SEASONAL POLYMORPHISM. 197 



and are usual, or whether they are aberrations, only met with 

 accidentally, must be some grade of a series of forms leading up 

 to it from the average, and going beyond it to more extreme ones, we 

 must contrive some way of designating these series, and then of desig- 

 nating each form as a grade of it by a comparative number. Tutt 

 used compound names as a first attempt, but this is a cumbersome 

 method, when several characters are combined in one specimen. I 

 believe the Linnean fancy name will have to be restricted to species, 

 races, and generations, infinitely less numerous, and that individual 

 forms will have to be designated by some sort of formula drawn from 

 letters and figures, indicating the various parts of the wing pattern, 

 and conveying at once to one's mind the position of that individual 

 form in the totality of variation of the species or genus. I have worked 

 at this problem considerably, and some day I hope to publish some 

 results. Contrary to my expectation, however, I have found that the 

 laborious task of making out characters by the statistical method does 

 not lead, in practice, to any useful result in fixing races ; the standards, 

 best characterised, are so obvious by simply comparing adequate series 

 of specimens, collected at random, that no elaborate method is necessary 

 to prove their existence, and for the present I see no object for long 

 calculations on intermediate and transitional ones. Statistical data, 

 however, give interesting results in other ways, and it is an excellent 

 training for the eye in the analysis of characters, besides being con- 

 ducive to the drawmg out of descriptions in a methodical and rational 

 way. 



In these various studies I have found it of great help to use series 

 of specimens set in glass mounts in the way I have described at length 

 in the Bull. Soc. Kntom. de France, 1917, p. 312, which enables one to 

 compare a large number at a glance, and on either surface. I have 

 thus detected many characters that would probably have escaped ray 

 notice among the thousands of specimens mounted for me by Querci and 

 his family, and which I now in many cases preserve as a " typical " 

 series. 



Last, but not least, what has been fundamental in my contribu- 

 tions to the study of variation has consisted in my having been able to 

 avail myself of the enormous amount of material collected each year 

 by the Quercis in Peninsular Italy (about 25,000 specimens yearly), 

 and of large series received by him from all sorts of regions, not to 

 speak of his observations and his forty years experience in the field. 



I must end ixp these introductory remarks by a few words of 

 explanation of the classification adopted in the List which follows. I very 

 much appreciate the efforts made of late by several, and chiefly by 

 British entomologists, to substitute a more rational and homogeneous 

 classification for the very rough and incongruous ones in use during the 

 last century. It is an extremely difficult task, and it still requires much 

 work, but my impression is we are on the right track. First of all, it 

 was quite the right thing to adopt the general principle of dealing with 

 the simplest groups and to gradually rise to the more complex and 

 highly specialised. I have followed this order, like Tutt, Wheeler and 

 others, in the large Divisions or Sections, and I have, besides, 

 endeavoured to apply it to the smaller groups more thoroughlj^ than 

 they seem to have done, if I am not mistaken. I need scarcely say 

 that the classification along a single line, which one is compelled to 



