9

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

should be addressed to M.T. (email: mtidiere@gmail.com) received: 10 June 2016 accepted: 30 September 2016 Published: 07 November 2016 OPEN www.nature.com/scientificreports/ Scientific Reports | 6:36361 | DOI: 10.1038/srep36361 2 free-ranging populations have not yet been performed. Indeed, it is probably difficult to gather accurate demographic estimates in these two contrasted environments for a large range of species. In mammals, comparisons between wild and zoo populations published so far were based on a small number of species (consistently less than 25) and did not control for confounding effects of phylogeny12,13 or were restricted to a narrow taxonomic range (e.g. the mammalian order Artiodactyla14). In addition, these studies have led to conflicting results because they both failed12,13 and succeeded14 to detect the expected lower actuarial senescence rate (i.e. the rate of decrease in annual survival with increasing age) in zoos. Lastly, none of these studies included survival metrics other than senescence rate, such as longevity or age at the onset of senescence. Therefore, whether the common belief that mammals in zoos outlive their wild counterparts holds true remains unknown. To address this question, we compared a set of survival metrics derived from life tables available from the literature for males and females of free-ranging populations of 59 mammalian species (including eight different orders, see Supplementary Figure S1) to those derived from the data on captive specimens of the same species from the Species360 database13,15 (formerly named International Species Information System database, ISIS). Based on these sex-specific life tables, four metrics describing the survival pattern of each species were calculated: longevity, baseline annual mortality, age at the onset of senescence and rate of senescence (Fig. 1). We compared these metrics between free-ranging and captive populations using linear models and controlled for phylogenetic relatedness among species using a mammal super-tree16. An expected higher longevity in zoos can originate from a lower baseline mortality, a later onset of senescence, a lower rate of senescence, or any combination of these measures (Fig. 1). Results In 84% of species we analyzed (85% for males and 83% for females, including all carnivores), longevity was higher in zoos than in the wild for both sexes (Figs 2 and 3A). The positive relationship between longevity in the zoo and in the wild had a slope less than 1 (Table 1), indicating that short-lived species benefited from living in zoos to a higher extent than long-lived ones (Fig. 3A). In about 69% of the species (76% for males and 63% for females), the age at the onset of senescence was identical or delayed in zoos compared to the wild (Fig. 3B). The positive relationship between zoo and wild data of onset of senescence also had a slope less than 1 (Table 1), again indicating that species with an early onset of actuarial senescence delayed this onset in captivity to a larger extent than species with a late onset of actuarial senescence. For these latter species, often no difference in response to captivity occurred, and some species even displayed an earlier onset of senescence in zoos (Supplementary Figure S2). The slopes of the relationship between the baseline mortality (Fig. 3C) or the rates of actuarial senescence (Fig. 3D) at the zoo and in the wild were close to zero (Table 1), indicating that these metrics did not strongly covary between zoo and wild populations. While the baseline mortality was lower in zoos for about 62% of the species (61% for males and 64% for females) and the rate of senescence was lower in zoos for about 73% of the species (76% for males and 71% for females), the nearly horizontal slopes underline the importance of the species’ pace of life for these two metrics: species with a high baseline mortality and high rate of senescence in the wild (i.e. species with a faster pace of life) typically had lower values at the zoo. In contrast, species with a low baseline mortality and Figure 1. Graphical displays of the metrics of survival and actuarial senescence analyzed in this study. Data from female lions (Panthera leo) in zoo (in brown) and free-ranging (in green) conditions are used for illustrative purposes. Female lions in the zoo population live longer (age in years) and have a lower baseline annual mortality (in log%), a later onset of senescence (in years) and a lower rate of actuarial senescence (measured as the exponential rate of mortality increase per year). www.nature.com/scientificreports/ Scientific Reports | 6:36361 | DOI: 10.1038/srep36361 3 Figure 2. Longevity in free-ranging and zoo conditions for males (triangles) and females (circles) of each species of Artiodactyla, Carnivora, Primates and other orders (Diprotodontia, Lagomorpha, Perissodactyla, Rodentia and Scandentia). Species living longer in zoos are indicated with solid lines and species living shorter in zoos are indicated with dotted lines. Full species names are given in Supplementary Table S1. Animal pictures: nebojsa78©123RF.com. www.nature.com/scientificreports/ Scientific Reports | 6:36361 | DOI: 10.1038/srep36361 4 a low rate of senescence in the wild (i.e. species with a slower pace of life) typically had higher values at the zoo. Notably, mammals in zoos displayed less variation in both baseline mortality