THE CYTOGENETICS OF GOSSYPIUM AND THE ORIGIN OF NEW WORLD COTTONS

Abstract

It has long been accepted by students of the genus Gossypium that the New World cottons (G. hirsutum L., G. barbadense L., and G. tomentosum Nutt.) are natural amphidiploids combining an A genome from a taxon of the Asiatic diploid group and a D genome from a taxon of the American diploid group (Skovsted, 1937). Beasley (1940b) and Harland (1940) provided confirming evidence on the origin of the New World type by synthesizing amphidiploids from A x D hybrids that were similar in morphology and cytogenetically compatible with the natural amphidiploids. More recent research on the general question of the origin of New World cottons has been directed toward pinpointing the taxa that were involved in the original hybrid and to the problem of when and how they were brought together so that hybridization could take place. These two questions form the basis of this paper, and they will be treated in turn. Triploid hybrids of New World (genome formula, AD) x Asiatic (A) and New World x American (D) show 13 II and 13 I at metaphase I (MI) of meiosis (this evidence supported Skovsted's hypothesis of amphidiploid origin). But since each of the two AD x A triploids and all of the seven possible AD x D triploids show essentially the same number of bivalents at MI, there is little basis here for identifying the parents of the New World amphidiploid. A method described by Stephens (1950) and used extensively in this laboratory (see review in Phillips, 1962) takes advantage of the segregation frequencies of marker genes in synthetic allohexaploids to discern interrelationships among the A, D, and AD taxa. A brief summary of the results of these studies (Gerstel, 1956; Gerstel and Phillips, 1957; Phillips and Gerstel, 1959; Phillips, 1962) appears in table 1. The relationships indicated in the table show that G. herbaceum and G. arboreum are very closely related to the taxon that contributed the A genome to New World cotton and that G. raimondii is rather closely related to the contributor of the D genome. The gap in this picture is the absence of the cross AD x D3 from table 1. Neither G. klotzschianum Andersson (D3-K) nor G. davidsonii Kellogg (D3-D), which some students of the genus consider to be conspecific, can be successfully hybridized with any of the New World species, and they will form vigorous hybrids with only two other species in the genus (G. anomnalum Wawra and Peyr. and G. longicalyx Hutchinson and Lee). There is thus no direct genetical or cytological method for assessing the relative chromosome homologies of D3 and DnW (D genome of New World cotton). However, by means of a bridging-species hybrid, the D3 and DnW genomes have been brought together in the same hybrid and their homologies compared by assessing chiasma frequency. By comparing chiasma frequency in this hybrid with that of a similar hybrid combining D5 and D,wy 'Contribution from the Crop Science Department, North Carolina Agricultural Experiment Station, Raleigh, N. C. Published with approval of the Director of Research as Paper No. 1509 of the Journal Series and with approval of the Crops Research Division, Agricultural Research Service, U.S.D.A.; supported in part by Project S-1 of the Research and Marketing Act of 1946. Aided by grant G-14203 of the National Science Foundation. 2 The genome nomenclature used in this paper follows Beasley (1940a and 1942) and Brown and Menzel (1950 and 1952).