PARTULA RADIOLATA. 4] 
On the higher ground of the Ylig territory the shells become much larger, which is 
just the opposite of the first-noted relation, namely, that on the high ground of 
Fonte and Macajna the shells were the shortest in the whole primary region. 
Thus there is no consistent relation between the variations in the average length 
of the shells of the several associations and the differences in altitude. The same 
negative result is the outcome when other ecological circumstances are examined. 
Taking a brief general survey of the whole material as defined by the statistics 
and as illustrated by figures 20 to 56, plate 11, the longest shells as indicated by the 
colonial average are those of Umatac Madog (fig. 27, plate 11); in the same locality 
the shells with the largest apertures occur. The colony with the shortest shells is 
that from the western part of Cabras Island, but this is followed by the series from 
Orote and not by the collection from the adjacent area of eastern Cabras; in fact 
the shortest shell of the entire series is an Orote specimen (fig. 41, plate 11). Pro- 
ceeding outward from the Apra Region, the colony which comes next in the matter 
of shell length is that of Inarajan at the extreme southeast; in the intervening 
territory of the Southwest Region the shells are long and large, as we have noted. 
So also in the proportionate measures the colonies vary without any discover- 
able relation to ecological conditions. The stoutest specimen (fig. 33, plate 11) 
is from Dededo, while the highest colonial average in the quality of relative width 
is displayed by the outer Presidio association. A slender specimen is shown in 
figure 30, plate 11, from Saucio. 
Reverting to the character of length, the figures and the illustrations show a 
truly remarkable range of variation. The longest shell (fig. 27, plate 11) is 20.4 
mm. long, while the shortest (fig. 41, plate 11) is only 13.6 mm. Thus the total 
range amounts to 6.8 mm., which is very large in comparison with the general 
average of this character in the whole collection, namely, 16.62 mm. When it is 
recalled that this measure is but one of the three general diameters of the animals, 
it is evident that the total bulk is subject to an extraordinary degree of variation 
within the species as a whole. 
Turning to the regional groupings, in the nature of the case the Northeast, 
West Central, and Southeast Regions are each represented by only a single associa- 
tion. The smaller collections of the East Central and Coast Central Regions are 
included in the combined populations of those districts, and hence the total 
numbers of snails in each instance exceeds the added numbers of the groups 
enumerated and defined in detail. While the regional data in question are derived 
from very uneven geographical assemblages, they are valuable so far as they go as 
empirical descriptions for the larger areal divisions of the entire population. Their 
significance is the same as before, namely, that the environmental diversities of 
the several regions, such as those of altitude, nature of the soil, and the like, are not 
causal with reference to the structural characters of the shells. While this con- 
clusion is readily apparent from an inspection of tables 12 to 14, it is perhaps more 
evident from the figures of table 15, which gives the numerical differences in mean 
values in the progress from one region to another; when the differences are not 
statistically significant they are inclosed in brackets. Taking the quality of shell 
length, a marked increase is noted in passing from the Northeast territory (Tarague) 
