I {278}believe there is one way, and one way only, in which primitive man was at all likely to become familiar with these facts. I shall try to show that all the operations of primitive agriculture very forcibly point to this strange and almost magical origin of cultivation; that all savage agriculture retains to the last many traces of its origin; and that the sowing of the seed itself is hardly considered so important and essential a part of the complex process as certain purely superstitious and bloodthirsty practices that long accompany it. In one word, not to keep the reader in doubt any longer, I am inclined to believe that cultivation and the sowing of seeds for crops had their beginning as an adjunct of the primitive burial system. Up to the present time, so far as I know, only one origin for cultivation has ever been even conjecturally suggested; and that is a hard one. It has been said that the first hint of cultivation may have come from the observation that seeds accidentally cast out on the kitchen-middens, or on the cleared space about huts, caves, or other human dwelling-places, germinated and produced more seeds in succeeding seasons. Very probably many savages have observed the fact that food-plants frequently grow on such heaps of refuse. But that observation alone does not bring us much nearer to the origin of cultivation. For why should early man connect such a fact with the seeds more than with the bones, the shells, or the mere accident of proximity? We must rid our minds of all the preconceptions of inductive and experimental science, and throw ourselves mentally back into the position of the savage to whom nature is one vast field of unrelated events, without fixed sequence or physical causation. Moreover, a kitchen-midden is not a cleared space: on the contrary, it is a weed-bed of extraordinary luxuriance. It brings us no nearer the origin of clearing.