References and Notes 1. M. Saginor and R. Horton, Endocrinology 82, 627 (1968); R. M. Rose, T. P. Gordon, I. S. Bernstein, Science 178, 643 (1972); K. Purvis and N. B. Haynes, J. Endocrinol. 60,429 (1974); F. Macrides, A. Bartke, F. Fernandez, W. D'Angelo, Neuroendocrinology 15, 355 (1974). 2. C. B. Katongole, F. Naftolin, R. V. Short, J. Endocrinol. 50, 457 (1971). 3. F. Macrides, A. Bartke, S. Dalterio, Science 189, 1104 (1975); J. A. Maruniak, A. Coquelin, F. H. Branson, Biol. Reprod. 18, 251 (1978). 4. J. A. Maruniak and F. H. Bronson, Endocrinol- ogy 99, 963 (1976). 5. I. Martin, Psychol. Bull. 61, 35 (1964); R. F. Thompson and W. A. Spencer, Psychol. Rev. 73, 16 (1966); R. A. Hinde, in Short-term Changes in Neural Activity and Behavior, G. Horn and R. A. Hinde, Eds. (Cambridge Univ. Press, Cambridge, 1970), p. 3. 6. In our cannulation procedure, an atrial cannula (inner diameter, 0.30 mm; outer diameter, 0.64 mm; 45-cm Silastic tubing, Dow Corning Corp.) exits via a backpack sutured to the mouse and is protected by an extension spring. The entire unit is supported from above and swivels freely. During the 5-day surgical recovery period, 0.33 units of heparin in 0.3 ml of saline is continuous- ly infused per 24 hours. Data to be presented elsewhere (10) document that cannulated males exhibit normal aggressive and sexual behaviors, when compared with males that do not have surgery. 7. The CF-1 mice used in this experiment were reared in our laboratory colony. At weaning (21 to 23 days of age), the males were isolated in 29 by 18 by 13 cm polypropylene cages in a room without females. The ambient conditions were 23? ? 10C, 14:10 hour light-dark cycle, lights on at 0600 hours, and Purina mouse chow and wa- ter always available. Experimental males were 70 to 80 days old. After cannulation, they were housed in 29 by 14 by 14 cm wooden boxes and remained in the same room. All testing began between 0900 and 1100 hours. 8. After each 25-jitl blood sample was obtained, the cannulas were rinsed and the blood was re- placed with heparinized saline (10 units per mil- liliter). The 10-^1 samples of plasma were stored in 40-fil radioimmunoassay (RIA) buffer at ?80?C before assay. Plasma LH concentrations were determined with the NIAMDD rat radioim- munoassay kit verified for measuring mouse go- nadotropins by W. G. Beamer, S. M. Murr, and I. I. Geschwind [Endocrinology 90, 823 (1972)]. The reference curve was fitted and the unknown concentrations were interpolated by using the computer analysis described by D. Rodbard and D. M. Hutt [in Symposium on Radioimmunoas- say and Related Procedures in Medicine (Inter- national Atomic Energy Agency, Vienna, 1974), p. 165]. Within-assay variation was 10.4 percent and between-assay variation 13.9 percent; the minimum detectable amount was 0.125 ng. Re- sults are expressed as nanogram-equivalents of NIAMDD-RAT-LH-RP-1. 9. H. R. Nankin and P. Troen, J. Clin. Endocrinol. 33, 558 (1971); A. Bartke, R. E. Steele, N. Mus- to, B. V. Caldwell, Endocrinology 92, 1223 (1973). 10. A. Coquelin and F. H. Bronson, in preparation. 11. Analysis of variance with repeated measures was performed on the peak LH values which were observed during the four sampling periods. Within both patterns of repetitive female ex- posure, (-tests were used to compare the peak LH values observed during the first and fourth sampling periods; the /-test between the con- tinuous exposure and repetitive exposure groups compared the peak LH values during the last sampling period. 12. R. E. Whalen, cited by J. R. Wilson, R. E. Kuehn, F. A. Beach, J. Comp. Physiol. Psy- chol. 56, 636 (1963). 13. J. O. Almquist and E. B. Hale, Proc. Congr. An- im. Reprod. 3, 50 (1956); H. Fowler and R. E. Whalen, J. Comp. Physiol. Psychol. 54, 68 (1961); T. E. McGill, J. Genet. Psychol. 103, 53 (1963); R. P. Michael and D. Zumpe, Science 200, 451 (1978). 14. We are grateful to C. Desjardins for his consid- erable effort and direction during development of the cannulation procedure and to M. Graham for his assistance with computer analysis of RIA. This investigation was supported by Public Health Service grant HD-03803 to F.H.B. from the National Institute of Child Health and Human Development and by an NSF graduate fellowship to AC. 30 January 1979; revised 11 May 1979 SCIENCE, VOL. 206, 30 NOVEMBER 1979 Inbreeding and Juvenile Mortality in Small Populations of Ungulates Abstract. Juvenile mortality of inbred young was higher than that of noninbred young in 15 of 16 species of captive ungulates. In 19 of 25 individual females, belong- ing to ten species, a larger percentage of young died when the female was mated to a related male than when she was mated to an unrelated male. An ever increasing number of the world's ungulate species exist only in relatively small populations in which some degree of inbreeding will inevitably occur. Extensive studies of laboratory and domestic mammals and birds in- dicate that inbreeding leads, in the ma- jority of cases, to increased mortality in young animals and reduced fertility in adults (1-3). Inbred animals are usually "less able to cope with their environ- ment than are noninbred animals" (2, p. 215) and are often more susceptible to various diseases and environmental stresses (3, 4). The limited data from nat- ural populations suggest that close in- breeding has the same deleterious con- sequences in the wild (5). Table 1. Juvenile mortality in inbred and noninbred young. Non- In- X2 test Sign Species inbred young bred young TV X2 P test* Elephas maximus (Indian elephant) Lived 11 2 19 + Died 2 4 4.997 .025! Equus burchelli (zebra) Lived 20 3 32 .413 .528! + Died 7 2 Choeropsis liberiensis (pygmy hippopotamus) Lived 139 23 235 17.28 .ooot + Died 45 28 Muntiacus reevesi (muntjac) Lived 18 12 40 1.212 .270 + Died 4 6 Cervus eldi thamin (Eld's deer) Lived 13 0 24 11.679 .001! + Died 4 7 Elaphurus davidianus (Pere David's deer) Lived 15 19 39 .030 .857 + Died 2 3 Rangifer tarandus (reindeer) Lived 19 9 50 2.538 .107 + Died 10 12 Girajfa camelopardalis (giraffe) Lived 11 2 19 2.537 .107 + Died 3 3 Tragelaphus strepsiceros (kudu) Lived 10 8 25 .005 .941 - Died 4 3 Tragelaphus spekei (sitatunga) Lived 15 31 75 9.012 .006t + Died 1 28 Hippotragus niger (sable) Lived 18 3 32 8.183 .005! + Died 4 7 Oryx dammah (scimitar-horned oryx) Lived 35 0 42 28.378 .ooot + Died 2 5 Connochaetes taurinus (wildebeest) Lived 6 29 48 .680 .419 + Died 1 12 Madoqua kirki (dik-dik) Lived 10 7 32 .473 .499 + Died 7 8 Gazella dorcas (Dorcas gazelle) Lived 36 17 92 9.288 .003! + Died 14 25 Capricornis crispus (Japanese serow) Lived 52 27 135 10.585 .002! + Died 21 35 *For the sign test, + = juvenile mortality higher in inbred than noninbred young (P = .0003). at .05 level; one degree of freedom in all cases; probabilities are rounded to three places. 0036-8075/79/1130-1101$00.50/0 Copyright ? 1979 AAAS tSignificant 1101 Table 2. Breeding success of individual females which were mated to both unrelated and related males. Mating with Mating with unrelated male related male Female (No.) Sign test* Species Births Juvenile deaths Births Juvenile deaths (No.) (No.) (No.) (No.) Pygmy hippopotamus 61 1 0 9 8 + 87 7 0 6 2 + 102 3 1 3 2 + 112 10 3 1 1 + Muntjac 101,432 10 1 3 1 + 34,847 8 0 2 0 0 M00,510 1 1 4 2 - Eld's deer M00.391 1 0 2 2 + M00.281 1 1 3 3 0 Pere David's deer M00.346 1 0 3 1 + Reindeer M00,029 2 0 3 2 + Oryx M00.262 4 0 1 1 + MO0,263 4 0 1 1 + Wildebeest 28-375A 5 0 2 2 + Dik-dik M00,540 4 1 1 1 + M00,901 1 0 1 1 + Dorcas gazelle 28,147-B 10 5 3 1 - 28,918 8 0 11 6 + 30,168 5 3 6 5 + Japanese serow TYM-1 9 1 2 0 - KYT-7 6 2 1 1 + JSC-2a 2 1 4 4 + JSC-2d 1 1 2 2 0 OMC-12 6 2 1 1 + KOB-lcc 2 0 2 1 + *For the sign test, + = juvenile mortality higher when mated with related male (P = .0082). Although the deleterious effects of in- breeding are well known to geneticists and students of animal husbandry, they have been ignored, or even denied, by many workers in conservation and wild- life management. In an article on the captive breeding of rare and threatened deer, Whitehead (6) recently expressed the view of many skeptics: "Adverse comment about the ill-effects of in- breeding is, in my opinion, often exag- gerated, and provided the area is well looked after and not overstocked, there should be no harmful effects. . . . No liv- ing mammal could be more inbred than the Pere David's deer, for the whole world stock is derived from the initial Woburn stock, yet it seems to have changed little over the years." Skepticism as to the harmful effects of inbreeding in the majority of ungulates probably stems from a number of sources: the existence of a few success- ful highly inbred species?such as the Pere David's deer?ignorance of the se- vere problems with increased juvenile mortality and decreased fertility encoun- tered during the development of the modern inbred strains of laboratory ani- mals (7) and livestock breeds (8), and the rarity of recorded instances of dele- terious effects of inbreeding in captive exotic species (9, 10). Seal (//) has sug- gested that the rarity of such reports in- dicates not that exotic species tend to be more resistent to the harmful effects of inbreeding than common laboratory and domestic animals, but only that few zoos have maintained and analyzed adequate records. We report here on the relation be- tween inbreeding and juvenile mortality in 16 species of captive ungulates (12). Because of the small sample sizes, we compared only two levels of inbreeding: "noninbred," which included all young with unrelated parents, and "inbred," which included all young with an in- breeding coefficient greater than zero (13). Of the inbred young for which it was possible to calculate an inbreeding coefficient, 75 percent had inbreeding co- efficients of 0.25 or more. Young that survived 6 months or more were consid- ered to have "lived." The "died" cate- gory consisted of all young surviving less than 6 months, including stillbirths and those born prematurely. The number of inbred and noninbred young that lived and died in each of the 16 species is shown in Table 1. Juvenile mortality was significantly higher in in- bred than noninbred young in the four species with the largest sample sizes (pygmy hippopotamus, Japanese serow, Dorcas gazelle, and sitatunga) and four species in which the sample size was small but the mortality of inbred young was very high (elephant, Eld's deer, sable, and oryx). The increased mortali- ty rate of inbred young in the four spe- cies with the largest sample sizes ranged from a low of 28 percent in the Japanese serow to a high of 41 percent in the sita- tunga. Inbred young had a higher mortal- ity in 15 of 16 species (P = .0003, one- tailed sign test), which suggests that in- creased juvenile mortality in inbred young is a general phenomenon in ungu- lates and that the failure to show a sig- nificant difference by the chi square test in some cases may be due to an in- sufficient sample size per cell. In 19 of 25 individual females, belonging to ten spe- cies, a larger percentage of young died when the female was mated to a related male than when she was mated to an un- related male (Table 2) (P = .008, one- tailed sign test). We have not yet analyzed most of our data with respect to the many other vari- ables that might influence juvenile mor- tality, such as birth season, management improvements, birth order of the young, and possible differences between wild- and captive-born females. However, a detailed analysis of the data on the spe- cies with the largest sample size of those for which we had medical records (Dor- cas gazelle) showed that none of these significantly affected the high mortality rates of inbred young (10). Furthermore, causes of death in inbred and noninbred gazelle calves were strikingly different. Once past the immediate postnatal peri- od, noninbred calves were remarkably sturdy. Only five died after the age of 4 days, all of traumas sustained during capture or transport. Inbred calves often succumbed to inanition and a variety of miscellaneous medical problems and in- fections not found in noninbred calves. Our results support Seal's (//) claim that deleterious effects of inbreeding are widespread in captive populations of ex- otic animals and have remained unrecog- nized only because the majority of zoos have not maintained detailed records. To date, conservationists have paid relative- ly little attention to the problem of pre- serving genetic variability in small popu- lations of ungulates, whether these are captive breeding stocks, translocated or reintroduced populations, or small popu- lations in isolated reserves. The degree of inbreeding depression in a small ungu- late population of a given size presum- ably varies with many factors, such as the mating system of the species, its ge- netic load, and the length of time the population has been genetically isolated (14). However, because inbreeding has been shown to reduce viability and fertil- ity in such taxonomically distant orga- 1102 SCIENCE, VOL. 206 nisms as insects (3, p. 87) and plants (3, p. 6), it seems reasonable to predict that these effects may be expected in the ma- jority of normally outbreeding ungulate species. The evidence that inbreeding of- ten leads to increased juvenile mortality and other deleterious effects in ungulates is at hand and the time has come to insti- tute sound genetic management of small ungulate populations wherever practi- cable, without waiting for evidence of such effects in each species or popu- lation. (CATHERINE RALLS KRISTIN BRUGGER JONATHAN BALLOU Office of Zoological Research, National Zoological Park, Smithsonian Institution, Washington, D.C. 20008 References and Notes 1. D. S. Falconer, Quantitative Genetics (Ronald, New York, 1961), pp. 247-254; E. J. Warwick and J. E. Legates, Breeding and Improvement of Farm Animals (McGraw-Hill, New York, ed. 7, 1979), pp. 201-229. 2. J. F. Lasley, Genetics of Livestock Improve- ment (Prentice-Hall, Englewood Cliffs, N.J., ed. 3, 1978). 3. S. Wright, Evolution and the Genetics of Popu- lations, vol. 3, Experimental Results and Evo- lutionary Deductions (Univ. of Chicago Press, 1977), pp. 44-96. 4. L. L. Cavalli-Sforza and W. F. Bodmer, The Genetics of Human Populations (Freeman, San Francisco, 1971), pp. 368-77, 517-20. 5. P. J. Greenwood, P. Harvey, C. M. Perrins, Na- ture (London) 271, 52 (1978); C. Packer, Anim. Behav. 27, 1 (1978). 6. G. K. Whitehead, in Threatened Deer (Inter- national Union for the Conservation of Nature and Natural Resources, Morges, Switzerland, 1978), p. 353. 7. L. C. Strong, in Origins of Inbred Mice, H. C. Morse, Ed. (Academic Press, New York, 1978), p. 45. 8. C. Darwin, The Variation of Animals and Plants under Domestication (John Murray, London, ed. 2, 1899), p. 96; R. Wallace, Farm Livestock of Great Britain (Oliver & Boyd, Edinburgh ed. 5, 1923); S. Wright, Heredity 14, 405 (1923). 9. J. Bouman, Int. Zoo Yearb. 17, 62 (1977); N. R. Flesness, ibid., p. 77; V. D. Treus and N. V. Lobanov, ibid. 11, 147 (1971); H. M. Slatis, Genetics 45, 275 (1960); K. Rails et al. (10). Although many authors have been concerned about possible problems due to the loss of ge- netic variability in small populations, both in the wild and in captivity, most treatments have been primarily theoretical: F. Bunnell, in Threatened Deer (International Union for the Conserva- tion of Nature and Natural Resources, Morges, Switzerland, 1978) p. 264; K. W. Corbin, in Endangered Birds: Management Tech- niques for Preserving Threatened Species, S. A. Temple, Ed. (Univ. of Wisconsin Press, Madi- son, 1978) p. 264; C. Denniston, in ibid., p. 281; T. E. Lovejoy, in ibid., p. 275; U.S. Seal, in ibid., p. 303; N. R. Flesness, far. Zoo Yearb. 17, 77 (1977); I. Franklin, in Conservation Biology, M. E. Soule and B. Wilcox, Eds. (Sinauer Asso- ciates, Sunderland, Mass., in press); M. E. Soule, in ibid.; J. Senner, in ibid.; R. I. Miller, Environ. Conserv. 5, 191 (1978). 10. K. Rails, K. Brugger, A. Click, Int. Zoo Yearb., in press. 11. U. S. Seal, in Endangered Birds: Management Techniques for Preserving Threatened Species, S. A. Temple, Ed. (Univ. of Wisconsin Press, Madison, 1978), p. 303. 12. The elephant data were obtained from the Wash- ington Park Zoo in Portland. Ore., and the Japanese serow data from Atushi Komori, keep- er of the studbook for this species. The sitatunga data were taken from a pedigree published by E. M. Lang [Zool. Gart. N.F. 49, 8-16 (1978)] and the pygmy hippopotamus data from the pub- lished studbook [E. M. Lang, Das Zwergfluss- pferd (Ziemsen, Wittenberg, Lutherstadt, 1975)]. Part of the sable data were obtained from the Baltimore Zoo. Additional sable data and all the data on the 11 other species were collected from SCIENCE, VOL. 206, 30 NOVEMBER 1979 the records of the National Zoological Park (NZP), Washington, D.C. 13. Inbreeding coefficients (F) were calculated by hand from the formula Fx = '/z X [('/%)" (1 + Fa)] (2, p. 240). We assumed that the indi- viduals used to found a population were unre- lated, although we knew that this assumption was unjustified in some cases. Unless we knew the manner in which individuals subsequently obtained from outside sources were related to existing stock, we assumed they were unrelated. In some cases, it was not possible to calculate an exact inbreeding coefficient although we knew it must be greater than zero. 14. F. Bunnell, in Threatened Deer (International Union for the Conservation of Nature and Natu- ral Resources, Morges, Switzerland, 1978), p. 264; G. Bush, S. M. Case, A. C. Wilson, J. L. Patton, Proc. Natl. Acad. Sci. U.S.A. 74, 3942 (1977). 15. We thank W. Iliff for the elephant data, A. Ko- mori for the Japanese serow data, S. Graham for part of the sable data, the many NZP staff mem- bers who recorded data over the years, and J. Block for her care in maintaining the NZP rec- ords and her assistance in using the files; J. Spit- zer for checking the accuracy of the data; L. A. Hayek and M. Anderson for statistical advice; and M. Bush, J. Eisenberg, N. Flesness, N. Muckenhirn, and C. Wemmer for comments on an earlier draft of the manuscript. We also thank the Fluid Research Fund of the Smithsonian Institution and Friends of the National Zoo for financial support of this project. 23 May 1979; revised 17 September 1979 Relatedness and Inbreeding Avoidance: Counterploys in the Communally Nesting Acorn Woodpecker Abstract. Acorn woodpeckers (Melanerpes formicivorus) live in family groups within which more than one female may lay eggs communally in a single nest. Com- munally nesting females are usually closely related and share evenly in nesting activ- ities. Although birds of either sex may breed in their natal territory, reproductive inhibition of offspring by the presence of their parent of the opposite sex and dis- persal by unisexual sibling units ensure that inbreeding between close relatives is rare. True communal nesting, in which more than one female regularly lays eggs in the same nest, is known from some ten species of birds (/). In none of these has either the genetic relatedness among such females or the consequences of communal nesting on amount of in- breeding been identified. We report such data for the acorn woodpecker. These data are difficult to gather and some of our sample sizes are small; but the re- sults bring out significant components of social dynamics acting to prevent in- breeding and needing attention in studies of all cooperative breeding birds. In California, the acorn woodpecker typically lives in permanently territorial family groups of 2 to 15 birds (2). Only a single nest is attended at any one time by a group, and most or all group members help to incubate and feed the young. Un- like many group-living species, either males or females may breed in their natal territory (5), thus apparently presenting unusually great opportunities for in- breeding among close relatives. As part of a continuing long-term study of the social behavior of this spe- cies at the Hastings Natural History Res- ervation, we recorded intergroup trans- fers of marked birds and clutch size in relation to group composition (3). On the basis of the deposition of two eggs on each of one or more days in a nest (4), we found that two females were nesting to- gether in at least 3 of 27 group breeding efforts in which the nest was found be- fore hatching and two or more females were known to be members of the group. Evidence from clutches of groups with differing compositions permits an analy- sis of the restrictions placed on repro- duction by females and the conditions under which more than one female may lay eggs in a nest. Our results suggest that (i) large sets of eggs are the result of true communal nesting rather than intra- specific nest parasitism by females from outside the group, (ii) females do not breed in their natal group as long as their known or presumed father is still in the group, and (iii) communally nesting fe- Table 1. Relatedness of acorn woodpeckers immigrating in unisexual units of either sex. Origin Number of units Number of individuals Siblings from the same group 8 (42 percent) 18 (40 percent) Birds from the same group, two of which were 3 (16 percent) 9 (20 percent) known siblings* Birds from the same group, one or none of which 7 (37 percent) 16 (36 percent) was known to have been born there* Birds from different groups 1 (5 percent) 2 (4 percent) Total 19 (100 percent) 45 (100 percent) *These units are also likely to have been siblings. 0036-8075/79/1130-1103$00.50/0 Copyright? 1979 AAAS 1103