514 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1962 



South America. Not only is the closest known New World wild 

 relative confined to the Peruvian coast, but the whole group of New 

 World wild species of Gossypium is distributed west of the continen- 

 tal divide. Nevertheless, certain other fragments of evidence seem 

 to point toward an eastern South American origin for the cottons. 

 Engel [8] has reported that the crops of the ancient preceramic 

 cotton-using cultures of the Peruvian coast include groundnuts 

 (Arachis), which originated east of the Andes, probably [20] in 

 northwestern Argentina, suggesting that the cultivators readied 

 the coast from the momitains, and not from the sea. If they came 

 over the mountains, they presumably brought their cotton with them. 



Archeological evidence from preceramic cultures in Peru is critical. 

 If it could be shown that the people to whom these cultures belonged 

 came there by sea, the recent-origin theory of the development of 

 the New World cottons would be greatly strengthened, and a search 

 of natural vegetation in remote valleys of the Peruvian coast should 

 be undertaken in the hope of discovering a wild diploid species related 

 to G. raimondii with characters more nearly those appropriate to the 

 ancestor of the New World cottons. If, on the other hand, it turns 

 out that the earliest Peruvian cotton farmers reached the coast from 

 the mountains, the wild American diploid ancestor of the New World 

 cottons might be sought east of the Andes, and not on the Pacific 

 slopes. One would naturally look in arid areas for such a plant, and 

 in the regions in which the other crops of the early cotton users 

 originated. 



To accept an eastern South American origin for the New World 

 cottons would involve accepting also an ancient and not a recent 

 origin for the allopolyploids. The cultivation of cotton on the Atlantic 

 coasts of Africa is very recent, and it would be necessary to suppose 

 that the parents of the allopolyploid were both wild and came in con- 

 tact by natural spread. The present distribution of the wild species 

 of Gossypium has been interpreted [15] in terms of the distribution 

 of the major land masses before the continents drifted apart. If the 

 main species distinctions in Gossypium were already established at 

 that time, and if a suitable wild diploid existed east of the Andes, a 

 type akin to the South African G. lierhaceum race afrkanuin might 

 have come naturally in contact with its counterpart, and the allopoly- 

 ploid might have arisen in that way. Then Stebbins's contention 

 would be upheld, that the New World cottons are among the oldest, 

 and not the youngest, of the known allopolyploid species. 



The trail of speculation about an ancient Atlantic origin for the 

 New World cottons is as long and as scantily supplied with beacons as 

 that about a recent Pacific origin. Nevertheless, speculation of this 

 kind is worthwdiile. For example, the original speculation influenced 

 Bird in the choice of the Huaca Prieta as a site for archeological study. 



