EXPEDITION TO TAHITI 205 



radially by more or less regular valleys, some of them being more 

 that a half-mile in width at their mouths, where they debouch upon 

 the low coastal plain of alluvial soil upon which grow the palms 

 and fruit trees characteristic of the tropics. Nearly every one 

 of the eighteen districts of the whole island contains at least one 

 large valley, and often many other lesser valleys, through each of 

 which flows a stream that rises in the higher central part of the 

 island, where the precipitation of rain is almost constant through- 

 out the day. Luxuriant vegetation fills the bottom of each 

 well-watered valley, even up to high barometric levels, and it is 

 upon certain of the plants in the higher and moister parts of the 

 valleys that the Partulse were to be found. And because of the 

 high and barren intervening ridges certain species are absolutely 

 isolated from their neighbors, while in some cases, where the moist 

 zone extends down quite or almost to the alluvial plain, a certain 

 amoimt of migration from one valley to another is possible. The 

 facts relating to the connection between geographical isolation 

 and specific differentiation are therefore of the greatest interest. 

 More than thirty collecting trips were made up the valleys, 

 of which twenty-one were explored, over 400 miles being traveled 

 in their course. Of the 120 miles of coastal circumference, over 

 70 were covered; of the remainder the peninsula, a uniform 

 region, formed the greater part, and the rest, the northeast part 

 of the island, was practically inaccessible without the expenditure 

 of an inordinate amount of time. The material collected con- 

 sists of more than ten thousand individuals, adults and ado- 

 lescents, a number that may rise to twenty thousand when the 

 young are dissected out from the adults. The distance from the 

 sea, the barometric level and other data relating to food plant, 

 etc., were determined in all cases. The snails belong to the 

 species P. hyalina, an invariable and widely distributed form; 

 P. nodosa, from the southern area, where within recent years it 

 has arisen and spread, at the same time undergoing wide varia- 

 tion ; P. filosa from a single valley on the northwest side ; and 

 the several forms, amahilis, sinistrorsa, sinistralis, lignaria, 

 mhescens and crassa, that are by some regarded as varieties of 

 P. otaheitana, still another form, while by others they are re- 

 garded as distinct species. For our present purposes, it is the 



