Effects of N Fertility on Plant Water Relations and Stomatal Responses to Water Stress in Irrigated Cotton'

Cotton (Gossypium hirsutum L.) was grown in Phoenix, AZ, in 1981 and 1982. In both years it was irrigated with either 10 or 15 cm of water per application; N was supplied at three rates within each water treatment. Stomatal conductances of the most recently expanded leaves were followed during irrigation cycles to determine effects of N fertility on stomatal responses to water stress. Differences in N status between treatments were small in 1981 but were much greater in 1982. In July 1982 (early in the fruiting period), stomata of low-N plants closed as leaf turgor neared zero, but stomata of high-N plants remained open well past this point. As the season progressed, the differential response to water stress was lost as stomata of high-N plants became more responsive to declining water potentials. Nitrogen had little effect upon predawn osmotic potential (determined by pressure-volume procedures as the water potential for zero turgor), but low N eliminated the diurnal osmotic cycling seen in high-N plants. Diurnal oscillations of osmotic potential in highN plants had an amplitude of 0.4 MPa early in the fruiting period, then declined to zero only to reappear as fruits matured and vegetative growth resumed. Nitrogen concentrations of the most recently expanded leaves also increased during regrowth. Seasonal changes in stomata] responses to water stress seemed unrelated to either osmotic potentials or overall leaf N status. The loss of the differential response to stress did coincide roughly with a decline in petiole nitrate-N concentrations to very low levels in all treatments. Seedcotton yields were significantly affected by water in 1981 and by water and N in 1982. In the 2nd year, full irrigation increased yields only on high N; on low N, full irrigation did not affect yield but greatly decreased wateruse efficiency. The effects of N fertility on stomata explain in part the complex effects of N on water use. Additional index words: Gossypium hirsutum L., Stomata] conductance, Osmotic adjustment, Water-use efficiency, Leaf area index. T HE growth and yield of cotton (Gossypium hirsuturn L.), like most crops, depend strongly upon the availability of N and water during the season. Management of these two inputs has received much attention. Several investigators have reported a significant interaction of N and water on yield (2, 5, 16). The primary effect of N may be to prolong the productive season by delaying maturity (5). However, Scarsbrook et al. (16) also found an interaction of N and water on water-use efficiency (WUE), suggesting that the two factors may affect the relationship between transpiration and biomass production (or harvest index). The fundamental plant water relations which may underlie such changes have not been reported. Recent work showed that in controlled environments, N nutrition exerts complex effects on cotton plant water relations. Suboptimal N increases stomatal closure in response to water stress (12, 14, 15). This increased closure during drought results from both greater accumulation of abscisic acid (ABA) in 'Contribution from the USDA-ARS, Western Cotton Research Lab, Phoenix, AZ 85040. Received 23 Jan. 1984. 2 Plant physiologists, USDA-ARS, Western Cotton Research Laboratory, Phoenix, AZ 85040. stressed leaves and greater stomatal response to increments of leaf ABA (12, 15). Similar effects of N on stomata] behavior have been seen in bean (Phaseolus z'ulgaris L.) (17), tea (Cae/ha sinensis L.) (10), and Panicurn maximum grown i m n controlled environments (8), but not in wheat (Triticum aestivum L.) (9), coffee (Coffea arabica L.) (18), or P. maximum grown outdoors (8). This paper describes experiments to explore the effects of N fertility on water relations and stomatal behavior of irrigated cotton in the field, and to compare those effects to effects on yield and WUE. MATERIALS AND METHODS Experiments were conducted at the Univ. of Arizona Cotton Research Center, Phoenix. The soil is an Avondale clay loam (a member of the fine-loamy, mixed, hyperthermic family of Typic Torrifluvents). Preplant irrigations were applied to wet the soil throughout the rooting zone. 'Deltapine 70' cotton was planted in rows 1 m apart on 2 Apr. 1981 and 7 Apr. 1982 and thinned after emergence to populations of about 100 000 plants ha. After thinning, berms were constructed around basins 10 rows wide X 30 m long arranged in a 6 X 6 Latin square. During subsequent irrigations, water was pumped into each basin through gated pipe. In 1981, all plots were irrigated on 21 May, 12 and 26 June, 9 and 24 July, and 7 and 24 August. In 1982, all plots were irrigated on 26 May, 18 June, 1, 15, and 29 July, and 13 August. The intermediateand high-N plots also received water on 25 Aug. 1982. Plots received eifher 10 or 15 cm at each irrigation. The latter amount allowed evapotranspiration to proceed at essentially the potential rate through most of the 2-week cycle (3). Total water applied during the season (including rainfall) was 137 and 102 cm for the fully irrigated and stressed plots in 1981, and 140 and 104 cm for the same plots in 1982 (except that the low-N plots received 125 and 94 cm in 1982). In both years, three N rates were studied. In 1981, plots received either a preplant application of N (low N), a preplant plus one postplant application (intermediate N), or a preplant plus two postplant applications (high N). The first and second postplant applications were on 3 June and 22 July. The total N applied (as urea) was 77, 161, and 218 kg ha' in the three treatments. In 1982, the low-N treatment received no applied N. In the fully irrigated plots, the intermediate and high-N treatments received 161 and 240 kg ha' as preplant plus postplant applications. In the stressed plots, the N rates were 84 and 163 kg ha' applied postplant in the intermediateand high-N treatments, respectively. These rates include 22 kg ha' urea-N applied to high-N plots as foliar sprays during August. The surface soil contained 24 mg kg-' NO-N before 1982 planting (mean across all treatments). No soil tests were run in 1981. Crop N status was followed in both years by analysis of petiole NO-N. Petiole samples were collected weekly, dried, ground to pass a 40-mesh screen, extracted with hot water, and analyzed by a modification of the technique of Cataldo et al. (1). In 1982, reduced N was also determined in expanded leaves at the top of the canopy. Sampling was