One more remark. Fresh food is absolutely necessary to larvæ of the Hymenopteron. If the prey were stored intact, in four or five days it would be a dead body, given up to decay, and the newly hatched grubs would find no food but a corrupted mass. Touched by the sting it can live two or three weeks—a period more than sufficient for the egg to hatch and the grub to develop. The paralysis has thus a double result—immobility, so as not to endanger the life of the delicate larvæ, and long preservation of the flesh to assure wholesome nourishment for them. Even when enlightened by science human logic could find nothing better. My two other ephippigers, stung by the Sphex, were kept in darkness with food. To feed inert creatures, differing only from dead bodies by the perpetual oscillation of their long antennæ, seems at first an impossibility; however, the play of the mouth organs gave me some hope, and I made the attempt. My success surpassed my expectations. There was no question, of course, of offering them a lettuce leaf or any other green thing on which they might have browsed in their normal condition; they were feeble invalids, to be nourished with a feeding-cup, so to say, and broth. I used sugar and water. The insect being laid on its back, I put a drop of sugared liquid on its mouth with a straw. Instantly the palpi stirred, mandibles and jaws moved; the drop was consumed with evident satisfaction, especially if the fast had been somewhat prolonged. I renewed the dose till it was refused. The repast took place once or twice a day at irregular intervals, as I could not devote myself very much to a hospital of this kind. Well, with this meagre diet one of the ephippigers lived twenty-one days. This was little longer than the life of the one which I allowed to die of inanition. It is true that twice the insect had had a bad fall, having dropped from the experiment table to the floor through some awkwardness of mine. The bruises consequent may have hastened its end. As for the other, exempt from accidents, it lived six weeks. As the nourishment offered, sugar and water, could not indefinitely replace the natural food, it is very probable that it would have lived longer still had its customary diet been available. Thus the point which I had in view is demonstrated: victims pierced by the sting of the Hymenopteron die from inanition and not of their wound. XII THE IGNORANCE OF INSTINCT