But here is something more remarkable; and it is this with which I wanted to conclude the present experiment. After repeatedly withdrawing the White-edged Sphex’ prize from the mouth of the pit and compelling her to come and fetch it again, I take advantage of her descent to the bottom of the shaft to seize the prey and put it in a place of safety where she cannot find it. The Sphex comes up, looks about for a long time and, when she is convinced that the prey is really lost, goes down into her home again. A few moments after, she reappears. Is it with the intention of resuming the chase? Not the least in the world: the Sphex begins to stop up the burrow. And what we see is not a temporary closing, effected with a small flat stone, a slab covering the mouth of the well; it is a final closing, carefully done with dust and gravel swept into the passage until it is filled up. The White-edged Sphex makes only one cell at the bottom of her shaft and puts one head of game into this cell. That single Locust has been caught and dragged to the edge of the hole. If she was not stored away, it was not the huntress’s fault, but mine. The Wasp performed her task according to the inflexible rule; and, also according to the inflexible rule, she completes her work by stopping up the dwelling, empty though it be. We have here an exact repetition of the useless exertions made by the Languedocian Sphex whose home has just been plundered. EXPERIMENT IV It is almost impossible to make certain whether the Yellow-winged Sphex, who constructs several cells at the end of the same passage and stacks several Crickets in each, is equally illogical when accidentally disturbed in her proceedings. A cell can be closed though empty or imperfectly victualled, and the Wasp will none the less continue to come to the same burrow in order to work at the others. Nevertheless, I have reason to believe that this Sphex is subject to the same aberrations as her two kinswomen. My conviction is based on the following facts: the number of Crickets found in the cells, when all the work is done, is usually four to each cell, although it is not uncommon to find only three, or even two. Four appears to me to be the normal number, first, because it is the most frequent and, secondly, because, when rearing young larvæ dug up while they were still engaged on their first joint, I found that all of them, those actually provided with only two or three pieces of game as well as those which had four, easily managed the various Crickets wherewith I served them one by one, up to and including the fourth, but that after this they refused all nourishment, or barely touched the fifth ration. If four Crickets are necessary to the larva to acquire the full development called for by its organization, why are sometimes only three, sometimes only two provided for it? Why this enormous difference in the quantity of the victuals, some larvæ having twice as much as the others? It cannot be