Cotton-top tamarins (Saguinus (o.) oedipus) in a semi-naturalistic captive colonyкод для вставкиСкачать
American J o u r n a l of Primatology 2O:l-12 (1990) RESEARCH ARTICLES Cotton-Top Tamarins (Saguinus (0.) oedipus) in a Semi-Naturalistic Captive Colony ELUNED C. PRICE AND W.C. McGREW Department of Psychology, University of Stirling, Stirling, Scotland To test the prediction that the breeding success of captive cotton-top tamarins (Saguinus (0.)oedipus) could be improved by maintaining them in groups whose size and age-sex composition resembled those of wild groups, data were collated from 6.5 years of records from a breeding colony that otherwise had housing and husbandry procedures similar to those of other successful colonies. Group size and composition in the colony closely resembled those of wild groups, and infant survival was the highest yet reported for the species, with 69% of the 124 infants born reared by their parents to adulthood, and a mean surviving litter size of 1.5 infants. Abortion, stillbirth, and parental neglect of infants were rare. Parity had several effects on reproduction: mean litter size decreased, but percentage infant survival increased; interbirth intervals decreased in length; and seasonality in reproduction was more pronounced for the first four litters born to breeding females than for their subsequent litters, with a birth peak in the spring. Although a spacious and complex physical environment, retention of offspring in their natal families until experience of several sets of infant siblings had been obtained, and non-invasive husbandry and research techniques may all have contributed to the colony’s success, it seems possible that the improvement over other colonies is due to the resemblance of group composition to those of wild tamarins. Key words: breeding, h u sb an d ry , naturalistic housing, ecological validity INTRODUCTION The cotton-top tamarin, Saguinus (0.1 oedipus, is a n endangered species, but has become relatively well established in captivity [Tardif & Colley, 19881. However, variable success in breeding the species has been reported, and some colonies have suffered high rates of infant mortality. These have often resulted from frequent stillbirths (e.g,, 32%: Kilborn et al. ; 20%:Tardif et al. [1984b]), or from parental rejection of offspring [Carroll, 1983; Evans, 1983; Kilborn et al., 1983; Kirkwood et al., 1983; Scullion, 19871. Wild-caught parents have been reported to be more successful than captive-born individuals [Kilborn et al., 1983; Tardif et al., 1984bl. Recently, however, other colonies have been more successful, Received for publication July 13, 1989; revision accepted November 6, 1989. Address reprint requests to Eluned C. Price, Department of Psychology, University of Stirling, Stirling, FK9 4LA, Scotland, UK. 0 1990 Wiley-Liss, Inc. 2 I Price and McGrew particularly since the widespread recognition that experience with younger siblings is needed for adequate parental behavior, especially for females [Kirkwood et al., 1983; Tardif et al., 1984a; Clapp & Tardif, 1985; Snowdon et al., 19851. Parity appears to lead to increased infant survival [Kirkwood et al., 1985; Snowdon et al., 19851, but some data also suggest that primiparous mothers can achieve comparable success to that of more experienced females if paired at a n older age and with a n experienced male [Tardif et al., 19861. Breeding success also seems to be greater in colonies with more spacious and complex housing [Snowdon et al., 19851. From its inception in 1982, the Stirling colony adopted policies similar to those that have led to successful breeding of this species in other colonies. However, as the main research emphasis in the colony is on the social behavior and reproductive strategies of cotton-top tamarins [e.g., McGrew, 19881, we have also tried to increase ecological validity by maintaining groups of similar size and composition to those in the wild, a s far as is possible in a captive environment. Fortunately, Stirling can provide large amounts of space, allowing us to maintain larger groups than elsewhere, and thus to keep offspring in their natal families well into adulthood. In other colonies, in contrast, adult offspring are typically removed once groups have about eight members. Since Garber et al. [19841 produced evidence suggesting that infant survival in wild Saguinus mystax was correlated with the number of adult males in a group, we predicted that keeping several adults in each group would lead to enhanced breeding success. To test this, we collated data from the colony from its arrival in Stirling in May 1982, until October 1988, and used them to make comparisons with the available field data [Neyman, 1978,19801 and with other breeding colonies. METHODS Background The monkeys were donated by the Department of Zoology, University College of Wales, Aberystwyth, in May 1982, and are housed in the Primate Unit of the Department of Psychology. Conversion of the unit, originally designed for macaques [Chamove, 19811, is described by Evans . The tamarin colony initially consisted of four breeding pairs, and has grown to a maximum of eight breeding groups at a time, with new breeding adults being either offspring of the original pairs or brought in from outside to reduce inbreeding. Two of the original breeding pairs were wild-caught, but were imported before the U.K. ratified the Convention on International Trade in Endangered Species in 1976. All other monkeys are captive-born. The colony is used for teaching, and for non-invasive research. No major health problems have been encountered. Housing Each of the four largest groups is housed in its own room (mean dimensions: 2.91 m high x 3.45 m wide x 3.54 m deep). The smaller groups are housed in one or two cage units (each unit measuring 1.97 m high x 1.19 m wide x 1.68 m deep) in a single large colony room and have regular access to four outdoor areas (mean volume 29.3 m3) in good weather. Large groups have more intermittent access to the outdoor areas. Groups are maintained in visual and tactile isolation, but can hear and probably also smell each other. Flexible ducting 15 cm in diameter (Crystaflex; Flexible Ducting Ltd., Milngavie, Glasgow) is used to connect different areas of the monkeys’ living quarters together, and allows us to move groups to new housing easily. The tamarins are thoroughly habituated to this apparatus and will readily travel distances of 30 m in it. Studies that require moving monkeys Breeding Cotton-Top Tamarins I 3 from one part of the unit to another [e.g., McGrew & McLuckie, 19861 also use the ducting, obviating any need to catch or handle the tamarins. All rooms have large Plexiglas skylights (1.15 m wide x 1.75 m long) which permit natural light to enter at all times. Additional lighting is provided by fluorescent strip lights operated by automatic time switches. Until December 1987 the lights were on a constant 12 hour light/dark cycle running from 0700 to 1900, but since then, seasonal changes have been made quarterly to mimic changes in day length in the species’ natural habitat in northwestern Colombia (spring and autumn: 0700-1900; summer: 0645-1930; winter: 0715-1845). Temperature indoors is maintained between 20 and 25”C, and humidity is kept at 40-60%. Extractor fans in each room change the air approximately 12-25 times a n hour during working hours. All living areas are furnished with a wide variety of natural branches and other furnishings, with one to three sleeping boxes. In indoor housing, a layer of woodchips covers the floor. Further details of furnishings are given by McKenzie et al. [19861. Plants are grown seasonally in the outdoor areas, and year-round indoors in view of the monkeys. From time to time the tamarins are given various devices designed to enrich their environment, e.g., food hidden in holes (Chamove, unpublished data). Cages are cleaned with detergent and disinfectant weekly, and rooms every 8 weeks. Outside areas are hosed weekly and cleaned thoroughly every 8 weeks. All personnel wear lab coats and paper masks, and disinfectant foot baths are placed outside each room housing monkeys. The tamarins are caught up only if they require medical treatment, or when they are marked for identification purposes. Each tamarin over the age of 12 months wears a light-weight, loose-fitting chain necklace holding a colored disc; infants and juveniles are marked on their crests with picric acid, a yellow dye. Diet The tamarins are fed a mixed diet of commercially prepared foods and fresh food. The components of the first feed, given in the morning, vary from day to day and include dry and soaked chow (S.D.S. Rat and Mouse Diet No. 3 and S.D.S. Mazuri Chow; Special Diet Services, Witham, Essex), peanuts in their shells, marmoset jelly (S.D.S.), commercial baby foods, rosehip syrup, eggs, peas, and beans. Vitamin supplements (D3, approx. 1000 IU per individual; B,; multivitamins) are given weekly. A small amount of milk is given daily, and water is available ad lib. At midday a mixture of chopped fresh fruit is given with some fruit-flavoured yogurt, and on weekdays a small amount of chopped apple is provided in the late afternoon. Recently, meals have been split between two sets of dishes, which appears to reduce competition for food. RESULTS Group Size and Composition Except for “ i n t r o d u ~ t i o n noted ~ ~ ’ below, our groups consist of monogamous breeding pairs with their offspring. Breeding adults average 7.5 5 1.9 years old (males: 8.4 * 1.8; females: 6.5 _t 1.6), excluding the four wild-caught individuals, who are all a t least 12 years old. All but one breeding adult was either wild-caught or was known to have had experience with younger siblings. The exception, a n adult male (London, see Table I) acquired from another colony, proved to be a n inadequate parent and was later replaced. The mean group size in Stirling at present is 6.9, and ranges from two to 11. We have maintained families as large as 16. Tamarins are usually removed from their families only if involved in severe aggression, or occasionally if families in the smaller cages grow too large. Sometimes we have removed adult offspring if a new breeding adult was needed for 4 I Price and McGrew TABLE I. Breeding History of Ten Cotton-Top Tamarin Females at Stirling Litter sizes Female Mate Roxanne Elvis Elsa Mario Sioux Lesley Sybille Mohican Bilbo Bilbo Pixie Bilbo Genevieve 1. Jim Date paired 1 2 Infants born Infants surviving (>3 months) 3 Total M F ? Total M F Total 1977 1 8 3 24/1/79 0 11 0 12 13 11 2 11 14 26 11 8 19 7 1 22 14 6 20 15 8 0 Comments Both parents wild caught Bothparents wild caught Sioux died 3/6/88 Lesley died 30110185 Sybille removed 23110179 0 212182 0 1111185 0 9 1 3 2 0 1 10 5 1 9 12 0 7 4 1 1 2 0 21 12 3 7 4 0 8 4 0 2616186 0 2 0 2 1 3 0 4 0 1 24/6/86 1 Pixie removed 2015188 20110183 0 0 0 0 0 0 0 0 0 0 0 Jim removed 21/11/84 0 1 4 5 5 9 0 14 2 5 7 First litter sired by Jim, but born after he was replaced by Reading 2619184 0 24110185 0 2214186 0 2 2 3 0 0 1 4 3 10 6 3 2 2 0 5 4 7 6 1 4 6 0 2 4 0 1 2 0 0 0 2 2 3 2 0 2 2619184 2. Reading Hopi Jim Delaware Arnold Erica 1. London 2. Fideaux Totals B 1812187 1 1 0 2 40 14 4 71 25 56 10 59 6 1 4 124 44 4 1 48 49 3 75 67 London removed 5/2/87 after injuring infants and fighting Erica 85 69 another group, particularly when the colony was first being built up. Those offspring (19 males and 20 females) still in their natal groups are, at the time of writing, a mean of 20.2 (217.6) months old (range 3-71 months). The mean age of the females (15.9 L 12.7 months) is less than that of the males (24.7 t 21.0 months), but this difference is not statistically significant (Mann-Whitney U test, n, = 20, n2= 19, z = -1.29, P = 0.20). Birth a n d Infant Survival Table I gives data from the ten breeding females maintained at Stirling either currently or in the past. Three of the pairs (Roxanne and Elvis, Elsa and Mario, and Sioux and Mohican) began to breed in Aberystwyth; details of their reproductive history before their arrival at Stirling are given by Evans . Only two suspected miscarriages have occurred, both to the same female (Lesley);blood, but no evidence of fetuses, was found in both cases. They have not been included in further analyses, so the data presented in the rest of this paper refer only to full-term pregnancies. Since we do no hand-rearing, all data on infant survival represent parent-reared offspring. The seasonal distribution of the 56 full-term births at Stirling until October 1988 is shown in Figure 1. Since interbirth interval is longer for the first four litters than for later litters (see below), the data for these births were graphed Breeding Cotton-Top Tamarins / 5 7 Litters 5-15 Litters 1-4 6 v) 5 5 .- L. a 4 c 0 & Q a E z 3 r 2 1 0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Month Fig. 1. Distribution of births of cotton-top tamarins a t Stirling over the calendar year. Litters 1-4 = the first four litters born to each breeding female. Litters 5-15 = subsequent litters born to each breeding female. separately. Overall, there were fewer births in the second 6 months of the year (July to December) than in the first half of the year, but there is some indication that early births are more likely than subsequent births to be seasonally distributed. A more detailed analysis of seasonality will be given elsewhere [McGrew and Webster, submitted]. There is also evidence of a circadian rhythm: births in the colony are most frequently observed in the evening between 1830 and 1930 [Price, 19871. The 48 interbirth intervals (IBIS)ranged from 189 to 383 days. The modal IBI length was 194 days (eight IBIS), and the median was 206 days. Intervals between pairing and first parturition (PPIs) were longer than IBIS (range 225 to 498, median 271 days, n = 7). Interbirth interval tended to decrease with parity (Fig. a), with a dramatic drop in length and variance to a n asymptote a t about 200 days after the third interval. The increase in mean IBI length a t IBI 11 is due to one female who had a n IBI of 355 days. There was also a trend for intervals to become longer with increased numbers of infants reared in the previous litter (0:median = 197 days, n = 2; 1: median = 205 days, n = 12; 2: median = 207.5 days, n = 34), but this was not statistically significant (Kruskal-Wallis one-way analysis of variance, x2 = 1.6, n = 48, P = 0.45). Table I shows that only 4% (2 of 56) litters were singletons, 71% were twins, and 25% were triplets. There was no difference in the number of male and female infants born (59 males, 61 females, 4 unknown sex). Twin litters in which the sex of both infants was known (n=37) did not differ significantly from a 1:2:1 ratio (MM:MF/FM:FF, 10:18:9; x2 = 0.08, d.f. = 2, Pc0.98). Triplet litters in which the sex of all infants was known ( n = 13) were more likely to be MFF (10) than MMF (3),but this was not statistically significant ( x 2 = 2.76, d.f. = 1, R O . 1 0 ) . Infants were born into groups with a n average of 7.0 (k3.2) members, excluding the infants themselves, with a mean of 3.0 (k2.4) older brothers and 2.0 (21.4) older sisters. Table I1 presents data on infant survival to various ages. The biggest loss was 6 / Price and McGrew 320 2 0 N 6 4 5 5 8 6 4 3 3 3 3 12 10 3 3 3 3 14 2 2 IBI # Fig. 2. Mean interbirth intervals (IBIS) and parity of cotton-top tamarins a t Stirling. Bars represent +SEM; IBI #1 = interval between litters 1 and 2, etc. Sample sizes are given below the x-axis. TABLE 11. Survivorshin of Infant Cotton-ToD Tamarins Born at Stirlinn N Males Females Unknown Total (%) N Day 0 Day 1 Week 1 Month 1 >3 months 59 61 56 (95) 56 (92) 4 (100) 54 (92) 55 (90) 4 (100) 47 (80) 44 (72) 1(25) 45 (76) 4 1 (67) 0 (0) 44 (75) 41 (67) 0 (0) 92 (74) 86 (69) 85 (69) 4 124 116(94) 113 (91) from live-born infants who died. during the first week. All offspring who reached the age of 3 months survived to adulthood. Almost all pairs showed competent parental caretaking from birth onwards [Price, 19871, including those rearing their first litters; nevertheless, 28% of the total infants lost were the offspring of primiparous mothers, and infants of primiparous mothers were significantly more likely to die than the infants of multiparous mothers (x2=4.83, d.f. = 1, P<0.05). There was no difference between the sexes in survival rates past 3 months (males: 75%, females: 67%; x2=0.47, d.f.= 1, P<0.50). Of 39 infants lost before the age of 3 months, five (13%) died after developing respiratory disorders; six (15%)died after being seriously injured when rejected by their families; another four (10%) died shortly before their mothers also died; one (see below) was stillborn; and cause of death in the remaining 23 cases was not established. However, 36% of infants lost were surplus triplets (n = 14 litters, assuming one infant per litter could not be reared by its parents). Infants found dead on the first day (day 0 losses) all appeared to be full-term, based on size, weight, and development. Although in most cases no information on lung flotation was Breeding Cotton-Top Tamarins I 7 TABLE 111. Parity and Survival of Infant Cotton-TopTamarins Born at Stirling Litter No. ~~ No. of litters born No. of infants born No. of infants surviving ( > 3 mo) % survival Mean litter size born Mean litter size surviving 1 2 3 4 7 18 7 39 2.6 1.0 6 14 9 64 2.3 1.3 5 12 9 75 2.4 1.8 5 11 8 73 2.2 1.6 5 6-15 Total 5 10 8 28 59 44 75 2.1 1.6 56 124 85 69 2.2 1.5 80 2.0 1.6 available, post-mortem examinations were done on three of eight infants that were never seen alive; two of these showed that the infants were live-born. Even assuming that all other infants found dead were stillborn, this gives a maximum of only six stillbirths (4.8%) from full-term pregnancies. Table I11 gives data on birth and infant survival (either to 1 year or to the time of writing if less than 1 year old) in relation to parity. There was an increased percentage survival rate with parity. The mean number of infants surviving per litter also increased with parity, while mean litter size at birth tended to decrease. All measures of breeding success levelled off after the third litter. Three primiparous females had helpers for their first litters (one was housed with her twin sister; the other two with previous offspring of the breeding male). We therefore compared their breeding success with that of primiparous females who only had their mate’s assistance. However, there was no difference between the two groups: both reared a mean of one infant per litter (Mann-Whitney U-test; n1 = 3, n2= 4, U = 6, P>0.05). Deaths Five (5.9%) of the 85 Stirling-born tamarins who survived to adulthood died before leaving the colony. Four tamarins (two males and two females) died after aggression in their families; three of these deaths (one female and two males) occurred in one family over a single weekend in the absence of research staff. In no case were the wounds inflicted severe enough to have caused death. The final case was a male who died 3 weeks after being removed from his family after bullying his siblings. Postmortem examination showed a heavy growth of Klebsiella pneumoniae. In addition, two breeding females who were not Stirling-born died shortly after giving birth. Postmortem showed no conclusive cause of death in either case. Introductions On only three occasions have newcomers been introduced into established groups. All cases involved the introduction of a “stepmother” into a family whose breeding female had died or been removed. In each case, the new female showed sexual behavior both with the original breeding male and with one or more of his sons. Further details of these “departures from monogamy” are given elsewhere [Price and McGrew, submitted]. Departures For the 44 Stirling-born monkeys who have left their natal groups, mean age at departure was 29.4 (211.7) months, range 12-64 months. There was no difference in the number of males and females who left (24 males and 20 females, P=O.45, binomial test). Monkeys who left their families had experience with a 8 I Price and McGrew TABLE IV. Comparative Data on Group Composition of Cotton-Top Tamarins From the Stirling Colony and From the Wild Stirling" Wildb Group size Mean Range 6.9 2-16 5.7 3-13 No. adult males Mean Range 2.3 1-4 2.7 2-4 No. adult females Mean Range 1.9 1-4 1.5 1-4 Total no. adults Mean Range 4.1 2-7 4.3 3-6 Mean 1.5 1.7 JanJulv JanJulv Infants per litter surviving to 1 year Birth oeak "As of October 1988, except group size range (whole study period). bFromNeyman 119801,Figure 16, pp. 60-61; data for August 1975, except group size range (whole study period). mean of 3.1 (k1.7) subsequent litters of infants before they departed (males: 3.4 5 1.7; females: 2.8 1.7). Mean group size a t leaving was 9.0 3.0 overall, excluding the monkey who left (males: 9.4 3.2; females: 8.6 2.8). The most common cause for a monkey's being removed was aggression from other members of its family, and 43%(19 of 44) of those monkeys who left were evicted in this way. Victims of aggression were removed unless they were immature, in which case the aggressor was taken out. Of the others, 11 (25%)were removed after being aggressive to other family members, and one subsequently died (see above), 10 (23%) were removed in order to be paired, and four (9%)died after severe aggression in their families (see above). In all but the cases of death, females were likely to leave a t a slightly younger age than males. None of these differences, however, are statistically significant (Mann-Whitney U tests; evicted: z = 1.45, P = 0.15; died: z = -0.77, P = 0.44; aggressive: z = -0.24, P = 0.81; paired: z = 0, P = 1.0; overall: z = -1.09, P = 0.28).However, males were significantly more likely than females to be removed for bullying siblings (9:2, P=0.033, binomial test). There were no significant sex differences in the other categories (evicted: 11:8, P=0.32; died: 2:2, P = 0.69; paired: 2:8, P = 0.055; binomial tests). More details of these processes are given elsewhere (McGrew, in prep.). * * * * ~ Comparison With Wild Groups A comparison of several parameters between Stirling and the only published field study of the cotton-top tamarin [Neyman, 1978, 19801 is given in Table IV. Group sizes, number of adults of each sex, and infant survival are comparable to those in the wild. Since no data are available on the mating patterns of cottontop tamarins in the wild, we cannot assess whether our nuclear families are typical ofwild groups. However, all the data given by Neyman [1978,19801 are compatible with either monogamy or polyandry. Comparison With Other Colonies Table V presents data on breeding success for four successful colonies from which data for at least 40 litters were available. Stirling's percentage survival rate (69%)is higher than that reported for other colonies, and the overall difference in survival rates between colonies is significant (x2= 23.9, d.f. = 3, P<O.OOl).Post-hoc pair-wise comparisons showed that Stirling's survival rate was significantly Breeding Cotton-Top Tamarins I 9 TABLE V. Comparative Data on Breeding Success Across Four Colonies of Cotton-Top Tamarins Colony No. litters born No. infs. born No. infs. survive % Mean litter size survival Born Survive Source Stirling 56 124 85 69 2.2 1.5 This paper Wisconsin 55 101 46 46 1.8 0.8 Snowdon et al. Oak Ridge 42 77 39 51 1.8 0.9 Tardif et al. [19851  Bristol 52 108 41 38 2.1 0.8 Kirkwood et al. [19851 higher than that of each of the other three colonies (Stirling vs. Wisconsin: x2=11.5, d.f.=l, P < O . O O l ; vs. Oak Ridge: x2=6.45, d.f.=l, P<0.02; vs. Bristol: x2=21.4,d.f. = 1,P<O.OOl), but there were no significant differences between any of the other colonies. However, because litter-size ratios vary from colony to colony, and therefore (because triplets are rarely reared unaided) percentage survival also varies, we have also calculated mean surviving litter size as a further basis for comparing the breeding success of different colonies. Although insufficient information was available to analyze the data statistically, the mean surviving litter size from Stirling is considerably higher than for other colonies. DISCUSSION The reproduction of the cotton-top tamarins at Stirling is similar in many respects to that described recently for other successful colonies. The mild degree of seasonality observed in our colony is consistent with other reports that cotton-top tamarins housed in natural lighting conditions have a birth peak in the spring months [Brand, 1980; Kirkwood et al., 19851. Interbirth intervals in Stirling are consistent with a gestation period of about 183 days, with a post-partum estrus occurring up to 4 weeks after parturition [Ziegler et al., 19871. No effect on interbirth intervals of the number of infants reared in the previous litter was found, supporting previous work indicating that there is no lactational anestrus in this species [French, 1983; Kirkwood et al., 1985; but see Ziegler et al., 19871. Despite these similarities, some differences in reproduction were seen. Interbirth intervals after the fourth litter were shorter than intervals reported elsewhere [Evans, 1983; Tardif et al., 1984b; Kirkwood et al., 19851. This is perhaps because few other reports included data from more than five litters for any female. These shorter interbirth intervals may explain the decreased effect of seasonality found for later litters. As in most other colonies [Kirkwood et al., 1985; Snowdon et al., 1985; though see Tardif et al., 19861, infant survival increased with parity, reaching an asymptote with the third litter. This effect is likely to be due to two interacting factors: a decrease in the average litter size born (in contrast with Tardif et al. , who found an increase), and an increase in the number of infants surviving from each litter. The delayed improvement occurred despite the fact that all breeding females had had considerable experience with infant siblings before being paired. At least two hypotheses might explain this: 1) Experience with one’s own infants is necessary in order to develop completely adequate parental skills [Snowdon et al., 19851; 2) helpers may be required to rear twins successfully, even in the good conditions in captivity. Recent field studies have suggested that single pairs of tamarins may 10 I Price and McGrew be unable to rear twin infants without such help [Garber et al., 1984; Goldizen, 19871. In the case of primiparous mothers, the data we have presented do not allow us to draw firm conclusions. Even in those infants who died, signs of rejection such as injury were not usually apparent, and the six infants known to be rejected were from only two families. On the other hand, the presence of helpers did not appear to affect the breeding success of primiparous mothers. Nevertheless, significant differences were found between the overall survival rates of various colonies. The higher survival rate for Stirling tamarins may be explained by group composition. A comparison with the field data provided by Neyman [ 1978, 19801 demonstrated that Stirling groups were similar in size, agesex composition, and infant survival to wild groups. Neyman  found that no parous cotton-top female in her study population appeared to be less than 5 years old. In the wild, therefore, young tamarins may stay in their natal family (or perhaps in other groups) for several years before beginning to breed, and therefore experience the birth and rearing of several sets of infants. Similarly, tamarins in Stirling were born into families in which several older offspring (on average about five), including adults, were available to help care for them. The initial success of Stirling’s breeding colony could be attributed to the fact that the founder members were already known to be good parents [Evans, 19831. Other important factors may include the complexity of the physical environment, many features of which were designed to mimic the natural habitat of the species, and husbandry and research techniques which minimize handling and stress [Snowdon et al., 19851. However, such practices are not unique to Stirling: the three colonies with which we have compared our results all provide varied diets, maintain visual isolation between groups, and either have cages at least 3m3 in volume or use exercise cages to increase the space available to each group [Kirkwood et al., 1983, 1985; Clapp & Tardif, 1985; Snowdon e t al., 19851. The main difference between Stirling and these other facilities therefore seems to be the age a t which older offspring are removed from their families and the resulting group composition. Although detailed demographic data are unavailable for other colonies, in most cases the normal practice appears to be to remove the eldest offspring in a group once they have had experience with one or two sets of younger siblings, typically when they are 18-24 months old. Families are therefore unlikely to contain many adult offspring. In contrast, the policy in Stirling is to retain offspring in their natal families for as long as possible. Monkeys typically do not leave until they are 2.5 years old and have helped rear three sets of infants, and many stay considerably longer: the oldest offspring still in his natal group a t present is nearly 6 years old. As a result, families contain on average two fully adult helpers in addition to the breeding pair. We therefore suggest that the Stirling colony’s continuing success may be due to the extensive experience the tamarins gain in helping to care for younger siblings, resembling that of their wild counterparts, and to the presence of several experienced adult caretakers in most families; and consequently, that success in breeding this species can be increased by attempting to keep groups of similar size and composition to those in the wild. CONCLUSIONS 1. Cotton-top tamarins were successfully bred in a colony which combined a spacious and complex physical environment with husbandry and research methods that minimized stress. The composition of groups in the colony, and their reproductive characteristics, resembled those of wild groups. Breeding Cotton-Top Tamarins I 11 2. Mean surviving litter size and percentage infant survival increased with parity, while mean litter size at birth and interbirth intervals decreased. Abortion, stillbirths, and parental neglect of infants were rare. 3. Breeding success, measured by percentage survival and mean surviving litter size, was the highest so far reported for the species. It is suggested that breeding success can be increased by maintaining groups of similar size and composition to those in the wild. ACKNOWLEDGMENTS We are immensely grateful to Dr. Sign Evans, who brought the tamarins t o Stirling and established many of the procedures which have led to the subsequent success of the colony. Dr. Evans was in charge of the colony from 1982 to 1983, when W.C.M. took over as Director of the Unit. Dr. Arnold Chamove has made many important contributions to colony practices. We thank Ian Rodgerson and Evelyn Halloren for taking excellent care of the monkeys over the years, J. Barnett, J. Russell, and R. Scott for technical assistance, and R. Anderson for veterinary advice. We are grateful to Arnold Chamove, Sign Evans, Anna Feistner, Eleanor Moodie, and four anonymous reviewers for helpful comments on earlier drafts of the manuscript. E.C.P. was supported by a research scholarship from the Carnegie Trust for the Universities of Scotland for 1986-89. REFERENCES Brand, H.M. Influence of season on birth distribution in marmosets and tamarins. LABORATORY ANIMALS 14:301-302, 1980. Carroll, J.B. Breeding the cotton-topped tamarin Suguinus oedipus oedipus a t Jersey Wildlife Preservation Trust. DODO, JOURNAL OF THE JERSEY WILDLIFE PRESERVATION TRUST 20:48-52,1983. Chamove, A S . Establishment of a breeding colony of stumptailed monkeys (Mucucu urctoides). LABORATORY ANIMALS 15: 252-259,1981. Clapp, N.K.; Tardif, S.D. Marmoset husbandry and nutrition. DIGESTIVE DISEASES AND SCIENCES 30:17%23S, 1985. Evans, S. Breeding of the cotton-top tamarin Suguinus oedipus oedipus: A comparison with the common marmoset. ZOO BIOLOGY 2:47-54, 1983. Evans, S. Captive management of marmosets and tamarins. Pp. 250-252 in STANDARDS IN LABORATORY ANIMAL MANAGEMENT, PART 2. Laboratory Animal Science Association; Universities Federation for Animal Welfare, eds. Potters Bar, Herts, UFAW, 1984. French, J.A. Lactation and fertility: An examination of nursing and inter-birth intervals in cotton-top tamarins (Suguinus 0. oedipus). FOLIA PRIMATOLOGICA (BASEL) 40:276-282, 1983. Garber, P.A.; Moya, L.; Malaga, C. A preliminary field study of the moustached tamarin monkey (Saguinus mystax) in north- eastern Peru: Questions concerned with the evolution of a communal breeding system. FOLIA PRIMATOLOGICA (BASEL) 42:17-32, 1984. Goldizen, A.W. Facultative polyandry and the role of infant-carrying in wild saddleback tamarins (Suguinus fuscicoZZis).BEHAVIORAL ECOLOGY AND SOCIOBIOLOGY 20:99-109, 1987. Kilborn, J.A.; Sehgal, P.; Johnson, L.D.; Beland, M.; Bronson, R.T. A retrospective study of infant mortality of cotton-top tamarins (Saguinus oedipus) in captive breeding. LABORATORY ANIMAL SCIENCE 331168-171, 1983. Kirkwood, J.K.; Epstein, M.A.; Terlecki, A.J. Factors influencing population growth of a colony of cotton-top tamarins. LABORATORY ANIMALS 17:35-41, 1983. Kirkwood, J.K.; Epstein, M.A.; Terlecki, A.J.; Underwood, S.J. Rearing a second generation of cotton-top tamarins (Suguinus oedipus oedipus) in captivity. LABORATORY ANIMALS 19:269-272, 1985. McGrew, W.C. Parental division of infant caretaking varies with family composition in cotton-top tamarins. ANIMAL BEHAVIOUR 36:285-286, 1988. McGrew, W.C.; McLuckie, E.C. Philopatry and dispersion in the cotton-top tamarin, Suguinus (0.)oedipus: An attempted laboratory simulation. INTERNATIONAL JOURNAL OF PRIMATOLOGY 7: 399-420, 1986. McGrew, W.C.; Webster, J. Birth seasonality 12 I Price and McGrew in cotton-top tamarins (Suguinus(0.1 oedipus) despite constant food supply and body weight. JOURNAL OF ANIMAL ECOLOGY (submitted). McKenzie, S.M.; Chamove, AS.; Feistner, A.T.C. Floor-coverings and hanging screens alter arboreal monkey behavior. ZOO BIOLOGY 5339-348, 1986. Neyman, P.F. Aspects of the ecology and social organization of free-ranging cotton-top tamarins (Suguinus oedipus) and the conservation status of the species. Pp. 39-71 in BIOLOGY AND CONSERVATION OF THE CALLITRICHIDAE. D.G. Kleiman, ed. Washington DC, Smithsonian, 1978. Neyman, P.F. ECOLOGY AND SOCIAL ORGANIZATION OF THE COTTON-TOP TAMARIN (Suguinus oedipus). PhD thesis, University of California, Berkeley. Ann Arbor MI, University Microfilms International, 1980. Price, E.C. Birth and perinatal behaviour in captive cotton-top tamarins (Suguinus 0. oedipus). INTERNATIONAL JOURNAL OF PRIMATOLOGY 8:482, 1987. Price, E.C.; McGrew, W.C. Departures from monogamy: Survey and synthesis from colonies of cotton-top tamarins. FOLIA PRIMATOLOGICA (BASEL) (submitted). Scullion, F.T. Husbandry, breeding and maintenance of a viable population of cotton-top tamarins (Suguinus 0. oedipus). ANIMAL TECHNOLOGY 38:167-174, 1987. Snowdon, C.T.; Savage, A.; McConnell, P.B. A breeding colony of cotton-top tamarins (Suguinus oedipus). LABORATORY ANIMAL SCIENCE 35477-480, 1985. Tardif, S.D.; Carson, R.L.; Clapp, N.K. Breeding performance of captive-born cotton-top tamarin (Suguinus oedipus) females: Proposed explanations for colony differences. AMERICAN JOURNAL OF PRIMATOLOGY 11:27 1-275,1986, Tardif, S.D.; Colley, R. INTERNATIONAL COTTON-TOP TAMARIN STUDBOOK. Oak Ridge, TN, Oak Ridge Associated Universities, 1988. Tardif, S.D.; Richter, C.B.; Carson, R.L. Effects of sibling-rearing experience on future reproductive success in two species of Callitrichidae. AMERICAN JOURNAL OF PRIMATOLOGY 6:377-380, 1984a. Tardif, S.D.; Richter, C.B.; Carson, R.L. Reproductive performance of three species of Callitrichidae. LABORATORY ANIMAL SCIENCE 34:272-275,1984b. Ziegler, T.E.; Bridson, W.E.; Snowdon, C.T.; Eman, S. Urinary gonadotrophin and estrogen excretion during the postpartum estrus, conception, and pregnancy in the cotton-top tamarin (Suguinus oedipus oedzpus).AMERICAN JOURNAL OF PRIMATOLOGY 12:127-140, 1987.