Large granaries and pre-domestication cultivation

Pre-domestication granaries in Jordan: Kuijt & Finlayson (2009) in PNAS report preserved archaeological structures for granaries—round buildings which would have had raised wooden floors and lifespans of ca. 50 years. (This has also been reported by Michael Balter on the ScienceNOW blog.) They suggest these structures might have contained morphologically wild barley. It is a pity the archaeobotanical evidence from this site, which they elude to, is not actually presented in any detail (nor is whoever has done the work mentioned in the acknowledgements).

The quality of the excavation and the reconstruction of the architectural remains is exemplary, and leaves little doubt that raised floor granaries were constructed at Dhra (and presumable other Southern Levant sites) in the PPNA, by ca. 9200 BC. This is important evidence, since storage has more often been inferred than structurally demonstrated. They associate these structures with evidence from other sites for morphologically wild cereals (Gilgal: oats and barley; Netiv Hagdud: barley) and suggest that large-scale storage was a key part of the transition to reliance on pre-domestication cultivation. This evidence for large scale storage starting only in the PPNA seems to challenge our expectations that storage amongst hunter-gatherers ought to be a prerequisite for cultivation. Instead it appears that large-scale storage occurs alongside, or even results from (occurring after), the move to early cultivation. Undoubtedly there would have been smaller-scale precursors, such as the more limited evidence of possible Natufian silos and caching that they review.

Also of interest is the observation that storage structures move from public spaces to the inside of houses at ca. 8500 BC in the PPNB and then to special rooms by 7500 BC. This shift seems to parallel the move from early pre-domestication cultivation to more intensive cultivation as domestication traits, like non-shattering that began to increase in cereal populations from c.8500 BC and became predominant after 7500 BC (as documented for the northern Levant by Tanno & Willcox, Science 2006; and for the broader region, including Jordan, in Fuller 2007 Ann. Bot.—most of the evidence from Jordan is from S. Colledge (2001) Plant exploitation on epipalaeolithic and early neolithic sites in the Levant, BAR). This also parallels a trend in the find spots of food processing tools (grinding stones), documented by Karen Wright (UCL), which move from public spaces in the Natufian to inside houses in the PPNA-Ea. PPNB and into well-hidden back rooms by the end of the PPNB (Wright, K, 2000. The social origins of cooking and dining in early villages of western Asia, Proceedings of the Prehistoric Society 66, 89-12)

There is something of a southern Levant bias reflected in this paper, as archaeobotanical evidence for cultivation, and even archaeological evidence for possible storage structures from the Northern Levant (Syria) seems to be ignored. As reflected especially by the work of Willcox (e.g. Science 2006; the Balter News Focus piece of 2007), as well as work by Hillman and Colledge, pre-domestication cultivation is demonstrated earlier by ca. 10,000 BC at Mureybit and Qaramel. Perhaps upto 500-1000 years earlier at Abu Hureryra. Many northern sites (Mureybit, Tell Abr, Jerf el Ahmar) have large sunken floor ‘public buildings’. While these are usually regarded as structures of ritual use, they may also have been used for storing grain as suggested recently by Willcox et al. (2008) based on grain finds at Tell Abr., Vegetation History & Archaeobotany 17, p. 315). Oddly, this paper’s weakness is that it has a rather limited view of the archaeobotanical evidence, ignoring the broader macro-regional pattern for the shift to cultivation that accompanies increasing sedentism and architectural investment (both in the northern Levant/upper Euphrates area—probably first, and then the Jordan valley, a little later). It ignores archaeobotanical data from elsewhere in Jordan (e.g. publications by S. Colledge, J. Meadows). The examples it draws on are some of the least well-documented cases for pre-domestication cultivation, as they lack arable weed flora or sequences of change towards morphological domestication (by contrast to northern Levant cases), and despite some headline pieces in Science have never been published in any detail. (I nevertheless agree that they almost certainly are pre-domestication cultivation). In the end this is a nice archaeological find, with some great illustrations, but if feels somehow that they rushed to press without too much account for the archaeobotany.