Accessory gland size changes during recovery
We here tested how developmental temperature affected accessory gland (AG) growth in young males. As developmental temperature affected male body size (GLM with a gamma error distribution: Deviance = 1.075,df = 4, F = 228.33, P < 0.0001) and this scales with organ size, we corrected for body size in further analyses using wing length as a proxy. Overall, males were smaller on day 6 after eclosion when they developed at higher temperatures (29: 1.25mm ± 0.007; 29R: 1.23mm ± 0.007; 31: 1.21m ± 0.011; 31R: 1.17mm ± 0.009) than control males (1.37mm ± 0.007, see Supplementary table S4).
We observed a clear reduction in AG growth during the first six days after eclosion in males exposed to elevated developmental temperatures (Fig. 4A, table 2). The reduction in growth was gradual: the higher the temperature, the slower the increase in AG size, resulting in significant differences on day 6. Surprisingly, the growth trajectories did not differ for males raised at 29°C and kept at the growth temperature and those allowed to recover (see post-hoc test in Supplementary table S5). Interestingly, AG sizes of recently eclosed males (day one) were fairly similar. In a second analysis we only compared AG size of treated males and hence, excluding the control group from the analysis, using larval temperature and the possibility of recovery or not as fixed variables. This revealed that the 31°C recovery treatment resulted in males having significantly smaller AGs compared to males not allowed to recover (GLM with a gamma error distribution, recovery: F = 18.88, df = 1, P < 0.0001). By day six, the reduction in size was 12.8% for males of both treatments exposed to 29°C and 40.3% for those exposed to 31°C compared with the control group.
We next assessed whether the allometric relationship between the gland and body size is also altered. The regression analysis between both factors (see Supplementary table S6 and Fig. S3) indicates that in the two elevated treatments a hypometrical relationship prevails (b<1) highlighting that the gland is smaller than expected for the body size, while it tends to be isometrical in males grown at 25°C (b=1). Thus, differences in AG size can be attributed to altered growth patterns due to temperature. Overall, these results suggest that AG maturation is inhibited by elevated temperatures with males ultimately having smaller glands.