Introduction
Generation time is a key parameter describing the pace of key biological processes. In age-structured populations, generation time is determined by age-specific survival and reproduction (Cochran & Ellner 1992), so it has been suggested as a good descriptor of the pace of life-history strategies (Gaillard et al. 2005). Generation time is also in allometric relation with key characteristics of organisms such as its body size and its metabolic rate (Brown et al. 2004). It is related to the rate of mutation (Lehtonen & Lanfear 2014) and the time a population needs to replace itself (Bienvenu & Legendre 2015). Evolutionary responses to selection of a trait per unit time depend upon a population’s generation time (Lande 1982), and as such it is a measure that connects the demographic and phenotypic characteristics of a population with the potential rate of evolutionary change. Furthermore, generation time is related to the susceptibility of organisms to stochastic fluctuations in the environment (Sæther et al. 2005) and it is a key component of evolutionary rescue models (Chevin et al. 2010). Understanding the ecological processes affecting the generation time is therefore essential for predicting the rate at which organisms can adapt to environmental change.
Generation time is generally defined as a population-level attribute. There are several definitions of generation time for age-structured populations, all of which can be calculated from survival and reproduction projection matrices (Cochran & Ellner 1992; Bienvenu & Legendre 2015). For instance, generation time has been defined as the time it takes a population to grow by a factor of its net reproductive rate (Coale 1972), and also as the average time between birth events in an ancestral lineage going from daughter to mother (Bienvenu & Legendre 2015). Other definitions relate to the average age of mothers of newborns in the population (Charlesworth 1994), such as “mean age of the parents of offspring produced in a particular time period” (Cochran & Ellner 1992), “the mean age at birth events for all individuals in a cohort” (Caswell 2001) or a generation (Steiner et al. 2014). Importantly, all formulations provide similar relative metrics when applied to a range of scenarios (Ellner 2018). In this study, we focus on understanding the evolutionary potential and ecological factors shaping generation time of individuals, measured as the weighted mean age of successful contribution to the breeding population (i.e. the mean age of recruit-producing parents weighted by the number of recruits they produce at each age; Cochran & Ellner 1992; McGraw & Caswell 1996).
Among-species comparisons have shown that generation time predicts the position of species on the fast-slow continuum of life-history covariation (Gaillard et al. 2005, 2016). At the fast end are organisms prioritizing current reproduction; that mature early, have high reproductive rates, short lifespans and short generation times. At the slow end of the continuum are organisms prioritizing survival over reproduction; characterized by high survival rates, low reproduction rates and long generation times. Generation time thus reflects how organisms resolve life-history trade-offs between current versus future reproduction, such as the trade-off between reproduction and survival (Gaillard et al.2005; Wright et al. 2019). Despite its evolutionary importance, surprisingly little research has focused on understanding the sources of within-species variation in generation time. This is probably because generation time has generally been considered as a static attribute at the population or species level. However, generation time as an individual-level life-history trait can be estimated from an individual’s age dependent reproduction (McGraw & Caswell 1996), and has been suggested as a potential measure of an individual’s position on the fast-slow continuum of life-history covariation (Araya-Ajoyet al. 2018). Studies quantifying the sources of variation in generation time at the among-individual and among-population levels should provide important insights into its evolutionary potential and how it is shaped by ecological conditions (McGraw & Caswell 1996; Araya-Ajoy et al. 2018; Wright et al. 2019).
Ecological conditions are expected to shape the generation time of organisms. The role of density dependence in determining the tempo of life-history strategies has been a long-standing research topic in evolutionary ecology (Pianka 1970; Stearns 1976; Boyce 1984; Wrightet al. 2019). Density dependence was introduced as a driver of life-history strategies in the context of r - versusK -selection after island colonization (MacArthur & Wilson 1967). The general idea was that when populations are growing, density-independent selection will favor fast life-history strategies, but as populations approached their carrying capacity, density-dependent selection will favor slower life-history strategies. More recently, age-structured models of density-dependent evolution have been shown to provide general predictions concerning the factors that may affect the optimal age of reproduction (e.g. Engen & Sæther 2016). Therefore, while generation time may be constrained by reproductive and developmental trade-offs, it can also be shaped by selection in response to ecological pressures.
Our aim in this study is to determine the evolutionary potential of generation time as a measure summarizing patterns of life-history covariation, and to assess the role of population dynamics in determining generation time. We studied variation in generation time in a metapopulation of house sparrows (Passer domesticus ) in northern Norway. We first estimated individual generation times based on individual projection matrices and decomposed its variation into genetic and environmental sources. Next, we studied the relationships between individual generation time and first age of successful reproduction, lifespan, reproduction rate, lifetime reproductive success and individual expected growth rates (McGraw & Caswell 1996). We then studied how age- and density-dependent survival and reproduction in turn affected generation time. Finally, we studied how fluctuations in population growth affected the mean age at reproduction to test the prediction that when populations are expanding, generation times should be shorter, whereas when populations are decreasing or near their carrying capacity generation times should be longer.