The objective of this work was to estimate the average length of phenological subperiods and of soybean development cycle as affected by maturity groups and sowing dates, and to determine the probability distribution functions that best characterize the distribution of this data. It was evaluated the exponential, gamma, lognormal, normal, and Weibull functions, and determined the one with the best fit to the simulated crop development data, using meteorological data collected from October 1968 to July 2012, at the main climatological station of Santa Maria, RS, Brazil. Crop development was simulated for different sowing dates, considering the maturity groups (GM): 5.9-6.8, early/semiearly cycle; 6.9-7.3, medium cycle; and 7.4-8.0, semilate/late cycle. The lognormal function has the best fit for the length of soybean subperiods and cycle. This length varies according to the sowing date, with bigger cycles in early sowings (October), and smaller ones at late sowings (December).
Glycine max; sowing dates; development models; probability density function; thermal summation
