Seed banks, salmon, and sleeping genes: effective population size in semelparous, age-structured species with fluctuating abundance.

Previous studies reached contrasting conclusions regarding how fluctuations in abundance affect Ne in semelparous species with variable age at maturity: that Ne is determined by the arithmetic mean N among the T years within a generation (Ne approximately = T(N)t; monocarpic plants with seed banks) or the harmonic mean (Ne approximately T[symbol: see text]; Pacific salmon). I show that these conclusions arise from different model assumptions rather than inherent differences between the species. Sequentially applying standard, discrete-generation formulas for inbreeding Ne to a series of nominal generations accurately predicts the multigenerational rate of increase in inbreeding. Variability in mean realized reproductive success across years (kt) is the most important factor determining Ne and Ne/N. When abundance is driven by random variation in kt, Ne < or = T[symbol: see text] < T(N)t. With random variation in Nt and constant per capita seed production (C), variation in kt is low and Ne approximately T[symbol: see text]; however, if C varies among years, Ne can be closer to T[symbol: see text]. Because population regulation affects the genetic contribution of entire cohorts of monocarpic perennials, Ne for these species may be more closely approximated by T[symbol: see text] than by T(N)t. With density-dependent compensation, Cov(kt, Nt) < 0, and Ne is further reduced because relatively few breeders make a disproportionate contribution to the next generation.