Both lines showed transitions from non-diapausing offspring to diapausing offspring (Fig. 2). Comparing diapause incidence within lines and between the different T-cycles, we noticed in the northern line that for all T-cycles a steep transition from non-diapausing offspring to diapausing offspring occurred at longer photoperiods (Fig. 2a, b), although 100% diapause was not reached in T-18 h, T-27 h and T-30 h (Fig. 2a, b). In T-18 h, the critical photoperiod was the shortest of all T-cycles, followed by T-21 h. T-24 h and T-27 h showed a nearly identical CPP and in T-30 h the CPP was the longest. In summary, the northern line showed diapause response at all T-cycles.

Photoperiodic response curve for diapause induction of Nasonia vitripennis lines from a a, b northern and a southern c, d European origin in different light regimes of different T-cycles and photoperiods. Diapause incidence is plotted as percentage diapause as a function of day length (a, c) or as a function of night length (b, d)

In the southern line, diapausing offspring was only observed in T-21 h and T-24 h. T-21 h expressed a shorter CPP than T-24 h (Fig. 2c, d). The other T-cycles resulted for all LD regimes in less than 50% diapause. In T-30 h, we measured 4–12% diapause incidence for all LD cycles (Fig. 2c, d).

Comparing both lines with each other, we see that the northern line responded stronger and with distinct response to the different T-cycle regimes than the southern line. The southern line did not show diapause response when T-cycles deviation from 24 h was large, especially when T was longer than 24 h.

Vaze and Helfrich-Förster (2016) raised the question whether the absolute day length or rather the night length was measured by the photoperiodic system. Therefore, we replotted our data also as a function of absolute night length expecting that when absolute night length is measured, the curves would superimpose. A logistic regression model was used within each strain and day length or night length analysis to determine significant differences in inflection points between T-cycles, using identical slopes. For the northern strain, the analyses showed a highly significant dependency of diapause on day length, when different constants were estimated for each T-cycle treatment (Fig. 2a, b; n = 24, df = 18, dev. = 103.79, p < 0.0001). For the southern strain, the analyses showed a highly significant dependency of diapause on day length, when different constants were estimated for each T-cycle treatment (Fig. 2c, d; n = 24, df = 18, dev. = 113.28, p < 0.0001). The variation in inflection points for each T-cycle did not differ within each strain between the day length and night length analysis (Northern strain: sdday = 2.82, sdnight = 2.24; F4,4 = 1.59, p > 0.33; Southern strain: sdday = 4.05, sdnight = 4.21; F4,4 = 1.08, p > 0.47). Thus, diapause incidence plotted against night length (Fig. 2b, d) does not show better superposition than when diapause is plotted against day length (Fig. 2a, c), suggesting that both day length and night length influence diapause induction in both Nasonia strains.

In his complete Nanda-Hamner study, David Saunders used a Nasonia line originating from Cambridge, United Kingdom (52°12′19.213’’N, 0°7′18.541’’E; Saunders 1968, 1974), an intermediate latitudinal location between Oulu, Finland (65°3′40.16’’N, 25°31′40.80’’E) and Corsica, France (42°22′40.80’’N, 8°44′52.80’’E). Re-plotting Saunders results (Fig. 3) from T-cycles similar to cycle lengths that we used, we find similarities to both our northern and the southern data (Figs. 2a, c & 3). T-16 h, T-21 h, T-24 h, T-28 h show diapause induction at short photoperiods. T-32 h does not show more than 20% diapause induction, similar to results shown by the southern line (Fig. 2c). We conclude that the results of Saunders also support our results since those data show intermediate diapause induction responses when compared to our northern and southern data, while the geographical origin of the Saunders strain was also intermediate to the strains we used.

Circadian photoperiod landscapes for diapause response—comparison with Saunders 1974. Diapause response data for the northern (a) and southern (c) line are replotted as contour landscapes against T-cycle duration and photoperiod and compared with Saunders’ 1974 data (b), which were obtained from a Nasonia vitripennis line that originated from an intermediate latitude (Cambridge, UK; only data were plotted from similar T-cycle and photoperiod ranges as we used in Fig. 2)

To enable interpretation of the differences in photoperiodic responses between the northern and southern line, we investigated whether both lines may differ in their circadian light sensitivity. Five different white light pulse durations (0.3, 1, 4, 8, 16 h; Fig. s1, s2) and three different light intensities (9.37*1013 (low); 2.62 · 1014 (intermediate); 2.10 · 1015 (high) photons·cm-2·s-1) at 1-h (Figs. s3, s4) and 4-h (Figs. s5, s6) light pulse duration in an Aschoff type II protocol. Light pulses were provided at 12 time points spread out over a 24-h time axis and circadian phase shifts were quantified to provide 44 PRCs for males and females of the northern and southern strain. The resulting PRCs were also plotted as phase transition curves to enable classification as weak (type 1) or strong (type 2) circadian phase resetting. With increasing light pulse duration, males and females of the Northern line consentingly switch to strong resetting at lower stimulus strength than the southern line, indicating that the circadian system of the northern line is more sensitive to light (Table 1).

Likewise, at different light intensities, we see that males from only the northern line show strong circadian resetting to a 1-h light stimulus, while females from only the northern line show strong resetting to a 4-h light stimulus. This also indicates higher circadian light sensitivity in the northern line (Table 2).

To provide a better quantitative comparison of circadian light sensitivity for the different strains, we integrated the PRCs between ZT0 and ZT16 to provide a single phase shifting capacity value for each PRC. To compare the response capacity value to a single currency value for each light stimulus, we calculated the total photon dose provided by integrating light intensity over the duration of the stimulus. This data set now allows us to plot integrated circadian phase response against integrated stimulus strength for males and females of the different strains (Fig. 4).

Photon dose response curves for females and males from northern and southern lines. Each point is the integration of a phase response curve (area-under-curve over ZT0-16). Integrated values originate from light pulse duration experiments at high light intensity (Figs. s1 & s2, grey), light pulse intensity experiments using 1 h or 4 h light pulses (Figs. s3– s6, black), or both (grey filled black circle). Dashed lines mark the curve inflection points, indicating higher response magnitude, but lower circadian light sensitivity in females (a, c) than in males (b, d), and higher circadian light sensitivity in the northern line (a, b) than in the southern line (c, d). The sigmoidal curve fits for sex and strain described the data significantly (F10,26 = 4.82, p < 0.0001)

These results indicate that Nasonia females are less light sensitive than the males, but have a higher circadian response capacity than males (Fig. 4a, c vs. b, d). For both males and females, we see that northern line is more light sensitive than the southern line with about one order of magnitude differences (Fig. 4a, b vs. c, d; F6,30 = 10.64, p < 0.001). General oscillator theory would predict that for the northern strain this will result in less variation in phase angle of entrainment under different LD dark schedules. The consequences of this finding for the photoperiod responses in our partial Nanda-Hamner protocol will be further evaluated in the discussion (Fig. 5).

Graphical external coincidence – phase angel of entrainment (EX-PA) model explaining latitudinal variation in diapause induction using differences in light sensitivity. Higher light sensitivity (and thus stronger resetting) in the northern line results in a narrower distribution of phase angle of entrainment over the different T-cycles and photoperiods. This results in a narrower phase distribution of the light-sensitive phase (red box) over the T-cycles and as a result diapause occurs in all T-cycles (white asterisk). The southern line shows a wider phase angel of entrainment due to its lower light sensitivity and a wider distribution of the light-sensitive phase across the T-cycles. Consequently, diapause occurs in just two T-cycles with a period close to the intrinsic circadian period around 24 h