14 
more ill-defined higher grades (genera, tribes, etc.) on the other. 
Much that has been said on these subjects has been repeated over and 
over again at different Societies and in different connections, and is in 
danger of palling; but there are certain points which, though I cannot 
at all claim them as original, have not often been brought out 
prominently in this room—at least, not of recent years; and I will 
venture to inflict upon you my views of these. 
The question, “ What is a Species ? ” will never be fully answered 
unless we abandon evolution as our “ working hypothesis.” If new 
forms may, by any evolutionary process whatsoever, arise out of old, 
it is surely manifest that incipient species of various grades will have 
to be faced by the systematist, and that no one criterion can serve him 
for his differentiations in all cases. Even if it can be proved that 
certain variations do arise per saltum, it would be hard to convince 
careful and logical thinkers that these become instantly and finally 
segregated, and so become entitled to the designation “species.” 
Indeed, many of Bateson’s examples of “discontinuous variation,” 
many of the dimorphs which are supposed to act as “ Mendelian ” in 
cross-breeding, are recognised to breed together so freely that the most 
ardent species-monger would not think of separating them specifically. 
This question of breeding together (the “ Syngamy ” of Prof. 
Poulton s presidential address on the present subject) and the cognate 
one of tracing faithfulness to type in pedigrees (“ Epigony ” of the 
same paper), probably suggest the best criteria that we can hope to 
obtain for a decision of our species-limits ; but even if we had all the 
knowledge of these points—which is, of course, an impossible sup¬ 
position—I believe Ave should find untold difficulties and irregularities. 
To give one or two purely conjectural instances : A and B perhaps 
would be found to breed freely together in a state of nature, and so 
would B and C, but A and C would, on account of some subtle 
geographical or physiological barrier, fail to do so; or some individuals 
of A would produce offspring so mixed in characters, that we should 
call them hvo forms (A and B) while other individuals would produce 
offspring indistinguishable from the parent type; or, in a word, a 
“ species ’ in one locality, or in one generation, might become a 
“variety” in another. That there are any cases of this kind yet 
proved to absolute demonstration, I cannot positively assert, though it 
seems to me highly probable, that the extreme forms of Papilio 
dardanus, a species discussed in Dr. Jordan’s important Die Gegensatz 
zwischen geographisclier und nichtt/eographischer Variation (1905), would 
be found incapable of cross-pairing ; in any case, some species within 
my own domain, as a Geometrid specialist, might plausibly, I think, 
be regarded as suggesting problems of this nature. Even Cidaria 
( Dysstroma ) truncata and citrata, on which I discoursed in this room 
not so very long ago, have relatives in Asia and in North America, 
which might easily be regarded as specifically distinct from both ; yet, 
the British ones are amply distinct species by any and every criterion. 
Similarly, Tephrosia ( hctropis ) bistortata and crepuscularia, which are 
certainly neither syngamic nor synepigonic, in Southern Britain, may 
possibly be so in some parts of Europe, as so very few non-British 
entomologists seem to have noticed their distinctness in dates of 
appearance, and constancy of facies in accordance therewith ; or, at 
xix. 
