Repeatability analysis on morphological descriptors in the early stages of development

The insufficiency of morphological descriptors reveals the importance of studies related to this topic for the genetic improvement of soybean, with attention to the possible descriptors measured in early stages of plant development which allows obtaining fast results. The aim of this work was to estimate the coefficient of repeatability of some morphological descriptors in the early stages of development of soybean and the minimum number of evaluations necessary to predict the real value of genotypes. Five (5) experiments were carried out with 124, 93, 90, 16 and 16 genotypes of soybean for the first, second, third, fourth, and fifth experiment, respectively, with characteristics combined at the stage V3. It was concluded that the length of the first internode requires fewer measurements compared to other measured characteristics. With six measurements, it was possible to obtain 95% and 90% reliability for plant height in V3 and first internode length, respectively, by the methods ANOVA, CP (correl), CP(cov) and AE(correl). With seven measurements, it was obtained 90% reliability for the epicotyl length for all methods used; 85% for the petiole length of unifoliate leaf by method of CP(cov) and 90% by methods ANOVA, CP(correl), AE(correl). With 15 measurements, it was possible to obtain 90% reliability for the hypocotyl length, petiole length of trifoliate leaf and the angle formed by the insertion of the petioles of the unifoliate leaf for all methods used; and, for the rachis length of the first trifoliate leaf, 37 measurements would be necessary for reliability of 90% by methods of ANOVA, CP(correl), CP(cov) and AE(correl). Key words: Glycine max (L.), soybean breeding, number of evaluation, stability.