Dendrolimus tabulaeformis overwinters as third to fourth instar larvae at short days in autumn. Using 24-h light–dark cycles, the photoperiodic response curves were similar at 24 and 28▫C. The critical night length was 9 h 20 min at 24▫C and 9 h 50 min at 28▫C. Under non-24 h light–dark cycles, duration of scotophase proved crucial in the determination of diapause. In night interruption experiments using 24-h light–dark cycle, a 1-h light pulse falling 8 h in the darkness strongly averted diapause in comparison with other light pulses. Nanda–Hamner experiments showed two weak troughs of diapause inhibition, suggesting the possible involvement of the circadian system. However, Bünsow experiments did not support the evidence of the involvement of circadian oscillatory system in photoperiodic time measurement. These results suggest that photoperiodic time measurement in this moth shows a non-oscillatory ‘hourglass-like’ response model or a rapidly damping oscillator model.