PARTULA OTAHEITANA. 181 
In the judgment of the present writer the balance of evidence is in favor of (1) 
a Mendelian order of inheritance and (2) the dominance of red over yellow when 
the individual cases are weighted according to their reliability, and when the con- 
clusions of Lang in the analogous case in H/elix are taken into account. Clearly, 
however, actual breeding experiments are necessary before the final word can be 
said. 
SUMMARY. 
Having completed the survey of the whole primary variety, P. 0. rubescens, 
we may now bring together the general results which have been obtained through 
the detailed analysis of its several valley associations. This summary is more 
interesting than in the case of P. 0. amabilis, for many reasons: principally because 
this subspecies is exclusively sinistral, it is simpler in its make-up because it ranges 
over a far wider territory, and above all because in most of the valleys of its range 
it exists side by side with another subspecies, P. 0. affinis, with which it has no 
reproductive relations. 
I. The relative numbers of rubescens in the otaheitana population and in the total 
Partula population vary from valley to valley. ‘The figures referring to the actual 
collections may be taken only as approximations to the true values without rendering 
this statement incorrect. In a few localities included in the whole area of habita- 
tion, viz, Faaripoo, Mahaena, Utuufai, Faone, Tehoro, and Aiurua, the numbers 
are so scanty as to indicate a real danger of extermination. The prevalence of 
affints in those valleys proves that the conditions in themselves are not adverse to 
Partule; hence there are positive and not merely negative reasons for concluding 
that variations in congenital vigor are displayed by the colonies of rubescens. 
Il. The general composition of the several rubescens associations varies likewise. 
The relativenumbersof “yellow” and “red” shells exhibit many different values, even 
where large numbers of snails were secured. In Haavinithe former kind greatly pre- 
ponderates, while in the next large valley of Tehoro the relation is entirely reversed. 
Ahonu and Papenoo are also near together, yet their rwbescens populations show 
essentially different combinations of the two principal color-classes. 
Ill. The structural characters of the shells, defined in quantitative terms, differ in 
the several colonies without revealing any consistent relations to environmental conditions. 
The eastern sector, in which most of the colonies are found, is uniformly moist, like 
the eastern part of Taiarapu and the southeast corner, where Oopu Valley is located. 
It is theoretically possible that undiscovered influences might vary from valley to 
valley, with increasing or decreasing power toward the extremes of the range, in 
such a way as to manifest their effects in the dimensions and proportions of the 
shells. Nevertheless, there are no empirical indications of such influences or of 
their effects. 
Specifically, the figures given in summary in table 115, the statistical differ- 
ences with their errors as given in table 116, and the graphic representations of 
table 117, which show the progressive changes in average values in geographical 
