138 VARIATION, DISTRIBUTION, AND EVOLUTION OF THE GENUS PARTULA. 
PIRAI VALLEY. 
Pirai Valley, as the sole area inhabited by P. filosa, has been described at some 
length at an earlier juncture (pp. 82, 83), where its distinctive features and the 
recent changes in the aspect of its vegetation in the lower portion have been noted. 
The representatives of P. 0. amabilis which have been collected here are of 
considerable interest, especially because this is the only valley in which collections 
were made in all four years of field work. The four annual series were obtained in 
different parts of the long valley, those of 1906 from near the mouth because the 
cyclonic disturbance previously mentioned made progress into the interior all but 
impossible; in 1907 further ingress was feasible, and in later years collections were 
made still farther inland, very nearly if not quite to the highest limit of the area 
inhabited by Partula. It is possible, therefore, to compare different sections of 
the whole population, and it appears that, as in jilosa, these sections are statistically 
different in certain respects. 
Only sinistral specimens were found in a series of 1,467 adult and adolescent 
individuals brought back for detailed study, in addition to some 400 others which 
were taken in 1909 and after examination were returned to their environment. We 
may be reasonably sure, then, that dextral specimens do not occur at present, and 
also that none have occurred for many decades, because Mayer obtained 131 adults 
which were all sinistral, while Garrett explicitly states that no dextral shells were 
found by him. It will be interesting to ascertain the state of affairs in Pirai Valley 
after the lapse of 25, 50, or 100 years, to see if the direct shells make their appear- 
ance; and, if they do, whether they arise spontaneously by mutation or migrate 
from another valley. 
Although the four color-classes of the previously described valley recur here 
(plate 26, figs. 1 to 16), the general complexion of the colony as a whole is decid- 
edly distinctive, for the darker-colored shells of Classes III and IV (plate 26, figs. 
g to 16) are greatly diminished in numbers. Among those of Class I (plate 26, 
figs. 1 to 8) a well-marked group of nearly pure white shells exists (Ia) (figs. 1 to 4) 
as a result of general decortication. Such specimens are very few in Hamuta and 
still rarer in Fautaua. 
CoMPARISON OF THE ANNUAL SERIES. 
The four annual series represent the populations of as many sections of the 
inhabited area, in regular order from the coastward to the most inland. The pro- 
portionate numbers of the several color-classes change greatly in passing from the 
outer to the inner sections (table 72). Class I is the most richly represented, and it 
increases steadily from 60.5 to 84.2 per cent, although subclass Ia remains about 
the same; it follows, then, that decortication occurs in a larger number of Class I 
shells in the lower part of the valley than it does in the interior. Class III is very 
poorly represented, while only one entirely brown snail was secured in the highest 
part of the valley. Banded shells (plate 26, fig. 17) occur sparsely, on the whole in 
about the same relative numbers as in Hamuta; they are all sinistral. 
