I have received a recent query about early cotton in Nubia and South Asia, vis-a-vis the early occurrence of apparent cotton fibres in plaster from Dhuweila in Jordan (published by Betts et al 1994, J. Archaeological Science). This came in response to discussion of the find in the recent review I co-wrote with Nicole Boivin on the Holocene prehistory of Arabia (including several discussions of crops). I reproduce here my emailed reply as to why I have not committed on the possibility suggested by Betts (et al.) of a Nubia OR South Asian source. I have added links to references cited.
The Nubian source is based on reported cotton from Chalcolithic (A-Group) Nubia, first reported by Chowdhury & Buth in 1970 in Nature. The Indian Expedition to Nubia has preliminary reports on the archaeology in Indian Archaeology: A Review 1961-62 (which can be downloaded), and more extensively in Fouilles en Nubie (1961-1963). The A-Group Nubian cotton is problematic, with criticism published previously be me and by Zohary & Hopf. First it comes from goat dung, which does not necessarily indicate textile processing (admitted in the original Chowdhury publication). Second it comes from an undated excavation where it is alleged to be associated with A-Group ceramics (i.e. 3300-3100 BC-- mid-Third M. BC as cited by Betts  is simply a mis-dating by the Indian archaeologists of 1962 who never got to grips with the Nubian sequence-- the dating of which was sorted out later, e.g. Trigger 1965; but especially Nordstrom 1972, Williams 1986; the region is largely depopulated at 2600 BC except for the Old Kingdom outpost at Buhen), but I am very worried that goat dung containing cotton seeds in Lower Nubia is more likely intrusive from later (Late Meroitic/Post-Meroitic) when cotton was known to be widely cultivated in the region. Stratigraphy in Nubia is always problematic. (I speak from experience in the 4th cataract), due to heavy deflation and reburial with windblown sand. For a published case-in-point consider Wadi Kubbaniya: Late Palaeolithic with charred wild plant remains AMS-dated to 16,000 bp and un-carbonized barley AMS'd at 500 bp [although not from the same excavation season] ! Things get mixed and the depth of deposit at Afyeh where it was found (the Indian Expedition to Nubia) was not very far beneath the surface. If it were from Nubia, how come there are no further cotton finds for 3000 years in the region? (Nevertheless, it would be nice to have more details about the Afyeh excavations, which remain largely unpublished and somewhere in the ASI archives in Delhi). I would note that Chowdhury also identified the Afyeh cotton as G. herbaceum based on a cross section of the seed coat. Perhaps this method works, but I have been unable to replicate it on the SEM (in modern material), which is a pity as it would be great to be able to track the two cotton species from archaeobotanical finds.
In addition, Naive African cotton, Gossypium herbaceum, is reported wild only much further south in Africa. It is of course possible it has gone extinct in Nubia, but if it were in 4th M. BC Nubia then it should on biogeographic grounds have been in Egypt too-- there is nothing in the Lower Nubian flora which is not in Egypt. (Is it likely the Egyptians would have overlooked using a species like this?). But, if one does postulate extinct wild Gossypium herbaceum as far north as Nubia then why not in mid-Holocene (wetter) Arabia as well?
Mehrgarh Period II has cotton seeds and fibres preserved in the earliest copper beads (Moulherat et al. 2004). On present evidence it seems the best, and only reasonable candidate. There can be presumed (but not proven) to come from Gossypium arboreum.
I must admit to remaining perplexed about the Dhuweila find. If textiles were traded long-distance one would not expect them to be valued, not used in tempering plaster. But perhaps this was not a common temper but a rare inclusion where someone tore a frayed tunic or something; the Dhuweila team got lucky-- one would need some systematic and quantitive approach to plaster temper to establish this.
My (obscure) article on Indian cotton and flax archaeology may be of some interest, as it reviews the archaeobotanical finds of these two species in India and Pakistan, with some notes on elsewhere [pdf]. For an updated review of cotton in Nubia see the chapter by Clapham and Rowley-Conwy in the recently published Hillman Feitschrit.