MAMMALS ON OCEANIC ISLANDS. 391 



sometimes by species found elsewhere, striking instances of 

 which have been given by Dr. A. A. Gould in relation to the 

 Pacific. Now it is notorious that land shells are easily killed 

 by sea-water ; their eggs, at least such as I have tried, sink 

 in it and are killed. Yet there must be some unknown, but 

 occasionally efficient, means for their transportal. Would 

 the just-hatched young sometimes adhere to the feet of 

 birds roosting on the ground and thus get transported ? It 

 occurred to me that land-shells, when hibernating and hav> 

 ing a membranous diaphragm over the mouth of the shell, 

 might be floated in chinks of drifted timber across moder- 

 ately wide arms of the sea. And I find that several species 

 in this state withstand uninjured an immersion in sea-water 

 during seven days. One shell, the Helix pomatia, after hav- 

 ing been thus treated, and again hibernating, was put into 

 sea-water for twenty days and perfectly recovered. During 

 this length of time the shell might have been carried by a 

 marine current of average swiftness to a distance of 660 

 geographical miles. As this Helix has a thick calcareous 

 operculum I removed it, and when it had formed a new 

 membranous one, I again immersed it for fourteen days in. 

 sea-water, and again it recovered and crawled away. Baron 

 Aucapitaine has since tried similar experiments. He placed 

 100 land-shells, belonging to ten species, in a box pierced 

 with holes, and immersed it for a fortnight in the sea. Out 

 of the hundred shells twenty-seven recovered. The presence 

 of an operculum seems to have been of importance, as out of 

 twelve specimens of Cyclostoma elegans, which is thus 

 furnished, eleven revived. It is remarkable, seeing how 

 well the Helix pomatia resisted with me the salt water, that 

 not one of fifty four specimens belonging to four other spe- 

 cies of Helix tried by Aucapitaine recovered. It is, how- 

 ever, not at all probable that land-shells have often been 

 thus transported; the feet of birds offer a more probable 

 method. 



ON THE RELATIONS OF THE INHABITANTS OF ISLANDS TO 

 THOSE OF THE NEAREST MAINLAND. 



The most striking and important fact for us is the affinity 

 of the species which inhabit islands to those of the nearest 

 mainland, without being actually the same. Numerous 

 instances could be given. The Galapagos Archipelago, situ- 

 ated under the equator, lirss at the distance of between 500 



