Page:The Gall Wasp Genus Cynips.pdf/20

 and from England, France, Germany, Denmark, Italy, Hungary, Austria, Czecho-Slovakia, Finland, and Russia among European countries. This coöperation has been indispensable, and specific acknowledgment of my indebtedness to each collector is made with each item in the systematic portion of this paper. The accompanying map (Fig. 6) will show the extensive—albeit still inadequate—sources of our American material of Cynipidae.

The accumulation of these geographic data has received some impetus from repeated analyses we have made of our field technic. By using the automobile we reach twice as much territory in a given time as we were formerly able to reach by railroad transportation. We find that the densest populations of Cynipidae are, for some reason not yet apparent to us, located on isolated oak trees or at edges of woodlands, and it is there that we now concentrate most of our efforts while in the field. It was some years before we learned that the distribution of most species of gall wasps is very local, and that it is a waste of time to attempt to make collections of populations that are sparsely scattered over the countryside. If large populations are not readily available at the first stopping place, we take a sample and move a hundred yards to a half-mile across country, continuing the procedure until we have discovered some tree or thicket heavily laden with galls. Many a rural community will bear witness to the peculiarities of our conduct in their countryside, but our cabinets of insects and galls bear more eloquent testimony to the effectiveness of this method.

Because of the peculiar host relations of the higher Cynipidae, it is necessary to collect from every species of oak in each locality in order to obtain a satisfactory set of specimens. To avoid being misled, in our interpretations of species, by such local populations as might segregate into Mendelian races, we make it a point to secure material from separated trees of each host wherever possible. Our returns are more than doubled when two of us work together in the field. We have given from a few hours to several days to the exploration of each locality, but upon careful accounting we find that, with two or more of us working together, we can take a fair sample of a region in four or five hours.

In order to secure approximate data on the furthest extension of the range of each species, we do not make our