August 5, 2015

Weed evolution in the context of earliest proto-farming in Palestine

Ohalo II, at (often under) the Sea of Galilee, is probably the oldest site of sedentary proto-farming, dating to before the Last Glacial Maximum, experiment that may have been interrupted afterwards. A new study finds the first signals of not just many edible plants that would eventually become crops but also the signature of evolution of weeds already towards the forms we are familiar with nowadays.

A. Snir et al., The Origin of Cultivation and Proto-Weeds, Long Before Neolithic Farming. PLoS ONE 2015. Open accessLINK [doi:10.1371/journal.pone.0131422]


Weeds are currently present in a wide range of ecosystems worldwide. Although the beginning of their evolution is largely unknown, researchers assumed that they developed in tandem with cultivation since the appearance of agricultural habitats some 12,000 years ago. These rapidly-evolving plants invaded the human disturbed areas and thrived in the new habitat. Here we present unprecedented new findings of the presence of “proto-weeds” and small-scale trial cultivation in Ohalo II, a 23,000-year-old hunter-gatherers' sedentary camp on the shore of the Sea of Galilee, Israel. We examined the plant remains retrieved from the site (ca. 150,000 specimens), placing particular emphasis on the search for evidence of plant cultivation by Ohalo II people and the presence of weed species. The archaeobotanically-rich plant assemblage demonstrates extensive human gathering of over 140 plant species and food preparation by grinding wild wheat and barley. Among these, we identified 13 well-known current weeds mixed with numerous seeds of wild emmer, barley, and oat. This collection provides the earliest evidence of a human-disturbed environment—at least 11 millennia before the onset of agriculture—that provided the conditions for the development of "proto-weeds", a prerequisite for weed evolution. Finally, we suggest that their presence indicates the earliest, small-scale attempt to cultivate wild cereals seen in the archaeological record.


  1. Totally off topic but, if you have not already done so, you might be interested in reading

    It seems to lend support to some of your views.

    1. Thank you very much, it looks very interesting and I will no doubt comment it tomorrow. On first quick and partial read, fig. 2 (dated tree) is the most informative single image and does indeed tend more to my perception than most "molecular clock" estimates. One of my key calibration points would be the OoA, which should correspond with the E/CF split, and in this case it is estimated to be c. 100 Ka BP, what is at least plausible (100-125 Ka BP is my archaeologically based reference, so still in the low range but at least within the range). It's getting better and indeed the use of ancient specimens provides a great aid, dismissing the ilusions caused by pedigree rates, which are not subject yet to the unavoidable evolutionary constraints (a harmful mutation will not survive in the long run). Also big names on the paper: Trombetta, Myres and Cruciani.

      Cool. :)

    2. Notice that with a ruler on hand, Lochsbour appears to be 7000 y.o. (when it is actually c. 8000) and Ust' Ishim appears to be 44.4 Ka old, when it is actually c. 45 Ka calBP (close enough but this one is a less precise date, because of its extreme antiquity). So the data allows for a 14% more in all dates (using Lochsbour as main reference), what would make the E/CF split more realistically around 115 Ka ago (with the same error margins, which are large for that node).

      That would fit even better. Looking nice with minor adjustments, yeah.

    3. With fig.2 as such R1b1 could be like 9.7 Ka old (early Epipaleolithic or Mesolithic) and with the extra 14% it becomes 11 Ka old. However the detail is unclear and I can't be sure if this R1b1 is the true R1b1 or just R1b-L11, because it shows a simple bifurcation that can only correspond (ref: to U106 and P312 (S116). I mention because it always ends up on the issue of R1b and the data here is not clear enough.

      What is more apparent in the data is that the R1b main split R1b1 vs R1b2 (V88) is of c. 18-21 Ka BP, i.e. at the LGM, that the R1 split (R1b vs R1a is of 25-29 Ka ago) and that the R basal split is of c. 32-36 Ka BP. The P basal split (R vs Q) is of c. 38-43 Ka BP, i.e. early UP.

      However I'm a bit concerned about the 50 (or 57) Ka date for K2, which IMO is a bit too recent. True that Ust' Ishim is some sort of proto-NO but if this would be correct, it'd suggest that K2 would have expanded from West to East (i.e. with the UP, whose first evidence is soon after 50 Ka in Europe, West Asia and Central Asia), what is inconsistent with all the modern data we have from Karafet that clearly defines SE Asia as the origin of K2.

      So I'd still think of an earlier K2 split date, of at least 60 Ka (maybe even post-Toba, i.e. c. 70 Ka BP) but not sure how that can fit with molecular clock methods, which are more mechanic but less holistic.


Please, be reasonably respectful when making comments. I do not tolerate in particular sexism, racism nor homophobia. Personal attacks, manipulation and trolling are also very much unwelcome here.The author reserves the right to delete any abusive comment.

Preliminary comment moderation is... ON (your comment may take some time, maybe days or weeks to appear).