so the 93 species may be put into 28 groups and these in turn into 6 subgenera which constitute two main groups, each of which is a closely compacted unit in its geographic distribution.
Thus, while the genus cannot be based on any single diagnostic character, the remarkable coördination of so many characters in so many species and the absence of anything approaching this combination of characters among any of the other members of the family testify to the phylogenetic unity of Cynips. There seems no reason for believing that such a body of coördinated characters could have arisen independently in more than one time and place. In a single period, in a limited area, there must have existed a population from which all of the present-day species, with their varying grades of relationships, have developed. This history of the expansion of a single genetic stock into 93 distinct populations, by processes of mutation, isolation, and on occasion subsequent recombination into hybrid populations, is the story we have been unfolding in this study, and which we are now ready to fit into the geologic time and the geographic areas in which speciation probably proceeded in the genus.
Cynips is in every respect a highly specialized genus of the oak-inhabiting tribe Cynipini of the family Cynipidae. The fossil record of the Cynipidae is meager and without significance except to prove that the family was in existence in the Oligocene and Miocene (Kinsey 1919; Cockerell 1921). There seems no reason for believing that the Cynipidae have ever been associated with any plants except the Angiospermae on which the family occurs today. The diversity of the present-day genera of the primitive gall makers of the Aulacini indicates that that tribe must have had a long history antecedent to the origin of the Cynipini. The family could not have originated before the rise of the flowering plants in the late Cretaceous, and it was probably much later before the Cynipini developed such specialized genera as Cynips and Disholcaspis.
Our attempts to fathom the history of any of these higher genera must proceed on the assumption that all species of these groups have from the first been associated with oak, and our analysis of the cynipid history must do no violence to the known history of the sources and development of Quercus. To this end, the accompanying summary of the paleontological
