82 MR. A. MURRAY Otf THE GEOGRAPHICAL RELATIONS OF 



(the Bees) are dependent on flowers for subsistence in the larva 

 state ; and supposing the cold of the glacial epoch to stop the 

 flowering of plants in Japan without killing the plants, the bees 

 might be exterminated while the other classes still survived. 

 Were Japan, therefore, separated from the mainland, so that on 

 the restoration of a milder climate no fresh supply of species 

 could be received from the north, and united to Southern China, 

 so that it received its new inhabitants from it, and then finally 

 separated from it as it now stands, we should perhaps have an 

 explanation of the actual phenomena as regards bees ; but there 

 are other Hymenoptera to which this explanation will not apply, 

 and further research may show that the exclusion of northern 

 types is not so rigorous as at present appears. At any rate, it 

 seems to me that if the whole earth might be replenished by chance 

 colonization, then the presence and absence of particular classes 

 of insects in Japan is without explanation. 



I presume it will not be necessary for me to show that the 

 same distribution prevails throughout Europe and Asia in every 

 class of animals ; Dr. Sclater was the first to do so in the birds, I 

 have elsewhere done the same for the mammals. Dr. Grimther 

 has done it for reptiles, G-abriel Koch has done it for the Lepi- 

 doptera, Meyen and Hooker for plants. In fact every person 

 is at one upon it, each in his own speciality. 



The Europeo-Asiatic Beetle-fauna* does not stop even at 

 Japan ; it passes over into North America by Behring's Straits, 

 or rather, I should say, it is found in North America on the 

 other side of Behring's Straits. In Bussian America we have 

 a fresh crop of Europeo-Asiatic forms, genera and species ; and 

 here another noteworthy circumstance presents itself. It is 

 generally taken for granted that there is a uniform homogeneous 

 arctic fauna which extends all round the arctic circle. It is so, 

 and it is not so. It is so on the large scale, but not so on the 

 small. The arctic fauna is subject to the laws of spreading by 

 continuity and stoppage by barriers just the same as any other 

 fauna. I have elsewhere endeavoured to show that the mamma- 

 lian fauna of Greenland is Europeo-arctic as distinguished from 

 Americano-arctic. I maintain that the homogeneity of a fauna 



* I was unable, in my 1 Geographical Distribution of Mammals,' to adopt 

 Dr. Sclater's terminology of Palaearctic, Neoarctic, &c, because we did not agree 

 in the extent and limits of our regions ; and now, of course, in this paper I can 

 still less do so, as a principal effect of my hypothesis, if it be sound, must be to 

 still further break down their limits and destroy their solidity. 



