206 Genetical Studies in Moths 



Any exact or even approximate definition of what constitutes a sub- 

 species is, however, an enormously complicated matter, rendered the 

 more so by the uncertainty as to what a species itself really is. 



If we compare 0. autumnata and 0. filigrammaria we find that in 

 important structural characters they differ not at all ; moreover, they 

 are perfectly fertile when mated together. Nevertheless, when one 

 actually tabulates the differences (as I have done below, as well as those 

 between 0. dilutata and 0. christyi) we find that the insects fail to 

 agree in many minor points both in structure and in life history. In 

 addition each has the power of setting up its own peculiar local races — 

 races each one of which is characteristic of its own locality, but in fili- 

 grammaria never altering its fundamental details to show that its 

 degree of differential evolution passes from that of its fellows elsewhere 

 toward the 0. autumnata with which it is phylogenetically connected : 

 and the same remarks apply to 0. dilutata and 0. christyi. Probably 

 the best view to assume concerning the subspecies is to regard the term 

 as covering those cases in which we have groups, sufficiently differentiated 

 from those pronounced by systematists as good species for one to regard 

 them as possible species, but upon which one is forced to suspend one's 

 judgment, either because of a lack of clarity as to what a species is, or 

 because one cannot gain sufficient of the essential details upon which to 

 found a decision free from any tendency to waver. The very great 

 difficulties in the way of gaining the necessary decisive facts are em- 

 phatically shown when the perfect fertility between 0. autumnata and 

 0. filigrammaria and between 0. dilutata and 0, christyi is contrasted 

 with the puzzling behaviour in that respect of certain forms of the genus 

 Orgyia which I can only regard as subspecies; the matured judgment 

 of years of study will not permit me to treat them as being of specific 

 rank. These are those subspecies of Orgyia antiqua known as 0. hadia, 

 found in Pacific North America, and 0. nova in North America east of 

 the Rocky Mountains. In every character the two American forms are 

 widest apart, 0. hadia being almost unrecognisable as the same species 

 a^ 0. nova when one examines larva, pupa and imago. 0. antiqua from 

 Europe on the contrary is excessively near to 0. nova and thus differs 

 greatly from 0. hadia. 



Nevertheless, whilst the two American forms, when crossed together, 

 are reciprocally fertile and continue so as long as one cares to pursue 

 the experiment, when one crosses either of the American forms with 

 the European the cross is only fertile when the European insect provides 

 the female, but, as before, the brood so produced can be inbred in- 



