to find positive, encouraging exceptions. On the other hand, getting information from developed countries has become increasingly difficult, due to high labour costs.

3.3 Frequency for Different Species

It is also difficult to give clear rules for different species. In principle, threatened breeds and rapidly reproducing species need more frequent monitoring. However, extensive use of A.I. and crossbreeding can make frequent follow-up important for at least part of the population. The appropriate frequency depends also on the policies of governments and of local organizations working on the species in question.

4 Useful criteria to assess the threat to breeds

4.1 Need for Criteria

A choice for conservation is often necessary because of the great number of breeds in some areas, and of the limited resources available. This calls for making clear the criteria for choosing breeds for conservation. Such criteria have been discussed by several authors (e.g. 26, 2, 18, 24, 16, 17, 4).

4.2 Population Size

Population size and its trends are decisive factors in determining the vulnerability of a given population and the need and kind of special conservation actions. Population size can be expressed as:

1. total number of individuals of a breed alive at a time,
2. total number of animals of breeding age,
3. effective population size Ne = 4NmNf/(Nm+Nf), where Nm = no. of breeding males and Nf = no. of breeding females (30).

Alternative (c) makes it possible to compare the effects of various population sizes and structures on changes in gene frequencies. With Nm = Nf and random matings, Ne = N (=alternative B above).

Alternative (c) is preferable, since Ne largely determines the levels of random genetic drift and inbreeding in a breed both before and after conservation. Heterozygosity per generation starts to decrease at an accelerating rate after Ne falls below 100. Values below 50 decrease the heterozygosity by more than 1 %/generation. These changes were discussed by Maijala (17) and also illustrated by a figure. It has been calculated that:

• the S.D. of gene frequency due to chance increases with accelerating rate after Ne falls below 30 or 40 (17),
• ca.18, 10, 4, 1.6, and 0.8 % of genetic diversity is lost in 10 generations, when Ne = 25, 50, 125, 250 and 500, resp. (15),
• the probability of transferring both alleles at a 2-allelic locus, from the base population to the progeny, decreases with accelerating rate, when the number of founders goes under 30, (27),
• the loss of alleles at a many-allelic locus can be considerable, when the number of founders is less than 20 (27).

4.2.1 Sex ratio

Ne is also important for understanding the effects of varying numbers of males and females on genetic drift variance and inbreeding. Hence, it is important to know the numbers of breeding animals of each sex. For example, 4 males + 4 females gives the same Ne as 2 males + 100 females. Even 1000 females give the same result, so the number of males is very decisive. In fact, with large values of Nf (e.g. > 1000) Ne = approximately 4 Nm and is thus simple to calculate.

Assuming there is no variation in family size in a population, Ne is linearly dependent on the Nf/Nm ratio and can thus be simply estimated from this ratio. The following regressions can be used:

The Ne for each sex ratio (1st row) can be obtained by multiplying the Nf by the corresponding figures of the 2nd row.

A table giving details of these relationships is given by Maijala (17). Curves for Ne-values can easily be plotted and are available from the author.

4.2.2 Recent Trends in Population Size

Changes in population size can be caused by a general decrease or increase of demand for a certain product or by crossbreeding with a breed considered more competitive. The latter cause has increased in importance, because of the generality of A.I., frozen semen and embryos, of improved communications and more effective marketing of genetic material. The changes can be very abrupt and can threat also reasonably large breed populations. They are important to be estimated in assessing degrees of endangerment. Turner (29) has calculated the effects of various annual rates of decreases in population size on the total decreases in different time periods. She calculates, for example, that annual decreases of 5% and 10% lead to total decreases of ca. 18% and 33% in 5 years, respectively. Her paper gives a figure illustrating these trends.

Turner (29) also estimated the percentage decline in the number of females after 5 years in a breed, with varying proportions of females crossed with males of another breed. She shows that the initial number of purebred females can be reduced to 20% in five years, if all females are used for crossbreeding and there are five age groups of females. This can easily happen, when an exotic breed becomes popular in a country with effective A.I. service. In case half-bred males of the foreign breed are used to all females, ca. 87% of the females after five years would have at least ¾ of their genes from the original breed, but the corresponding figure after 10 years would be only 13%. These figures are also illustrated in a figure by Turner (29).

Thus, the recent trends of population can be used for predicting the population size in the future and can be very important in assessing the vulnerability. In the European surveys, the trends were estimated as three alternatives of direction in the last three years, but the speed of change was not recorded (17). It would be desirable to have a crude estimate of the speed, in order to adjust the Ne accordingly for risk assessment. In Germany, a breed is classified as endangered, if more than 10% of matings are performed with males of another breed (6).

4.2.3 Variations in family size

In practice, there is considerable variation in fertility and length of life, causing variations in family size and increasing the numbers of males and females needed for a given Ne. Accordingly, the numbers below which a population should be considered threatened are increased. Because there are difficulties in getting exactly enough numbers of breeding animals, it is not sensible to give exact values of Ne for different values of the standard deviation of family size. However, it pays to have some idea about this variation and to apply larger values of Ne in assessing the vulnerability of a breed, including cases with known artificial selection.

4.3 Number of Herds

When animals are located only in a few herds or flocks, there is a risk for that accidents, disease outbreaks, disposal of the herd for economic, health or age reasons, etc. will decrease the population size to a dangerous level. In Germany, less than 10 herds is considered to place a breed in a threatened state (6). The risk of dispersing a herd has increased due to over-production and economic depression, and may also affect big herds or flocks, even though small units are at greater risk. The critical number of herds depends on local economic circumstances and has to be assessed independently in different countries.

4.4 Existence of Cryogenic Stores of Genetic Material

The need and urgency of special conservation actions depends on whether frozen semen and/or embryos have already been stored and to what extent (19). In most cases nothing has been done, and in some cases the stores are too small and do not contain large enough numbers of males or matings. According to Smith (28) the numbers needed for securing 98% of the genetic variation would be semen from 25 unrelated sires or by 25 parental pairs with frozen embryos. Some increases in these amounts would be desirable for safety reasons. The number of males or matings stored can be added to the number of breeding males in estimating Ne for assessing vulnerability. In the case of low numbers of females, the number of matings can be added to the number of breeding females.

4.5 Criteria for Different Species

Assuming the above criteria concerning population size are applied, similar criteria can generally be used for different species. The relative importance of trends, variations in family size and number of herds may be somewhat different with small and large animals and should be considered separately by experts of the species in question. In poultry, the availability and utilization of hatching machines have to be considered. They increase the risk of extinction and thus the level of Ne needed for not being considered threatened.

4.6 Criteria for Different Global Regions

The use of Ne, adjusted by the various factors described above, diminishes the need for finding different criteria for different global regions. However, because of the greater uncertainties in the developing countries, the levels of threat (see below) have to be set differently from those in the developed countries. Regional and national expertise is required for considering the specific criteria.

4.7 Type of Information needed/desired

On the basis of paragraphs the above paragraphs, the following kinds of information are necessary:

1. General information: country, species, breed, main location.
2. Numbers of males and females in breeding age.
3. Trends in numbers in last 3 years: decreasing, stable or increasing. Estimates of the percentage annual decrease would be very desirable, if they are possible to obtain.
4. Percentage of females bred pure (or crossed with purebred/half-bred males of an unrelated).
5. Number of males from which frozen semen has been stored.
6. Number of matings, from which embryos have been stored.
7. Number of herds: <5, 5–9, >10
8. Estimate of health and other risks in the herds left.

In addition, it is desirable to get following information:

9. Note about outstanding traits of the breed for possible future utilization.
10. Note about the special environment to which the breed is especially adapted.
11. Note of special products, which the breed can produce.
12. An assessment of endangerment and priority for conservation, based on items a-k and made by a group of local experts.

4.8 Recommendations for Guidelines

It is recommended that the effective population size (Ne) be estimated on the basis of numbers of breeding males (Nm) and females (Nf), by using sex ratio (Nm/Nf) and Nf in small populations, and 4 Nm in case of large numbers of females. The estimates should be adjusted by known trends in population size and extent of crossbreeding (by using Turner's estimates of the effects), and by estimates of variation in family size. In addition, one should consider whether the number of herds is alarmingly small, what kinds of outstanding traits the breed has, and to what kind of special environment it has adapted itself. Finally, an assessment of the degree of threat and priority for conservation should be made by local experts.

5 Proposals on levels of risk

5.1 Earlier classifications of risk status

Previously, different levels of threat, based mainly on the numbers of breeding females, were suggested by some authors or working parties (Table 1).

Table 1. Criteria proposed by some working groups for considering breeds to be threatened. RBST = Rare Breeds Survival Trust, UK,(1): EAAP = Eur. Assn. Anim. Prod.(17): AMBC = American Minor Breed Conservancy (5): Bodó (4).

Notes to table above:

1. * = Ne-values of 0–333 were given for varying sex ratios.

2. RBST required in addition, that:
(a) an official herd book has existed over six generations,
(b) the breed “breeds true to type”
(c) other breeds have contributed less than 20% of the genetic make-up in the last six generations,
(d) the breed has been known for 75 years.

These objectives are mainly cultural-historical, hence the classification does not suit cases where the main motives for conservation are economic-biological. The requirements are especially difficult to fulfil in developing countries, where herd books are rarely kept and where it is difficult to identify even the immediate parents and certainly not six generation of ancestors.

3. The EAAP working party added that steadily decreasing numbers of breeding animals or increasing use of crossbreeding are reasons for increasing the numbers of animals below which a breed is considered threatened. FAO (12) recommended multiplying the EAAP figures by 2 or 3 for developing countries, because of the subdivisions and genetic isolation of nomadic populations, harsh climatic conditions and high risk of disease. It proposed as a working rule that "when a population size approaches 5000 breeding females, the survival risk of the breed should be studied and appropriate actions initiated. Specific recommendations are needed for each circumstance, based upon the local circumstances of the breed, its management system, the extent of crossbreeding, rate of decline in numbers and the certainty of the breed having unique qualities.

4. The AMBC classification is based on the number of registrations per year, and there are three categories. The AMBC system, like the RBST system, cannot be applied in developing countries, since registration of parents is rare.

5. Bodó (4) used five categories (normal, insecure, vulnerable, endangered, critical), of which “endangered” is included in Table 1. He gave also Ne-values for different sex ratios. For the “endangered” category they varied from 4 to 333.

5.2 Suggestion for a classification based on Ne

The RBST and EAAP systems need different criteria levels for different species and try to cover different situations within species. In order to make the classification more systematic, the use of Ne is suggested here. It is proposed that Ne (adjusted by trends in animal numbers, extent of frozen stores and crossbreeding, and variations in family size) be used as the main criterion for setting the threat levels. In addition, too few herds, the existence of outstanding traits and special environmental adaptations should be considered. The suggested levels of threat are presented in Table 2.

Table 2. Suggested levels of risk status based on Ne.

The main category to be found at the first stage E, which could be considered the threshold for inclusion in the Early Warning List and the indication of need for urgent conservation action.

Categories V and R indicate need for a close watch, since the speed of changes in the economic, political and commercial life is accelerating and the use of various techniques of reproduction is expanding. They can also be used for indicating breeds in which some selection could be carried out in addition to preservation. However, the question deserves to be discussed and a search for other alternatives should be continued. Since the number of herds is important, evaluation of the limits in Table 2 is needed to determine whether they are appropriate or if some other values would be more effective.

5.3 Recommendation for risk status assessment

It is suggested that a classification of breeds for threat level should be based on the concept of Ne, adjusted by trends in population size, extent of crossbreeding, frozen stores, and variability of family size. In addition, the number of herds and trends in numbers of herds, outstanding traits and environmental adaptations should be considered in assessing the risk status and conservation priority.

6 References

1. Alderson, G.L.H., 1981. The conservation of animal genetic resources in the United Kingdom. Anim. Prod. Health Paper 24: 53–76.

2. Barker, J.S.F., 1980. Animal genetic resources in Asia and Oceania - the perspective. Proc. SABRA Workshop on Animal Genetic Resources in Asia and Oceania, Tropic. Agric. Res. Center, Japan, Publ. No. 47: 13–19.

3. Barker, J.S.F. & Turner, H. (Eds.), 1980. Proceedings of the SABRA Workshop on Animal Genetic Resources in Asia and Oceania. Tropic.Agric.Res.Center, Japan, Publ.No.47: 555 pp.

4. Bodó, I., 1990. Methods and experiences with in situ preservation of farm animals. FAO Anim.Prod.Health Paper 80: 85–102.

5. Crawford, R.D., 1984. Assessment and conservation of animal genetic resources in Canada. Can. J. Anim. Sci. 64: 235–251.

6. DGZ, 1991. Ausschuss der Deutschen Gesellschaft für Züchtungskunde zur Erhaltung genetischer Vielfalt bei Lantwirtschaftlichen Nutztieren: Empfehlungen zur Erhaltung genetischer Vielfalt bei einheimischen Nutztieren. Züchtungskunde 63: 426–430.

7. FAO, 1981. Animal Genetic Resources Conservation and Management. FAO Anim. Prod. & Health Paper 24: 388 pp.

8. FAO, 1984. Animal Genetic Resources Conservation by Management, Data Banks and Training. FAO Anim. Prod. & Health Paper 44/1: 186 pp.

9. FAO, 1986a. Animal Genetic Resources Data Banks. 1. Computer systems study for regional data banks. FAO Anim.Prod.Health Paper 59/1: 18 pp.

10. FAO, 1986b. Same. 2. Descriptor lists for cattle, buffalo, pigs, sheep and goats. FAO Anim. Prod. Health Paper 59/2: 150 pp.

11. FAO, 1986c. Same. 3. Descriptor lists for poultry. Anim. Prod. Health Paper 59/3: 111 pp.

12. FAO, 1989. Preservation of Animal Genetic Resources. Report of the Tenth COAG Meeting held in April 1989. Document CL/95/9.

13. FAO, 1990a. Animal Genetic Resources. A Global Programme for Sustainable Development. FAO Anim. Prod. Health Paper 80: 300 pp.

14. FAO, 1990b. Manual on Establishment and Operation of Animal Gene Banks. FAO, May 1990: 67 pp.

15. Henson, E.L., 1990. The organization of live animal preservation programmes. FAO Anim. Prod. Health Paper 80: 103–117.

16. Maijala, K., 1986. Possible role of animal gene resource in production, natural environment conservation, human pleasure and recreation. FAO Anim. Prod. Health Paper 66: 205–215.

17. Maijala, K., 1990. Establishment of a World Watch List for Endangered Livestock Breeds. FAO Anim. Prod. Health Paper 80: 167–184.

18. Maijala, K., Cherekaev, A.V., Devillard, J.M., Reklewski, Z., Rognoni, G., Simon, D.L. & Steane, D., 1984. Conservation of animal genetic resources in Europe. Final report of an EAAP Working Party. Livest. Prod. Sci. 11: 3–22.

19. Maijala, K. & Kolstad, N., 1992. Gene banks for livestock conservation. Proc. Int. Conf. Conservation of Biodiversity for Sustainable Development 230–242.

20. Maijala, K. & Simon, D.L., 1987. Plans for European data banks on animal genetic resources. 37th Ann. Meet. EAAP, Lisbon, Oct. 1987: 24 pp.

21. Mason, I.L., 1988. World Dictionary of Livestock Breeds. 3rd Edit., C.A.B. Intern., U.K.: 348 pp.

22. Powell, R.L. & Norman, H.D., 1989. Animal germplasm information systems. In: L.Knudson & A.K. Stone (Eds.) Biotic Diversity and Germplasm Preservation, Global Imperatives (Beltsville Symp. Agric. Res., May 1988), Kluver Acad. Publ.:427–443.

23. Ruane, J., 1991. Summary of work carried out on global database. PM for a meeting of EAAP/FAO Working Party on Global Animal Genetic Resources, Berlin, Sept. 1991: 4 pp.

24. Simon, D.L., 1984. Conservation of animal genetic resources-reviewing the problem. Livest. Prod. Sci. 11: 61–64.

25. Simon, D.L., 1990. The Global Animal Genetic Data Bank. FAO Anim. Prod. Health Paper 80: 153–166.

26. Simon, D. & Schulte-Coerne, H., 1979. Verlust genetischer Alternativen in der Tierzucht - notwendige Konsequenzen. Züchtungskunde 51: 332–342.

27. Sirkkomaa, S., 1983. Calculations on the decrease of genetic variation due to the founder effect. Hereditas 99: 11–20.

28. Smith, C., 1984. Genetic aspects of conservation in farm livestock. Livest. Prod. Sci. 11: 37–48.

29. Turner, H.N., 1986. Principles for preservation of endangered species and breeds in the tropics. FAO Anim. Prod. Health Paper 66: 181–188.

30. Wright, S., 1931. Evolution in Mendelian populations. Genetics 16: 97–159.

7 Appendix 1. Summary of SABRA documentation of livestock breeds (3).

(B=buffalo, C=cattle, Ch=chicken, D=duck, G=goat, P=pig, S=sheep, T=turkey, v=verbal description)

Appendix 2. Summary of FAO-descriptors (9, 10, 11). (Species symbols as in Appendix 1).

Appendix 3. Summary of EAAP/FAO-forms (25), and shortened FAO-form

(= X) (23). (Species symbols as in Appendix 1, H=horses).