Diapause induction
The photoperiods experienced during the larval period strongly affected the induction of pupal diapause (analysis of deviance; treatment χ25=312; p<0.001), with asymmetric results of switching daylength regimes. Constant exposure to 15-hour days resulted in 100% diapause induction, whereas 0% diapause induction was attained in all three treatments that ended with larvae experiencing 21-hour days (Fig. 2). In other words, exposure to long days in the fourth and final instar was sufficient for consistently activating nondiapause development, regardless of previously experienced daylength. Meanwhile, the opposite change in daylength, from 21 hours to 15 hours, only resulted in approximately half of individuals entering pupal diapause, even when switched as early as the third instar. Two thirds of the diapausing individuals in this treatment were males, reflecting an overall effect of sex on diapause propensity (analysis of deviance; sex χ21=8.83; p=0.003). Diapause patterns were similar across populations; there were no significant interactions with other explanatory factors. A single male from the univoltine Stockholm population was the only individual that entered diapause upon being switched from long to short days in the fourth instar, and Stockholm pupae also showed higher diapause rates when switched to short days in the third instar (Fig. 2), although this population difference was not quite statistically significant (analysis of deviance; population χ22=5.31; p=0.07).