Photoperiod experiment
P. aegeria for the experiment were derived from field-mated females collected from three populations across Sweden in 2011. In late May/early June, females were collected from Stockholm (59.63°N, 18.52°E; univoltine population; 11 females) and Öland (56.62°N, 16.56°E; bivoltine population; 5 females); in August, females were collected from Skåne (56.29°N, 12.48°E; bivoltine population; 6 females). The experiment was carried out in two temporal blocks. The first block started in June, using first-generation offspring of the wild females from Stockholm (6 families) and Öland (5 families). The second block started in September, and used first-generation offspring of the wild females from Skåne (6 families), alongside second-generation offspring for Stockholm and Öland (1 family each). Apart from these differences, both experimental blocks used the same methods, and were analyzed together.
Shortly upon hatching from the egg, each larva was placed into an 0.5-liter plastic container containing a living tuft of bluegrass (Poa annua ). The grass in each rearing cup was replaced as needed, to ensure ad-lib access to food throughout the experiment. Cups were placed into climate cabinets (Termaks series KB8400L; Termaks, Bergen, Norway) set to 17°C and one of two photoperiods: short days (15 hours light / 9 hours dark) or long days (21 hours light / 3 hours dark). Each photoperiod was duplicated, for a total of four cabinets. To test the effects of changes in daylength information during development, larvae were assigned into six treatments (Fig. 1). The first two sets of larvae acted as control treatments: these were kept under constant daylength (either long or short) for the entire experiment. Another two sets of larvae were, upon molting into the third instar, moved to a cabinet set to the opposite daylength regime (from short to long, or long to short, respectively). The last two sets of larvae were likewise switched between daylength regimes, but not until later in development, at the molt to the fourth and final instar. After accounting for mortality, each combination of population and treatment was represented by 14–23 individuals (mean=18.9).
Experimental individuals were weighed at set points in development: on the day of hatching, using a Cahn 28 Electrobalance (Cahn Scientific, Irvine, CA, USA), as well as on the day of molting to the third instar, on the day of molting to the fourth instar, and two days after molting to the pupal stage, using a Precisa 205A balance (Precisa Gravimetric, Dietikon, Switzerland). The two-day wait for the pupae was to allow the pupal cuticle to harden, preventing damage during handling. In general, eggs and larvae were monitored daily to note the precise timing of hatching and molts. The exception was the Skåne population, where for logistical reasons the exact time (and hence also weight) at larval hatching could not be obtained. Individuals were sexed according to the number of genital slits in the pupal cuticle, and pupal development was monitored to determine whether diapause had been initiated. At 17°C, a non-diapausing P. aegeria pupa is expected to develop within 25 days or less (Nylin et al. 1989; Lindestad et al. 2020); here, eclosion occurred either after <23 days or >45 days, allowing the two developmental pathways to be clearly separated.