Introduction
Most organisms exhibit plastic responses to the developmental environment. These responses are adaptive if they generate phenotypes suited to conditions in spatially and/or temporally heterogeneous habitats (Lande 2009, Monaghan 2008, Arnold et al. 2019). Indeed, developmental plasticity (i.e., the capacity of a genotype to express multiple phenotypes in response to early-life environments) is one route by which organisms could overcome the challenges of environmental heterogeneity (West-Eberhard 2003, DeWitt and Scheiner 2004; Snell-Rood and Ehlman 2021). For example, some species develop defensive morphologies when predators are detected, but otherwise, do not expend energy on these traits (e.g. Daphnia waterfleas, Parejko and Dodson 1991; larval amphibians, Newman 1989). Such plasticity is described by a reaction norm, which is a mathematical function that describes phenotype values across different environments. Consequently, reaction norms are useful tools to calculate, visualize, and evaluate differences in plastic traits among environments, populations, and individuals (Gomulkiewicz and Kirkpatrick 1992, Brommer et al. 2005, Monaghan 2008).
Reaction norms may vary among individuals and populations, potentially due to genetic variation (Ellis and Boyce 2008; Scheiner 1993, Murren et al. 2015). This indicates an opportunity for natural selection to shape plasticity in adaptive ways (Levis and Pfenning 2016) like any trait with additive genetic variation (Hillesheim and Stearns 1991, Gavrilets and Scheiner 1993, Scheiner 2002). Environmental heterogeneity creates conditions that allow developmental plasticity to be adaptive (Gomulkiewicz and Kirkpatrick 1992, Lande 2009), as selection will favor reaction norms that match phenotypes to different environments, thereby enhancing fitness. Furthermore, adaptive plasticity should arise if the environmental cues that generate a phenotype also predict the future environment in which the phenotype is expressed and where its fitness consequences are realized (Casal et al. 2004, Beldade et al 2011).
Variation in reaction norms among individuals of a population provides an opportunity for selection to act on plasticity, whereas variation across populations may signify past selection that has shaped plasticity in response to local environments. Both theory (Gavrilets and Scheiner 1993; Lande 2009, Levis and Pfenning 2016) and empirical studies (Suzuki and Nijhout 2006) indicate that under strong selection, the magnitude (i.e., slope) of reaction norms should be maintained and become homogenized, decreasing within-population variation. Alternatively, if distinct populations experience different environmental pressures with varying levels of heterogeneity, then both the magnitude and variation of reaction norms might change (Duffy et al. 2015). Moreover, experimental evolution studies demonstrate that patterns of plasticity can vary between populations (e.g., populations with or without plasticity; van der Burg et al. 2020), and this process could be reversed with artificial or natural selection. Ultimately, a population’s genetic structure may contain individuals that are more (or less) plastic to environmental conditions. Consequently, when a population is faced with a major environmental change due to natural or human-induced causes, or resulting from invasion, plasticity may be amenable to selection.
The aim of this study is to quantify variation in developmental reaction norms among individuals and within populations of a non-native lizard, the brown anole (Anolis sagrei ). Past work on A. sagrei (Warner et al. 2012; Pearson & Warner 2018; Hall & Warner 2022), as well as studies on other reptiles (Mitchell et al. 2018; Noble et al. 2018; Warner et al. 2018), demonstrate strong effects of egg incubation environments on developmental rate, offspring body size, locomotor performance, behavior, physiology, and fitness; however, within- and among-population variation in reaction norms has never been examined inA. sagrei , but among-population variation in embryonic reaction norms of other Anolis species has been documented (Goodman 2008; Goodman & Heah 2010). Here, we collected lizards from two islands that differ in habitat structure and, thus, the predominant nest environments in which embryos develop (i.e., open-canopy island with warm, dry conditions versus closed-canopy with cool, moist conditions). Individuals from each island were bred in a common garden, and we incubated eggs in one of two regimes that mimic natural conditions on each island. Because each females’ eggs were divided between treatments (i.e., split-clutch design), we could quantify among-individual and among-population variation in reaction norms for a range of fitness-related phenotypes.
Our novel study design helps fill two important knowledge gaps in studies of developmental plasticity. First, most studies of developmental plasticity consider the isolated effects of factors like incubation temperature or moisture, but such factors usually co-vary in predictable ways in the wild (e.g. warmer nests are often drier; Pruett et al. 2020). We need a better understanding of the effects of real nest environments on plasticity. For example, warmer temperatures result in greater sprint speed at hatching (Pearson & Warner 2018), but dry incubation conditions can decrease hatchling performance (Gatto & Reina 2022). Moreover, the effect of temperature on sprint speed may depend upon thermal variation, not just mean temperature (Hall & Warner 2020). Thus, we need studies that combine multiple nest conditions that represent real habitats to better understand developmental plasticity in the wild.
Second, we test the hypothesis that variation in offspring phenotypes results from the influence of the environment (i.e., open- vs closed-canopy conditions), parentally induced variation in phenotypes (i.e., family-group reaction norm intercepts), and plasticity (i.e., family-group reaction norm slopes). We address this hypothesis by quantifying the slopes of reaction norms (for morphological, physiological, and performance traits) for each family group and comparing them among families and populations. Significant variation in reaction norm slopes among family groups (indicative of genetic x environment interactions) would support this hypothesis and indicate potential for plasticity to evolve in response to future pressures. Our results have important implications for understanding how natural developmental environments generate phenotypic variation, and the capacity for populations to adapt to changing environments.