This study documents the degree to which six populations of mosquitofish that shared common ancestors in 1905 had diverged by 1980. Seven hundred thirty-eight fish were raised to maturity in individual containers with controlled feeding and closely regulated water temperatures. The two populations from Hawaii inhabited reservoirs whose water level did not vary; the four populations from Maui inhabited reservoirs whose water level fluctuated dramatically on periods ranging from days to seasons. Island effects (effects of stable vs. fluctuating water levels) were significant for male and female age and length at maturity and for offspring weight. Females from stable reservoirs were larger and younger at maturity and had smaller offspring than females from fluctuating reservoirs. Males from stable reservoirs were smaller and younger at maturity than males from fluctuating reservoirs. Stock effects (effects associated with particular reservoirs) were significant for male and female age and length at maturity and growth rates, for the slope of the fecundity-length relation, and for offspring weights. No significant effects were observed for length-adjusted fecundity, interbrood interval, and longevity, for which sample sizes were smaller The degree of divergence among stock means ranged from 69% for female growth rate to 8% for female length at maturity, corresponding to a rate of evolution ranging from 0.49% to 0.06% of the value of the trait per generation, assuming 140 generations since 1905 and gradual change. The only significant correlation among stock means was between longevity and length at maturity Correlations between age at maturity and longevity, age at maturity and fecundity, and fecundity and weight of offspring were low and insignificant. These results do not fit neatly into any scheme that dichotomizes life-histories into a group of early-maturing, highly fecund organisms that have short lives and small young, and late-maturing, long-lived organisms that have a few large young. Correlations between measures of water level fluctuations in the reservoirs and the life-history traits of the fish whose parents had lived in those reservoirs were significant and consistent with field observations of the impact of fluctuations on mortality rates.