Discussion
The present results reinforce earlier findings from both P. aegeria and other species (Friberg et al., 2011) suggesting that photoperiodic control of the diapause/nondiapause developmental switch is asymmetric: activating diapause development required a consistent photoperiod signal for a longer time than activating nondiapause development (Fig. 2). In addition, we here demonstrate a similar asymmetry in pre-diapause development rate, a trait that has likely been under selection to match the diapause phenotype (Kivelä et al., 2013). An increase in daylength led to an immediate increase in development rate (further accelerated by prolonged exposure to long days), favorable for producing an additional generation; in contrast, the slow development typically seen in diapause-destined larvae was here only engendered by sustained exposure to short days (Fig. 3, Fig. 4). The results are largely consistent with P. aegeria possessing two alternative, overall modes of larval development, cued by photoperiod: slow, “diapause-track” development and fast, “nondiapause-track” development. It is evident that, while these two modes diverge early in life, they are not irreversibly locked states; instead, photoperiodic information is continually used throughout the larval period to update developmental plasticity. Switching between developmental modes does not appear to be instantaneous, but shows a degree of inertia, as the phenotype matching the new photoperiodic environment did not fully manifest until the next larval instar after the change in photoperiod had taken place (Fig. 3). This inertia may represent a delay in perceiving and acting on the changed photoperiod signal, and/or a delay in “resetting” the hormonal machinery that controls development and growth.
Meanwhile, a rather different result was obtained for another life history trait associated with the diapause switch: body size. Pupal weight has previously been shown to be larger for diapausing individuals at least in bivoltine P. aegeria populations, and especially in the Öland population (Aalberg Haugen et al., 2012; Aalberg Haugen & Gotthard, 2015). Here, differences between individuals reared under short versus long days were not visible until the pupal stage, suggesting that these inter-pathway differences, unlike those seen for development rate, do not emerge until late in larval development.