Embedded Files
be explained by higher susceptibility of males to environmental conditions. Carnivores show enhanced survival in zoos in our study, but are more susceptible to behavioral abnormalities7 , highlighting the need for husbandry techniques to reduce these abnormalities while simultaneously maintaining the survival benefits. Although zoos offer simplified environments, social interactions might be as complex and challenging as in the wild, considering the high frequency of non-antagonistic contacts with humans and other species. Do animals, even when born and raised in zoos, perceive their enclosures as a spatial constraint in terms of compressed home ranges, or as an actual restriction of freedom in terms of a limitation of their own choices? Alternatively, do animals perceive zoos as a safe habitat where potential predators, food scarcity, or extreme climatic conditions are absent, allowing them to drastically reduce vigilance24? Our mere comparison of survival metrics between wild and captive populations should not be interpreted as a conclusive ethical judgment. Our findings should rather be considered as evidence that zoos generally enhance the longevity of mammals, except in species where there is little potential for such an enhancement because of their slower pace of life, which is already linked to both a low mortality and a high longevity in the wild. Because species with a slow pace of life are particularly threatened by extinction25, maintaining ex situ insurance populations of such threatened species remains a crucial conservation strategy. Methods Life tables. Zoo and wild population life tables were compiled from the Species360 database and literature, respectively (see Lemaître et al.14 for more details). Concerning free-ranging populations, publications containing life-tables from semi-captive populations were excluded to allow a strict comparison between captive and free-ranging populations. For 25 species, we collected several life tables from the same or different populations. When available, we gave preference to life tables obtained from longitudinal data. When several life tables of a given quality were available, we averaged them. When life tables were given in months or not with an integer of years, a standardization was made to obtain the survival at each integer age. For 9 species, the total number of individuals followed or considered was not given in the focal study. In such cases, we arbitrarily assumed that 100 individuals were considered per sex (close to the median value of the number of individuals alive at 1 year of age in the life tables we used). For wild life tables with a known total number of individuals (N= 50 species), the lowest was found for females of Mustela vison for which the life table only included 30 individuals, and the highest is observed for males of Oryctolagus cuniculus with a total of 9,020 individuals (Supplementary Table S1). For captive populations, we only used extinct cohorts of animals for which the sex and both birth and death dates were known, implying that animals were born in captivity. Extinct cohorts were defined as all cohorts born before a given year, which is determined as 2013 minus three quarters of the maximum longevity recorded for the species (Supplementary Table S1). As in captivity the sex ratio could be biased due to the culling of some young males during the first year for management issues (mostly in ruminants), we only computed parameters when at least 25 individuals for each sex of each species were alive at 1 year of age to get accurate estimates of age-specific survival. We finally obtained a dataset of 52 species for which data for both females and males were available, with one additional species with male-only data (leading to 53 species in males) and 6 additional species with femaleonly data (leading to 58 species in females) (Supplementary Table S1). For both captive and wild populations, we made the same calculations to obtain exactly comparable life tables. For visual comparison only, we included data from females of the two elephant species26,27 and sex-combined data from the common hippopotamus21 in the resulting data plots but did not include them in the analyses, as their data did not correspond to the data selection criteria stated above. Metrics of survival. To measure species- and sex-specific patterns of survival and actuarial senescence in captive and free-ranging populations, we used four distinct but complementary metrics: the longevity, the baseline annual mortality, the age at the onset of actuarial senescence and the rate of actuarial senescence (see Fig. 1. for a graphical display of these metrics). The longevity was extracted from species-specific life tables for both males and females and for both captive and wild populations (Supplementary Table S1). We defined longevity as the age at which 90% of individuals from the initial cohort (alive at 1 year of age) had died (Fig. 1). This allows avoiding spurious estimates due to the exceptionally long life of a few individuals15. However, this trait (called ‘longevity’ hereafter) is not a direct measure of senescence because it does not include any explicit information www.nature.com/scientificreports/ Scientific Reports | 6:36361 | DOI: 10.1038/srep36361 6 about age-dependent decline in survival. For other metrics, we first measured the logit-transformed age-specific mortality from a given life
Google Sites
Report abuse