This experiment was conducted to evaluate the effect on evapotranspiration and yield of soybean according to different soil water conditions, and to find the optimum time and amount for irrigation in soybean cultivation. The difference between potential evapotranspiration (PET) and maximum evapotranspiration (MET) during growing season of soybean planted in lysimeter was higher during reproductive stage than during vegetative one. The maximum crop coefficient was obtained at beginning seed stage of soybean. Soil water coefficient of irrigation treatment was higher than that of non-irrigation treatment during soybean growth stage in field experiment. Grain yield was highest in lysimeter due to its high water use efficiency and evapotranspiration rate.

Water deficit is the primary constraint of soybean [Glycine max (L.) Merr.] yield, and a physiological understanding of processes affected by water deficit is a key step in identifying and improving drought tolerance in soybean. The objectives of this research were to evaluate biomass and nitrogen accumulation patterns and water use efficiency (WUE) as possible mechanisms associated with the drought tolerance of Jackson. Biomass accumulation of Jackson was contrasted with the PI416937, which also has demonstrated tolerance to drought. For water-deficit treatment, total biomass accumulation was negligible for PI416937, but biomass accumulation continued at approximately 64 % of the well-watered treatment of Jackson. Transpirational losses for Jackson and PI416937 were approximately the same for the water-deficit treatment, indicating that Jackson had superior WUE. Isotopic discrimination of 13 C relative to 12 C also indicated that Jackson had higher WUE. Results indicated that increased WUE for Jackson under water deficit showed it was tolerant to drought rather than had an avoidance mechanism.

Leaf area is critical for crop light interception, and thereby has a substantial influence on crop yield. This experiment was conducted to characterize the development of soybean [Glycine max (L.) Merr.] leaf area. Plastochron index and leaf relative growth rate of Jackson was contrasted with the PI416937, which also has demonstrated tolerance to drought. First, plastochron ratio (PR) and plastochron index (PI) were evaluated in greenhouse to compare the leaf growth rate between two genotypes under well-watered condition. There was reasonable constancy of PR between two genotypes. The PR means of Jackson and PI416937 were 0.41 and 0.44, respectively. A fairly smooth increase of PI during vegetative stage was observed. Second, the relative growth rates were graphed against leaf area, normalized with respect to final leaf area, under well-watered and water-deficit conditions. Leaf growth was sustained longer in well-watered condition than water-deficit condition and there was a sizable proportion of leaves which was ceased earlier their growth in water-deficit condition compared to well-watered condition. The leaf relative growth rate of Jackson until leaves had completed at 45% of their growth during water deficit period was higher than that of PI416937.

Extensive research has been conducted on effects of drought stress on growth and development of soybean but information is rather restricted on the limited-irrigation system by way of precaution against a long-term drought condition in the future. The experiment for limited-irrigation was conducted in transparent vinyl shelter at Asian Vegetable Research and Development Center (AVRDC), Taiwan in 1997. Two soybean varieties, Hwangkeum and AGS292, improved in Korea and AVRDC, respectively were used for this experiment. The relationships between normalized transpiration rate (NTR) and fraction of transpirable soil water (FTSW) in both varieties were similar that the NTR was unchanged until FTSW dropped to about 0.5 or 0.6. At FTSW less than those values, NTR declined rapidly. Days required to harvest in both varieties were significantly prolonged at IR6 treatment compared to any other treatments. Daily mean transpiration rate was significantly higher at IR5 treatment, as averaged over varieties. Similarly, water use efficiency was also high at 1R5 treatment. In both varieties, seed yield was the greatest at the IR5 treatment, as compared to any other limited-irrigation treatments, due to the increased seed number and high transpirational water use efficiency. The indices of input water and seed yield for the different limited-irrigation treatments against control indicated that Hwangkeum produced 59.6% or 60.7% of seed yield using 36.1% or 44.9% of input water, as compared to control, by irrigation at only R5 or R6 stages, respectively. The AGS292 produced 56.1% of seed yield with 35.4% of input water of control, when irrigated at R5 stage. The results of this study have elucidated that the limited irrigation at R5 stage in soybean can be minimized yield loss with such small quantity of water under the environment of long-term drought stress and the expected shortage of agricultural water in the future.

The effect of water deficits on soybean [Glycine max (L.) Merr.] could appear on seed quality through changes of morphological plant characteristics. Two Korean genotypes, Hwangkeum (determinate growth habit) and Muhan (indeterminate growth habit), were used to examine the influences of treatment stage and method of water deficit during reproductive growth period on yield and seed quality of soybean. Water deficit at R5 or R6 stages was as damaging to seed quality as double water-deficit treatments at R2+R5 or R2+R6. However, seed from double water-deficit treatment tended to have lower oxidation-reduction potential compare to the corresponding single water-deficit treatment. In comparison with Muhan, Hwangkeum had significantly greater oxidation-reduction potential value. Seed yield per plant in both genotypes depended greatly on seed yield of branches. However, the proportion of number of branch seed to total seed umber in Hwangkeum was increased as the water deficit was applied during later reproductive stage, whereas, in Muhan the proportion was lower. Water-deficit treatments including the single and double water-deficit treatments and non-stressed treatment were able to be classified into five groups for Hwangkeum and four groups for Muhan based on the influences on yield components, number of pod, number of seed, and single seed weight, using principal component analysis. In both genotypes, R2+R5 water-deficit treatment decreased number of pod and seed, but increased single seed weight. On the contrary, R6 or R2+R6 stress increased the pod and seed number, but decreased single seed weight