Some additional localities for a wild tetraploid cotton, G. musteliinum Watt, in northeast Brazil are reported, together with field observations on this cotton in its natural habitats. The history of cotton cultivation in northeast Brazil is reviewed and G. mustelinum compared with cottons cultivated in the region from early Colonial times onward. G. mustelinum appears to be specifically distinct from the other tetraploid cottons, and it is not likely that it is a feral derivative of any commercially cultivated cotton. This little-known tetraploid may help in reconstructing the origin of the New World polyploid cottons. THE NEW WORLD CULTIVATED COTTONS, Gossipium hirsutum and G. barbadense, are tetraploids (2n=4x_52) which contain the A genome found in the Old World diploids G. herbaceum and G. aiboreum and the D genome found in American wild diploids. In both the cultivated tetraploids and the A genome diploids the hairs on the seed coat are differentiated into long lint hairs, which become twisted on drying and can thus be spun into thread, and an undercoat of short hairs (fuzz), which have more extensive cellulose thickening and which do not twist on drying. The D genome diploids have seed hairs of various sorts but never bear lint and have never been used by man for fiber. At one time it was thought that man had introduced a domesticated, linted A genome cotton from the Old World to the New, where it hybridized with a native D diploid to produce the tetraploids. Gerstel (1953) showed that G. herbaceuin chromosomes were more similar to those of the New World A genome than were those of G. arboreum, although even G. herbaceum chromosomes differed from the A genome of the tetraploids by two translocations. At present it seems that the diploid most similar to the A genome progenitor of the tetraploids is G. herbaceum var. africanum, a linted, hard-seeded cotton that occurs wild in southern Africa. The D genome diploids are today confined to the western sides of the American continents. The Mexican species, together with G. klotzschianum var. davidsonii from the Galapagos Islands, form a group fairly closely related among themselves but more distantly related to the Peruvian species G. raimondii (Phillips 1966). G. rairnondii is the species most similar to the D genome ancestor of the tetraploids on the criteria of morphology of the seed hairs (Hutchinson, Stephens, and Dodds 1945) and chromosome pairing and chiasma formation in hybrids with the tetraploids (Phillips 1963, 1966). Furthermore, electrophoretic studies on seed proteins have shown that a synthetic mixture of proteins from G. raimondii and G. herbacweum var. africanum gave very similar banding patterns to proteins from G. hirsutum and G. barbadense (Cherry, Katterman, and Endrizzi 1970). However, although G. raimondii is the most similar extant diploid to the D genome ancestor of the tetraploids, there are more dissimilarities between the chromosomes of G. raimondii and the tetraploid D genome than there are between G. herbaceum and the tetraploid A genome (Phillips 1963). Today the ranges of the A and D genome diploids are separated by some thousands of miles (see map, fig. 1), and the origin of the tetraploids poses a considerable phytogeographic problem. Cytogenetic data (Phillips 1963) suggested that the tetraploids are derived from a common ancestor (i.e. originated monophyletically) and that speciation occurred after the tetraploids were established. The different genome groups in Gossypiurm may have diverged when the range of an early prototype cotton became fragmented when Gondwanaland broke up into the land masses which correspond to the southern continents of today. This separation probably began in the early Cretaceous (Raven 1972). The first records of tetraploid cottons are archaeological specimens of domesticated (not wild) cot1 Present address: Department of Botany, University of California, Berkeley, California 94720 U.S.A. 42 BIOTROPICA 7(1): 42-54 1975 This content downloaded from 157.55.39.110 on Tue, 20 Sep 2016 05:48:05 UTC All use subject to http://about.jstor.org/terms
Read full abstract