But, while the interest is great, the difficulty of observation is not small. The solitary habits of the Ammophilæ, their distribution one by one over wide areas, the fact that one almost always comes across them merely by chance: all this makes it hardly possible to carry out premeditated experiments with them, anymore than with the Languedocian Sphex. You have to be on the look-out a long time for an opportunity, to wait for it with untiring patience, and to know how to profit by it at the very moment when at last it presents itself, a moment when you were not thinking of it. I watched for that opportunity for years and years; then one day it suddenly appeared before my eyes, offering a facility of examination and a clearness of detail that compensated me for my long waiting. At the beginning of my investigations I was twice enabled to witness the murder of the caterpillar, and I saw, as far as the swiftness of the operation permitted, the Wasp’s sting applied once and for all to either the fifth or the sixth segment of the victim. To confirm this result, I thought of ascertaining which ring had been stabbed on caterpillars which I had not seen sacrificed, but which I had taken from their captors while they were being dragged to the burrow. It was no use employing a magnifying-glass, for no magnifying-glass enables one to discover the least trace of a wound upon the victim. The method adopted is the following: when the caterpillar is quite still, I try each segment with the point of a fine needle and thus measure the amount of sensibility by the more or less manifest signs of pain in the insect. When the needle pricks the fifth segment or the sixth, even piercing it right through, the caterpillar does not stir. But if you prick even slightly a second segment, behind or in front of that insensible segment, the caterpillar wriggles and struggles with a violence which increases in proportion to the distance of the point attacked from the original segment. At the hinder end in particular, the least touch provokes wild contortions. There was only one sting, therefore, and it was administered to the fifth or sixth ring. What peculiarity then do these two segments possess that one or other of them should be the target of the assassin’s weapon? None whatever in their organization; but their position is another matter. Leaving the Silky Ammophila’s Measuring-worms on one side, I find that the prey of the others is organized as follows, the head being counted as the first segment: three pairs of real legs on the second, third and fourth rings; four pairs of membranous legs on the seventh, eighth, ninth and tenth rings; lastly, a final pair of membranous legs on the thirteenth and last ring, making in all eight pairs of legs, of which the first seven form two vigorous groups, one of three, the other of four pairs. These two groups are separated by two legless segments, which are precisely the fifth and sixth.