Responses of early and late potato cultivars to moderate drought conditions: agronomic parameters and carbon isotope discrimination

Our main objective was to test under moderate drought conditions the hypothesis that early cultivars behave better than late cultivars under conditions of drought occurring late in the season whereas late cultivars would be less susceptible to a drought occurring early in the season. Agronomic parameters [tuber dry weight, harvest index, dry matter content, tolerance to a decrease in water supply (TDWS) and carbon isotope discrimination (delta)] were measured in six potato genotypes (Solanum tuberosum L.) subjected in two field trials (1995 and 1996) to three levels of water supply (treatments: 1, rainfall + irrigation; 2, rainfall only; 3, rainfall reduced to stemflow during early growth). In both years, treatment 3 (lower water supply) reduced significantly tuber dry yield and harvest index in comparison with treatment 1 (higher water supply). Tuber dry matter content was increased significantly only by treatment 3 in 1996. For these parameters cultivars did not differ in their response to the lower water regimes except in the case of tuber dry matter content during one year (1996). TDWS defined as tuber dry weight of treated plots relative to the tuber dry weight of treatment 1 allowed more discrimination among the behaviour of cultivars in case of moderate drought than yield and delta. Our TDWS results confirm partly the hypothesis tested: in 1996, TDWS values of mid-late cultivars were higher than those of early cultivars in the case of early drought (treatment 3), indicating that later cultivars support better a drought stress in the early part of the season. Differences among cultivars in the responses of delta to the different water regimes could not be shown, and there was no difference between the values of delta observed for early and late cultivars. In our moderate drought conditions an escape mechanism of earlier cultivars could not be confirmed through the consideration of the delta values observed. The linear regression of delta on TDWS gave a positive slope significantly different from zero (n = 24) but the R(2) values were low (R(2) = 0.41).