Evolution 37(3): 637-639 (May 1983)
Quantitative Genetic Response to Environmental Heterogeneity in Tribolium confusum
Dave F. Zirkle and Russel A. Riddle

One of the most fundamental and challenging problems in population biology concerns the relationship between genetic variation and environmental heterogeneity. Levene (1953) and Dempster (1955) were the first to examine theoretically the conditions under which diversifying selection in space or time, respectively, could maintain genetic variation. However, present interest in this problem has largely been stimulated by the discovery of vast amounts of allozyme variation at electrophoretically detectable loci (Harris, 1966; Lewontin and Hubby, 1966). Since that time, a substantial body of theoretical work concerning the effects of diversifying selection on the genetic structure of populations has been developed (reviewed by Christensen and Feldman, 1975; Felsenstein, 1975; Hedrick et al., 1976). As a result of these theoretical investigations, it is now generally accepted that diversifying selection in space and/or time could potentially account for a major portion of the observed allzyme variation.

In contrast, the relationship between quantitative genetic variation and environmental heterogeneity has rarely been considered. This seems rather odd since natural populations contain large amounts of quantitative genetic variation (reviewed by Istock, 1981) and polygenic characters often exhibit genotype x environment interactions (reviewed by Pani and Lasley, 1972). Thus, the genetic structure of many polygenic characters is such that diversifying selection could potentially maintain a significant portion of the observed quantitative genetic variation. In order to assess the relationship between genetic and environmental variability, we have estimated genetic variation for female pupa weight in a series of flour beetle (Tribolium confusum) populations subjected to either constant or variable environments.


The analysis of population means and additive genetic variances demonstrates that evolutionary forces have affected allele frequencies at loci that influence female pupa weight. Therefore, it is appropriate to ask whether or not the observed patterns of differentiation are consistent with a model based on diversifying selection. In the case of additive genetic variances, temporal diversifying selection that occurs within a generation should produce an increase in the variance of an additively based character (Slatkin, 1978). However, the estimates of additive genetic variance for female pupa weight do not follow this prediction. The population subjected to the temporally variable selection regime (W/C) contained an intermediate amount of genetic variation, while populations subjected to constant selection regimes (C and W) contained low and high amounts of genetic variation, respectively.

In the case of population means, diversifying selection should produce extensive genotype x environment interactions. However, in a previous factorial analysis involving replicate populations and selection regimes, we were unable to detect significant genotype x environment interactions for any of the ten polygenic characters measured, of which pupa weight was one (Riddle et al., 1982). Thus, the patterns of differentiation based on population means and additive genetic variances suggest that diversifying selection as a result of the selection regimes has not been a significant factor in determining the levels of quantitative genetic variation for female pupa weight in the populations.

Although the results of previous studies generally indicate that the genetic variance of electrophoretic (e.g., Powell, 1971; McDonald and Ayala, 1974; Powell and Wistrand, 1978), karyotypic (Da Cunha and Dobzhansky, 1954), and quantitative characters (Beardmore and Levine, 1963; Mackay, 1981) is greater in populations subjected to variable environments than control populations, there are several reports of no relationship between environmental heterogeneity and genetic variability (e.g., Gooch and Schopf, 1973; Ayala and Valentine, 1974; Somero and Soul6, 1974; Minawa and Birley, 1978; Mitter and Futuyma, 1979). Furthermore, in those reports of a significant association, there is the possibility that the environmental gradient is confounded with other factors known to influence genetic variability, such as population size. Finally, the presence of a genotype x environment interaction component, which would indicate diversifying selection rather than some other form of balancing selection, has not been established. As noted by Mackay (1981), all reported examples of a significant association between environmental heterogeneity and genetic variation are consistent with other types of balancing selection, such as heterozygote superiority. Thus, given the possibility of confounding factors and the failure to distinguish between various types of balancing selection, we are led to the conclusion that what exactly is being selected as the environment varies remains unanswered at the present time.

Genetic Polymorphism bibliography