BRRI dhan55 (V₁), BR6976-2B-15 (V₂) and tolerant check Hashikalmi (V₃) were collected from BRRI (Bangladesh Rice Research Institute).

Seven drought conditions were used as treatments that started from 20 days of seedling age. The pot experiment was done with Randomized Complete Block Design (RCBD). There were seven treatments of water deficiency, three replications, and three genotypes (63 pots) were used. Seven treatments were:

T₀ = Throughout the experimental period continuously irrigated (control).

T₁ = When the seedlings were 20 days old water deficit imposed for 7 days,

T₂ =When the seedlings were 35 days old water deficit imposed for 7 days,

T₃ = When the seedlings were 55 days old water deficit imposed for 7 days,

T₄ = When the seedlings were 75 days old water deficit imposed for 7 days,

T₅ = When the seedlings were 95 days old water deficit imposed for 7 days,

T₆ = When the seedlings were 115 days old water deficit imposed for 7 days.

Uniform size and shape of seeds were cured with Bavistin (5 g in 1/2 liter of water) for 20 minutes. Cured seeds were spreading in the Petridis with water. In March, 2014 sprouted seeds were scattered in pots.

Pots were filled up with 10 kg sandy loam soil. Earthen pots were used of 38 cm × 25 cm in size, sandy and sandy loam. The soil was fertilized 160-150-150 kg urea, triple super phosphate (TSP) and muriate of potash (MP) per hectare, respectively [11] .

Three uniform and vigorous seedlings were permitted to grow in the pots after seedling establishment. For confirming normal growth, the germinated seeds were usually irrigated when the seedlings were 20, 35, 55, 75, 95 and 115 days old water deficit was forced for seven days.

Data were collected about yield and yield attributing character at maturity stage.

Dry matter

Total dry matter (Figure 1) (root, shoot, panicle etc.) was measured after oven drying for 72 hours at 72˚C.

Weight of 1000-grain

After sun and oven drying thousand cleaned grains weight (g) was calculated with an electronic weighing scale.

Filled and unfilled grain

Filled and unfilled grain was counted up.

Spikelet sterility percentage

Spikelet sterility percentage was recorded from the main stem panicle. The calculation is given below-

Spikeletsterility ( % ) = Emptyspikelet/panicle Totalspikelet/panicle × 100

Harvest index (HI)

HI is the percentage of grain yield and biological yield [12] . HI was given below

HI ( % ) = Grainyield Biologicalyield × 100

All the data were analyzed and the means were separated by DMRT at 5% level of significance using MSTAT-C [13] .

Leaf dry weight was presented in Tables 1-3. There was 0.26 g in V₁ followed by 0.25 in V₃ and the lowest was 0.23 in V₂ in varietal effect. Leaf dry weight under different treatments varied significantly and ranged from 0.17 in T₁ and T₅ to 0.21 in T₀. In the interaction effect of variety and treatment the highest leaf dry weight (0.26 g) was recorded in V₁T₀ and the lowest weight (0.1267 g) was recorded in V₂T₁. Water stress significantly decreased plant total dry mass, but the proportion of changes differed among root, stem, and leaf, whereas leaf dry mass decreased [14] .

Weight of thousand grains (g) was shown in Tables 1-3. The highest weight of thousand grains was 16 g recorded in V₁ followed by V₃ and the lowest weight of thousand grains was 11.67 g in V₂ in varietal effect. The highest weight of thousand grains 22.00 (T₀) and the lowest weight 14.76 g in T₆ In combination effect (variety and treatment) thousand grains 22.66 g in V₃T₅ and 11.67 g in V₁T₃.

Weight of a thousand grains was different which depends on the individual grain weight. In this study, BRRI dhan55 and Hashikalmi possess the highest weight of thousand grains.

Values followed by some letter(s) did not differ significantly at 5% level of probability.

Values followed by some letter(s) did not differ significantly at 5% level of probability.

Values followed by some letter(s) did not differ significantly at 5% level of probability.

Considering the treatment effect the highest weight in T₀ and in combination the height weight thousand grains was in V₁T₀. Under drought conditions BR 6976-2B-15 was mostly source limited during the grain filling stage as a result, grain weight decreased.

The result has infirmity with the results of [15] that showed that weight of 1000 grains was reduced depending on soil moisture levels. [16] showed that water stress after flowering decreased the individual grain weight. [17] advocated that water stress reduced grain weight.

Total number of spikelets/panicles was shown in Tables 4-6. Total number of spikelets was 176.7 in V₁ followed by V₃ (169.0) and the lowest found 158 (V₂). Number of spikelets was highest 158 in T₀ and was lowest 139.6 in T₁. In combination effect the spikelet 200 in V₀T₁ and 120 in V₃T₃.

When photosynthesis became lower, all the spikelets did not get sufficient assimilates, as a result decreased the number of spikelets per panicle. The spikelets recorded highest in BRRI dhan55 (V₁). The spikelets found much lower in BR 6976-2B-15. Due to drought stress, the number of spikelets was decreased. After drought conditions the tolerant genotype would quickly recover their biomass, leaves and take no longer time to recover and then develop new growth. The less tolerant genotypes would lose their biomass, leaves and take much longer time to recover. Decreased spikelet’s might be due to inhibition of stomatal conductance, translocation of assimilate to the grains.

Number of empty grains/panicles was shown in Tables 4-6. The empty grains/panicle 11 in V₃ followed by 10.00 in V₂ and the lowest unfilled grain 9.58 inV₁. The highest unfilled grain was 12.38 in T₀ and the lowest 10.38 in T₁. The highest unfilled grain was found at 18.88 in V₃T₀ and the lowest at 10.27 in V₂T₃.

Values followed by some letter(s) did not differ significantly at 5% level of probability.

Values followed by some letter(s) did not differ significantly at 5% level of probability.

Values followed by some letter(s) did not differ significantly at 5% level of probability.

In this study, it was found that drought stress greatly reduced filled grain and increased the number of unfilled grain. Due to water stress, the current stomatal conductance decreased, as a result the current photosynthesis became lower, insufficient assimilates production was seen and its distribution to grains was insufficient, all the spikelets did not get sufficient assimilates which resulted increased the number of empty grains and decreased the number of filled grains ultimately causes yield losses [15] . The results have the similarity with the results of [16] observed that after flowering increased the number of empty spikelets per panicle under water stress. Before panicle initiation water stress reduces potential spike number and decreases translocation of assimilates to the grains, which results low in grain weight and increases empty grains [9] . [15] stated that under lower soil moisture levels reduced grain yield due to inhibition of photosynthesis and less translocation of assimilates towards grain due to soil moisture stress.

Reduction percent of filled grains is in Tables 4-6. The reduction percent of filled grains found 6.51% in V₃ followed by 6.33% in V₂ and 5.42 in V₁. The highest reduction percent filled grain was 8.00 in T₆ and the lowest was 6.74 in T₄. In combination effect the reduction percent of filled grain 10.97% in V₃T₄ and the lowest reduction percent filled grain 6.02% in V₁T₁.

In this study, it was found that the lowest reduction percent of filled grains was 5.42 in BRRI dhan55 (V₁) and the highest reduction of filled grain was found in Hashikalmi (V₃) and in case of treatment effect the highest reduction percent of filled grains was 7.84, 8.00 in T₀, T₆ respectively.

Therefore it is suggested that sterility percentage increased with increasing drought duration and number of unfilled grains, decrease in filled grains per plant, tiller number, panicle number, leaf number, plant height. A significant decrease in panicle number and filled grain per plant, increased in number of unfilled grain were the main causes of sterility percentage increase due to drought treatment. These results conform with the results of [10] , who observed increased sterility in rice under water stress conditions. This result also agrees with [16] , who observed that water stress after flowering, increased the number of empty spikelets per panicle. Increased unfilled grains per panicle under lower soil moisture level occurs which decreases translocation of assimilates to the grains, ultimately which results in low gain weight and increases empty grains [9] .

The data on total dry weight per plant (g) was presented in Tables 7-9. In varietal effect the highest dry weight was 64.79 g recorded in V₃ followed by 57.08 g in V₁ and the lowest weight was 44.82 g in V₂. In treatment effect the highest total dry weight per plant was 58.05 recorded in T₀ and the lowest was 34.79 in T₃.

Values followed by some letter(s) did not differ significantly at 5% level of probability.

Values followed by different letter(s) indicate significantly different from each other by DMRT at 5% level.

Values followed by some letter(s) did not differ significantly at 5% level of probability.

In combination effect the highest found 65.05 in V₁T₀ and the lowest found 28.66 in V₂T₄.

Due to drought stress conditions root, shoot, leaf and panicle dry weight decreased, as a result the total dry matter became lower. In this study, the highest total dry weight per plant was in V₁ and V₂ and the lowest weight was in V₂. This might be due to reduction in tiller number, panicle number and filled grain per plant, plant height, leaf area etc. All of this ultimately affected the grain yield under water stress treatment. The results also agree with the results of [14] who stated that drought stress significantly decreased plant total dry mass, but the proportion of changes differed among root, stem, and leaf, whereas leaf dry mass ratio was decreased. [17] stated that water stress reduced grain weight. [9] observed that water stress at or before panicle initiation reduces potential spike number and decreases translocation of assimilates to the grains, which results low in grain weight and increases empty grains. [15] stated that reduced grain yield under lower soil moisture levels might be due to inhibition of photosynthesis and less translocation of assimilates towards grain due to soil moisture stress. [18] advocated that drought affect rain fed rice systems. Root characteristics such as root length density, root thickness, changes in root dry matter.

The results of harvest index (%) were shown in Tables 7-9. Significant differences found among the varieties and the treatments for harvest index. Harvest index 0.40 was recorded in V₁ followed by 0.329 in V₃ and the lowest harvest index was found as 0.325 in V₂. In treatment effect the highest harvest index 0.40 was recorded in T₀ and 0.32 in the T₁ treatment. Considering the combination effect, harvest index 0.46 was recorded in V₁T₀ and 28.66 in V₃T₃.

In this study, the highest harvest index found was 0.40 in V₁ and the lowest harvest index in V₂. The results have the similarity with the results that harvest index was significantly influenced by moisture level in all rice genotypes [15] . Where water shortages occurred, harvest index was more conservative than biomass accumulation; harvest index was reduced only when water deficits severely decreased grain-yield [19] .

Yield/plant (g) was shown in Tables 7-9. Yield/plant 23.80 g was recorded in V₁ followed by 21.77 g in V₃ and 21.33 g in V₂. Yield/plant 22.08 g was recorded in T₀ control and the lowest yield/plant found as 14.44 g in the treatment T₆. In combination effect (variety and drought treatment) the highest yield/plant was found as 23.80 in V₁T₀ and the lowest yield/plant found 12.20 g in V₂T₃.

In this study, the highest yield/plant in BRRI dhan55 followed by tolerant check Hashikalmi and the lowest yield/plant in BR 6976-2B-15. The lowest grain yield per plant was recorded in V₂ genotypes. The results also have the similarity with the results of [15] who stated that reduced grain yield under lower soil moisture levels. Yield parameters were decreased with the increase of water stress in different growth stages of the crop. As a result of drought, the stomatal conductance and gas exchange were decreased. All of this ultimately affected the grain yield under water stress treatment. The yield components like grain number and grain size were decreased in wheat [20] . Water deficit during vegetative, flowering and grain filling stages reduced grain yield.