Demographic and population-genetic tests provide mixed support for the abundant centre hypothesis in the endemic plant Leavenworthia stylosa ANDREA L. DIXON,* CHRISTOPHER R. HERLIHY † and JEREMIAH W. BUSCH* *School of Biological Sciences, Washington State University, PO Box 644236, Pullman, WA 99164, USA, †Department of Biology, Middle Tennessee State University, PO Box 60, Murfreesboro, TN 37132, USA Abstract The abundant centre hypothesis (ACH) assumes that population abundance, popula- tion size, density and per-capita reproductive output should peak at the centre of a species’ geographic range and decline towards the periphery. Increased isolation among and decreased reproductive output within edge populations should reduce within-population genetic diversity and increase genetic differentiation among edge relative to central populations. The ACH also predicts asymmetrical gene flow, with net movement of migrants from the centre to edges. We evaluated these ecological assumptions and population-genetic predictions in the endemic flowering plant Leav- enworthia stylosa. Although populations were more spatially isolated near range edges, the geographic centre was surrounded by and not coincident with areas of peak population abundance, and plant density increased towards range edges. Per-capita seed number was not associated with distance to the range centre, but seed number/m 2 increased near range edges. In support of ACH predictions, allelic diversity at 12 microsatellite loci declined with distance from the range centre, and pairwise F ST values were higher between edge populations than between central populations. Coalescent analyses confirmed that gene flow was most infrequent between edge populations, but there was not an asymmetric pattern of gene flow predicted by the ACH. This study shows that among-population demographic variability largely did not support the ACH, while patterns of genetic diversity, differentiation and gene flow were generally consistent with its predictions. Such mixed support has frequently been observed in tests of the ACH and raises concerns regarding the generality of this hypothesis for species range limits. Keywords: demography, endemism, niche, range limit, species range Received 4 April 2012; revision received 7 December 2012; accepted 11 December 2012 Introduction Geographic range limits are a fundamental property of species and are thought to reflect limits on abundance, dispersal or local adaptation in nature (Antonovics 1976). The range of environmental conditions that can support natural populations of a particular species determines its fundamental niche, while the ability of individuals to disperse into an area and persist in the face of biotic interactions determines the extent of a species’ distribution (Sexton et al. 2009). Species ranges have therefore been defined as all the populations of the species within the realized niche (Hutchinson 1958; Holt et al. 2005). Within the realized niche, the environ- ment is seldom homogenous and commonly contains abrupt and/or gradual changes in abiotic and biotic pressures (Kirkpatrick & Barton 1997). These changes are predicted to influence individual performance, and in turn, population density, and these effects may Correspondence: Andrea L. Dixon, Fax: 509 335 3184; E-mail: andrea.dixon@email.wsu.edu © 2013 Blackwell Publishing Ltd Molecular Ecology (2013) 22, 1777–1791 doi: 10.1111/mec.12207