l 9ß Geographische Verbreitung, Reisen. 



though replaced by a distinct species in the next. Thus the important factor 

 appears to be the actual geographica! isolation and not difference in food or 

 habits. In the same way certain insects frequenting one and the same species 

 of food-plant, but isolated in separate islands, are so modified as to be considered 

 distinct species. Among insects the cases where a species inhabiting one island 

 are replaced in several others by closely allied species are too numerous to 

 mention. Some groups of species are very abundant in certain islands closely 

 adjacent to one another, but are very poorly represented or even totally absent 

 in the more remote ones. In many cases isolation appears to be directly pro- 

 portional to the distance between the islands: thus in certain genera the closely- 

 adjacent isles of Lanai, Molokai and Maui have many species in common, while 

 the more remote Hawaii and Oahu have few such, and the most remote island 

 of all (Kauai) has none in common with the others. It is sometimes apparent 

 that there are two distinct Clements in an endemic group, one consisting of 

 wide-ranging forms found throughout the archipelago, the other consisting of a 

 number of species restricted each to a single island. In such cases the wider 

 distributed forms usually have the greater distributional powers: e. g. in the Ca- 

 rabid beetles, the majority of endemic species of which are wingless, it is mostly 

 the fully-winged forms that are found in moi-e than one island. But this rule has 

 many exceptions, witness the native. crow and buzzard, endemic birds with great 

 powers of flight, but neither of them occurring outside the island of Hawaii. 

 "We may see" writes Perkins "forms on the various islands in various stages 

 of becoming distinct species, from the earliest stage, where merely the ränge of 

 Variation of a species is different on two islands, to that where every individual 

 is different but the differences are so slight as to be hardly specific, and finally 

 to that where the separating characters are too important to be considered racial 

 or varietal, when we may feel fairly confident, that even should a fresh immi- 

 gration of the original species take place from the one island to the other, inter- 

 breeding would not follow." He notes also that when isolation has taken place, 

 secondary sexual characters are frequently the first to be modified. 



There are also many cases in which allied forms are geographically isolated 

 though inhabiting the same island. The best Illustration is provided by the 

 multiplicity of endemic forms of land Mollusca. Perkins describes the case of 

 some of the arboreal Achatinellidae in Molokai, where distinct forms occupy 

 mountain-ridges almost adjoining one another. It cannot be told whether the 

 parent form once occupied the intervening areas, and subsequently died out in 

 them and so became discontinuously distributed. But as Perkins shows by de- 

 finite observations on particular specimens, the sluggishness of these creatures is 

 so excessive, that a valley between two ridges would be almost as effective a 

 barrier as a sea-channel, and the individuals on either side would be almost as 

 isolated as though on two sej)arate islands. 



There remains a quite different category of cases, in which distinct but 

 allied forms inhabit not only the same island but the same area. This 

 has probably come about by isolation due to change of habits. For example, 

 there are three species of Longicorn Coleoptera of the genus Plagiihmysus, which 

 are sometimes all found in an area of a few Square yards. But though so simlar 

 and closely related they keep quite apart: each keeps to its own food-plant and 

 though they sometimes occur on adjoining trees they neither mix nor interbreed. 

 How and why such changes of habits occur is not exactly known, but once they 

 have occurred it is easy to see how efficiently the forms are isolated from one 

 anotlier. For instance there are two forms of a single species of beetle {Plagith- 



