Research Note
Limits to captive breeding of mammals in zoos John Alroy Department of Biological Sciences, Macquarie University, Sydney NSW 2109, Australia, email [email protected]
Abstract: Captive breeding of mammals in zoos is the last hope for many of the best-known endangered species and has succeeded in saving some from certain extinction. However, the number of managed species selected is relatively small and focused on large-bodied, charismatic mammals that are not necessarily under strong threat and not always good candidates for reintroduction into the wild. Two interrelated and more fundamental questions go unanswered: have the major breeding programs succeeded at the basic level of maintaining and expanding populations, and is there room to expand them? I used published counts of births and deaths from 1970 to 2011 to quantify rates of growth of 118 captive-bred mammalian populations. These rates did not vary with body mass, contrary to strong predictions made in the ecological literature. Most of the larger managed mammalian populations expanded consistently and very few programs failed. However, growth rates have declined dramatically. The decline was predicted by changes in the ratio of the number of individuals within programs to the number of mammal populations held in major zoos. Rates decreased as the ratio of individuals in programs to populations increased. In other words, most of the programs that could exist already do exist. It therefore appears that debates over the general need for captive-breeding programs and the best selection of species are moot. Only a concerted effort could create room to manage a substantially larger number of endangered mammals. Keywords: conservation planning, IUCN Red List, population dynamics, zoo history Los L´ımites para la Reproducci´ on en Cautiverio de Mam´ıferos en Zool´ ogicos Alroy
Resumen: La reproducci´on en cautiverio de mam´ıferos en zool´ogicos es la u ´ ltima esperanza para muchas de las especies en peligro mejor conocidas y ha sido exitosa en el rescate de la extinci´ on segura de algunas especies. Sin embargo, el n´ umero de especies seleccionadas para el manejo es relativamente peque˜ no y est´ a enfocado en mam´ıferos grandes y carism´ aticos que no est´ an necesariamente bajo una fuerte amenaza de extinci´ on y que no siempre son buenos candidatos para la reintroducci´ on a la vida silvestre. Dos preguntas interconectadas y m´ as fundamentales siguen sin respuesta: ¿Han tenido ´exito la mayor´ıa de los programas de reproducci´ on en el nivel b´ asico de mantener y expandir poblaciones y hay espacio para expandirlas? Us´e conteos publicados de nacimientos y muertes desde 1970 hasta 2011 para cuantificar las tasas de crecimiento de 118 poblaciones de mam´ıferos reproducidos en cautiverio. Estas tasas no variaron con la masa corporal, contradiciendo las fuertes predicciones hechas en la literatura ecol´ ogica. La mayor´ıa de las poblaciones de mam´ıferos con mayor manejo se expandi´ o consistentemente y muy pocos programas fallaron. Sin embargo, las tasas de crecimiento declinaron dram´ aticamente. La declinaci´ on fue predicha por los cambios en la proporci´ on del n´ umero de individuos dentro de los programas al n´ umero de especies de mam´ıferos dentro de la mayor´ıa de los zool´ ogicos. Las tasas disminuyeron conforme a la proporci´ on de individuos en los programas a la de individuos en las poblaciones. En otras palabras, la mayor´ıa de los programas que podr´ıan existir ya existen. Por esto parece que los debates sobre la necesidad general de programas de reproducci´ on en cautiverio y una mejor selecci´ on de especies son irrelevantes. Solamente un esfuerzo coordinado podr´ıa crear el espacio para manejar un n´ umero sustancialmente mayor de mam´ıferos en peligro. Palabras Clave: din´amicas poblacionales, historia de los zool´ogicos, Lista Roja UICN, planeaci´on de la conservaci´ on
Paper submitted June 5, 2014; revised manuscript accepted November 18, 2014.
1 Conservation Biology, Volume 00, No. 0, 1–6 C 2015, Society for Conservation Biology. DOI: 10.1111/cobi.12471
2
Introduction Researchers interested in captive breeding programs have previously focused most of their efforts on answering 3 questions: are captive breeding programs costeffective (e.g., Rahbek 1993; Balmford et al. 1996), how are species selected for being kept and bred in zoos (e.g., Frynta et al. 2013; Martin et al. 2013) and are breeding populations sufficient to maintain substantial genetic diversity over long periods (e.g., Mace 1986; Earnhardt et al. 2001)? For example, it has been shown that captive mammal species have relatively high body mass (Frynta et al. 2013), are aesthetically attractive to humans (Frynta et al. 2013), and are less often threatened than their close relatives (Martin et al. 2013). Although there have been some general discussions of whether zoos have enough capacity to foster a meaningful number of breeding programs (Conway 1986; Soul´e et al. 1986), there have been no rigorous analyses focused on the basic question of whether large programs have succeeded in the specific sense of increasing captive populations. It also has not been shown whether full captive-breeding capacity has already been reached.
Methods Since 1972, the International Zoo Yearbook (IZY) has documented substantial international efforts to manage captive populations by regularly publishing a list of studbooks. The list includes the number of births and deaths each year and end-of-year total counts of individuals. This list is incomplete and omits regional studbooks, but it seems representative of major international programs based on a comparison with overall counts of species held in zoos by International Union for Conservation of Nature (IUCN) threat category (Conde et al. 2013). Data are available for 1970–2011 (IZY volumes dated 1972– 2013). I focused on mammals because few studbooks pertaining to other groups existed prior to the 1980s. I considered all censused mammal species and treated subspecies that were tracked independently as separate populations for the purpose of data analysis. I accounted for changes in taxonomic nomenclature in collating data records from different years. I omitted data records when the number of births and deaths was omitted, the total number of individuals was omitted, or totals did not differentiate between males and females. Populations were included if the sum of end-of-year totals y was at least 250. In other words, if a species’ population averaged 20 and data were available for 20 years then y was 400 and the species was included. This figure was chosen to allow computation of precise estimates of population growth rates. The growth rate was defined as b/n – d/n, where b is the number of births, d is the number of deaths, and n = y – b/2 + d/2.
Conservation Biology Volume 00, No. 0, 2015
Limits to Captive Breeding
I quantified median numbers of species and individual animals in large zoos to test the hypothesis that crowding is the reason for changes in the rate of growth of breeding programs. I identified the 12 zoos that held the most tetrapod species as of 1991 (middle of the study period). In descending order, these zoos were the ones in Berlin (Zoologischer Garten und Zoo-Aquarium), San Diego, Berlin (Tierpark), Moscow, Antwerp, St. Louis, London, the Bronx, Cincinnati, Los Angeles, Philadelphia, and Amsterdam. Other metrics are less useful for this purpose because most large zoos hold roughly the same number of mammal species, counts of fishes and invertebrates are highly variable, and so are counts of individual animals. I interpolated counts of mammal species and individuals when data were missing by assuming that trends were exponential and then computed the medians. I used an identical protocol to compute median attendance and staffing level values, which are also reported by the IZY. The most obvious explanation for a downward trend in ratios of population counts to individual counts would be that zoos have reached their physical capacity with respect to the number of individuals associated with breeding programs, perhaps because the number of large zoos has increased only very slowly, and have therefore intentionally decreased birth rates. However, alternative explanations exist, such as the possibility that finances have become limited in general or that staffing levels have not been able to keep up with the growth of breeding programs. I tested these alternative hypotheses by predicting the 3 key demographic variables: median net per-capita rate of population growth across all the breeding programs that existed in a given year, square root of the median birth rate, and square root of the median death rate. My use of square root transformation was appropriate because birth and death rates are zero bounded and therefore have skewed distributions. By contrast, log transformation would have greatly increased the variance due to error in small counts. I employed 4 predictor variables: median mammal species counts, median individual counts, median staffing levels, and median attendance. Respectively, these variables stand for the hypotheses that growth rates are controlled by capacity for maintenance of species in the abstract, capacity for maintenance of individuals, available labor, and finances. Attendance is a proxy for financing only because zoos draw substantial income from other sources including concessions and government support. The variables were all log transformed because they were zero bounded and presumably generated by a multiplicative process. Prior to log transformation they were multiplied by the number of large zoos in order to get a rough proportional estimate of total numbers across all zoos. This scaling made no difference to the results because changes in the number of zoos were very slow.
Alroy
3
Rate of increase
0.20
a
all species
0.15
0.10
0.10
0.05
0.05
0.00
1970
0.4
Rate of increase
b birth rate death rate
0.15
0.00 1980
1990
2000
2010
c
1970
0.20
0.3
0.15
0.2
0.10
0.1
0.05
0.0
0.20
1980
1990
2000
2010
1990
2000
2010
1980
1990
2000
2010
1980
1990
2000
2010
d
e
1970
0.20
0.15
0.15
0.10
0.10
0.05
0.05
0.00 1970
1980
0.00
1970
Rate of increase
0.20
f
0.00 1980
1990
2000
2010
1970
Year
To avoid statistical artifacts related to the crosscorrelation of autocorrelated time series, I detrended all of the variables with a standard first-order autoregressive model as implemented in the R function arima. After detrending, I exponentiated the independent variables, divided them by the breeding population estimate to create per-individual ratios, and transformed them to log values again. For the purpose of showing whether rates of population increase scale with body size, mass estimates were obtained from a standard literature compilation (Smith et al. 2003).
Results There were gaps in the records for most populations, and some studbooks have lapsed. However, enough information existed to compute an overall rate of increase for 118 populations, of which 81 constituted species and 37 subspecies (Supporting Information).
Year
Figure 1. Historical trend in birth and death rates of captively managed mammalian populations. Values are smoothed by taking means across 3-interval windows. (a) Median values for all populations (values are only shown for years for which at least 10 estimates are available; estimates are only computed when at least 100 individuals of a given population are censused in a given year). (b–f) Rates for (b) Choeropsis liberiensis (pygmy hippopotamus), (c) Chrysocyon brachyurus (maned wolf), (d) Okapia johnstoni (okapi), (e) Pan paniscus (bonobo), and (f) Rhinoceros unicornis (Indian rhinoceros). These species were chosen because they are taxonomically disparate and present unusually complete data sets. The y-axis in (c) is compressed.
The median rate of population increase/year was 0.0281, and only 17 populations (14.4%) had a negative value. Five of these 17 (Dendrolagus matschiei, Dicerorhinus sumatrensis, Gymnobelideus leadbeateri, Panthera tigris sumatrae, and Pygathrix nemaeus) were endangered or critically endangered according to the IUCN. Except for these species, positive growth rates were maintained for a large majority of populations in all categories save least concern (Supporting Information). Historical trends in net population increase (Fig. 1) consistently showed very subtle changes in death rates and either no change in birth rates or higher birth rates in the 1970s and 1980s than later. The opposite pattern occurred in just a few cases (e.g., Okapia johnstoni) (Fig. 1d), suggesting only isolated recent improvements in breeding techniques. The consistency of trends was easier to show by illustrating changes in birth:death ratios between the early and late periods of coverage (Fig. 2). Very few birth:death ratios increased, and there was a clearcut tendency for very high ratios to decline (Fig. 2a). More rigorously,
Conservation Biology Volume 00, No. 0, 2015
Limits to Captive Breeding
4
a
5.0
1970−1990
Births/deaths
1991−2011
2.0
1.0
0.5 0
10
20
30
40
50
Change in log births/deaths
Species rank b
1.5 1.0 ●
0.5
● ●
●
0.0 −0.5
● ●● ●● ●● ● ● ● ● ● ● ● ●●● ● ●● ● ● ● ● ●● ● ●● ● ● ● ● ● ● ● ● ● ● ●
−1.0
● ● ● ●
−1.5 0.5
1.0
2.0
5.0
(one-tailed, parametric and nonparametric p = 0.006 and 0.056), birth rates by number per individual in general (p = 0.032 and 0.036), and death rates by number per visitor (p = 0.014 and 0.018) (Supporting Information). The pattern for growth and birth rates is easily interpreted as showing that zoos intentionally slowed the growth of breeding programs due to space limitation. The prediction of death rates by attendance is harder to explain, but attendance is the one strongly predictable driver of financial income for zoos. Therefore, the result might indicate zoos are able to support better veterinary care when their finances are stronger. The multiple regression results implied that the availability of labor was not a real constraint on population demographics. This observation might have to do with the fact that zoo staffs have steadily expanded whereas collection sizes have decreased (Supporting Information), indicating that a smaller and smaller proportion of employees have actually been involved in animal care. Rates of increase in zoo populations did not scale with body mass (Fig. 3 & Supporting Information). Even the data for 1970–1991 failed to show any relationship (Fig. 3a: Spearman’s ρ = −0.106, p = 0.406). The median rate of population increase in this period (0.0543) was much higher than in 1992–2011 (0.0208), and the difference was significant according to a Wilcoxon ranksum test (W = 5125.5, p < 0.001). Both medians were at the low end of the distribution of intrinsic rates for all mammals, but they were consistent with older data for species weighing 1 kg or more.
Births/deaths (1970−1990) Figure 2. Historical changes in log birth:death ratios for managed captive populations of mammals. In (a) arrows connect ratios for 1970 through 1990 with ratios for 1991 through 2011 and species are rank-ordered based on the ratios for the first period. In (b) points represent populations and the y-axis is the log ratio for 1970–1990 minus the ratio for 1991–2011. there was a strong negative correlation between the magnitude of the change and the initial value (Fig. 2b: Spearman’s nonparametric, rank-order correlation ρ = −0.784, p < 0.001). I computed an equilibrium value by finding the y-intercept after regressing the change in the ratio against the initial ratio. This equilibrium point was 0.1017 (significantly different from zero: p = 0.031). Although this kind of statistic is not directly comparable with a net rate of population increase because it is not standardized for population size, the actual value was generally consistent with the median rate reported above. Stepwise multiple regression analyses showed that net population growth rates were best predicted by the median number of managed individuals per species
Conservation Biology Volume 00, No. 0, 2015
Discussion Because breeding programs expanded rapidly in the 1970s and 1980s, the high rates of increase in that period likely represent biological maxima in the absence of intraspecific competition, predation, environmental variability, and most communicable diseases. A large literature suggests that these intrinsic rates of increase should scale negatively with body size (Fagan et al. 2010 and references cited therein). This inference rests mostly on theoretical rates computed from empirical life history parameters (Henneman 1983; Thompson 1987; Fagan et al. 2010) because there is little previously published information on captive mammalian populations (Fenchel 1974) apart from survivorship parameters (Lynch et al. 2010) and because growth rate data based on field studies appear to provide underestimates (Fagan et al. 2010; Lynch et al. 2010). Non-linearity of the scaling relationship is not only demonstrated here (Fig. 3) but visually obvious in these earlier data sets (Henneman 1983; Fagan et al. 2010). Thus, a strong scaling relationship may not exist for medium- and large-sized mammals. This conclusion challenges the idea that growth rates for all animals are
Rate of increase (1970−1990)
Alroy
5
0.25
a ●
0.15
● ● ●
0.05
●
● ●
● ● ●
●
●
●●●
●
●
● ● ●
● ● ●
● ●
● ● ●
●
● ●
−0.05
●
●
●
−0.15 0.01
0.1
1
10
100
1000
10k
Rate of increase (1991−2011)
Body mass (kg) 0.25
b
0.15 ●
0.05 ● ● ●
−0.05
● ●● ● ●● ● ●● ● ● ●● ● ● ● ● ● ● ● ●● ● ● ●● ●● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ●
● ●
● ● ●
−0.15 0.01
0.1
1
10
100
1000
10k
Body mass (kg) Figure 3. Size scaling of rates of increase of managed captive populations relative to body mass and conservation status seen in data for (a) 1970–1990 and (b) 1991–2011 (open squares, endangered species, critically endangered species, and species extinct in the wild [IUCN 2014]; closed circles, all other species).
strongly controlled by universal metabolic constraints (Savage et al. 2004) and that the relationship obeys a power law with a slope of −0.25, much less −0.5 (Fagan et al. 2010). Counts of species with positive rates of growth (Supporting Information) were small relative to the number of mammal species in general (Conde et al. 2011). However, the fact that captive populations of mammals continue to grow refutes the argument that breeding programs are ineffectual for the purpose of maintaining populations (Snyder et al. 1996), putting aside the question of whether reintroduction programs are feasible. It is also important to keep in mind that the IZY data do not account for successful breeding of species lacking international
studbooks (Conde et al. 2011) and that the data I analyzed applied to less than half of all high-risk species actually held in zoos (Conde et al. 2013). Thus, the counts are conservative in absolute terms. Furthermore, many of these species are highly distinct in a phylogenetic sense and of unquestioned high conservation value. Examples include multiple species of lemurs, great apes, and rhinoceroses. However, rates of growth have slowed (Figs. 1 & 2, & Supporting Information) as the number of studbooks has climbed over the past few decades, tracking the expansion in the number of species survival plans (overseen by the Association of Zoos and Aquariums) and European Endangered Species Programmes (overseen by the European Assocation of Zoos and Aquaria). In other words, individual zoos seem to have already set aside as much space for captive breeding of mammals as they deem fit. Along with the need to maintain genetic diversity, this fact helps explain why the AZA’s Taxon Advisory Groups set specific population size targets (Earnhardt et al. 2001). These self-imposed limits may be reasonable. For example, at the end of 2011, zoos reporting to the IZY housed 47,873 individual mammals (Zoological Society of London 2013). The 101 reported mammal populations with international studbooks by themselves totaled 30,904 individuals, and many of the other populations that are not under management are simply too small to maintain adequate genetic diversity (Earnhardt et al. 2001). Thus, the evidence suggests there is literally not enough space in the world’s zoos right now to substantially increase the number of truly successful captive breeding programs. There are 2 key counterarguments. First, many zoos have systematically decreased their holdings over the last several decades and may be able to increase them again. For example, in 1970 the London Zoo held 809 individual mammals, but by 2011 this figure had fallen to 586. Husbandry considerations presumably drive such trends: zoos are making more space for individual animals to improve their care and general well-being. However, favorable birth:death ratios in the 1970s (Figs. 1 & 2, & Supporting Information) suggest that if populations were increased, there would be no great cost from the standpoint of captive breeding. Second, much effort is expended on species that are not endangered or are otherwise not good candidates for eventual restocking of wild populations (Snyder et al. 1996). Indeed, of the 101 studbooks listed in 2011 only 50 pertained to endangered species or subspecies, and only 14,393 individuals belonged to those taxa. Most of these populations are of large-bodied species such as primates, carnivores, and ungulates (Supporting Information) that are drivers of attendance but are also expensive to maintain in captivity (Conway 1986). Refocusing attention on smaller taxa could be a cost-effective way to expand captive breeding programs of greater relevance for conservation.
Conservation Biology Volume 00, No. 0, 2015
6
Acknowledgments The author is the recipient of an Australian Research Council Future Fellowship (project number FT0992161). I thank the editors and anonymous reviewers for helpful feedback.
Supporting Information Figures illustrating historical trends in the number of captive breeding programs, demographic parameters, and key variables related to the size of 12 major zoos; rates of increase of individual species in individual years; and trends in numbers of censused individuals and the number of large zoos (Appendix S1) plus a list of species included in the analysis (Appendix S2) are available online. The authors are solely responsible for the content and functionality of these materials. Queries (other than absence of the material) should be directed to the corresponding author. Figure S1. Historical trends in the number of captive breeding programs and in demographic parameters. Time series start at 1977 because data were available for less than 10 programs in each year prior to that one. (a) Number of breeding programs. (b) Median net population growth rate (birth rate minus death rate). (c) Median per capita birth rate. (d) Median per capita death rate. Figure S2. Historical trends in key variables related to the size of the 12 zoos having the largest collections of tetrapods in 1991. (a) Median number of species of mammal species held. (b) Median number of individual mammals held. (c) Median number of staff members. (d) Median annual attendance. Literature Cited Balmford A, Leader-Williams N, Green MJB. 1996. Parks or arks: Where to conserve threatened mammals? Biodiversity Conservation 4:595– 607. Conde DA, Colchero F, Gusset M, Pearce-Kelly P, Byers O, Flesness N, Browne RK, Jones OR. 2013. Zoos through the lens of the IUCN Red List: a global metapopulation approach to support
Conservation Biology Volume 00, No. 0, 2015
Limits to Captive Breeding
conservation breeding programs. PLOS ONE 8 (e80311) DOI:10.1371/journal.pone.0080311. Conde DA, Flesness N, Colchero F, Jones OR, Scheuerlein A. 2011. An emerging role of zoos to conserve biodiversity. Science 331:1390– 1391. Conway WG. 1986. The practical difficulties and financial implications of endangered species breeding programmes. International Zoo Yearbook 24/25:210–219. Earnhardt JM, Thompson SD, Marhevsky EA. 2001. Interactions of target population size, population parameters, and program management on viability of captive populations. Zoo Biology 20:169–183. Fagan WF, Lynch HJ, Noon BR. 2010. Pitfalls and challenges of estimating population growth rate from empirical data: consequences for allometric scaling relations. Oikos 119:455–464. Fenchel T. 1974. Intrinsic rate of natural increase: the relationship with body size. Oecologia 14:317–326. Frynta D, ˇSimkov´a O, Liˇskov´a S, Landov´a E. 2013. Mammalian collection on Noah’s ark: the effects of beauty, brain and body size. PLOS ONE 8 (e63110) DOI: 10.1371/journal.pone.0063110. Henneman WW III. 1983. Relationships among body mass, metabolic rate and the intrinsic rate of natural increase in mammals. Oecologia 56:104–108. IUCN (International Union for Conservation of Nature). 2014. The IUCN red list of threatened species. Version 2014.3. IUCN, Gland, Switzerland. Lynch HJ, et al. 2010. Survivorship patterns in captive mammalian populations: implications for estimating population growth rates. Ecological Applications 20:2334–2345. Mace GM. 1986. Genetic management of small populations. International Zoo Yearbook 24/25:167–174. Martin TE, Lurbiecki H, Joy JB, Mooers AO. 2013. Mammal and bird species held in zoos are less endemic and less threatened than their close relatives not held in zoos. Animal Conservation 17:89–96. Rahbek C. 1993. Captive breeding—a useful tool in the preservation of biodiversity? Biodiversity Conservation 2:426–437. Savage VM, Gillooly JF, Brown JH, West GB, Charnov E. 2004. Effects of body size and temperature on population growth. The American Naturalist 163:429–441. Smith FA, et al. 2003. Body mass of late Quaternary mammals. Ecology 84:3403. Snyder NF, et al. 1996. Limitations of captive breeding in endangered species recovery. Conservation Biology 10:338–348. Soul´e M, Gilpin M, Conway W, Foose T. 1986. The millenium ark: How long a voyage, how many staterooms, how many passengers? Zoo Biology 5:101–113. Thompson SD. 1987. Body size, duration of parental care, and the intrinsic rate of natural increase in eutherian and metatherian mammals. Oecologia 71:201–209. Zoological Society of London. 2013. Zoos and aquariums of the world. International Zoo Yearbook 47:231–288.
Limits to captive breeding of mammals in zoos John Alroy Department of Biological Sciences, Macquarie University, Sydney NSW 2109, Australia, email [email protected]
Abstract: Captive breeding of mammals in zoos is the last hope for many of the best-known endangered species and has succeeded in saving some from certain extinction. However, the number of managed species selected is relatively small and focused on large-bodied, charismatic mammals that are not necessarily under strong threat and not always good candidates for reintroduction into the wild. Two interrelated and more fundamental questions go unanswered: have the major breeding programs succeeded at the basic level of maintaining and expanding populations, and is there room to expand them? I used published counts of births and deaths from 1970 to 2011 to quantify rates of growth of 118 captive-bred mammalian populations. These rates did not vary with body mass, contrary to strong predictions made in the ecological literature. Most of the larger managed mammalian populations expanded consistently and very few programs failed. However, growth rates have declined dramatically. The decline was predicted by changes in the ratio of the number of individuals within programs to the number of mammal populations held in major zoos. Rates decreased as the ratio of individuals in programs to populations increased. In other words, most of the programs that could exist already do exist. It therefore appears that debates over the general need for captive-breeding programs and the best selection of species are moot. Only a concerted effort could create room to manage a substantially larger number of endangered mammals. Keywords: conservation planning, IUCN Red List, population dynamics, zoo history Los L´ımites para la Reproducci´ on en Cautiverio de Mam´ıferos en Zool´ ogicos Alroy
Resumen: La reproducci´on en cautiverio de mam´ıferos en zool´ogicos es la u ´ ltima esperanza para muchas de las especies en peligro mejor conocidas y ha sido exitosa en el rescate de la extinci´ on segura de algunas especies. Sin embargo, el n´ umero de especies seleccionadas para el manejo es relativamente peque˜ no y est´ a enfocado en mam´ıferos grandes y carism´ aticos que no est´ an necesariamente bajo una fuerte amenaza de extinci´ on y que no siempre son buenos candidatos para la reintroducci´ on a la vida silvestre. Dos preguntas interconectadas y m´ as fundamentales siguen sin respuesta: ¿Han tenido ´exito la mayor´ıa de los programas de reproducci´ on en el nivel b´ asico de mantener y expandir poblaciones y hay espacio para expandirlas? Us´e conteos publicados de nacimientos y muertes desde 1970 hasta 2011 para cuantificar las tasas de crecimiento de 118 poblaciones de mam´ıferos reproducidos en cautiverio. Estas tasas no variaron con la masa corporal, contradiciendo las fuertes predicciones hechas en la literatura ecol´ ogica. La mayor´ıa de las poblaciones de mam´ıferos con mayor manejo se expandi´ o consistentemente y muy pocos programas fallaron. Sin embargo, las tasas de crecimiento declinaron dram´ aticamente. La declinaci´ on fue predicha por los cambios en la proporci´ on del n´ umero de individuos dentro de los programas al n´ umero de especies de mam´ıferos dentro de la mayor´ıa de los zool´ ogicos. Las tasas disminuyeron conforme a la proporci´ on de individuos en los programas a la de individuos en las poblaciones. En otras palabras, la mayor´ıa de los programas que podr´ıan existir ya existen. Por esto parece que los debates sobre la necesidad general de programas de reproducci´ on en cautiverio y una mejor selecci´ on de especies son irrelevantes. Solamente un esfuerzo coordinado podr´ıa crear el espacio para manejar un n´ umero sustancialmente mayor de mam´ıferos en peligro. Palabras Clave: din´amicas poblacionales, historia de los zool´ogicos, Lista Roja UICN, planeaci´on de la conservaci´ on
Paper submitted June 5, 2014; revised manuscript accepted November 18, 2014.
1 Conservation Biology, Volume 00, No. 0, 1–6 C 2015, Society for Conservation Biology. DOI: 10.1111/cobi.12471
2
Introduction Researchers interested in captive breeding programs have previously focused most of their efforts on answering 3 questions: are captive breeding programs costeffective (e.g., Rahbek 1993; Balmford et al. 1996), how are species selected for being kept and bred in zoos (e.g., Frynta et al. 2013; Martin et al. 2013) and are breeding populations sufficient to maintain substantial genetic diversity over long periods (e.g., Mace 1986; Earnhardt et al. 2001)? For example, it has been shown that captive mammal species have relatively high body mass (Frynta et al. 2013), are aesthetically attractive to humans (Frynta et al. 2013), and are less often threatened than their close relatives (Martin et al. 2013). Although there have been some general discussions of whether zoos have enough capacity to foster a meaningful number of breeding programs (Conway 1986; Soul´e et al. 1986), there have been no rigorous analyses focused on the basic question of whether large programs have succeeded in the specific sense of increasing captive populations. It also has not been shown whether full captive-breeding capacity has already been reached.
Methods Since 1972, the International Zoo Yearbook (IZY) has documented substantial international efforts to manage captive populations by regularly publishing a list of studbooks. The list includes the number of births and deaths each year and end-of-year total counts of individuals. This list is incomplete and omits regional studbooks, but it seems representative of major international programs based on a comparison with overall counts of species held in zoos by International Union for Conservation of Nature (IUCN) threat category (Conde et al. 2013). Data are available for 1970–2011 (IZY volumes dated 1972– 2013). I focused on mammals because few studbooks pertaining to other groups existed prior to the 1980s. I considered all censused mammal species and treated subspecies that were tracked independently as separate populations for the purpose of data analysis. I accounted for changes in taxonomic nomenclature in collating data records from different years. I omitted data records when the number of births and deaths was omitted, the total number of individuals was omitted, or totals did not differentiate between males and females. Populations were included if the sum of end-of-year totals y was at least 250. In other words, if a species’ population averaged 20 and data were available for 20 years then y was 400 and the species was included. This figure was chosen to allow computation of precise estimates of population growth rates. The growth rate was defined as b/n – d/n, where b is the number of births, d is the number of deaths, and n = y – b/2 + d/2.
Conservation Biology Volume 00, No. 0, 2015
Limits to Captive Breeding
I quantified median numbers of species and individual animals in large zoos to test the hypothesis that crowding is the reason for changes in the rate of growth of breeding programs. I identified the 12 zoos that held the most tetrapod species as of 1991 (middle of the study period). In descending order, these zoos were the ones in Berlin (Zoologischer Garten und Zoo-Aquarium), San Diego, Berlin (Tierpark), Moscow, Antwerp, St. Louis, London, the Bronx, Cincinnati, Los Angeles, Philadelphia, and Amsterdam. Other metrics are less useful for this purpose because most large zoos hold roughly the same number of mammal species, counts of fishes and invertebrates are highly variable, and so are counts of individual animals. I interpolated counts of mammal species and individuals when data were missing by assuming that trends were exponential and then computed the medians. I used an identical protocol to compute median attendance and staffing level values, which are also reported by the IZY. The most obvious explanation for a downward trend in ratios of population counts to individual counts would be that zoos have reached their physical capacity with respect to the number of individuals associated with breeding programs, perhaps because the number of large zoos has increased only very slowly, and have therefore intentionally decreased birth rates. However, alternative explanations exist, such as the possibility that finances have become limited in general or that staffing levels have not been able to keep up with the growth of breeding programs. I tested these alternative hypotheses by predicting the 3 key demographic variables: median net per-capita rate of population growth across all the breeding programs that existed in a given year, square root of the median birth rate, and square root of the median death rate. My use of square root transformation was appropriate because birth and death rates are zero bounded and therefore have skewed distributions. By contrast, log transformation would have greatly increased the variance due to error in small counts. I employed 4 predictor variables: median mammal species counts, median individual counts, median staffing levels, and median attendance. Respectively, these variables stand for the hypotheses that growth rates are controlled by capacity for maintenance of species in the abstract, capacity for maintenance of individuals, available labor, and finances. Attendance is a proxy for financing only because zoos draw substantial income from other sources including concessions and government support. The variables were all log transformed because they were zero bounded and presumably generated by a multiplicative process. Prior to log transformation they were multiplied by the number of large zoos in order to get a rough proportional estimate of total numbers across all zoos. This scaling made no difference to the results because changes in the number of zoos were very slow.
Alroy
3
Rate of increase
0.20
a
all species
0.15
0.10
0.10
0.05
0.05
0.00
1970
0.4
Rate of increase
b birth rate death rate
0.15
0.00 1980
1990
2000
2010
c
1970
0.20
0.3
0.15
0.2
0.10
0.1
0.05
0.0
0.20
1980
1990
2000
2010
1990
2000
2010
1980
1990
2000
2010
1980
1990
2000
2010
d
e
1970
0.20
0.15
0.15
0.10
0.10
0.05
0.05
0.00 1970
1980
0.00
1970
Rate of increase
0.20
f
0.00 1980
1990
2000
2010
1970
Year
To avoid statistical artifacts related to the crosscorrelation of autocorrelated time series, I detrended all of the variables with a standard first-order autoregressive model as implemented in the R function arima. After detrending, I exponentiated the independent variables, divided them by the breeding population estimate to create per-individual ratios, and transformed them to log values again. For the purpose of showing whether rates of population increase scale with body size, mass estimates were obtained from a standard literature compilation (Smith et al. 2003).
Results There were gaps in the records for most populations, and some studbooks have lapsed. However, enough information existed to compute an overall rate of increase for 118 populations, of which 81 constituted species and 37 subspecies (Supporting Information).
Year
Figure 1. Historical trend in birth and death rates of captively managed mammalian populations. Values are smoothed by taking means across 3-interval windows. (a) Median values for all populations (values are only shown for years for which at least 10 estimates are available; estimates are only computed when at least 100 individuals of a given population are censused in a given year). (b–f) Rates for (b) Choeropsis liberiensis (pygmy hippopotamus), (c) Chrysocyon brachyurus (maned wolf), (d) Okapia johnstoni (okapi), (e) Pan paniscus (bonobo), and (f) Rhinoceros unicornis (Indian rhinoceros). These species were chosen because they are taxonomically disparate and present unusually complete data sets. The y-axis in (c) is compressed.
The median rate of population increase/year was 0.0281, and only 17 populations (14.4%) had a negative value. Five of these 17 (Dendrolagus matschiei, Dicerorhinus sumatrensis, Gymnobelideus leadbeateri, Panthera tigris sumatrae, and Pygathrix nemaeus) were endangered or critically endangered according to the IUCN. Except for these species, positive growth rates were maintained for a large majority of populations in all categories save least concern (Supporting Information). Historical trends in net population increase (Fig. 1) consistently showed very subtle changes in death rates and either no change in birth rates or higher birth rates in the 1970s and 1980s than later. The opposite pattern occurred in just a few cases (e.g., Okapia johnstoni) (Fig. 1d), suggesting only isolated recent improvements in breeding techniques. The consistency of trends was easier to show by illustrating changes in birth:death ratios between the early and late periods of coverage (Fig. 2). Very few birth:death ratios increased, and there was a clearcut tendency for very high ratios to decline (Fig. 2a). More rigorously,
Conservation Biology Volume 00, No. 0, 2015
Limits to Captive Breeding
4
a
5.0
1970−1990
Births/deaths
1991−2011
2.0
1.0
0.5 0
10
20
30
40
50
Change in log births/deaths
Species rank b
1.5 1.0 ●
0.5
● ●
●
0.0 −0.5
● ●● ●● ●● ● ● ● ● ● ● ● ●●● ● ●● ● ● ● ● ●● ● ●● ● ● ● ● ● ● ● ● ● ● ●
−1.0
● ● ● ●
−1.5 0.5
1.0
2.0
5.0
(one-tailed, parametric and nonparametric p = 0.006 and 0.056), birth rates by number per individual in general (p = 0.032 and 0.036), and death rates by number per visitor (p = 0.014 and 0.018) (Supporting Information). The pattern for growth and birth rates is easily interpreted as showing that zoos intentionally slowed the growth of breeding programs due to space limitation. The prediction of death rates by attendance is harder to explain, but attendance is the one strongly predictable driver of financial income for zoos. Therefore, the result might indicate zoos are able to support better veterinary care when their finances are stronger. The multiple regression results implied that the availability of labor was not a real constraint on population demographics. This observation might have to do with the fact that zoo staffs have steadily expanded whereas collection sizes have decreased (Supporting Information), indicating that a smaller and smaller proportion of employees have actually been involved in animal care. Rates of increase in zoo populations did not scale with body mass (Fig. 3 & Supporting Information). Even the data for 1970–1991 failed to show any relationship (Fig. 3a: Spearman’s ρ = −0.106, p = 0.406). The median rate of population increase in this period (0.0543) was much higher than in 1992–2011 (0.0208), and the difference was significant according to a Wilcoxon ranksum test (W = 5125.5, p < 0.001). Both medians were at the low end of the distribution of intrinsic rates for all mammals, but they were consistent with older data for species weighing 1 kg or more.
Births/deaths (1970−1990) Figure 2. Historical changes in log birth:death ratios for managed captive populations of mammals. In (a) arrows connect ratios for 1970 through 1990 with ratios for 1991 through 2011 and species are rank-ordered based on the ratios for the first period. In (b) points represent populations and the y-axis is the log ratio for 1970–1990 minus the ratio for 1991–2011. there was a strong negative correlation between the magnitude of the change and the initial value (Fig. 2b: Spearman’s nonparametric, rank-order correlation ρ = −0.784, p < 0.001). I computed an equilibrium value by finding the y-intercept after regressing the change in the ratio against the initial ratio. This equilibrium point was 0.1017 (significantly different from zero: p = 0.031). Although this kind of statistic is not directly comparable with a net rate of population increase because it is not standardized for population size, the actual value was generally consistent with the median rate reported above. Stepwise multiple regression analyses showed that net population growth rates were best predicted by the median number of managed individuals per species
Conservation Biology Volume 00, No. 0, 2015
Discussion Because breeding programs expanded rapidly in the 1970s and 1980s, the high rates of increase in that period likely represent biological maxima in the absence of intraspecific competition, predation, environmental variability, and most communicable diseases. A large literature suggests that these intrinsic rates of increase should scale negatively with body size (Fagan et al. 2010 and references cited therein). This inference rests mostly on theoretical rates computed from empirical life history parameters (Henneman 1983; Thompson 1987; Fagan et al. 2010) because there is little previously published information on captive mammalian populations (Fenchel 1974) apart from survivorship parameters (Lynch et al. 2010) and because growth rate data based on field studies appear to provide underestimates (Fagan et al. 2010; Lynch et al. 2010). Non-linearity of the scaling relationship is not only demonstrated here (Fig. 3) but visually obvious in these earlier data sets (Henneman 1983; Fagan et al. 2010). Thus, a strong scaling relationship may not exist for medium- and large-sized mammals. This conclusion challenges the idea that growth rates for all animals are
Rate of increase (1970−1990)
Alroy
5
0.25
a ●
0.15
● ● ●
0.05
●
● ●
● ● ●
●
●
●●●
●
●
● ● ●
● ● ●
● ●
● ● ●
●
● ●
−0.05
●
●
●
−0.15 0.01
0.1
1
10
100
1000
10k
Rate of increase (1991−2011)
Body mass (kg) 0.25
b
0.15 ●
0.05 ● ● ●
−0.05
● ●● ● ●● ● ●● ● ● ●● ● ● ● ● ● ● ● ●● ● ● ●● ●● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ●
● ●
● ● ●
−0.15 0.01
0.1
1
10
100
1000
10k
Body mass (kg) Figure 3. Size scaling of rates of increase of managed captive populations relative to body mass and conservation status seen in data for (a) 1970–1990 and (b) 1991–2011 (open squares, endangered species, critically endangered species, and species extinct in the wild [IUCN 2014]; closed circles, all other species).
strongly controlled by universal metabolic constraints (Savage et al. 2004) and that the relationship obeys a power law with a slope of −0.25, much less −0.5 (Fagan et al. 2010). Counts of species with positive rates of growth (Supporting Information) were small relative to the number of mammal species in general (Conde et al. 2011). However, the fact that captive populations of mammals continue to grow refutes the argument that breeding programs are ineffectual for the purpose of maintaining populations (Snyder et al. 1996), putting aside the question of whether reintroduction programs are feasible. It is also important to keep in mind that the IZY data do not account for successful breeding of species lacking international
studbooks (Conde et al. 2011) and that the data I analyzed applied to less than half of all high-risk species actually held in zoos (Conde et al. 2013). Thus, the counts are conservative in absolute terms. Furthermore, many of these species are highly distinct in a phylogenetic sense and of unquestioned high conservation value. Examples include multiple species of lemurs, great apes, and rhinoceroses. However, rates of growth have slowed (Figs. 1 & 2, & Supporting Information) as the number of studbooks has climbed over the past few decades, tracking the expansion in the number of species survival plans (overseen by the Association of Zoos and Aquariums) and European Endangered Species Programmes (overseen by the European Assocation of Zoos and Aquaria). In other words, individual zoos seem to have already set aside as much space for captive breeding of mammals as they deem fit. Along with the need to maintain genetic diversity, this fact helps explain why the AZA’s Taxon Advisory Groups set specific population size targets (Earnhardt et al. 2001). These self-imposed limits may be reasonable. For example, at the end of 2011, zoos reporting to the IZY housed 47,873 individual mammals (Zoological Society of London 2013). The 101 reported mammal populations with international studbooks by themselves totaled 30,904 individuals, and many of the other populations that are not under management are simply too small to maintain adequate genetic diversity (Earnhardt et al. 2001). Thus, the evidence suggests there is literally not enough space in the world’s zoos right now to substantially increase the number of truly successful captive breeding programs. There are 2 key counterarguments. First, many zoos have systematically decreased their holdings over the last several decades and may be able to increase them again. For example, in 1970 the London Zoo held 809 individual mammals, but by 2011 this figure had fallen to 586. Husbandry considerations presumably drive such trends: zoos are making more space for individual animals to improve their care and general well-being. However, favorable birth:death ratios in the 1970s (Figs. 1 & 2, & Supporting Information) suggest that if populations were increased, there would be no great cost from the standpoint of captive breeding. Second, much effort is expended on species that are not endangered or are otherwise not good candidates for eventual restocking of wild populations (Snyder et al. 1996). Indeed, of the 101 studbooks listed in 2011 only 50 pertained to endangered species or subspecies, and only 14,393 individuals belonged to those taxa. Most of these populations are of large-bodied species such as primates, carnivores, and ungulates (Supporting Information) that are drivers of attendance but are also expensive to maintain in captivity (Conway 1986). Refocusing attention on smaller taxa could be a cost-effective way to expand captive breeding programs of greater relevance for conservation.
Conservation Biology Volume 00, No. 0, 2015
6
Acknowledgments The author is the recipient of an Australian Research Council Future Fellowship (project number FT0992161). I thank the editors and anonymous reviewers for helpful feedback.
Supporting Information Figures illustrating historical trends in the number of captive breeding programs, demographic parameters, and key variables related to the size of 12 major zoos; rates of increase of individual species in individual years; and trends in numbers of censused individuals and the number of large zoos (Appendix S1) plus a list of species included in the analysis (Appendix S2) are available online. The authors are solely responsible for the content and functionality of these materials. Queries (other than absence of the material) should be directed to the corresponding author. Figure S1. Historical trends in the number of captive breeding programs and in demographic parameters. Time series start at 1977 because data were available for less than 10 programs in each year prior to that one. (a) Number of breeding programs. (b) Median net population growth rate (birth rate minus death rate). (c) Median per capita birth rate. (d) Median per capita death rate. Figure S2. Historical trends in key variables related to the size of the 12 zoos having the largest collections of tetrapods in 1991. (a) Median number of species of mammal species held. (b) Median number of individual mammals held. (c) Median number of staff members. (d) Median annual attendance. Literature Cited Balmford A, Leader-Williams N, Green MJB. 1996. Parks or arks: Where to conserve threatened mammals? Biodiversity Conservation 4:595– 607. Conde DA, Colchero F, Gusset M, Pearce-Kelly P, Byers O, Flesness N, Browne RK, Jones OR. 2013. Zoos through the lens of the IUCN Red List: a global metapopulation approach to support
Conservation Biology Volume 00, No. 0, 2015
Limits to Captive Breeding
conservation breeding programs. PLOS ONE 8 (e80311) DOI:10.1371/journal.pone.0080311. Conde DA, Flesness N, Colchero F, Jones OR, Scheuerlein A. 2011. An emerging role of zoos to conserve biodiversity. Science 331:1390– 1391. Conway WG. 1986. The practical difficulties and financial implications of endangered species breeding programmes. International Zoo Yearbook 24/25:210–219. Earnhardt JM, Thompson SD, Marhevsky EA. 2001. Interactions of target population size, population parameters, and program management on viability of captive populations. Zoo Biology 20:169–183. Fagan WF, Lynch HJ, Noon BR. 2010. Pitfalls and challenges of estimating population growth rate from empirical data: consequences for allometric scaling relations. Oikos 119:455–464. Fenchel T. 1974. Intrinsic rate of natural increase: the relationship with body size. Oecologia 14:317–326. Frynta D, ˇSimkov´a O, Liˇskov´a S, Landov´a E. 2013. Mammalian collection on Noah’s ark: the effects of beauty, brain and body size. PLOS ONE 8 (e63110) DOI: 10.1371/journal.pone.0063110. Henneman WW III. 1983. Relationships among body mass, metabolic rate and the intrinsic rate of natural increase in mammals. Oecologia 56:104–108. IUCN (International Union for Conservation of Nature). 2014. The IUCN red list of threatened species. Version 2014.3. IUCN, Gland, Switzerland. Lynch HJ, et al. 2010. Survivorship patterns in captive mammalian populations: implications for estimating population growth rates. Ecological Applications 20:2334–2345. Mace GM. 1986. Genetic management of small populations. International Zoo Yearbook 24/25:167–174. Martin TE, Lurbiecki H, Joy JB, Mooers AO. 2013. Mammal and bird species held in zoos are less endemic and less threatened than their close relatives not held in zoos. Animal Conservation 17:89–96. Rahbek C. 1993. Captive breeding—a useful tool in the preservation of biodiversity? Biodiversity Conservation 2:426–437. Savage VM, Gillooly JF, Brown JH, West GB, Charnov E. 2004. Effects of body size and temperature on population growth. The American Naturalist 163:429–441. Smith FA, et al. 2003. Body mass of late Quaternary mammals. Ecology 84:3403. Snyder NF, et al. 1996. Limitations of captive breeding in endangered species recovery. Conservation Biology 10:338–348. Soul´e M, Gilpin M, Conway W, Foose T. 1986. The millenium ark: How long a voyage, how many staterooms, how many passengers? Zoo Biology 5:101–113. Thompson SD. 1987. Body size, duration of parental care, and the intrinsic rate of natural increase in eutherian and metatherian mammals. Oecologia 71:201–209. Zoological Society of London. 2013. Zoos and aquariums of the world. International Zoo Yearbook 47:231–288.
