Original scientific paper

Characterization and interrelationships of soybean [Glycine max (L.) Merrill] yield components during dry and humid seasons

2018, 19 (2)   p. 466-481

Ivan VARNICA, Sonja PETROVIĆ, Andrijana REBEKIĆ, Sunčica GUBERAC, Katarina JUKIĆ, Goran JUKIĆ


Global climate changes have caused a significant weather oscillation. The objectives of this study were to determine the influence of weather conditions on the association of soybean seed yield and yield components, to find out the magnitude and effect of each component and their reaction to environmental stress. Seventy-four soybean varieties from nine geographical origins were studied during two seasons (2015 with less rainfall and increased temperatures and 2016 with increased humidity and moderate temperatures) at the Croatian Centre for Agriculture, Food and Rural Affairs in Osijek, Croatia. Correlation and path analysis were used to examine the association of the studied traits. The variance analysis revealed significant (P<0.01) effect of genotype, year and their interactions on all examined traits. During both seasons the seed yield was in significant and positive correlation with seed number per plant, pods number per plant, seed number per pod and 1,000-seed weight. All the traits mentioned had a stronger correlation with the yield in the dry 2015 compared to the year 2016. The seed number per plant had the highest correlation coefficients and a high and positive direct effect on seed yield in both years. Therefore, the selection of high yielding genotypes based on this trait can be done directly regardless of the variable weather conditions. The hierarchical clustering of varieties resulted in eight clusters in both years, confirming high genetic variability of the examined varieties. In 2016 one cluster that mainly consisted of varieties typical for the breeding programs of this region was singled out.


correlation, hierarchical clustering, path analysis, seed yield, soybean, weather oscillation, yield component

 Download      Find similar articles

Share article

email    facebook    twitter

  • Sign in

    If you are an existing user, please sign in. New users may register.

Cookies help us deliver our services. By using our services, you agree to our use of cookies. Got it