While moths responded the fastest under cool phosphor-coated LEDs (on average 12 s faster than under other treatments, figure 2 b ), the interaction between light CCT and emission spectrum was only a near-significant predictor of latency ( F 1,326 = 3.68, p = 0.06), but, in isolation, both were highly significant predictors (CCT: F 2,326 = 6.65, p = 0.002; spectrum: F 1,327 = 9.13, p = 0.003).
Temperature and humidity interacted as near-significant predictors of harmful response behaviour ( X 2 = 2.790, p = 0.095; individually temperature X 2 = 5.595, p = 0.018; humidity X 2 = 4.913, p = 0.027); there was a weak non-significant positive correlation between temperature and response ( t = 1.214, d.f. = 383, p = 0.226, r = 0.062) and a non-significant weak negative correlation between humidity and response ( t = −0.946, d.f. = 383, p = 0.345, r = −0.050).