PARTULA. 161 



P. faba and the terrestrial P. radiata-, a number 'between 

 P. faba and P. fusca, and many between the latter and P. 

 navigatoria, as well as many between the latter and P. faba. 

 I failed to detect hybrids between the Tahitian species, and 

 found none at Huaheine. ' ' 



As a general proposition, apparent evidence of hybridism 

 among land snails is so unusual in a state of nature that 

 one is inclined to believe that well-established species very 

 rarely hybridize. In the few species tested experimentally 

 (such as Lang's crosses between Helix nemoralis and 

 hortensis) the progeny are not abundant and crosses between 

 them are almost sterile. It must be remembered, however, 

 that the Moorean and Baiatean species mentioned as hybrid- 

 izing by Garrett v are very closely related forms, which have 

 for the most part scarcely reached the "specific" stage of 

 differentiation; hence it is not inherently improbable that 

 hybrids occur. 



In the case of species between which supposed hybrids are 

 "common," I would rather explain the intermediate forms 

 occurring on the overlapping confines of the two "pure" 

 races as an undifferentiated remnant of the parent stock 

 which has elsewhere become "specifically" differentiated. 

 Pending experimental evidence to the contrary, this view 

 seems in accordance with what seems to occur in a multi- 

 tude of other cases, whereas hybridism on an extensive scale 

 is certainly a very rare condition. 



Variation in Partula. 



Extended study of variation does not come within the plan 

 of the present work, the purpose of this Manual being the 

 definition of species and their systematic classification. 

 Moreover, the materials for a study of variation large, un- 

 selected series of shells are not accessible to me; my per- 

 sonal knowledge of them is restricted to museum specimens, 

 which as usually preserved and labelled give little idea of the 

 association of forms or the conditions of their existence. 

 Some general considerations bearing upon variation may how- 

 ever be in order. 



