Years ago, when we found that many of the orb–weavers laid enormous numbers of eggs (A. cophinaria from 500 to 2000), we wondered what became of the thousands of spiderlings. An acquaintance with Trypoxylon has shown us their fate, and has given us an illustration of how closely the two groups are related. To make a very modest estimate there must have been twenty wasps at work in our straw–stack. During the six weeks which make the busiest part of their working season each of these must have stored, at the very least, thirty cells, putting an average of ten spiders into a cell. It may then be considered certain that the straw–stack, with its working surface of twelve by twenty feet, was the mausoleum of six thousand spiders, and it is very probable that twice as many were interred within its depths. It must be remembered, too, that before the spiders have grown large enough to be interesting to rubrocinctum, bidentatum has had her turn at them, and that those that are allowed to grow too large for rubrocinctum are preyed upon grade after grade, first by albopilosum and finally by Pelopæus, Pompilus, and other genera. The wasps of this genus lose their interest in family affairs about the second week in August, though after this time they may still be seen taking their well–earned holiday on the blossoms of the aster and the golden–rod. Chapter IX THE SPIDER–HUNTERS WHILE Ammophila provides caterpillars for her larva, and Bembex, after the manner of the social wasps, feeds her young from day to day on dead flies, the Pompilidæ, so far as their habits are known, all prey upon spiders. The family is a large one in the United States, one hundred and twenty–seven species having been described. The members of the group differ in size, color, and habits, and the individuals of the same species show the very considerable amount of variation which seems common to all those groups of animals which have been carefully studied. Happily the old notion that habits and instincts, unlike structural peculiarities, are always uniform, is no longer insisted upon, and there is ample evidence for the opinion that functional variations are as common as morphological. We have studied five species of this family, and have found their respective rôles of great interest. According to Fabre, the French members of this genus, although they do not make their own nests, still exercise some foresight in the matter by selecting a suitable crevice before catching their prey. Among the species that we have studied, quinquenotatus, biguttatus, fuscipennis, marginatus, and interruptus first catch the spider and then make the nest; while calipterus and scelestus prepare the nest before capturing their prey. [Illustration: TORNADO WASP (POMPILUS QUINQUENOTATUS) DIGGING NEST] Quinquenotatus is usually rather less than half an inch in length and is black, the abdomen having a variable number of white bands and a white tip.