52 VARIATION, DISTRIBUTION, AND EVOLUTION OF THE GENUS PARTULA. 
period in its history. Moki (sometimes spelled Mauke) is similar in structure, but 
the central area is not so deeply depressed as in Mangaia, nor are there any cliffs 
surrounding the now dried lagoon. Aitutaki is likewise an uplifted limestone mass. 
Garrett visited Rarotonga, Mangaia, Atiu, and Aitutaki, and reported hyalina 
from Mangaia only. His thorough exploration of other islands would certainly 
enable him to discover the species if it occurred. In Voy’s collection, now in the 
Philadelphia Academy of Natural Sciences, there are some hyalina shells labeled 
Rarotonga. In view of Garrett’s statements, and on account of my own inability 
to find hyalina in a particularly favorable valley in the island, I conclude that Voy’s 
shells were obtained from natives who had brought them to Rarotonga from 
another place. 
My own material from Mangaia consists of “dead shells” taken from the ground 
near the cliffs bordering the lagoon-like valley. Only here, where the orange and 
other trees grow thickly, is there a situation sufficiently moist and shaded where 
hyalina can live. The absence of living examples must be attributed to the exces- 
sively dry weather of the previous weeks that had driven the animals to deep shelter. 
In Moki living individuals were taken in dense Pandanus thickets a mile inland, at 
an altitude of 150 feet, aneroid barometer. Here and in two other forest localities 
dead shells were found, but no living specimens were seen at the latter places. 
Characteristic Cook Island shells are shown in figures 11 to 13, plate 20. Figure 
II shows a small example from Moki, and figure 12 presents a large specimen from 
the same island. ‘The shell of figure 13 is an average specimen from Mangaia. 
The statistics relating to these collections are given in tables 16 and 17. The 
conclusions drawn from the tabulated data are the following: 
(1) The adult shells of hyalina from Mangaia are longer, wider, and stouter than 
those of Moki; their apertures are also /onger and wider, but more narrowly oval; the 
aperture is relatively /onger in proportion to the total length of the shell. These 
conclusions are based upon a comparison of the dead shells from the two localities 
in question. From inspection it is clear that weathering of a dead shell reduces the 
apex more rapidly than the lower whorls or the margin of the aperture, but in the 
above comparisons such changes can be justly ignored. 
(2) The shells of living hyalina from Moki are larger on the whole than the 
dead examples from the same locality; the differences observed serve to indicate the 
amount of the change produced by the weathering of the apex and margin of the 
aperture. 
(3) Combining the above results, it would appear that living snails from Mangaia 
would undoubtedly be correspondingly larger than the dead shells; and if the 
available material were more abundant it would be profitable to determine the 
probable values of the several characteristics of the “living” shells from that island 
for comparison with forms from other islands. 
(4) Comparing the characteristics of adult shells from Moki with the whole 
population of Tahiti, it appears that the former are significantly different in several 
characters. The precise differences with their probable errors are given in table 17. 
