Effect of Water Deficits on Seed Development in Soybean

Abstract

Water deficits during seed filling often decrease seed size in soybean (Glycine max L.). The physiological basis for this response is not known but may result from direct effects of low seed water potential (',) on the seed filling process. To determine whether low l' occurred in reproductive tissues of soybean, we monitored the water status (',I, ls, and I,p) of leaf, pericarp, and seed (embryo and testa) tissue of greenhouse-grown plants subjected to a brief water deficit during the linear period of seed growth. Water deficits were imposed by withholding water and monitored in the reproductive tissues by thermocouple psychro-metry. When water was abundant, leaf, pericarp, and seed I, were-0.5 to-0.7 megapascal at midday. When water was withheld, leaf l' decreased to-2.3 megapascals within 6 days. Pericarp 1' also decreased to-1.9 megapascal during this time. Pericarp ls followed the decline in *', but osmotic adjustment was not evident as the pericarp lost turgor completely by day 6. However, seed ',, I,, and 'ip were not significantly different from the controls. These results indicate that the water status of the developing seeds of soybean is not altered by short-term water deficits severe enough to inhibit the metabolic activity of the maternal plant. Maintenance of a favorable water status may be important for the conservation of seed growth rate exhibited by soybean under dry conditions. Water deficits during reproductive development often decrease seed size in soybean (Glycine max L.) (18, 21, 26). Recent evidence indicates that the reduction in seed size is due primarily to a shortening of the seed filling period rather than an inhibition of seed growth rate (20, 24). Since low TwI2 are known to affect the metabolic activity of the vegetative plant (5, 10, 12, 25), the maintenance of seed growth rate implies either that seed metabolism is relatively insensitive to low TI or that the water status of developing seeds is not affected by water deficits that develop in the maternal tissues. To determine whether low T,, occurred in reproductive 'Contribution from the USDA-Agricultural Research Service-MWA, North Central Soil Conservation Research Laboratory, in cooperation with the University of Minnesota West Central Experiment Station, Journal Series No. 16,809. 2 Abbreviations: T', water potential; T', osmotic potential; p turgor; HID, high intensity discharge; PAR, photosynthetically active radiation. tissues of soybean, we examined the water status (TI, Is, and T,) of leaf, pericarp, testa, and embryo tissues of soybean in response to a water deficit imposed during linear seed growth. In a companion paper (29), we investigated the relative effects of this water deficit on the metabolic activity of maternal and zygotic tissues simultaneously to ascertain whether seed growth might be limited by assimilate supply or by sink activity at low *I'. MATERIALS AND METHODS Culture Conditions Soybean (Glycine max L. cv Evans) plants were grown in pots containing approximately 18 kg of soil mix (2:3:1 [v/v] soil:sand:vermiculite) in the greenhouse (day/night air temperature , 28 ± 5/18 ± 2°C; RH, 50/80 ± 20%). Ambient light was supplemented for 15 h with HID (sodium vapor) lamps that supplied approximately 500 ± 100 uEm-2 s-' (PAR) at the level of the upper leaves. At planting, the soil mix was saturated with a modified Hoagland solution No. 1 (16) containing one-half the total N and supplemented with mi-cronutrients according to media No. 5 of Hewitt (15). Each pot was thinned to three uniform plants at the V2 stage (11). One L of additional fertilizer solution was applied 2 weeks after emergence and at weekly intervals thereafter. Water deficits were imposed by withholding water and nutrients at the R5.5 stage (11) approximately 30 d after anthesis at node 7. This corresponds to the period of rapid seed growth and initial seed dry weights of 60 to 70 mg. Control plants received adequate water throughout development. Water Potential Measurements Preliminary measurements indicated that variability in seed size, seed number per fruit, and node number did not affect pericarp or seed *I'. However, the sampling procedure was standardized to 3-seeded fruit from nodes 7 to 11 to ensure uniformity between experiments. At 0, 3, or 6 d after water was withheld, one or two fruits were sampled from each plant in the control and treatment groups. Fruit from each pot were pooled so that each sample represents the average of three to six fruits from three plants. To minimize water loss from the tissue during sampling, excised fruits were enclosed immediately in an airtight plastic bag and transferred to a humid 975