Effect of pre-diapause temperature on summer and winter diapause intensity was examined under both laboratory and field conditions. Under short photoperiods of 8L : 16D and 10L : 14D, all pupae entered diapause at 15, 18 and 20°C and the incidence of diapause dropped to 82.3% and 85.5% at 22°C, respectively. Under long photoperiods of 14L : 10D and 16L : 8D, the incidence of diapause decreased with increasing temperature and there were significant differences among temperatures. The incidence of diapause at 16L : 8D was significantly lower than that under14L : 10D at 20 and 22°C. By transferring diapause pupae induced under various temperatures (18, 20 and 22°C) at a short day of 10L : 14D or a long day of 14L : 10D, to 12.5L : 11.5D, 20°C, the duration of summer diapause induced under 22°C (mean 76.1 days) was significantly shorter than those under 20°C (mean 85.9 days) and 18°C (mean 90.9 days), showing that the incidence of summer diapause was positively linked to the intensity of summer diapause; whereas the duration of winter diapause induced under 10L : 14D was similar at 18°C (89.2 days), 20°C (88.7 days) and 22°C (89.2 days) and there were no significant differences. Field experiments also showed that the high rearing temperatures significantly decreased the incidence and intensity of summer diapause, but had no significant affect on the intensity of winter diapause. When the naturally aestivating pupae from the first spring generation (formed on 24 April) and second spring generation (formed on 15 May) were kept under summer conditions, the diapause duration of the first generation lasted for 107-166 days (mean 146 days), about twenty days longer than that of the second generation [lasted for 92-151 days (mean 126 days)]. All results reveal that the sensitivity to temperature prior to aestivation and hibernation was quite different.