Nanda-Hamner curves show huge latitudinal variation but no circadian components in Drosophila montana photoperiodism

Abstract Insect species with a wide distribution offer a great opportunity to trace latitudinal variation in the photoperiodic regulation of traits important in reproduction and stress tolerances. We measured this variation in the photoperiodic time-measuring system underlying reproductive diapause in Drosophila montana, using a Nanda-Hamner (NH) protocol. None of the study strains showed diel rhythmicity in female diapause proportions under a constant day length (12 h) and varying night lengths in photoperiods ranging from 16 to 84 h at 16°C. In the northernmost strains (above 55°N), nearly all females entered diapause under all photoperiods and about half of them even in continuous darkness, while the females of the southern strains showed high diapause proportions only in the circadian 24 h photoperiod. Significant correlation between the strains’ mean diapause proportions in ≥ 24 h photoperiods and critical day length (CDL; half of the females enter diapause) suggests at least partial causal connection between the traits. Interestingly, females of the northern strains entered diapause even in ≤ 24 h photoperiods, where the night length was shorter than their critical night length (24 h — CDL), but where the females experienced a higher number of Light:Dark cycles than in 24 h photoperiods. NH experiments, performed on the control and selection lines in our previous selection experiment, and completed here, gave similar results and confirmed that selection for shorter, southern-type CDL decreases female diapausing rate in non-circadian photoperiods. Overall, our study shows that D. montana females measure night length quantitatively, that the photoperiodic counter may play a prominent but slightly different role in extra short and extra long photoperiods and that northern strains show high stability against perturbations in the photoperiod length and in the presence of LD cycles. These features are best explained by the quantitative versions of the damped external coincidence model