The survival of captive-born animals in restoration programmes – Case study of the endangered European mink Mustela lutreola Tiit Maran a,b, * , Madis Põdra b , Merje Põlma a , David W. Macdonald c a Species Conservation Lab, Tallinn Zoological Gardens, Paldiski Road 145, 13522 Tallinn, Estonia b Institute of Mathematics and Natural Sciences, Tallinn University, Narva Road 25, Tallinn 10120, Estonia c Wildlife Conservation Research Unit, Department of Zoology, University of Oxford, Tubney House, Abingdon Road, Tubney, Oxon OX13 5QL, UK article info Article history: Received 22 May 2008 Received in revised form 17 February 2009 Accepted 3 March 2009 Available online 5 April 2009 Keywords: Survival analysis Estonia Re-introduction Failure time analyses Kaplan–Meier estimates Cox proportional regression abstract As part of a conservation initiative, captive-bred individuals of the endangered European mink Mustela lutreola were released to a Baltic island ‘sanctuary’, Hiiumaa Island (Estonia), and their survival was mon- itored over eight years. Altogether, 54 released mink were equipped with radio-collars between 2000 and 2003 and their survival and causes of death were studied as a function of sex, age, housing conditions before release and number of generations in captivity. Mortality was greatest ( 50%) during 1–1.5 months following release. The survival of males was significantly higher than that of females (25% decline in 21 days, and 10 days, for males and females, respectively). Releasing pregnant females appeared to be an unsuccessful strategy. There was no evidence that the number of generations for which the lineage of the released individuals had been bred in captivity had any effect on survival. The main cause of death was other carnivores and raptors, although this broad categorization may conceal a diversity of fatal sce- narios. We recommend for the future that the European mink being prepared for release should be maintained in large naturalistic enclosures beforehand, that a preponderance of females should be released, and that the indications that younger animals make better candidates for release should be investigated. Ó 2009 Elsevier Ltd. All rights reserved. 1. Introduction Restoration is a major element of practical conservation, and the release of captive-born individuals into the wild can be an important component of it. Indeed, re-introduction biology is an important emergent field (Seddon, 1999; Seddon et al., 2007) and there is a growing number of reviews of release methodologies and their effectiveness (e.g. Breitenmoser et al., 2001; Griffith et al., 1989; Reading and Clark, 1996; Seddon et al., 2005). In a classic review by Griffith et al. (1989), 700 release opera- tions conducted in 1973–1986 were considered with respect to such variables as the number of founders, availability of shelters, origin of founders (captive versus wild animals), taxonomic origin of the animals, repeated releases versus a single operation. Jule et al. (2008) recently showed that the use of wild animals in carni- vore restoration is, in general, significantly more successful than the use of captive-bred animals. Further, analysing the releases of carnivores, Reading and Clark (1996) concluded that too often the focus is on biological variables, while sociological and organiza- tional variables are unwisely neglected (see also Macdonald et al. 2002a,b). However, Breitenmoser et al. (2001) express doubts as to the validity of such generalisations as they are often based on incom- plete information, and Seddon (1999) emphasises difficulties and inconsistencies in defining re-introduction success. Despite these caveats, the survival of the released animals is surely a fundamen- tal benchmark of success, and evidence as to the factors associated with their fates will contribute usefully to the adaptive manage- ment of future releases. The factors potentially influencing the survival of released ani- mals can be divided into four classes: (a) biological variables, e.g. age and sex; (b) pre-release factors, e.g. maintenance conditions, pre-conditioning, experience with humans; (c) release methodol- ogy, e.g. along a continuum between hard and soft releases; (d) re- lease site characteristics, e.g. availability of suitable shelter and habitat, abundance of predators, availability of food resources and level of disturbance. These categories may interact with one another in their influence on survival. Systematic studies of the post-release survival of captive- born mammals, and especially Carnivores, are few. The effect of 0006-3207/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.biocon.2009.03.003 * Corresponding author. Address: Species Conservation Lab, Tallinn Zoological Gardens, Paldiski Road 145, 13522 Tallinn, Estonia. Tel.: +372 6943318; fax: +372 6578990. E-mail addresses: tiit.maran@tallinnlv.ee (T. Maran), madis.podra@yahoo.es (M. Põdra), merje.polma@tallinnlv.ee (M. Põlma), david.macdonald@zoo.ox.ac.uk (D.W. Macdonald). Biological Conservation 142 (2009) 1685–1692 Contents lists available at ScienceDirect Biological Conservation journal homepage: www.elsevier.com/locate/biocon