Abstract
A study was undertaken to determine if phenotypic selection based upon the number of kernels would be effective. The second objective was to determine the most effective means of increasing the number of kernels per panicle. Three different populations were used in this study. Population 1 was derived from a cross of a line possessing a long panicle X a line with many kernels per panicle. Population 2 was derived from a cross between a line with many rachis branches per panicle X a line with many kernels. Population 3 was derived from a cross between a line with a long panicle X a line with many rachis branches per panicle. Phenotypic selection based upon number of kernels was not effective in any of the three populations. Differences in length of panicle, number of rachis branches, number of kernels, and yield per panicle were observed. Within populations correlation coefficients were calculated to determine which characters had the greatest affect on yield. Number of kernels per panicle was the most important component of yield in each population. Length of panicle was highly correlated with number of kernels and yield in two of the three populations. However, in the third population the length of panicle was negatively and significantly correlated with number of kernels and yield. Data from this study indicate that populations should be analyzed separately since they behave differently. Phenotypic selection for number of kernels is greatly affected by environmental conditions and selection procedures involving this character were not effective. However, selection procedures based upon length and number of rachis branches should be considered.
Pearson, Jimmy Dale (1973). Effectiveness of phenotypic selection based upon selected components of yield in grain sorghum, Sorghum bicolor (L.) Moench. Doctoral dissertation, Texas A&M University. Texas A&M University. Libraries. Available electronically from
https : / /hdl .handle .net /1969 .1 /DISSERTATIONS -157610.