tainly does not apply. That it is inapplicable to their phylogeny I am not prepared to say, although I am unable to think of any nonparasitic insects that show evidence of descent from parasitic species. There can be no doubt, however, that parasites are still able to give rise to new specific forms. This capacity is without doubt very feeble or languishing in the permanent parasites of the tape-worm and Sacculina types as compared with that of the insects. Indeed, there is much evidence to show that in insects, parasitism, far from interfering with the process of species formation, may actually have a tendency to favor or accelerate it. Sharp estimates the number of species of parasitic Hymenoptera on our globe at 200,000, and of this vast number probably 80,000 belong to a single family, the Chalcididæ, of which only some 6,000 species have been described. Another parasitic family, the Tachinidæ, belonging to the great order Diptera, seems to be in such an active stage of species formation that the most diligent and thoughtful students of the group flounder about in it with a dazed and almost ludicrous helplessness. And not only is practically the whole enormous group of moths and butterflies to be regarded as parasitic, but the same is true also of untold legions of plant-lice, scale-insects and beetles. Hyperparasitism, which may be regarded as a kind of permutation of parasitism, must also be mentioned in this connection, because it gives us a glimpse of the virgin fields which the holometabolic insects, owing to their peculiar method of development, are beginning to invade.
I believe that the foregoing discussion of the peculiarities of insect parasites adequately supports the view that these organisms are eminently fitted to function in controlling the depradations of injurious insects. That they can not be regarded as instruments of extermination is obvious from the fact that under natural conditions the complete extinction of the host species involves the destruction of the parasitic species, unless the later is able to live on more than one host. Although it is not improbable that during geological time such joint extermination of host and parasite has repeatedly occurred, we are unable to cite any case that has fallen under the observation of the entomologist. Purely local extermination of injurious hosts by their parasites has, however, been observed.
Before bringing my lecture to a conclusion two matters must be briefly discussed. One of these, which is mainly of theoretical interest, relates to the development of the parasite's association with its host, the other, of more practical significance, to the methods of greatest promise in the study of insect parasitism. We need not stop to consider cases of the tape-worm type which reach their hosts by chance. In the two other types which I have distinguished, we have the association with the host established through the initiative of the larval parasite itself (Sacculina type) or through the parasite's mother (insect type). While the former type does not seem to call for any special ex-
