Weekly maximum, minimum and mean temperatures during the two cropping seasons are given in Fig. 1 together with the time ranges of anthesis and maturity of the three sowing dates. Temperatures in the second season were 3 to 6°C higher than in the first season. This was dearly affected grain yield and other related traits. Grain yield in the second season was reduced by 25 % compared with the first season. In agreement with Fischer and Maurer (1976), a 1°C rise in temperature reduced grain yield by about 4%.
In the first season, a reduction of 16% was found in grain yield of early sowing compared with the optimum sowing (Table 1). As well as grain yield, biomass, grains/m2, grain growth rate, biomass growth rate, vegetative growth rate, spikes/m2, days to anthesis and plant height were all reduced in different degrees by early sowing. On the other hand, grain filling duration was significantly longer in the early sowing. Harvest index, grains/spike, thousand kernel weight and days to maturity remained unaffected by the early heat stress.
The effect of late heat stress was clearly reflected in the great reduction in almost all traits measured (Table 1). Grain yield, biomass, grains/m2, spikes/m2, grain growth rate, biornass growth rate and vegetative growth rate were more heat sensitive. A reduction of 32% was found in grain yield of the late sowing compared with the optimum sowing. Similar reductions were also found in biomass, grains/m2, grain growth rate, biomass growth rate and vegetative growth rate. Reduction in harvest index, spikes/m2, thousand kernel weight, days to maturity, grain filling duration and plant height were all significant.
During the second season, the grain yield of the optimum sowing was the same as the early sowing, and only 8% higher than that of the late sowing (Table 1). This was due to that the grain yield of the optimum sowing in the second season was reduced by 35% compared with that of the first season while yields of the early and late sowings were reduced by 23 and 12%, respectively.
Significant reductions in the early sowing biomass, grains/m2,grain growth rate, biornass growth rate and vegetative growth rate were observed. On the other hand, harvest index, grains/spike, thousand kernel weight, days to maturity and grain filling duration were significantly better under the early sowing (Table 1). With the late sowing, all traits measured, except harvest index, spikes/m2, grains! spike and thousand kernel weight were significantly reduced compared with the optimum sowing. Across the two seasons, biomass, grains/m2,grain growth rate, biomass growth rate, vegetative growth rate and plant height were more sensitive to heat stress during early stages of growth.
High temperatures late in the season were more detrimental to most of the traits during the two seasons. Biomass, grains/m2, grain growth rate, biomass growth rate, vegetative growth rate, days to maturity, grain filling duration and plant height were negatively affected during both seasons. Harvest index and thousand kernel weight were significantly lower only in the first season. In the second season, the high temperatures during the grain filling period of the optimum sowing (Fig. 1) negatively affected the two traits. Limitation in the source to meet the demand of high sink might have been important beside the big reduction in the grain filling duration especially when compared with that of the early sowing.