Article — Artikel Genetic variation in the feral horses of the Namib Desert, Namibia E G Cothrana, E van Dykb and F J van der Merwec loci. Genetic variation in this population ABSTRACT is a concern, owing to a probable small Genetic variation at 7 blood-group and 10 biochemical genetic loci was examined in 30 founding population size and continuous horses from a feral herd from the Namib Desert of Namibia, Africa. The observed genetic low population numbers due to the harsh variability was extremely low compared with that found in domestic horse breeds. The low conditions of the environment. Also, variation was most probably a result of recent small population size and a small founding genetic comparisons of the Namib horses population size. Genetic comparison of the Namib horses, which were of unknown origins, to other domestic horse breeds were to domestic horse breeds, showed that the Namib horses had the highest genetic similarity made in hope of shedding some light to Arabian type horses, although they did not closely resemble this type of horse in confor- upon the origins of these populations. mation. Key words: biochemical genetics, blood groups, feral horses genetic distance, genetic MATERIALS AND METHODS variation. Blood was collected by jugular veni- Cothran E G, Van Dyk E, Van der Merwe F J Genetic variation in the feral horses of the puncture from a total of 30 Namib Desert Namib Desert, Namibia. Journal of the South African Veterinary Association (2001) 72(1): 18–22 horses. Eighteen of these horses were (En.). Department of Veterinary Production and Ethology, Faculty of Veterinary Science, born in the desert, while the remainder Private Bag X4, Onderstepoort, 0110 South Africa. was born in captivity as part of a breeding herd. One of the captive-born horses had desert-born parents that were not INTRODUCTION from gravel slopes, desert crust, desert sampled. This horse was included among Feral horse populations frequently live pavement plains and sand dunes. The the desert-born horses for analysis. The in harsh environments, because they are region is described as a cool desert with captive herd was formed in 1987 and forced into these areas to prevent compe- average temperature around 18 °C. Rain- maintained at the Onderstepoort Veteri- tition with grazing livestock such as cattle fall ranges from 0 to 200 mm with an nary Institute. Some of the samples were and sheep. The feral horse population of average around 65 mm per year. The area from horses captured during a drought the Namib Desert perhaps occupies the where the horses occur is an ecotone of in 1992 and now in private ownership in most inhospitable environment of any the Desert biome and the Succulent Potchefstroom. group of horses. Karroo biome18. The horses compete for Standard immunological procedures The horse population is concentrated grazing with gemsbok (Oryx gazella) and involving haemagglutination and com- around a well (the Garub water hole, the springbok (Antidorcas marsupialis). plement-mediated haemolysis20,21 were only permanent source of water in the The origins of these horses are not used to detect variation of red cell allo- area) about 20 km from the village of Aus known. One hypothesis is that early antigens at 7 blood-group loci. Starch and in Namibia22. Although no written re- immigrants from the Cape Colony moved polyacrylamide gel electrophoresis and cords exist, the horses are known to have through this area from the south, bring- isoelectric focusing were used to detect been in this area for at least 80 years and ing their Cape Horses with them. Another variation at 10 serum and red blood cell appear to have adapted well to these diffi- is that a cargo steamer ran aground on the lysate protein loci4,5,11,12,19,16. cult conditions. During favourable years Namib coast with a load of Thorough- The blood-group loci examined were the horses appear to be in good condition breds destined for Australia and that the A, C, D, K, P, Q, and U horse blood- and show evidence of being well-bred, some of these horses reached shore and group loci and the biochemical protein quality horses, although they show evi- survived. Another suggestion is that loci were alpha-1 $-glycoprotein (A1B), dence of stress in dry years. during a World War I campaign in this albumin (ALB), serum esterase (ES), According to Greyling10 and Rutherford region, military troops abandoned or lost vitamin D binding protein (GC), glucose and Westfall18, the area where the horses horses in the desert. Finally, there is spec- phosphate isomerase (GPI), alpha-hae- are found varies in topology, geology and ulation that the horses came from the moglobin (HBA), 6-phosphogluconate climate. The area is characterised by low farm of Baron von Wolf, who bred horses dehydrogenase (6-PGD), phospho- mountain islands. Elevation ranges from for the army up to just after the start of glucomutase (PGM), protease inhibitor 760 m near Garub to 1100 m in the east. World War I. This farm was situated on (PI) and transferrin (TF) loci. Nomen- Dry rivers penetrate the area and vast the edge of the Namib Desert about clature for variants at all 17 loci is in plains covered with sand expanses are 150 km from the Garub water hole. Von accordance with internationally stand- common. The surface of the region varies Wolf died in action in 1916, and it is possi- ardised usage for horses2,3 except for ble that horses from his farm found their variants at some loci, which have not yet aDepartment of Veterinary Science, University of Ken- way to the Garub water hole during the received international recognition. tucky, Lexington, KY, USA. b turmoil following the German defeat in All analyses were calculated for the 19 Department of Veterinary Production and Ethology, 22 Faculty of Veterinary Science University of Pretoria, colonial South West African . desert-born horses and for the total Onderstepoort, 0110 South Africa. In this study we examine genetic varia- sample. Gene frequencies for biochemical cDepartment of Agricultural Development, Private Bag X9087, Cape Town, 8000 South Africa. tion in the Namib horses based on 7 loci were calculated by direct count. Received: August 2000. Accepted: November 2000. blood-group and 10 biochemical genetic Frequencies of alleles at blood-group 18 0038-2809 Tydskr.S.Afr.vet.Ver. (2001) 72(1): 18–22 loci were calculated by the allocation and mean values based upon data from 99 Table 1: Allele frequencies for 17 genetic method1. Genetic variation was measured domestic horse breeds. There was statisti- loci tested in the Namib Desert feral horse as observed heterozygosity (Ho), Hardy- cally significant evidence for a recent pop- herd. HBG stands for Horse Blood Group. Only alleles found in this herd are shown. Weinberg expected heterozygosity (He), ulation bottleneck in the Namib herd 14 unbiased expected heterozygosity (Hu) , based upon the method of Cornut and Locus Allele Namib Desert herd effective number of alleles (Ae), and the Luikart6, although caution must be used total number of variants found in each in this interpretation because only 13 vari- Trf D 0.033 population (Na). Ho was calculated for able loci could be analysed and 20 are F2 0.950 biochemical loci only because of the pres- recommended for this test. The distribu- O 0.017 ence of recessive alleles and/or ambigu- tion of allele frequencies for all 17 loci is A1B K 1.000 ous genotypes at blood-group loci. shown in Fig. 1. Est I 1.000 Therefore, for direct comparison, He and Valuesof S and Da for the Namib horses Al A 0.900 Hu were calculated for biochemical loci compared to a number of domestic breeds B 0.100 Gc F 0.500 only (in an ideal population, He (Hu) are shown in Table 3. The domestic breeds S 0.500 should equal Ho), for blood-group loci, are grouped according to relationships PGD F 1.000 and for all 17 loci. Tests for the genetic among the breeds and mean S and Da PGM F 0.700 effects of a population bottleneck were values for the Namib population com- S 0.300 examined using the programme BOT- pared to each group also are given. The GPI F 0.083 TLENECK6. In addition, population in- genetic associations of the Namib horses I 0.917 breeding level was estimated by Wright’s with domestic breeds are summarised in Hb BI 0.183 Fis =1–(Ho/He). Values of genetic varia- the dendogram of Fig. 2. The tree shown BII 0.817 tion of the Namib horses were compared is a majority rule, strict consensus tree Pi L 0.083 to those of domestic horse populations from 30 separate RML trees. The breeds S 0.667 T 0.050 and other feral horse populations that that make up the branches other than q 0.200 have been tested at the University of those for the Oriental breeds’ branch is EAA adf 0.933 8 Kentucky. Genetic relationship of the not shown (see Cothran and Van Dyk for b 0.067 Namib herd to these other domestic a complete tree). EAC a 0.815 breeds was investigated using Rogers’ – 0.185 genetic similarity coefficient (S)17 and DISCUSSION EAD dk 0.517 Nei’s modified distance (Da)15. Restricted Genetic variation in the Namib horse deo 0.033 maximum likelihood analysis (RML)9 was was extremely low compared to other bcm 0.350 used to construct the dendogram of Fig. 2. horse populations that have been exam- cgm 0.100 ined. Of domestic horse populations, EAK – 1.000 EAP ac 0.034 only the Bedouin strain of Arabian horses RESULTS ad 0.428 Allele frequencies for the variants found (also known as the Blue Star Arabians) – 0.538 at the 17 loci are shown in Table 1.