The first wasp that we saw was just alighting with a medium–sized green caterpillar near a partly closed nest. When disturbed she flew away, but soon returned, dropped her prey half an inch from the nest, proceeded to clear the opening, ran inside to see that all was right, and then backed in with the caterpillar. Emerging after a few minutes, she placed a small pebble in the doorway, which was thus partly closed, and flew away. She brought three more caterpillars at intervals of thirty minutes, and then, after wedging a pebble into the neck of the opening, she began to fill it in solidly, scratching in dirt and packing in lumps of earth which were brought in her mandibles. We did not allow her to complete this operation, as it would have made excavation more difficult, but caught her and dug out the nest. The tunnel ran down obliquely for five inches, being two inches below the surface at the pocket. In it we found a wasp larva, which was at least three days old, and four caterpillars. There were no signs of the banqueting which must have already taken place. We carried this larva home with us, and it ate the caterpillars up clean, finishing with a fifth which we supplied from another nest, and going into its cocoon on September sixteenth. The caterpillars all wriggled about on the slightest stimulation, and remained in this lively state until they were eaten. They belonged to four different species. In a second nest to which food was being carried, we found four caterpillars and a larva about three days old, all the conditions being like those in the other example. Evidently the larva had been fed from day to day, since four or five days must have elapsed since the making of the nest. Westwood states that Ammophila, when she has captured her prey, walks backward, dragging it after her;[1] but in all the cases that came under our notice she went forward, the caterpillar being grasped near the anterior end, in her mandibles, and either lifted above the ground or allowed to drag a little if long and heavy. It is usually held venter up, but in one case, in which the wasp, while carrying it to her nest, frequently laid it down and picked it up again, it was held with the venter down or up indifferently. The all–important lesson that Fabre draws from his study of the Ammophiles is that they are inspired by automatically perfect instincts, which can never have varied to any appreciable extent from the beginning of time. He argues that deviation from the regular rule would mean extinction. For example, if the wasp should sting ever so little to one side of the median line the prey would be imperfectly paralyzed and the egg would consequently be destroyed; or a sting in the wrong place might cause the death of the caterpillar and thus the death of the wasp larva, which, he thinks, can be nourished only by perfectly fresh food.