The influence of natural incubation environments on the phenotypic traits of hatchling lizards

Ecology, Dec, 1997 by Richard Shine, Melanie J. Elphick, Peter S. Harlow

INTRODUCTION

The role of direct environmental influences on phenotypic traits has attracted increasing scientific attention in recent years, and reptiles have proved to be excellent model systems for research in this field. Research on phenotypic plasticity in these animals was stimulated by the discovery of temperature-dependent sex determination (e.g., Bull 1980), leading investigators to ask if incubation regimes (especially, thermal and hydric environments during incubation) affected aspects of the animal's phenotype other than sex determination. Many such effects have now been documented, using a diversity of reptile taxa (turtles, crocodilians, snakes, lizards) and examining a number of different organismal traits involving morphology, locomotor performance, and general behavior (see reviews in Rhen and Lang 1995, Shine and Harlow 1996, Roosenburg 1996). This work has led to suggestions that incubation-induced phenotypic plasticity may play an important role in many biological processes, including ecological phenomena (e.g., nest-site selection, micro- and macrogeographic variation in life histories: Gutzke and Packard 1987, Viets et al. 1993, Resetarits 1996, Roosenburg 1996, Shine and Harlow 1996) as well as evolutionary phenomena (e.g., shifts in reproductive mode, maternal behavior, mode of sex determination: Rhen and Lang 1995, Shine 1995, Shine et al. 1995, Tousignant and Crews 1995, Quails and Shine 1997).

This literature has shown an increasing methodological sophistication. Most early experiments relied on constant-temperature incubation, but more recent work has attempted to measure and then simulate conditions experienced in natural nests (e.g., Shine and Harlow 1996). Inevitably, this is difficult to do: natural nests show enormous spatial and temporal variation in traits such as the means and variances of temperature and soil moisture levels (Packard et al. 1977, Packard and Packard 1988, Palmer-Allen et al. 1991). Thermal variances as well as mean temperatures can affect hatchling phenotypes (Shine and Harlow 1996). Even when experiments simulate appropriate regimes with respect to one variable (e.g., temperature), they may introduce artifacts with another (e.g., moisture). For example, it is almost impossible to ensure that eggs incubated at different temperatures are maintained at identical water potentials (Packard 1991, Shine 1995). Although similar functional dependencies among physical conditions may also occur in natural nests, we do not know if this is the case.

In summary, laboratory experiments show that incubation conditions can influence the phenotypes of hatchling reptiles, but it is difficult to extrapolate this result to the field. For example, natural nests within a population may not vary enough in incubation conditions to engender significant phenotypic variance in hatchlings. Even if nests vary in such ways, successful hatching may only occur from a subset of nests - potentially, those with a restricted set of incubation conditions. Alternatively, even if factors such as temperature and soil moisture levels vary among natural nests to a degree that would influence hatchling phenotypes if each factor acted alone, the overall impact of these factors may be reduced by the patterns of covariation of temperature and moisture in natural nests.

The only way to overcome laboratory artifacts such as these is to study eggs in natural nests. This technique has been adopted in studies of temperature-dependent sex determination, and has demonstrated that clutches from natural nests display sex-ratio biases consistent with those predicted from laboratory studies (Bull 1980, Janzen and Paukstis 1991, Roosenburg 1996). However, effects of incubation conditions on other aspects of the hatchling phenotype (morphology, behavior, locomotor performance) are difficult to separate from maternal (including genetic) effects on these attributes, if all that is available is information on the phenotypes of naturally incubated hatchlings. The central problem is that in nature, hatchlings emerging from a single nest will be similar in their genetic constitution as well as in the physical conditions that they have experienced during incubation. Thus, any consistent among-nest differences in phenotypic traits of hatchlings may be due to genes (or more generally, to "nest-of-origin" effects) rather than to incubation regimes. In order to separate out these confounding factors, we moved eggs from one nest to another. By randomizing the genetic constitution of the hatchlings emerging from each nest, any consistent differences in the phenotypes of hatchlings emerging from different nests must be due to the nest environment rather than to maternal factors (Packard et al. 1993, Cagle et al. 1993).

This experimental design also allowed us to quantify the proportion of phenotypic variance in each trait attributable to incubation effects vs. maternal (including genetic) factors, and to examine the validity of several conclusions from laboratory experiments - for example, the notions that hatchling phenotypes are influenced by thermal variance as well as by the mean incubation temperature (Shine and Harlow 1996), and that the sexes respond differently to the conditions they experience during incubation (Shine et al. 1995). Lastly, we were able to test the idea that among-clutch differences in the reaction norms of embryos influence maternal nest-site selection, so that mothers deposit their eggs in nests best-suited to their own embryos (Shine and Harlow 1996).