58 
might have been chosen, but it is one that might easily and does 
happen. We have a well known instance of size in our Selenia 
hilunorta and juliaria. It is true that in this case the food-plant is 
not the reason of the difference, but the altered time of appearance. 
Do eggs laid by the summer brood alicays produce a spring brood 
or do they repeat the summer emergence only occasionally ? And also 
do imagines as a rule, or under any circumstances, that hatch in the 
spring, only produce the spring brood '? The experiments of Mr. Merri- 
field give us an answer to the latter question. He has found that with a 
leduced temperature the second brood fails. In nature this is exempli¬ 
fied in northern and western Scotland, and probably elsewhere. An 
answer is not forthcoming to the former query. One surmises that 
the spring brood yields a summer brood, and the summer brood the 
spring brood. But it is possible under certain conditions that the 
summer brood might fail to produce the spring brood. This might 
commence sporadically, but in course of time become regular. Should 
we then, as soon as this summer emergence fails to yield spring 
emergences, call it a species ? It would depend upon our arbitration. 
My personal answer would be no—not until the isolation had caused 
a greater amount of divergence. The divergence would be certain to 
take place at a later period. That the form should be named is 
indisputable, but not named with species rank, but as subspecies or 
some other such title. The use of the term subspecies shows now 
very evidently that there are some of our workers who feel that the 
number of denominations is too limited. Even with our species— 
subspecies—variety—and aberration, we are by no means able to 
relegate all the individuals to these four headings without difficulty. The 
trouble is, however, not comparable to the forming of genera, because we 
can by breeding test our theories very often as to species. Two, three 
and even a dozen so-called species can sometimes be proved to be only 
one species proper, with varying facies, called forms, by rearing from the 
egg. We are unable to demonstrate the same with genera for a collec¬ 
tion of species. In common with many others I at one time thought 
that perhaps we might have a definition of a species. That is totally 
impossible when we consider the various paths by which a species 
obtains its fixity or differentiation from others. But it would be 
possible to have a long series of categories, any one or more of which 
we might accept as valid for a species taking that rank. Supposing 
we fix on ten salient features from which characters we accept a species. 
By a small algebraical calculation we should have 1022 groups, from 
any one of which a species could take its standing as fulfilling one or 
more of those ten features. This is admittedly only taking ten main 
points. We might raise that number, and then the groups would 
be still more numerous. I cannot but think that as time goes on and as 
we still get nearer and nearer to exactitude, that the results to which 
our classification leads us will compel us to adopt some such scheme. 
If we look upon classification as an artificial arrangement for placing 
together the forms most like each other, and do not accept classification 
as illustrating the genealogy of the beings, there can be nothing to be 
said against one artificial arrangement within an artificial method. 
But even if we take the vein that I have been assuming this evening, 
that our classification reveals the descent of living beings in nature, I 
can see no objection to our maintaining certain conditions for our 
