Chapter 2 - Welfare and Well-Being of Captive Elephants
Chapter 2 Welfare and Well-Being of Captive Elephants: Perspectives from Wild Elephant Life Histories Phyllis C. Lee1,2 and Cynthia J. Moss2 Behaviour and Evolution Research Group, Department of Psychology, University of Stirling1 and Amboseli Elephant Research Project, Amboseli Trust for Elephants2 There are a number of current debates in the zoo world focused on elephants and their well-being and welfare. One of these relates to enclosure size as discussed by Poole and Granli (Chapter 1); another to protected contact—its benefits and costs (e.g., Whittaker & Laule, Chapter 13); another to social needs. Yet others emphasize illness and causes of mortality in relation to natural or expected life span (see Clubb & Mason 2002). A final issue of controversy is that of captive breeding. The social and environmental factors affecting reproduction, survival and longevity in the wild are the focus of this chapter, with the aim of contributing to the development of an understanding of these issues for captive elephants. No single factor determines well-being, and wild elephants exhibit a huge range of behaviors in response to their constantly changing social and physical environments. We aim here to present some of this range, and explore their underlying dynamics. Some in the zoo community have suggested that the environment of captive elephants is that of captivity, and thus there is little that knowledge of wild elephant behavior and life history can contribute to captive welfare (Hutchins 2006). It is also the case that relatively few studies have yet to address enrichment (environmental or behavioral) for captive elephants (Shepherdson 2003). There are few quantitative assessments of whether protected contact reduces stress and behavioral “anomalies,” although the qualitative evidence is compelling (see Whittaker & Laule, Chapter 13). A few studies have attempted to assess physiological markers of stress in captive elephants (Brown 2000; Stead, Meltzer & Palme 2000; Ganswindt, Heistermann, Palme, Borragan et al. 2003), but we do not yet have a sufficient baseline on the physiology of environmental or social stress to make firm conclusions about what causes stress or when it occurs. Indeed, these aspects of captive life need urgent attention and research if managers aim to provide rich and rewarding environments for existing captive elephants. Despite our lack of quantitative knowledge about the captive environment, elephant biology and behavior reflect their evolutionary ecology; therefore, what they do in the wild, how they do it, how often and why can inform us about best practices for elephants in captivity. Our goal here is to illustrate vital elements of the life cycle for wild African savanna elephants (Loxodonta africana) to provide a context for comparisons between captive and wild conditions. We also explore the differences between male and female elephants’ life cycles to illustrate sex differences and the complex dynamics of an elephant’s long life. 22
Phyllis C. Lee & Cynthia J. Moss
We have not made ecological comparisons of the use of space by either sex, or attempted to determine space “needs.” Rather we have focused on those social and physical factors that underlie well-being and thus which should inform spatial requirements. Suffice it to say that wild African elephants are active for approximately 18-20 hours each day, that the movements of males and females are determined both socially and ecologically and that elephants can travel as far as 50-80 kilometers in a day in a home range as small as 10 or as large as 20,000 km2 (see Douglas-Hamilton, Krink & Vollrath 2005; Leggett 2006; Poole & Granli, Chapter 1). We suggest that it may be more important to provide enough space for individuals to have the choice to be social or solitary as well as some landscape diversity, than to provide large but unchallenging areas lacking natural forage, shade or water features. Thus, the aims of this chapter are as follows: 1) To explore the differences between elephant male and female life cycles in order to illustrate the complexity of their long life span. The reproductive histories of males and females are placed into the context of their life cycles, specifically addressing those factors affecting reproductive onset and sexual function. 2) To discuss socialization and the compelling evidence that social learning opportunities are vital to the expression of normal behavior (e.g., foraging, infant care, dominance and mating). 3) Finally, to examine longevity and survival and ask why some elephants have such long lives in the wild. It is worth emphasizing that a significant difficulty in captive–wild comparisons of longevity has to do with cause of death; humans actively manage the life span of their captive elephants often for reasons unrelated to “normal” biology. The question for managers is why do so many captive elephants experience physical or behavioral abnormality, illness or injury to the point where euthanasia is a major cause of captive mortality, and why does this occur at a relatively young age? General methods The results we present here derive primarily from the Amboseli Elephant Research Project (AERP), and these are compared with some other wild African and Asian (Elephas maximus) elephant populations. The AERP was started by Cynthia Moss and Harvey Croze in 1972, with identification of individuals and records of basic demographic, social and ecological data. The last family was completely identified by 1982. Over 2,200 elephants have been individually recognized and the population is monitored continuously. Births and deaths, and, opportunistically, estrus, musth and copulations, have been recorded from 1972 to the present. Focal animal samples of behavior and interactions, as well as systematic scan samples of activities and associations, are used to present details of behavior. The analyses presented here are based on events up to the end of 2004 (C.J. Moss, unpublished long-term records). Age is known for individuals under the age of 35 (using known birth dates) and is estimated reliably for those born before 1972 based on shoulder height, back length, tusk eruption and growth; estimated ages are reassessed periodically by using early photographs and reference to the maturational changes among the known-aged sample. Finally, for jaws found post-mortem, estimated age was checked against tooth age (see Moss 2001). Animals with estimated ages represent only 15 percent of the current sample (346 individuals out of 2,173). Cause of death, both natural and human-related, was recorded and accuracy of knowledge about death cause and death dates was noted. Human-caused deaths have been separated in the analyses that follow (N = 348). 23
Chapter 2 - Welfare and Well-Being of Captive Elephants
We have used Kaplan-Meier proportional hazards analyses, assuming that hazards ratios were not constant over time, to examine mean and quartiles of survival time for first birth, first musth and longevity, with standard error and Brookmeyer-Crowley 95 percent confidence intervals. Where covariates were explored in survival analyses, we used Cox Regression and present Wald statistics. We used logistic regression analysis to explore probability of death between groups of individuals, and the Wald statistic was used as goodness-of-fit for the model functions. Univariate ANOVA was used to explore main effects and interaction of categorical variables on continuous measures. All data were tested to ensure normal distributions, and were log-transformed for normality before using parametric tests. Non-parametric Spearman correlations were used to test relationships between two variables where a normal distribution could not be approximated. All probabilities are two-tailed and set at p<0.05 for significance. Statistical analyses were made with SPSS version 13.0. Male life cycle One of the basic characteristics of the male life cycle (Figure 1) is the very long period of social and physical maturation (Lee & Moss 1999). Dependence on the family prior to social maturity and leaving the natal family can be as long as 18 years. In populations with natural age structures, including large, old males, an additional post-independence period of another 10 to 15 years is common before a male commences reproductive musth cycles. Reproductive onset
F IR S T MUS TH B irth-7yrs : s exually immature; with natal family; mean s urvival = 77%
8-16 yrs : s perm ons et c .10 yrs ; s perm in quantity 16+yrs ; depart family c . 8-16 yrs ; mean s urvival = 93%
17-25 yrs : interes t and as s oc iation with females ; mean s urvival = 90%
25-35 yrs : 1s t mus th c . 30 yrs ; as s oc iation w/ female groups alternating with bull groups ; mean s urvival = 81%
>35 yrs : Fully c ompetitive with other mature bulls ; more attrac tive to females ; mate guarding obs erved; mean s urvival 36-50 = 54%
S E XUAL L Y AC T IVE
0
10
20
30
40
Age Age (years)
50
60
Figure 1: Male life cycle based on 32 years of observations in Amboseli National Park, Kenya (Poole, Lee & Moss in press). Survival for each age group is based on the likelihood of surviving from age one to age two for natural mortality only. Male survival to age seven = cumulative probability of surviving to seven years for male calves. 24
70
Phyllis C. Lee & Cynthia J. Moss
(spermatogenesis) for males occurs around eight to 10 years of age, but at this point males are only two-thirds the size of a fully adult female, and thus might have some physical difficulty mating. By the time a male is regularly coming into musth (35+ years; Poole 1987) he is one and one-half times the size of a large adult female (Figure 2). Male African elephants exhibit indeterminate growth and increase in stature for the majority of their lives. Males grow faster than females from infancy and for longer over the life span (Figure 2), with obvious energy costs to sustaining this growth (Lee & Moss 1995). In the wild, “bigger” is better, with larger individuals more likely to survive and to commence reproduction earlier (Lee, Poole & Moss in progress). While there is considerable variation in final stature among individuals and across populations, the shape of the growth curve up to 15 years is consistent across a wide range of populations (Shrader et al. 2006). Growth in early life is probably an excellent indicator of the mental and physical health of individuals. In Amboseli, males leave their families at an average of about 14 years of age, with considerable individual variation (Figure 3); some will spend short periods away even as young as six years. In forest elephants (Loxodonta cyclotis), males who appear to be weaned only recently (four to six years of age) travel on their own or form groups with other bulls, both large and small (P. Lee, pers. observ—Petit Loango, Gabon Jan. 1997; Fishlock, pers. comm. April 2005—Mbeli, Congo). At this stage we cannot yet determine what causes some males to leave their families before they are
Figure 2: Growth in stature (shoulder height) for 548 measurements on 314 males and 436 measurements on 297 female Amboseli elephants (P.C. Lee, AERP long-term records).
25
Chapter 2 - Welfare and Well-Being of Captive Elephants
even half the size of a female and others to stay until they are as large as their mothers. Departure appears to be the result of a large number of factors such as maternal status, family size, male friendships with peers, relative and absolute size, confidence and other individual characteristics which interact with the local ecology and population densities. While it is not yet possible to describe “rules” governing male dispersal, wild males forcibly removed from a family social context and lacking experience with older males can become exceptionally aggressive to other elephants and other species (Slotow et al. 2000; Slotow & van Dyk 2001; Bradshaw et al. 2005). Male social life In Amboseli, males tend to live in areas that are ecologically distinctive, and which often are less central to female areas. Asian and African males appear to engage in risky behavior in general, wandering into areas where contact with humans is likely, crop raiding, and generally covering larger ranges, as well as encountering other males who may be competitors (Sukumar 1989; Tchamba et al. 1994; Osborn 1998). Despite the impression of “lone” bulls, Amboseli males are solitary for less than one-third of the time, and are typically (c. 20 percent of time, depending on age) found in small groups of two or three males (see also Evans 2006). All-bull groups can, however, be comprised of over 40 animals (J.H. Poole, AERP long-term records). Being solitary or in small groups may, again, increase males’ susceptibility to environmental risks or hazards from
110
100
90
Mean (95 CI)% Time Away
80
70
60
50
40
Mean (95 CI) % Time away 30 20
10 0 N =
52
50
65
79
91
94
93
84
77
71
69
67
46
44
5
6
7
8
9
10
11
12
13
14
15
16
17
18
Figure 3: The age-related patterns of departure from the family (Lee & Moss 1999; Lee, Poole & Moss in press). 26
Phyllis C. Lee & Cynthia J. Moss
humans. Newly independent males tend to spend most of their time (c. 70 percent; Lee, Poole & Moss in press) with female (non-mother) groups. Even males aged 25-50 years associate regularly with females, spending 25-50 percent of their time with these groups (Figure 4). Thus, males are not asocial; they associate regularly with female groups, form groups with other males and have consistent male associates in these groups (friends: Lee, Poole & Moss in press). Males can and do “get along” with other males, and are well tolerated by females. Male elephants in captivity For many social mammals with lengthy periods of development, such as primates or elephants, rearing in contexts that lack some proportion of the age-sex range typically found in social groups in the wild tends to produce individuals with inappropriately extended juvenile-type behavior. In particular, play, aggression and sexual interactions are frequent and misdirected (Slotow et al. 2000; La Prairie 2001; Slotow & van Dyk 2001; Suomi 2003). Thus it would seem reasonable to leave males with familiar or family females until the age at which they become sexually active. In the wild, females are relatively rarely aggressive towards family males before dispersal, but they do target non-natal males with escalated aggression, occasionally acting in concert with other females to poke, chase or drive away the young males. It might be suggested that some experience as the target of aggression by older animals may be vital to a male’s ability to learn to respond appropriately to aggressive interactions, suppressing hyperaggressive responses. In captivity, the rapid and prolonged growth of males could pose health risks. “Big” can become too large with attendant negative physical and physiological consequences for health. Obesity and joint and foot problems may be associated with reduced potential for movement, leading
60
% sightings by sex (N = 29894)
50 40 30 20 10
% sightings by sex (N = 29894) 0 Male S olitary
Bull Group
MixedGroup group Mixed
Cow-calf Cow-calf group Group
Figure 4: The average grouping patterns of Amboseli elephants as a percent of sex-specific sightings (all ages of independent males combined; AERP long-term records; N = Male groups = 18,490; N= female groups = 21,716).
27
Chapter 2 - Welfare and Well-Being of Captive Elephants
to further negative consequences for health. In captive contexts males may be seen as less “desirable”; they are more aggressive and therefore dangerous to keepers and other elephants. Given limited space, they can be difficult to house socially with females or other males once regular musth sexual activity commences. Males have both social needs and strong sexual and competitive motivations which make keeping them in naturalistic conditions problematic unless there is sufficient space for separate areas where males can withdraw from others (bull or “retirement” areas: Moss & Poole 1983) as well as opportunities and space to make social contacts when individuals seek companionship. As we emphasize, social dynamics underlie spatial needs especially for adult males. Welfare considerations require that the varied social and associated spatial needs of males be met in the design of captive enclosures. Female life cycle Elephant females in the wild can have extremely long reproductive and social lives (Figure 5). Females are less prone to mortality than are male elephants (Moss 2001; see below), although the energy costs of reproduction and lactation probably contribute to some increased natural mortality risks, especially during droughts. The average female born in Amboseli can survive until her late 40s or early 50s (Moss 2001; see below) and give birth to a calf every four to five years (median interbirth interval = 50 mo, range = 22-114 mo, N = 949: Moss & Lee in press a). Over 80 percent of these calves will survive to age five, and three generations of grandmothers, mothers, daughters and granddaughters form the basis of the stable savanna elephant family (see also Archie, Moss & Alberts 2006).
B irth-8 yrs: Probability of female survival to 8 = 83%
9-20 yrs: F irst birth; IB I to 2nd calf = 57.5 mo F irst calf survival to 8 = 62% ; Mean survival = 95%
20-50 yrs: E xperienced prime mothers IB I = 50 mo C alf survival to 8 = 79% Mean survival = 70%
50-65+ yrs: Old females, declining fertility & grandmothering IB I = 54.5 months C alf survival to 8 = 57%
S E XUAL L Y AC T IVE
0
10
20
30
Age
40
50
60
70
Figure 5: Female life cycle based on 32 years of observations in Amboseli National Park, Kenya (Moss & Lee in press a). Survival for each age group is based on the likelihood of surviving from age one to age two for natural mortality only. Calf survival to age eight = cumulative probability of surviving to eight years for both male and female calves for each category of female by parity or age. 28
Phyllis C. Lee & Cynthia J. Moss
In Amboseli, a female first gives birth at an average of 14 years, although females as young as nine have produced calves (Figure 6). Given a 22-month gestation, these younger females experienced their first fertile estrus at around seven years of age. As the figure suggests, the clustering around 14 years is quite marked; only four percent of females give birth when younger than 11 years, and seven percent when older than 16 years. What makes a female reproductively successful? In Amboseli, as elsewhere in Africa, females tend to show considerable seasonality in their fertility patterns (Lee, Lindsay & Moss in press). Low quality diets, poor body condition and time and energy spent searching for food all appear to reduce conception probabilities, and females take about three months to regain condition and conceive after a long dry season (Foley, Papageorge & Wasser 2001; Rasmussen, Wittemyer & Douglas-Hamilton 2006). In Amboseli, first-time mothers have reduced success in rearing infants, with 33 percent of firstborn calves dying early in life (under 24 months of age—the age of nutritional dependence from the mother) compared to only 13 percent for experienced mothers (Moss & Lee in press a). In the wild, a female at 12-14 years will be only 70 percent of her full adult height and will have another 20 years of growth ahead of her at the same time that she is reproducing, suggesting additional energetic and/or metabolic costs associated with growth for these young females. Inexperience also plays a role in the poor reproductive performance of first mothers; even in the wild, with extensive practice in caring for calves as an allomother or helper, they still have difficulties with calf suckling, with learning calf signals of hunger or distress, and with ensuring that calves remain close by at all times. As mothers age and gain experience with calves, their calves have higher survival rates and mothers appear to learn to discriminate in favor of behaviorally-demanding, faster-growing sons who receive a greater allocation of milk and are weaned later (Lee & Moss 1986; Lee & Moss in press). Thus, experienced mothers of sons tend to have longer intervals (55 mo) between successive births. Another critical factor in calf survival is whether the mother has helpers. Allomothers, nulliparous females aged five to 15, contribute significantly to calf survival in Amboseli, even Figure 6. Proportional hazards model of age at first reproduction for 326 females with known ages at first birth (see also Moss 2001); median age = 13.9 (95 percent CI = 13.6-14.2).
29
Chapter 2 - Welfare and Well-Being of Captive Elephants
when absolute family size is taken into account (Moss & Lee in press a). In the wild, a variety of predators—lions, hyenas, wild dogs and, throughout the millennia, humans—prey on vulnerable young calves. At birth a wild calf will only weigh 70-150 kg (Hanks 1979), making it easy prey for large social predators. Allomothers, grandmothers and the family—the social context for calves— are key to ensuring calf survival (Lee 1987; Lee & Moss in press). Along with their defensive functions, allomothers are gaining experience that can contribute to their ability to rear their own calves, although allomothering experience alone appears insufficient to ensure firstborn survival rates that are comparable to those for larger, older individuals. Female sociality Among savanna elephants, females tend to stay socially bonded and closely associated with members of their family from birth to death. These genetic and affiliative bonds are the bedrock of sociality upon which is layered a rich and dynamic domain of interactions and structures. Groups of female elephants, by contrast to family units, are fluid aggregations of individuals coming and going, including one to many adult males in almost half of occurrences (see Figure 4), where the individuals move together and apart, interact with and explore each other. Families can associate with other families in such large aggregations. In Amboseli, every family has been seen at least once with each of the other 50+ families in the population (Moss & Lee in press b). Social knowledge extends to almost all others within a population of some 1,200 individuals. There are occasions when family females are not all together in the same social group. Some members of a family will form a fragment or a sub-unit, and each fragment may associate separately with different larger groups. The fission and fusion of individuals occurs over a full 24-hour period (Figure 7). The tendency to fragment and associate with non-family groups contrasts with a tendency to remain in a single coherent and cohesive unit, either in a group with just other family members or to join with others in larger groups as an intact family. This latter inclination to group with others, to be gregarious, is also highly variable among different families. In Amboseli, Figure 7: The group size experienced by one female, Sandy, of the SA family over a 24-hour period. The SA family consisted of 11 individuals, including four adult females (Lee and Lindsay, pers. observ. March 16-17, 1984).
30
Phyllis C. Lee & Cynthia J. Moss
cohesiveness of families ranges from 100 percent to less than 50 percent of time together. Further, the larger the family and harsher the environmental conditions, the less cohesive families become (Moss & Lee in press b). The affinity for gregariousness—being with other elephants—is a basic elephant “need.” Finding a female completely on her own, when she is not ill or in estrus, is extremely rare in Amboseli. Even female forest elephants, known for their smaller family unit and group sizes, are likely to be found with at least one and often two other elephants (Theuerkauf, Ellenberg & Guiro 2000; Morgan & Lee 2007). Asian elephants tend to form groups of six to eight individuals (Vidya & Sukumar 2005). In some habitats, even savanna elephants also have very small but nonetheless cohesive families (e.g., 4-6 individuals: Mikumi, Tanzania; Mofulu 2005). The point here is that female elephants are highly dependent on relationships with other elephants; these relationships are based on genetic kinship (Archie et al. 2006) and extensive networks of knowledge (McComb et al. 2000) and persist over a female’s lifespan. Relationships are fundamental to negotiating elephants’ complex ecological and social environments (see Archie et al. 2006) as well as their success in rearing calves as discussed above. Female elephants in captivity Keeping an elephant female in captivity is a financial, management and spatial commitment that may outlast the housing institution’s future, and is certainly longer than the working life span of the average keeper. Given the importance of the family to reproduction and survival, it would seem essential to keep social units intact to act in some form as a family to ensure stable structures and reduce social stresses. Unfortunately, there appears to be relatively little debate in the zoo community about female elephants’ needs for social stability and continuity; they are moved about constantly (one might suggest like “furniture”) in the interests of institutions and exhibits, not elephants’ social needs. While arguments can be made about the need to manage genetic diversity, since males and females are almost never housed together for life, inbreeding seems unlikely. And, if as we argue here, females reproduce more effectively in families, then using arguments about genetic management could be in conflict with strategies for captive breeding. Can families be created? In the wild, females can be left on their own due to natural demographic processes (Moss & Lee in press b) or due to catastrophic human intervention such as in Queen Elizabeth National Park, Uganda (Nyakaana et al. 2001) or Akagera National Park, Rwanda (K. Fawcett, pers comm. Dec 2006). Under such conditions, unrelated females will form long-term cooperative bonds that act and appear structurally as families. Allowing opportunities for both contact (potentially over 24 hours) and sufficient space for individuals to choose to be apart as well as together can facilitate the formation of family-like sociality in captivity. Once established, however, the trauma of subsequent separation should be recognized and accounted for in long-term group management. It is vital to note that elephant social and environmental experiences in the wild are not confined to the hours that keepers work. One significant improvement to any management routine would be to allow social access, as well as opportunities for social escape, on a 24-hour basis. This means restructuring indoor and outdoor facilities to ensure group and individual access ad libitum with associated costs and security considerations. With regard to captive reproduction, we have noted a number of factors that affect fertility and calf survival in the wild. Fat females who exercise little, as in captivity, would not be expected to show any seasonality to conception. However, obesity and lack of exercise may result in other stressors or cause hormone imbalances that reduce estrus frequency or efficiency and thus fertility 31
Chapter 2 - Welfare and Well-Being of Captive Elephants
probabilities. Although the physiological controls on puberty are poorly understood for both wild and captive elephants, the relatively narrow age window within which we observe first reproduction in the wild might suggest why females in captivity have such difficulties conceiving when they are not given access to breeding opportunities until over the age of about 18. When institutions attempt captive breeding, we also suggest that considerable care needs to be taken to ensure that females have had some experience with calves prior to their first pregnancy. Given the relatively lower calving success of primiparous mothers compared to experienced mothers even in the wild, losses are to be expected in a captive context. Managers need to be clear about the costs and benefits of captive breeding, including calf death, poor publicity and sustained interventions for hand rearing which will perpetuate the cycles of calf mortality due to inexperience. Immature females—the typical and essential allomothers in the wild—have not been observed to threaten or hurt calves in Amboseli. Protective and defensive interactions on the part of allomothers are triggered by infant distress vocalizations, which cause the allomothers to crowd in near the calf and to contact the calf with their feet, trunks or body. Such contacts are clearly defensive rather than aggressive and act to ensure that the calf is on its feet, close to a protector and moving away from any potential threats. If such behavior can be anticipated, then successful allomothers could be created by allowing unrestricted access to new calves, rather than attempts to check what may appear to keepers to be overly enthusiastic interest in newborns. The social context for learning and development The social context is vital for much learning in elephants. Learning what vocalizations represent (their meanings) occurs within groups; spatial knowledge—where resources such as water and food are located, how to move on safe and unsafe routes, where areas of risk from predators or humans exist—all of these require repeated exposure which occurs initially in a family and in the larger group context (Moss 1988). Interestingly, males can and do wander into “unfamiliar” or high-risk areas after leaving their families (Lee, Poole & Moss in press), but as we discussed above, this may contribute to their greater vulnerability to mortality risks. Elephant calves also learn about foods socially: their temporal availability, processing and handling requirements, and quality (Lee & Moss 1999). Almost all foods are first experienced in a social context, since calves begin to forage at around six months but cannot sustain themselves energetically until after two years of age. Younger calves do not have the strength or height to forage on many of the foods in the wild, and thus depend on sampling (and scrounging from) others in the form of trunk-to-mouth contacts with adults while they are feeding (Lee 1987). Sexual interactions are first observed in a group context, and their performance is practiced from infancy via mounting and chase play (Lee 1986). Knowledge of familiar individuals, of friends, enemies, competitors and of one’s own size and strength are all gained through social interactions within and among families. Social experiences in the form of interactions and play have gender and age-specific patterns to their development, as in many species (Figure 8). As calves age, males in particular seek out novel partners for interactions, moving beyond the confines of their family and testing their strength and abilities with a variety of partners of different ages (Lee 1986; Lee & Moss 1999, in press). Deprive calves of social experience via play and greetings, of social novelty with partners of a variety of ages and sex, of the patterns of fission and fusion that make up a daily social experience for both males and females over the course of their development, and the likely result is an individual with deficient social, sexual or maternal skills. In effect, this produces an isolate relying on 32
Phyllis C. Lee & Cynthia J. Moss
forms of self-stimulation such as stereotypies, and unpredictable interactions with human keepers and visitors for social contact (Carstead 1996). Longevity and lifespan In the wild, maximum longevity is as yet unknown, but it is likely to be in the region of 70 years. Media reports in 2005 of a known-aged Asian zoo elephant in Thailand dying at age 86 due to tooth loss and a living Asian elephant in North America estimated to be 77 years old in 2005 (Weise & Willis 2005) make an estimate of 70 years maximum longevity reasonable for African elephants in the wild. Among African elephants, longevity is a key factor in reproductive success for both males and females (Poole 1989; Moss & Lee in press a; Poole, Lee & Moss in press). For an elephant, being “old” equates to being large, powerful, knowledgeable and dominant (Archie et al. 2006). However, after 50, physical aging begins to reduce fertility and paternity probabilities, and the costs of being social (group travel, food competition, interactions) have an influence on the success and status of aging elephants. One contentious issue with regard to welfare is that of longevity in captive elephants. Some analyses have suggested that mean age at death is low at about 16 years in captive European Male Social Contacts 7 6
Hourly Rate
5 4 3 2 1 0 0-12
13-24
25-36
37-48
49-60
37-48
49-60
Age (mo)
Female Social Contacts 7 6
Hourly Rate
5 4 3 2 1 0
0-12
13-24
25-36 Age (mo)
33
Figure 8. Hourly rates of aggressive contacts (poke, chase), friendly interactions (rub, greet, investigate food, invite play) and social play for male and female elephants between birth and five years old (Lee & Moss in press: N observation hours Male = 119; Female = 145)
Chapter 2 - Welfare and Well-Being of Captive Elephants
collections (Clubb & Mason 2002), while others have found an average lifespan of 33 years for North American African elephants (Weise & Willis 2005). While Clubb and Mason’s (2002) analysis was only on the age for dead animals, both analyses excluded individuals that died under one year of age, which is a substantial proportion of mortality in the wild (10.5 percent of calves die < 12 mo; 13.8 percent die <24 mo), and their exclusion will obviously increase average longevity. Here, in order to present some baseline data for expectations, we present average longevity from survival analyses in the Amboseli population. As mentioned above, males and females have different survivorship probabilities and different age-specific mortality hazard rates. It is therefore necessary to consider the average longevity of sexes separately. Even when animals died under one year of age, all known live births have been included in the analyses. In Moss (2001), average longevity was given as 41 for females and 24 for males. These figures included deaths known to result from human activities (control shootings, spearings, accidents due to human agency: N = 384). When these are excluded, the Cox proportional hazards model shows that median female longevity is 54.4 years (95 percent CI = 49.5-59.3, mean = 46.7, 95 percent CI = 44.1-49.4, N = 846), that for males is 42.5 (95 percent CI = 35.3-49.8, mean = 36.4, 95 percent CI = 33.1-39.8, N = 721) (Figure 9). Using different hazards models does not change the results to any extent; the data are robust and the sample sizes large. Even with environmental
A
1 0 .9 0 .8
Cumulatvie survival
0 .7 0 .6 F e m a le 0 .5
M a le
0 .4
0 .3 Cumulative survival 0 .2 0 .1 0 0
10
20
30
40
50
60
70
80
A g e (y e a r s )
B
0 .1 4
0 .1 2
0 .1
Hazard Rate
Figure 9. A) Survival probabilities for male and female Amboseli elephants excluding human-caused mortality. B) Hazard rate (smoothed lines; C.J. Moss, AERP long-term records).
0 .0 8 F e m a le M a le 0 .0 6
Hazard Rate 0 .0 4
0 .0 2
0 0
10
20
30
40
A g e (y e a r s )
34
50
60
70
Phyllis C. Lee & Cynthia J. Moss
events that cannot occur in captivity, such as catastrophic droughts or floods, or illnesses that are treated in captivity, and including the major first year mortality, elephants in the wild or at least in Amboseli are surviving far longer than the average captive elephant. Captivity does not appear to enhance an elephant’s chance of survival. Managers therefore need to ask why their elephants are dying so young. Other key points are the relatively higher chances of death for males by comparison to females between birth and weaning, along with increased rates of mortality throughout life. Male mortality hazard rates increase at dispersal and again with the onset of reproductive competition, associated with their increased exposure to social and environmental risks. This sex differential in mortality in later life may not be a major problem in a captive context, but the increased vulnerability of male calves in the first several years of life should be noted. Males grow rapidly early in life and their energy intake needs are high in order to sustain this growth. If access to milk or other foods is limited, their risk of death is higher than that of female calves. They could also potentially be more physiologically vulnerable to illness due to energy imbalance and high growth rates. Conclusions If elephants are to be maintained in captivity, then a set of best, rather than minimum, practices is obviously necessary to ensure their welfare. These best practices must surely go beyond providing scatter-feeding and a 200 m2 indoor and outdoor enclosure. Allowing elephants to lie down in some 40 m2 of space may be a considerable advance in well-being over nighttime chaining, but none of these practices addresses the full range of social or physical needs of elephants as biological organisms. Defining best practices, as recently attempted by the AZA, is obviously a significant step forward in captive elephant welfare, but some form of benchmark for assessing elephant “needs” is urgently required, and wild behavior can at least provide key insights into these needs. If we use just a few of the traits of wild elephants that we detailed above, we can suggest that elephants are allowed free access to indoor and outdoor areas over a 24-hour period; that they be given the choice as to whether and when to associate with their social companions over this entire period; that some attempt is made to provide mixed age and sex groups, which are stable over the very long-term, all of which facilitate family-like associations in a fission-fusion social context. Loss of individuals (e.g., through mortality) is a “natural” feature of elephant life cycles, but this does not imply that natural loss is without consequences for the well-being of even wild animals. Family disruption and fission after loss such as that of a matriarch is common (Moss 1988; Moss & Lee in press b), and places the surviving individuals at some greater risk of mortality. Whenever individuals can be maintained socially together, this should be a priority. As we note above, elephants depend on their social companions in almost all contexts; it is vital to their welfare that managers avoid breaking up established groups. Challenging spatial environments, timetables that are somewhat unpredictable and so mimic natural foraging and group dynamics, social consistency and cohesion, and a diverse social assemblage of others bringing experiences across a range of ages, interactions, activities and contexts—if these can be met in captivity, then a start towards the goal of best practices will have been made. Choice of space, choice of companions, choice of foods: what we emphasize is that priority should be given to elephant needs rather than to those driven by institutional fiscal requirements or keepers’ timetables. If elephant welfare conflicts with institutional priorities, then institutions need to consider whether keeping elephants in either the short- or the long-term is appropriate.
35
Chapter 2 - Welfare and Well-Being of Captive Elephants
Acknowledgments We thank the Government of Kenya and the Kenya Wildlife Service for research clearance and permission to work in Amboseli National Park, and the people of Amboseli for their hospitality. S. Sayailel, N. Njiraini and K. Sayailel contributed importantly to data collection. Funding for the long-term monitoring was provided by numerous individual donors, foundations and organizations through the Amboseli Trust for Elephants and the African Wildlife Foundation over the past 34 years. Support for the initial construction of the long-term database was provided by two residential meetings at the National Centre for Ecological Analysis and Synthesis (NCEAS, NSF-UC Santa Barbara). We thank all our project colleagues, especially Harvey Croze for figures and Joyce Poole and Keith Lindsay for their comments; Tufts University Cummings School of Veterinary Medicine’s Center for Animals and Public Policy, and the Coalition for Captive Elephant WellBeing for hosting the meeting on which this volume is based and inviting this contribution; the Symposium sponsors (Royal Society for the Prevention of Cruelty to Animals, Gary Fink, Phoenix Zoo, American Society for the Prevention of Cruelty to Animals, North Carolina Zoo and Oakland Zoo); and Lisa Kane, Paul Waldau, and Debra Forthman for organization and feedback. References
Archie E, Morrison T, Foley C, Moss CJ, Alberts S. 2006. Dominance rank relationships among wild female African elephants. Anim Behav 71:117-127. Archie E, Moss CJ, Alberts S. 2006. The ties that bind: genetic relatedness predicts the fission and fusion of social groups in wild African elephants. Proc Roy Soc Brit 273:513-522. Bradshaw G, Schore AN, Brown JL, Poole JH, Moss CJ. 2005. Elephant breakdown. Nature 433:807. Brown JL. 2000. Reproductive endocrine monitoring of elephants: an essential tool for assisting captive management. Zoo Biol 19:347-368. Carstead K. 1996. Effects of captivity on the behavior of wild mammals. In: Kleiman DG, Allen ME, Thompson KV, Lumpkin S, eds. Wild mammals in captivity: principles and techniques. Chicago, IL: University of Chicago Press. p. 317-333. Clubb R, Mason G. 2002. A review of the welfare of zoo elephants in Europe. Horsham, UK: RSPCA. Douglas-Hamilton I, Krink T, Vollrath E. 2005. Movements and corridors of African elephants in relation to protected areas. Naturwissenschaften 92:158-163. Evans K. 2006. The behavioural ecology and movements of adolescent male African elephants (Loxodonta africana) in the Okavango Delta, Botswana. Ph.D. Thesis. University of Bristol, UK. Foley C, Papageorge S, Wasser S. 2002. Non-invasive stress and reproductive measures of social and ecological pressures in free-ranging African elephants. Conserv Biol 15:1134-1142. Ganswindt A, Heistermann M, Palme R, Borragan S, Hodges JK. 2003. Non-invasive assessment of adrenal function in the male African elephant (Loxodonta africana) and its relation to musth. Gen Comp Endocrinol 134:156-166. Hanks J. 1979. A struggle for survival. London, UK: Hamlyn. Hutchins M. 2006. Variation in nature: its implications for zoo elephant management. Zoo Biol 25:161171. La Prairie G. 2001. Social and behavioural development in a group of wild-born orphan chimpanzees. Ph.D. Thesis. University of Cambridge, UK. Lee PC. 1986. Early social development among African elephant calves. Nat Geo Res 2:388-401. Lee PC. 1987. Allomothering among African elephants. Anim Behav 35:278-291. Lee PC, Moss CJ. 1986. Early maternal investment in male and female African elephants calves. Behav Ecol & Sociobiol 18:353-361.
36
Phyllis C. Lee & Cynthia J. Moss
Lee PC, Moss CJ. 1995. Statural growth in known-age African elephants (Loxodonta africana). J Zool 236:29-41. Lee PC, Moss CJ. 1999. The social context for learning and behavioural development among wild African elephants. In: Box HO, Gibson KR, eds. Mammalian social learning. Cambridge, UK: Cambridge University Press. p. 102-125. Lee PC, Lindsay WK, Moss CJ. In press. Ecological patterns of variability in demographic rates. In: Moss CJ, Croze H, eds. The Amboseli elephant: a long-term perspective on a long-lived species. Chicago, IL: University of Chicago Press. Lee PC, Moss CJ. In press. Calf development and maternal rearing strategies. In: Moss CJ, Croze H, eds. The Amboseli elephant: a long-term perspective on a long-lived species. Chicago, IL: University of Chicago Press. Lee PC, Poole JH, Moss CJ. In press. Male social dynamics: from independence to beyond. In: Moss CJ, Croze H, eds. The Amboseli elephant: a long term perspective on a long lived species. Chicago, IL: University of Chicago Press. Lee PC, Poole JH, Moss CJ. In progress. Size matters: long-term effects of growth on African elephant (Loxodonta africana) survival and success. Leggett KEA. 2006. Home range and seasonal movement of elephants in the Kunene Region, northwestern Namibia. Afr Zool 41(1):17-36. McComb K, Moss C, Sayialel S, Baker L. 2000. Unusually extensive networks of vocal recognition in African elephants. Anim Behav 59:1103-1109. Mofulu F. 2005. The elephant population and distribution in Mikumi National Park, Tanzania. MPhil Thesis. Anglia Polytechnic University, UK. Morgan B, Lee PC. 2007. Forest elephant group composition, frugivory and coastal use in the Réserve de Faune du Petit Loango, Gabon. Afr J Ecol (online March 2007) doi: 10.1111/j.1365-2028.2007.00762.x. Moss CJ. 1988. Elephant memories. New York, NY: William Morrow. Moss CJ. 2001. The demography of an African elephant (Loxodonta africana) population in Amboseli, Kenya. J Zool 255:145-156. Moss CJ, Lee PC. In press a. Female reproductive strategies: individual life histories. In: Moss CJ, Croze H, eds. The Amboseli elephant: a long-term perspective on a long-lived species. Chicago, IL: University of Chicago Press. Moss CJ, Lee PC. In press b. Female social dynamics: fidelity and flexibility. In: Moss CJ, Croze H, eds. The Amboseli elephant: a long-term perspective on a long-lived species. Chicago, IL: University of Chicago Press. Moss CJ, Poole JH. 1983. Relationships and social structure of African elephants. In: Hinde RA, ed. Primate social relationships. Oxford, UK: Blackwells Scientific Publications. p. 314-325. Nyakaana S, Abe EL, Arctander P, Siegismund HR. 2001. DNA evidence for elephant social behavior breakdown in Queen Elizabeth National Park, Uganda. Anim Conserv 4(3):231-237. Osborn FV. 1998. The ecology of crop-raiding elephants in Zimbabwe. Ph.D. Thesis. University of Cambridge, UK. Poole JH. 1987. Rutting behavior in African elephants: the phenomenon of musth. Behav 102:283-316. Poole JH. 1989. Mate guarding, reproductive success and female choice in African elephants. Anim Behav 37:842-849. Poole JH, Lee PC, Moss CJ. In press. Longevity, competition and musth: a long-term perspective on male reproductive strategies. In: Moss CJ, Croze H, eds. The Amboseli elephant: a long-term perspective on a long-lived species. Chicago, IL: University of Chicago Press. Rasmussen HB, Wittemyer G, Douglas-Hamilton I. 2006. Predicting time-specific changes in demographic processes using remote sensing data. J Appl Ecol 43:366-376. Shepherdson DJ. 2003. Environmental enrichment: past, present and future. Int Zoo Yrbk 38:118-124. Shrader AM, Ferreira SM, McElveen ME, Lee PC, Moss CJ, van Aarde RJ. 2006. Structural growth and age determination of savanna elephants. J Zool 270:40-48. 37
Slotow R, Van Dyk G, Poole J, Page B, Klocke A. 2000. Older bull elephants control young males. Nature 408:425-426. Slotow R, Van Dyk G. 2001 Role of delinquent young ‘orphan’ male elephants in high mortality of white rhinoceros in Pilanesberg National Park, South Africa. Koedoe 44:85-94. Stead SK, Meltzer DG, Palme R. 2000. The measurement of glucocorticoid concentrations in the serum and faeces of captive African elephants (Loxodonta africana) after ACTH stimulation. Tydskr S Afr Vet Ver 71(3):192-196. Sukumar R. 1989. The Asian elephant: ecology and management. Cambridge, UK: Cambridge University Press. Suomi SJ. 2003. Gene-environment interactions and the neurobiology of social conflict. Ann NYAS 1008:132-139. Tchamba MN, Bauer H, Hunia A, de Langh HH, Planton H. 1994. Some observations on the movements and home range of elephants in Waza National Park, Cameroon. Mammalia 58:527-533. Theuerkauf J, Ellenberg H, Guiro Y. 2000. Group structure of forest elephants in the Bossématie Forest Reserve, Ivory Coast. Afr J Ecol 38:262-264. Vidya TNC, Sukumar R. 2005. Social organization of the Asian elephant (Elephas maximus) in southern India inferred from microsatellite DNA. J Ethol 23:205-210. Wiese RJ, Willis K. 2005. Calculation of longevity and life expectancy in captive elephants. Zoo Biol (DOI 10.1002/zoo.20011).
Phyllis Lee has a Chair in Psychology at the University of Stirling, having been for many years a Reader at the University of Cambridge, UK. She has been carrying out field research on animal behaviour since 1975, and has been part of the Amboseli Elephant Research project since 1982. She has collaborated with a number of researchers working on forest elephants, and primates from around the world. She is the author of one book, four edited volumes, more than 35 primary journal publications and more than 30 chapters in edited volumes. She works with conservation attitudes and community conservation projects, as well as human-wildlife interactions. She may be contacted at Behaviour and Evolution Research Group, Department of Psychology, University of Stirling, Scotland FK9 4LA. Email:
[email protected]. Cynthia Moss has been the Director of the Amboseli Elephant Research Project, which she founded, since 1972. She gained her first elephant experience with Iain Douglas-Hamilton in 1968, and she now conducts research on the distribution, demography, population dynamics, social organization and behaviour of the Amboseli elephants. She supervises research and monitoring as well as training elephant researchers from African elephant range states. She is responsible for outreach to the Maasai community, carrying out surveys and training courses at other elephant study sites in Africa, disseminating scientific results, advocating for elephant welfare and promoting public awareness by writing popular articles and books, and making films about elephants. She was awarded a John D. & Catherine T. MacArthur Foundation Fellowship in 2002 and an honorary doctorate from Smith College, also in 2002, amongst numerous other awards and prizes for her contribution to understanding elephants. She may be contacted at Amboseli Elephant Research Project, Amboseli Trust for Elephants, P.O. Box 15135, Langata 00509, Nairobi, Kenya. Email:
[email protected]. 38
