Evaluation of Eleven Thermal Unit Methods for Predicting Soybean Development1

Abstract
Dates of emergence, flowering, beginning pod fill, termination of flowering, and physiological maturity were recorded for two soybean [Glycine max (L.) Merr.] cultivars from each of the maturity groups I to V. The cultivars were planted on five dates at three Missouri locations in 1971, 1972, and 1973, and at weekly intervals at Columbia, Mo., in 1972 and 1973. Daily maximum and minimum temperatures were used to calculate five thermal unit methods with no base temperature subtracted, four methods with a base temperature subtracted, and two methods using quadratic formulas.Thermal units with a base temperature subtracted, including the “growing degree day method” used by the corn industry, and thermal units using quadratic formulas were the best for predicting flowering. The number of thermal units was not better than calendar days for predicting post‐flowering development. Prediction of flowering was more accurate for early cultivars than for late cultivars, suggesting that another factor (probably day‐length) also influenced flowering and had its greatest effect on the later maturing cultivars.