6 MH. A. MUEEAT ON THE GEOGEAPHICAL EELATIONS OE 



the Pacific islands ; in the meantime I may briefly characterize it 

 as meagre to the last degree, most unequal in its proportions, 

 and all traceable to the shores of the nearest lands from which the 

 currents set. 



The supposition that the existence of the similarity in ques- 

 tion ia due to the facilities for migration to or from the northern 

 or southern hemispheres afforded by the low temperature of the 

 glacial epoch is open to various answers. But it is unnecessary 

 to discuss them at all ; for I shall presently show that the re- 

 semblances to which I have to refer were already in existence 

 before the glacial epoch commenced, consequently could not have 

 been caused by it. 



The hypothesis that similar forms occurring at distant places 

 are the remains of a general fauna (or, at least, of a more 

 general fauna than now exists), which had in former times ex- 

 tended over the whole or the greater part of the world, is more 

 attractive or more formidable. 



I used to think that in that hypothesis I had a satisfactory 

 explanation of all such anomalies as I speak of. Like Shak- 

 speare's barber's chair, it fitted all comers. If the similarity was 

 widely spread, it was due to universal prevalence in former 

 times. If found only in one or two isolated spots, then there 

 were solitary relics of a once universally distributed type ! But 

 I confess that my faith in my specific has latterly been a good 

 deal shaken. It costs me nothing to say so, for consistency is 

 a vice to which I have never been addicted. I believe it still 

 to be probably the true explanation of those cases (as in Terns, 

 for example) where the same type is very widely and gene- 

 rally distributed; but I have abandoned it for most isolated 

 instances, and for all specially localized faunas. In the first 

 place, although I do not dispute that in the earlier stages of the 

 history of our planet there was a greater homogeneity of type 

 than there is at present, it seems pretty well established now, 

 that there have been geographical regions with faunas and floras 

 difi'ering from each other, not indeed to the same degree as now, 

 but to some extent, from the very earliest times of which we have 

 any fossil record ; and in the next place, although it is not impos- 

 sible that a universally distributed form may have died out every- 

 where but in one or more specified spot or spots, the doctrine of 

 chances prevents us accepting the hypothesis whenever such 

 relics cease to be solitary. Species No. 1 may be a relic left at 



