88 the entomologist's eecord. 



whole of their area of distribution, and in fact one may say that most 

 of the work done by Lepidopterists in the last few years has consisted 

 in a more and more minute and exact analysis from this point of 

 view. Hitherto the most striking local variations have been noticed 

 and described. In several species however this part of the task has 

 been accomplished or is nearly coming to an end. It is now a question 

 of gradually establishing their distribution. In endeavouring to do so 

 we are usually confronted by the fact that they blend together in areas 

 intermediate between the regions, where the most striking races had 

 been observed. To complete our mental picture of the variations of a 

 species these shadings must be taken into account and a way of 

 recording them and memorising them must be adopted, or observations 

 about them will be entirely lost. I know this is a sore point for many, 

 but my own experience in more than one branch of science convinces 

 me always more and more that names, and nothing but names, can 

 fix a phenomenon in our brains and hold it in readiness there for 

 future use. On the other band it is perfectly true that an infinity of 

 fancy names must not be created or they will be more a hindrance 

 than a help ; they should be as expressive as possible and convey con- 

 nections and similarity as well as differentiation. I have already 

 shown in other papers, when dealing with several Bhopalocera and 

 Zyyaena, that in most cases geographical variation is due to quantitative 

 rather than to qualitative differences, and that the races we perceive 

 chiefly consist in the progressive grades of one or more lines of 

 variation, one of which is primary, in the latter case, and the others 

 collateral branches. When we have to deal with a race intermediate 

 between two already described and named I think the following policy 

 will be found practical : If the great majority of individuals are found 

 to belong to intermediate forms and only a few extreme ones, not ex- 

 ceeding a quarter of the total number, are similar to the named races, 

 there should be no hesitation in giving a new name to the series in 

 question, because we can regard it as a distinct grade. We thus 

 complete a series of grades, all, on the whole, quite distinct from each 

 other ; they will be numerous when individual variation is usually 

 small in each locality as compared with the extreme variations of the 

 species ; they will be few in species which in most localities produce 

 many individual forms ranging broadly in the specific variability. 

 When these distinct grades have been established as a base, cases of 

 intermediate-looking series are reduced to the following : — (a) Indi- 

 vidual variation is unusually broad and numerous examples are 

 observed identical with two or more known races. Series of this sort I 

 should designate by a compound name, joining the names of ihe latter 

 with a hyphen, (b) The series can on the whole be referred to one of 

 the named grades (A), but many individuals exhibit distinct signs of 

 variation towards the next grade (B). I should then use the designation 

 of race A tram, ad B. In other localities race B trans, ad A is sure to 

 be found and thus our nomenclature will cover all the shades of 

 variation. The following examples of filipendulae afford good 

 illustrations. 



Oriental sub-species or species allied to filipendulae: 

 ■ — I do not profess to know the oriental Zyijacnae and I will risk no rash 

 judgment on figures, descriptions, or a few specimens. I thus leave 

 aside the interesting ledereri, Stdgr. and Rebel, which Seitz says is 



