Effective size of harvested ungulate populations B.-E. Sæther 1 , S. Engen 2 & E. J. Solberg 3 1 Department of Biology, Centre for Conservation Biology, Norwegian University of Science and Technology, Trondheim, Norway 2 Department of Mathematical Sciences, Centre for Conservation Biology, Norwegian University of Science and Technology, Trondheim, Norway 3 Norwegian Institute for Nature Research, Trondheim, Norway Keywords effective population size; genetic drift; harvest strategies; moose; demographic stochasticity; age-structured population models. Correspondence Bernt-Erik Sæther, Department of Biology, Centre for Conservation Biology, Norwegian University of Science and Technology, Realfagsbygget, NO-7491 Trondheim, Norway. Email: bernt-erik.sather@bio.ntnu.no Received 25 February 2009; accepted 4 May 2009 doi:10.1111/j.1469-1795.2009.00278.x Abstract The harvest of ungulate populations is often directed against certain sex or age classes to maximize the yield in terms of biomass, number of shot animals or number of trophies. Here we examine how such directional harvest affects the effective size of the population. We parameterize an age-specific model assumed to describe the dynamics of Fennoscandian moose. Based on expressions for the demographic variance s 2 dg for a small subpopulation of heterozygotes Aa bearing a rare neutral allele a, we use this model to calculate how different harvest strategies influence the effective size of the population, given that the population remains stable after harvest. We show that the annual genetic drift, determined by s 2 dg , increases with decreasing harvest rate of calves and increasing sex bias in the harvest towards bulls 1 year or older. The effective population size per generation decreased with reduced harvest of calves and increased harvest of bulls 1 year or older. The magnitude of these effects depends on the age-specific pattern of variation in reproductive success, which influences the demographic variance. This shows that the choice of harvest strategy strongly affects the genetic dynamics of harvested ungulate populations. Introduction It has long been recognized that directing harvesting to- wards certain age or sex categories can increase the yield in terms of biomass or number of harvested animals (MacArthur, 1960; Beddington & Taylor, 1973; Slobodkin, 1973; Law, 1977). In general, the number which can be removed from an age class should be inversely related to their expected number of future offspring or the reproduc- tive value of Fisher (1930). Theoretical analyses of determi- nistic age-structured models often show that the greatest yield is obtained by harvest of only two age classes with small reproductive value (Beddington & Taylor, 1973; Bed- dington, 1974; Law, 1977). Similarly, simple theoretical analyses also showed that male-biased harvest will increase the yield dramatically (Caughley, 1977), provided that a sufficient number of males remain in the population to fertilize all sexually mature females. Many populations of ungulates across the world are subject to harvest either for consumptive use (e.g. meat or trophies) or for reducing the damage to agriculture or forestry (Hudson, Drew & Baskin, 1989). Such harvest often involves large changes in the social structure and the age and sex composition of the populations (Ginsberg & Milner- Gulland, 1994; Milner-Gulland et al., 2003), often with large demographic consequences (Coulson et al., 2004) that will influence the genetic dynamics (Allendorf et al., 2008) as well as evolutionary responses to changes in the environ- ment (Kuparinen & Meril¨a, 2007; Proaktor, Coulson & Milner-Gulland, 2007; Fenberg & Roy, 2008). Fennoscan- dian moose represents one of the most extreme examples of such harvest-induced structural changes of populations because the harvest has been directed towards young ani- mals and bulls over large areas for several decades (Solberg et al., 2002, 2006; Lavsund, Nygren & Solberg, 2003). Because annual variation in offtake is closely related to changes in quotas (Solberg et al., 1999), this has resulted in strongly skewed sex and age distributions of moose popula- tions in many parts of Fennoscandia (Solberg et al., 2002, 2005, 2006). Although the strong directional harvest has contributed to the large increase in yield over the past decades (Lavsund et al., 2003; Solberg et al., 2005), the changes in social structure and age composition affect the genetic dynamics in the populations (Ryman et al., 1981). The concept of effective population size is crucial for understanding changes in gene frequencies in finite populations because it determines the rate of random genetic drift and interacts with natural selection to influence the probability of fixation of deleterious and advantageous mutations. Wright (1931) defined the effective size of a diploid population n e as that which produces the same rate of random genetic drift as an ideal population of constant size, n, reproducing by random sampling of gametes with non-overlapping generations and Animal Conservation 12 (2009) 488–495 c  2009 The Authors. Journal compilation c  2009 The Zoological Society of London 488 Animal Conservation. Print ISSN 1367-9430