only have fallen into as great a danger, for the numerous birds that accompany the army of ants are ever on the outlook for any insect that may fly up, and the heavy flying locusts, grasshoppers, and cockroaches have no chance of escape.”[92] This locust resembled a green leaf which, as we have seen, was a very protective resemblance indeed. It might, however, had it been a smaller insect, have resembled one of the ants themselves, and in that case could have run about with them, pretending or appearing to forage, also with perfect impunity. Whether the _Eciton_ has such a double I know not, but various ants have. With some it is a spider that assumes their form. With others, as we have seen to a partial extent, a caterpillar, but the _Sauba_, or leaf-cutting ant—which is also the mushroom-growing one—is understudied, leaf and all, by an insect which, though the order to which it belongs has been determined, has not yet apparently received a name. “An example,” says Professor Poulton, “of protective mimicry, which I believe to be more wonderful in its detail and complexity than any which has been hitherto described, was observed and interpreted by my friend Mr. W. L. Sclater, in 1886, during his investigations in British Guiana. Mr. Sclater and his native servant had been collecting insects by shaking the branches of a tree over a sheet. The servant, although described as a very acute observer, saw an insect on the sheet which he mistook for one of the abundant Cooshie ants (perhaps the native name), carrying its little jagged segment of leaf over its back. Mr. Sclater looked more closely, and saw that it was an entirely different insect belonging to the order _Homoptera_. Its length was about that of an ant carrying its leaf. The leaf was represented by the thin flattened body of the insect which in its dorsal part is so compressed laterally that it is no thicker than a leaf” (or as we would say, which along the back is no thicker than a leaf), “and terminates in a sharp, jagged edge. The head and legs were brown, and suggested the appearance of that part of an ant which is uncovered by the piece of leaf. The jagged dorsal line, when seen in profile, evidently corresponds to the roughly gnawed edge of the fragment of leaf, for Mr. Sclater tells me that the contour of the latter is generally shaped by the mandibles of the ant rather than due to the natural margin.”[93]