Cytoge~letical studies designed to elucidate species interrelationships and to discern the evolutionary history of the genus GOSSY~~'ZLITT have been in progress in this laboratory for a number of years. These studies have been based pri~llarily on nlultivalent frequencies and segregation of illarlter genes in synthetic allohexaploids of hybrids between the tetraploid New World species and various wild diploid species of Gossypiun~. These amphidiploids exhibited meiotic cl~romosome pairing configurations, such as univalents and multivalents, which can lead to numerical irregularities in chron~osoine distribution, which in turn can result in unbalanced gametes and/or chromosome elimination (Sarvella, 1956; Phillips, 1962 and 1961). Randoill chromosome loss will have a non-' random influence on segregation frequency if one assumes that gametes deficient for a chromosoille are functional. Thus a gamete of the type An for a given iuarker gene becollles -a if the cl~romosome carrying the 4 allele is eliminated. Elimination of an n allele-carrying chromosome from gametes potentially An or nn will not be manifest in testcross progenies. The research reported here was designed to assess the effect of chrornosome loss on segregation frequences by ascertaining the frequency and pattern of variation in chromosome number, both in the direct products of illeiosis (microspores) and in testcross progenies of a series of synthetic allohexaploids. Results and Discussion The hesaploids used in this study were the colchicine doubled triploid hybrids between the New World natural amphidiploids, G. hi?-sz~tz/'/~? L., and G. bnrbade~lse L., and species from the h genome (G. arbo7.e~~~~ L.) and D genome IG. ~.ni1wo7zdii Ulb. and G. gossypioides (Ulb.) Standley1 groups of diploid cotto~ls. Three synthetic allohexaploids, chosen to represent a wide range for both univalent frequency and multivalent frequency, comprised the experimental materials:
Read full abstract