Degree Type


Date of Award


Degree Name

Doctor of Philosophy




Visual selection among and within S(,1) lines of 'Lancaster Composite', a genetically broad base maize (Zea mays L.) population, was used to develop 200 visually selected S(,2) lines. Traits visually evaluated included ear size, kernel set, maturity, stalk-rot resistance, stalk and root lodging resistance, disease and insect resistance, and general plant and line appearance. A randomly selected set of 200 S(,2) lines was also developed. The 400 S(,2) lines were yield tested in three Iowa environments in 1981, and their testcrosses to (B84 x B73) were evaluated in three Iowa environments in 1982 and one in 1983. Traits measured in the yield trials included grain yield, grain moisture, stalk and root lodging, stand percentage, dropped ears, and vertical root-pull resistance;Mean grain yield and stand percentage of the visually selected S(,2) lines, averaged over all environments, was 1.69 q/ha and 2.58% greater, respectively, than that of the randomly selected lines. Mean grain moisture and stalk lodging of the visually selected lines was greater and less, respectively, than that of the randomly selected lines in all environments, but no differences were significant. Differences between means of visually and randomly selected S(,2) lines were not significant for any other trait. No significant differences between means of testcrosses of visually and randomly selected lines were observed for any trait in the combined analyses. Genetic variance and heritability estimates were not consistently affected by visual selection;Correlations between S(,2) line traits and the same traits in testcrosses were 0.19, 0.59, 0.51, 0.48, and 0.29 for grain yield, grain moisture, stalk lodging, root lodging, and vertical root-pull resistance, averaged over both selection types. Only 12 of 50 entries were selected in common by Smith-Hazel selection indices utilizing S(,2) line and S(,2) line testcross progeny data. Few entries were selected in common by the two progeny types based on single-trait superiority. Rankings of entries in the two progeny types were dissimilar, making selection of lines with superior testcross performance unfeasible based on S(,2) performance per se.



Digital Repository @ Iowa State University,

Copyright Owner

Christopher Paul Clucas



Proquest ID


File Format


File Size

233 pages