Search

Cereal production is strongly influenced by flowering date. Wheat (Triticum aestivum L.) models simulate days to flower by assuming that development is modified by vernalization and photoperiodism. Cultivar differences are parameterized by vernalization requirement, photoperiod sensitivity, and earliness per se. The parameters are usually estimated by comparing simulations with field observations but appear estimable from genetic information. For wheat, the Vrn and Ppd loci, which affect vernalization and photoperiodism, were logical candidates for estimating parameters in the model CSM-Cropsim-CERES. Two parameters were estimated conventionally and then re-estimated with linear effects of Vrn and Ppd. Flowering data were obtained for 29 cultivars from international nurseries and divided into calibration (14 locations) and evaluation (34 locations) sets. Simulations with a generic cultivar explained 95% of variation in flowering for calibration data (10 d RMSE) and 89% for evaluation data (10 d RMSE), indicating the large effect of environment. Nonetheless, for the calibration data, the gene-based model explained 29% of remaining variation, and the conventional model, 54%. For the evaluation data, the gene-based model explained 17% of remaining variation, and the conventional model, 27%. Gene-based prediction of wheat phenology appears feasible, but more extensive genetic characterization of cultivars is needed.

Can observed life histories be successfully explained in terms of their current adaptive significance, or do we need historical data as well? This issue is examined for the case of the desert annual Eriogonum abertianum Torr. (Polygonaceae). Individuals of this species flower in the spring, or following the onset of summer rains, or both. In order to estimate survivorship probabilities and fecundities, large samples of marked plants were followed in each of two years at two sites. In both years the Sonoran Desert (lower summer rainfall, hotter) population experienced significantly more mortality during the foresummer droughts, and had a significantly greater proportion of spring-flowering plants, than the Chihuahuan Desert (greater summer rainfaill, cooler) population. This result was reproduced in a greenhouse study (Fox 1989). The model of selection comparing the spring-plus-summer flowering habit with spring-only flowering suggests that, in the years studied, expected summer fecundity may not have offset the risk of foresummer mortality in the Sonoran Desert population. Distributional data indicate that, rather than switching to a spring-only phenology as predicted by the model, the species' range ends where summer rainfall declines abruptly. Fossil and biogeographic evidence lend support to an adaptive interpretation of earlier flowering in the Sonoran Desert: They suggest a mainly summer-flowering origin for the species and an invasion of the Sonoran Desert during a period with greater summer rainfall than is the case at the present. However, the persistence of the spring-plus-summer habit is not explained by either the demographic or the historical data.