Back

Contrasted dissemination patterns of Old-World cottons Gossypium herbaceum L. and G. arboreum L. highlight how wet climate phases determined Levant Cotton domestication and spread

VIOT, C. R.

2026-01-15 evolutionary biology
10.64898/2026.01.14.699536 bioRxiv
Show abstract

Archaeological data show that the cultivation of Gossypium arboreum L. (Tree Cotton) began to spread inside south Asia in the mid-fifth millennium Before Present (BP), around three millennia after its first known textile use. Its sister species G. herbaceum L. (Levant Cotton) began its geographic spread at the end of the third millennium BP over northeast Africa, but its domestication history stays obscure. The initial geographic expansions of these two Old World cottons appear very contrasted. G. arboreum spread over typical tropical regions of southern, western and eastern Asia, with average speeds from 0.8 to 1.5 km/yr, close to the estimates for wheat during the Neolithic farming expansion northwards. G. herbaceum spread over sub-arid and arid regions of Africa and western and central Asia, at speeds from 1.6 to 6.7 km/yr. The dissemination of G. herbaceum coincided particularly with phases of increased rainfall over northeast Africa and south Levant during the Roman Warm Period at around 2200-1700 BP and over the Sahelo-Saharian region during the Medieval Climate Anomaly at around 1000-650 BP. Much earlier, the African Humid Period (AHP) from around the mid-15th to the 5th millennia BP could permit to explain the puzzling cotton remains dated to ca. 7000, 5500 and 4500 BP in three sites of southern Levant and the Nile Valley, as these remains could be posited to correspond to the earliest textile uses of G. herbaceum cotton growing naturally thanks to increased rainfall there during the AHP.

Matching journals

The top 5 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.