We consider theoretical predictions of bet-hedging theory using life-history and environmental data from six populations of Columbian ground squirrels. Contrary to the pattern predicted by bet-hedging theory, juvenile survival, pre-reproductive survival, and the percentage of squirrels surviving to maturity were positively associated, whereas the adult/juvenile survival ratio and variance in reproductive output were negatively associated with the degree of iteroparity, generation length, and/or the age at maturity. Survival of immatures and the degree of iteroparity were lower, whereas the adult/juvenile survival ratio was higher, in environments with more variable daily temperatures. Survival of immatures was higher, adult survival was lower, and the adult/juvenile survival ratio was lower where the length of the growing season was most variable. Most of our data show (P < 0.05) a pattern reversed to that predicted by bethedging theory. The possible effect of variability in length of growing season on survival of immatures versus adults could have precluded bet-hedging from existing among these populations.Several theoretical and empirical studies have focused upon life-history tactics (survival and reproduction) and the environmental factors that might shape them (see reviews by Parsons, 1982;Stearns, 1976Stearns, , 1977Stearns, , 1980. The first comprehensive theory of life·history evolution, rand K theory (MacArthur and Wilson, 1967), made specific predictions about the natural relation· ships among (mammalian) survival, reproduction, and various environmental conditions without regard to population age-structure (i.e., maturation is late and survival is high in stable environments). A subsequent theory, bet-hedging theory (Charnov and Schaffer, 1973;Murphy, 1968;Schaffer, 1974;Stearns, 1976), made predictions mostly contradictory to those of rand K theory and additionally it considers population age-structure (Le., maturation is late, immature survival is low, and adult survival is high in variable environments). Specifically, the theory of bet-hedging predicts that mortality of immatures, variance in reproductive output, variance in mortality of immatures, and/or the ratio of adult survival to juvenile survival should be positively associated with the degree of iteroparity (extent of repeated reproduction; see Stearns, 1976), generation length, and/or age at maturity in natural, age-structured populations (Abrams). Stearns (1976) coined the term bet-hedging theory and argued that variable environments which affect juvenile survival more than adult survival should lead to the above relationships. Conversely, Hastings and Caswell (1979) argued that the predictions of Schaffer (1974) and Schaffer and Gadgil (1975) (and therefore Stearns, 1976) can be completely reversed given different mathematical assumptions of their models and that environmental variability affecting juvenile survival should favor a decreased degree of iteroparity.Quantitative attempts to test bet-hedging theory have been made in several ...