PARTULA GIBBA, GUAM. 77 
reference to any other. It is conceivable that mitella is itself a recent product 
which has gained ascendancy over its parental type and its other associates. But 
whatever the situation in remote centuries, the fact remains that it is of long 
standing, for it occurs also in Saipan and in Tinian; the view that it has been 
evolved independently in these separated islands is less tenable than the conclusion 
that it is an ancient form, possibly distributed by passive colonization but more 
probably isolated in the several islands by the subsidence of a larger land-mass of 
earlier epochs on which it ranged widely. 
The mitella-rubra order is one which might have arisen independently from 
mitella in different localities, because its members are not everywhere identical in 
detail. The Lonfit and the Tarague representatives are especially noteworthy in 
this connection. 
Among the darkly colored classes, the phaea group is abundant in the north 
as compared with other areas. There are two tenable hypotheses regarding its 
present condition: firstly, that it originated somewhere in the north and subse- 
quently spread to the other areas now occupied, or secondly, it may have been far 
more general in its occurrence at an earlier time, only to disappear from much of 
its original territory. The same problem arises in connection with castanea, but 
in this case the latter hypothesis is supported by the fact that in Saipan there are 
types that are undoubted relatives of the castanea of Guam, even though they 
differ in distinctive details. 
With vespera the case is different. This is purely a Guam color-class so far 
as the facts are known. The present distribution is compact and indicates a center 
of origin and dispersal focusing about Mount Barrigarda, from which the lines 
of migration radiate along areas of thicket and forest which are relatively unin- 
terrupted. The relation of vespera to castanea is closer than that of vespera to 
any other form, as the immature stages demonstrate; whether or not the former 
was locally produced from the latter by mutation is impossible to determine with 
absolute certainty, but such an origin is at least probable. 
It is interesting to find the ruddy and purplish phases in the three distinct 
classes, phaea, castanea, and vespera. The same modes are displayed by many 
another species of Partula, such as P. rosea of Huaheine and P. hebe of Raiatea in 
the Society Islands. Apparently the two characteristic tints readily replace each 
other, or are interconvertible by the operation of a minor genetic factor. 
The present section must not conclude without reference to the clear demon- 
stration afforded by the foregoing facts that the environmental conditions are wholly 
devoid of any causative value as regards the qualities of coloration. We can not 
say that the general occurrence of the mitella type is due to external factors which 
are identical wherever the class occurs, and at the same time hold such assumed 
influences responsible for the diverse types of the same areas—that is, identical 
factors can not be the causes of different results in the way of coloration. Con- 
versely, such ecological diversities as are exhibited by areas like Barrigarda, Macajna 
second, and Aniguac could scarcely be the reasons for the existence in these places 
of one and the same color-class like vespera. Again the conditions at Dungcas 
