512 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1962 



In view of the difficulty of making crosses from which synthetic 

 tetraploids might be produced, interest turned in the direction of 

 genome analyses of hexaploids, made by doubling with colchicine Fi 

 hybrids between New World cottons and wild New World diploids. 

 Stephens [15] first reported on the cytology of a G. harhadenseXG. 

 raimondii hexaploid, and showed that pairing at meiosis was such 

 as would be expected if the G. raimondii set were rather closely homol- 

 ogous with one set of the G. harbadense complement. Gerstel [10] 

 conducted a genetic analysis in back-crosses of hexaploids to New 

 World cottons, and showed that segregation in G. raimondii crosses 

 gave ratios much nearer to those expected in autopolyploids than did 

 segregation in G. thurheri crosses. Such evidence as is available on 

 other wild New World species indicates that they are no more closely 

 related to the allopolyploid New World cottons than is G. thur- 

 heri [15]. 



Since the New World cottons originated from a hybrid between two 

 diploids, one of which was closely related to the Old World cottons, 

 it is not necessary to postulate a separate origin for their lint. On 

 the other hand, since the other diploid parent was related to the 

 Peruvian G. raimondii^ it is necessary to consider how an Old World 

 diploid and a New World diploid could have come together in western 

 South iVmerica so that the cross could take place. The first attempt 

 to account for the meeting of the two parents was by Harland [12], 

 who suggested that they came together on a land bridge across the 

 Pacific Ocean. Little evidence has been adduced for the existence 

 of such a land bridge, and it was later suggested [15, 19] that the link 

 was provided by civilized man, migrating eastward from the Old 

 World and taking his Old World cottons with him. Landing on the 

 Peruvian coast, he would find suitable conditions for his simple agri- 

 culture in the valleys of the rivers draining the Andes and crossing 

 the coastal desert to the sea. In those valleys, G. raimondii grows 

 naturally, and from natural crosses between the introduced Old World 

 cotton and an ancestor of G. rahnondii arose allopolyploids from 

 which the New World cottons sprang. The hypothesis fitted iha 

 known facts, and its usefulness as a working tool was demonstrated 

 when Bird [3] used it to help in the search which led to the excavation 

 of the Huaca Prieta and the discovery of the oldest known textiles 

 in the New World. 



An examination of seeds, carpels, and lint from the early Huaca 

 Prieta deposits revealed nothing to suggest the presence of an Old 

 World diploid cotton. Moreover, F. Engel [7] has since shown that 

 habitation sites of the Huaca Prieta type of culture occur over a great 

 stretch of the Peruvian coastline, far beyond the area in which G. 

 raimondii now exists. 



