116 THE CANADIAN ENTOMOLOGIST. 



the law of variation in representative species in the Owlet Moths. This 

 variation is first observed on the upper surface of the fore wings, then of 

 hind wings, while the whole under surface preserves its similarity longest. 

 The uniformity of the under surface in the Noctuidse seems to be corre- 

 lated with the habits of the insects themselves, to depend, in fact, upon 

 the conditions of its exposure to the light. I am not here arguing that 

 color in the wings is now dependent upon existing conditions of light. I 

 merely point out that variation both in color and marking proceeds appar- 

 ently more noticeably upon the more exposed surfaces of the insect in the 

 Noctuidse, from a comparison of related species inhabiting different parts 

 of the world. In the history of the Lepidoptera former geological con- 

 ditions have played a part in the evolution of species together with the 

 whole environment. In this paper I merely show the probability that the 

 first transverse markings were the effect of light, and that the more ex- 

 posed surfaces show most variation in representative species. From my 

 scattered writings I have in part brought these brief notices on this point 

 here together, so that the student may be spared that trouble, and in the 

 hope that the investigation may be carried further. 



In the investigation of this subject we must keep the phenomena of 

 color and pattern separate. The test of our theories must lie in the 

 observation of existing variations. In this direction the observations of 

 Mr. Edwards on the influence of cold in the pupa state upon the colors 

 of the imago, are of the utmost value. The class of facts bearing upon 

 the phylogeny of the species must be kept separate from those bearing 

 upon individual variation. But it must be remembered that varieties are 

 in the same sense evanescent species, that species are permanent varieties. 

 The crucial test of our modern idea of species lies in the demonstration 

 of the fact that, in the whole life history, the cycle of reproduction is now 

 distinct. To the establishing of this fact repeated observations are often 

 necessary. The whole conditions under which the form is produced must 

 be understood. This is a great field of work, and single instances, how- 

 ever carefully recorded, of breeding from the larva, only partially illustrate 

 the subject. The value of specific determinations from collections of per- 

 fect insects depends on the tact and experience of the naturalist and are 

 to this extent tentative. Only where the full round of insect life is known 

 can our determinations be absolutely reliable. The vista of entomological 

 labor is widening as we proceed, so that it is trite to say that the subject 



