Crop domestication didn't happen overnight or in one place

From the paper:

Recently, archaeobotanical evidence has overturned the rapid transition model that crop domestication was initiated and completed in a brief period at the Pleistocene/Holocene boundary (see Fig. 1). All three stages of the domestication process have been extended in timescale. Large plant assemblages have provided evidence for wild-cereal gathering as old as 23,000 years B.P. at Ohalo II (29), some 10,000 years earlier than previously thought. Evidence of predomestication cultivation has been established from 13,000–12,500 years B.P. (30, 31), during the Younger Dryas. Within the predomestication period, there appear to have been numerous beginnings of agriculture, with different species hailing from different localities rather than in a single Neolithic Package (32). The final stage of the domestication process in which the domestication syndrome traits are fixed has also been found to be a slow process; Tanno and Willcox (33, 34) argue that the tough-rachis mutant took over 3,000 years of cultivation to reach fixation and the syndrome traits themselves appeared in slow sequence, not together over a short period.

The apparent protracted process of domestication from plants of multiple origins is apparently at odds with the genetic data suggesting that domestication plants are monophyletic, the result of a single domestication event in a definite place. But:

The outcome of the final scenario of a single origin was most surprising because these simulations were slower than multipleorigin ones to reach monophyly. Consequently, we reach the superficially counterintuitive conclusion that, when viewed through multilocus systems, multiple-origin crops are actually more likely to result in monophyly than single-origin ones. In fact, it is quite reasonable to suppose that this conclusion might be the case. The underlying reason is that cultivated crops that originate from a single wild population are more similar to their wild progenitor population than an amalgamated cultivated population that will have additional genetic contributions from other wild sources: Crops of multiple origins are, from the outset, more differentiated from their wild sources than crops of a single origin.

From a press release:

Dr Robin Allaby says:

"This picture of protracted development of crops has major implications for the understanding of the biology of the domestication process and these strike chords with other areas of evolutionary biology."

"This lengthy development should favour the close linkage of domestication syndrome trait genes which may become much more important because linked genes will not be broken up by gene flow – and this makes trait selection and retention easier. Interestingly, as more crop genomes become mapped, the close linkage of two or more domestication syndrome genes has been reported on several occasions."

"This process has similarities to the evolution of 'supergenes' in which many genes cluster around a single locus to contribute to one overall purpose."

"We now need to move this research area to a new level. Domestication was a complex process and can now be viewed more legitimately as the paragon of evolutionary process that Darwin originally recognized. There are many interacting factors involved that we know about operating on a wide range of levels from the gene to the farmer and climate – the challenge is to integrate them into a single story."

PNAS - September 16, 2008 - vol. 105 - no. 37

The genetic expectations of a protracted model for the origins of domesticated crops

Robin G. Allaby et al.

Abstract

Until recently, domestication has been interpreted as a rapid process with little predomestication cultivation and a relatively rapid rise of the domestication syndrome. This interpretation has had a profound effect on the biological framework within which investigations into crop origins have been carried out. A major underlying assumption has been that artificial selection pressures were substantially stronger than natural selection pressures, resulting in genetic patterns of diversity that reflect genetic independence of geographic localities. Recent archaeobotanical evidence has overturned the notion of a rapid transition, resulting in a protracted model that undermines these assumptions. Conclusions of genome-wide multilocus studies remain problematic in their support of a rapid-transition model by indicating that domesticated crops appear to be associated by monophyly with only a single geographic locality. Simulations presented here resolve this conflict, indicating that the results observed in such studies are inevitable over time at a rate that is largely influenced by the long-term population size. Counterintuitively, multiple origin crops are shown to be more likely to produce monophyletic clades than crops of a single origin. Under the protracted transition, the importance of the rise of the domestication syndrome becomes paramount in producing the patterns of genetic diversity from which crop origins may be deduced. We identify four different interacting levels of organization that now need to be considered to track crop origins from modern genetic diversity, making crop origins a problem that could be addressed through system-based approaches.

Link