84 



hidiana University Studies 



conception of distribution facts. To designate these as 

 species, subspecies, and varieties is not as convenient a method 

 as to consider the group a genus. 



The distribution of the varieties of most of the species is 

 in strict accord with my findings for other Cynipidse, Hmiting 

 each variety to one faunal area. In Heteroecus hakeri there 

 is considerable variation within the single area, and not a dis- 

 tinct type in the San Bernardino range as would be expected. 

 Altho I have some hundreds of the insects and some thou- 

 sands of the galls, I do not yet have material enough to rec- 

 ognize all of the varieties which probably exist for any one 

 species. There are at least four, possibly six faunal areas in 

 which Qtiercus chrysolepis occurs, and each species is likely to 

 have as many varieties if it occurs over the entire range of its 

 host. Geographic isolation appears to have been a factor in 

 preserving these variations. 



The confinement of the genus in California to the single 

 host, Q. chrysolepis, suggests that isolation upon distinct hosts 

 has had nothing to do with the origin or preservation of the 

 species. Often several species occur on a single tree. 

 Whether species interbreed, whether each type will breed true 

 in successive generations, cannot be stated without experi- 

 mental data. Intermediates however do not exist in my col- 

 lection (except in hakeri, as noted), and each type of adult is 

 definitely connected with a particular type of gall. Exactly 

 the same and parallel conditions have held in the past, as 

 earlier collections show, collections made in some instances 

 (pacificiis and dasydactyli) thirty-five years previously. I am 

 inclined to expect to find that each species remains distinct. 



It is to be noted that this amount of variation occurs in an 

 agamic group. Field data would suggest that an alternate, 

 bisexual generation does not exist. Whether bisexual repro- 

 duction ever occurs is a matter of some importance genetically, 

 and one which some student on the field should determine as 

 soon as possible. 



I have attempted a key to the forms described, but can- 

 not effect an arrangement which is not very artificial, nor one 

 which would be intelligible without considerable series of sev- 

 eral species at hand. The following conspectus may prove 

 more satisfactory, and not too inconvenient for this number 

 of forms. 



