inedible often take root, and, somewhat later, form an arc of tall grass more or less closely approximating to a complete circle round the nest. Since these ants feed largely, though by no means exclusively, on grass-seeds, and since these particular seeds are a very common and favourite article of food, it is easy to see how their grass should often predominate in the circle. In reality, however, only a small percentage of the nests, and only those situated in certain localities, present such circles. Now to state that the ant, like a provident farmer, sows this cereal, and guards and weeds it for the sake of garnering its grain, is as absurd as to say that the family cook is planting and maintaining an orchard when some of the peach-stones which she has carelessly thrown into the backyard, with the other kitchen refuse, chance to grow into peach trees.”[75] Certainly such a thing should have been observed before the statement was made, and, if it has not been, the facts seem more probably accounted for on the above explanation. Professor Wheeler goes on to say that “there are several other facts which show that the special ring of grass about the nest is an unintentional and inconstant result of the activities of the ant colony. For instance, one often finds very flourishing ant-colonies that have existed for years in the midst of much-travelled roads, or in stone side-walls, often a hundred or more feet from any vegetation whatever (without any ant-rice on their mounds therefore). Again, it is very evident that even a complete circle of grass like those described by Lincecum and McCook would be entirely inadequate to supply more than a very small fraction of the grain necessary for the support of a flourishing colony of these ants. Hence they are always obliged to make long trips into the surrounding vegetation, and thereby wear out regular paths, which radiate in different directions, often to a distance of forty to sixty feet from the entrance of the nest. The existence of these paths, which are often found in connection with grass-encircled nests, is alone sufficient to disprove Lincecum’s statements.”[75] It certainly seems easier to suppose that Lincecum misinterpreted certain facts, not themselves in dispute, than that an explanation on which so many considerations seem to throw doubt is the correct one. One thing, at least, seems certain—if some of these ant communities grow grain of set purpose, all of them do not. This may be possible, but more proof of it than Lincecum has brought is demanded. If the ants really sow and reap the grain that grows upon their mounds, and, more especially, if they carefully keep the patch clear, it ought not to be difficult to see them doing so. This last would be decisive, whereas the other two are by no means so.