It was wearisome work, for the wasp stayed away a long time, and we dared not let our gaze wander lest she should slip in without our knowledge. At the end of thirty–five minutes she returned, but again we failed to see what she carried. She flew with great rapidity, and we scarcely caught sight of her before she vanished into her nest. We could not but wonder at the ease and certainty with which she recognized her own doorway among the hundreds of holes on the side of the stump. This power of localization, while it is one of the most common among wasps, is surely also one of the most remarkable. Our little Rhopalum pedicellatum, for that proved to be her name, made six more journeys within the next two hours. At the end of this time we opened the tunnel, and, after a great deal of sawing and cutting, succeeded in finding the nest five inches from the surface. It was nothing but a slight enlargement of the gallery, in the soft decaying wood. In it we found thirty–three gray gnats, all of them, except two, being dead. On one of the dead ones was the egg, which had probably been laid within a few hours. The egg hatched two days later, on July twelfth, but on the fifteenth the larva died. By this time many of the gnats looked very dry, although we had tried to arrange for both moisture and ventilation by packing the bottom of the tube with pith and covering the top with muslin. Further watching gave us one more wasp of this species, in the same stump. This time the nest was only two inches from the surface. It contained four dead gnats and two live ones, but no egg, showing that the egg is not always laid on the first ones stored. Much later in the season, toward the end of August, we found another species of Rhopalum which proved to be new, and for which Mr. Ashmead has proposed the name rubrocinctum, since it wears a red girdle around the front end of the abdomen, being otherwise dressed in black like pedicillatum. It makes its home in the stalks of raspberry bushes. We opened a stem which contained thirteen compartments, separated by partitions of pith. These were filled with black, gray, and green gnats, which were packed in so closely that they were doubled over and pressed out of shape. Each cell contained from twenty–five to thirty gnats. In some of them were cocoons, in others larvæ, and in one was an egg. The gnats were very carefully examined, and all of them, from the cells that had been filled last as well as from those provisioned earlier, were dead. Other species of Rhopalum are said to prey upon spiders and aphides. In studying the species that come in our way we are continually developing distinct likings for some kinds above others. The appearance of one of these favorites is always hailed with delight, and when the season’s work is over we remember them with lively pleasure. [Illustration: OXYBELUS QUADRINOTATUS]