dreds of adults from a single original egg by a process not unlike that employed in the egg-shaking experiments of our laboratories, this phenomenon, though restricted to comparatively few species, is nevertheless of considerable economic importance.
7. The development of social life among insects. This, as I have shown on former occasions, has its origin both ontogenetically and phylogenetically in the parasitism of the offspring on the parent.
Paleontology seems to show very clearly the conditions that have favored the enormous development of parasitism among insects especially within comparatively recent times. Some of these conditions are:
1. The diminution in insect stature which occurred in the late Carboniferous and during the Permian and seems to have been originally in great part an adaptation to increased reproduction and dispersal. Other things equal, a small animal will, for very obvious reasons, become a parasite more easily than a large one.
2. The development of metamorphosis. This was already clearly established in the earliest known insects, the Paleodictyoptera, which were predatory and amphibiotic like the may-flies of the present-day, living in the water during their apterous larval stages and spending their winged imaginal stage in the air. They show plainly the great peculiarity of insect development, i. e., metamorphosis succeeding growth and not preceding it as in the crustaceans, mollusks and anne-
- The occurrence of polyembryony was first clearly recognized and thoroughly investigated by Marchal in Eucyrtus fuscicollis ("Recherches sur la Biologie et le développement des Hyménoptères Parasites. I. La Polyembryonie Specifique on Germinogonie," Arch. Zool. Expér. Gén. (4), II., 1904, pp. 257-335, 5 pls., and an earlier paper: "La dissociation de l'oeuf en un grand nombre d'individus distincts chez l'Encyrtus fuscicollis," C. R. Acad. Sci. Paris, CXXVI., 1898, pp.-662-664), although Bugnion ("Recherches sur le développement postembryonnaire, l'anatomie et les mœurs de l'Encyrtus fuscicollis," Rec. Zool. Suisse, V., 1891, pp. 435-534, 6 pis.) had previously studied the same insect. Silvestri has published several valuable papers on polyembryony, the most important being "Contribuzioni alia Conscenza Biologica degli Imenotteri Parassiti. I. Biologia del Litomastix truncatellus (Dalm.)," Ann. B. Scuola Sup. d'Agric. Portici, VI., 1906, pp. 1-51, 5 pls.
- Wheeler, "Ants, their Structure, Development and Behavior," Columbia Univ. Press, 1910. Recently Holmgren ("Termitenstudien, I. Anatomische Untersuchungen," R. Svensk. Vetensk. Handl., XLIV., No. 3, 1909, 216 pp., 3 pls., 76 text-figs.) and Eseherich ("Termitenleben auf Ceylon," Gustav Fischer, Jena, 1911, 262 pp., 3 pis., 68 text-figs.) have accumulated much evidence to support the conclusion that the mutual attraction among the individuals and the development of the castes of the termite colony are due to the habit of these insects of feeding on the fatty exudates of one another's bodies and on that of their queens. This may also be true of ants and other colonial insects. A very similar method of feeding on the surface secretions of their host-ants is adopted by certain myrmecophiles (Oxysoma, Attaphila and Myrmecophila) and certain parasitic ants (Leptothorax emersoni).