552 



SCIENCE. 



[N. S. Vol. XXII. No. 666. 



since it is thinkable at least that the subspecies 

 might have become ' isolated ' in the same locality 

 by adopting different times of propagating. Might 

 not such cases occur among fishes? I have also 

 a sort of suspicion that it may occur, or may have 

 occurred earlier, in some migratory birds on both 

 sides of the equator. Altogether, I wish to be 

 very reserved in my answers and I would have 

 you understand an ohligato ' so far as I know ' 

 added to all of them. The fact is, that the 

 records when it comes to distinguish between the 

 propagation habitat and the general habitat of 

 the species are so defective that no fully reliable 

 conclusions can be drawn from them. 



1 and 2. I know of no well-founded subspecies 

 propagating simultaneously in the same locality 

 (I would avoid the word region as not restricted 

 enough ) . 



As a matter of fact I can not well conceive of 

 two subspecies propagating simultaneously in the 

 same locality except perhaps during a short time 

 under the following supposition as shown by an 

 example : 



Suppose the so-called Lanius major of Siberia 

 and Lanius excubitor were only subspecifically 

 distinct, i. e., that they had not yet developed be- 

 yond a certain degree of character stability along 

 the original territorial line separating them some- 

 where in Central Asia (a supposition I do not 

 admit, though I do not deny the possibility, the 

 data at hand being inconclusive ) . Now sup- 

 pose both species extended their range westward, 

 L. excuhitor to central Europe thence north to the 

 Scandinavian peninsula via Denmark, while L. 

 major pushed northwestward over Finland to 

 northern Norway (and these suppositions I be- 

 lieve are correct ! ) . Suppose further that the 

 breeding ranges of both subspecies met and finally 

 overlapped, say in some localities in Finmarken 

 (and there are indications that they do, though 

 the records are anything but satisfactory). The 

 chances are, of course, that in these localities the 

 two subspecies would mix, but for some time at 

 least it is probable that a certain percentage of 

 both might continue to breed pure alongside of 

 each other. Though all this is mere speculation 

 it is probably legimate to carry the supposition 

 a step further. Finmarken is the most extreme 

 end of the range of both (supposed) subspecies, 

 it is then not unnatural to conclude that in the 

 specimens meeting there the characters might 

 have become so fixed that the two forms would 

 react on each other as two distinct species, 

 though at their original dividing line they might 



still remain in the imperfectly differentiated 

 stage. 



3. No, as a rule not. But if as I hinted 

 above the isolation were one of time of spawning, 

 for instance, concomitant with which a certain 

 amount of structural differentiation had taken 

 place, then I can conceive of such a state as indi- 

 cated in question 3, viz., that the ranges (in space) 

 of both subspecies might be coextensive. 



4. There are a few cases where apparently 

 two species inhabiting exactly the same region 

 (these terms restricted as above), are apparently 

 more closely interrelated than to any other. As 

 an example I may quote two Japanese birds, Gettia 

 cantans and C. cantillans, the chief and possibly 

 the only difference being the constantly greater 

 size of the former. Yet we are told that they 

 occur in exactly the same localities. I used the 

 word apparently above because in the first place 

 the exact relationship of the two species is not 

 ascertained, and in the second, the records of the 

 breeding ranges of both are not as complete as 

 might be desired. That the two forms are dis- 

 tinct and do not intergrade seems certain. Alto- 

 gether I do not pretend to understand the case. 



There are a few other similarly puzzling cases, 

 for instance, that of Phyllopneuste horealis and the 

 so-called subspecies xanthodyas but apart from 

 the fact that in this case the latter is not coex- 

 tensive with the former only occupying part of 

 the territoiy, the records are so incomplete and 

 defective the chances are the puzzle may be easily 

 solved when all the facts are known. 



The ease of various forms for crossbills might 

 also be cited. In their case it almost looks as if 

 each band is kept isolated by their gipsy-like 

 habits, though it is just possible that they may 

 have evolved different breeding periods or dif- 

 ferent food making them dependent on the latter 

 for the selection of their temporary breeding 

 locality. At any rate, the isolation does not seem 

 to be geographical or topographical. 



I hope I have been able to make my standpoint 

 clear, viz., that two geographical subspecies can 

 not propagate in the same locality (except in rare 

 cases illustrated by the Lanius example ) , but that 

 there are other kinds of isolation — or possibly 

 better: segregation, which — though more rarely — 

 may produce subspecies — eA'entually species when 

 the incomplete differentiation has become com- 

 plete, — capable of living side by side in the same 

 locality. 



WILLIAM BREWSTER. 



I have no knowledge based on personal experi- 

 ence that two well-marked subspecies of the same 



