One aspect of prehistoric crop agriculture which has received little attention is the purity of the crops when ready for either storage or consumption. For two reasons, this topic deserves consideration. First, the number of weed seeds and other impurities in a fully-processed crop is an indication of the efficiency of the prevailing technology for grain cleaning. There are several reasons why it would have been desirable to remove as many weeds as possible from a crop. Many for example are poisonous and should be removed from a crop before it is eaten. The sowing of a pure seed is also one way of controlling the growth of weeds, and thus maintaining economic crop yields. One pound of herbage seed for example, containing 1 per cent by weight of the seeds of mouse-eared chick-weed (Cerastium arvense L.) could result in as many as 43,360 plants of chick-weed per acre and cause severe reduction in the yield of the main crop (Gill and Vear 1958). Barley yields in one experiment were reduced from 21½ to 7½ bushels per acre through the growth of bind-weed (Leonard and Martin 1967). Since the artefactual evidence for prehistoric grain cleaning is meagre, the composition of the crops themselves must be our chief source of evidence. A second reason why the composition of prehistoric crops when ready for storage or consumption deserves study is that it affords a way of knowing whether crops consisted of one or several cultigens at a given period.