Glands, Gossypol Content, and Tobacco Budworm Development in Seedlings and Floral Parts of Cotton

Abstract

Larvae of tobacco budworm, Heliothis virescens (F.), were placed on seedlings and fed on detached fresh flower buds (squares) of cotton, Gossypium hirsutum L. Cotton entries included four parents, carrying all truebreeding combinations of the duplicate gland‐determining alleles Gl 2 , gl 2 , Gl 3 rat , and gl 3 , and all 12 F 1 hybrids, including reciprocals. Data were obtained on larval preference for seedlings; weight of larvae fed on fresh squares; gland density in seedlings; stigmas; and bolls; and gossypol content of cotyledons, seedling leaves, and squares. Gland density and gossypol content were highly correlated with the number of gland‐determining alleles present and with each other. Both were highly negatively correlated with larval seedling preference and larval weight. Gl 2 , the “A” subgenome allele, was more expressive in seedlings, while Gl 3 rat , the “D” subgenome allele, transferred from G. raimondii Ulbr., was more expressive in stigmagland density and square‐gossypol content. In fact, the number of stigma glands was as high, and larval weights were as low, in the monomeric gl 2 gl 2 Gl 3 rat Gl 3 rat as in the dimeric Gl 2 Gl 2 Gl 3 rat Gl 3 rat , thus suggesting the predictive value of stigma‐gland counts in these cottons, and the possible practical value of the monomeric as a breeding stock. Diallel analyses showed that most of the genetic variance in gland density, gossypol content, and larval response was additive. Dominance was virtually absent. Epistasis varied considerably, but contributed substantially to the total genetic variance of gland density in seedlings, stigmas, and bolls. Reciprocal and maternal effects were generally small and nonsignificant.