36 VARIATION, DISTRIBUTION, AND EVOLUTION OF THE GENUS PARTULA. 
VARIATION IN NUMERICAL FREQUENCY. 
In the volume of the Partule of Tahiti, the significance of variations in numer- 
ical abundance was discussed at considerable length in the case of P. hyalina, 
taken as a representative species, and more briefly in the chapters of the other 
species of that island. In the present connection a succinct statement will suffice, 
as the facts are identical in character and their significance is the same in principle. 
Partula radiolata occurs almost everywhere in the island where suitable vege- 
tation exists; it is sometimes abundant and sometimes scarce in regions that are 
ecologically identical as far as the evidence goes. Apparently it is the innate 
congenital vigor of the members of a given association that determines their numer- 
ical abundance. Where no suitable vegetation exists, the species can not live; 
otherwise there seems to be no direct relation between the ecological diversification 
of Guam and variations in the colonial frequency of radiolata, or indeed of any other 
species. 
The general census table given in Chapter II (table 4) shows the absolute and 
relative numbers of radio/ata in the several localities where collections were made. 
Clearly the figures of relative frequency must be regarded as approximations where 
only a few snails were taken, owing to their actual scarcity or to transitory adverse 
conditions at the time of the survey, such as drought, shortness of the time in the 
field, and the like. Then, also, relative abundance is determined in part by the 
actual numbers of the other Partula associates of the same locality. If the time 
devoted to field collection were the same in all places, and if the circumstances of 
moisture, heat, and density of vegetation were identical in all instances, then and 
only then would the figures be entirely trustworthy at their face values; obviously 
it is virtually impossible to fulfill these conditions. 
Yet when all allowances are made, certain interesting facts emerge from the 
returns, and a somewhat definite conclusion is reached. The northern (or north- 
eastern) half of Guam is practically uniform in its ecology, while radiolata is actually 
as well as relatively scarce in the eastern areas and more abundant in the western 
portion. A comparison of the collections from Santa Rosa, Asados, and Lolo with 
those from Dededo and Saucio discloses this point. But in the southern (or south- 
western) half of the island, radiolata is quite as abundant in the inhabited areas to 
the east as it is on the west, although the scarcity of gibba obscures this fact. The 
point is that the paucity of specimens in the eastern half of the northern portion 
can not be referred to climatic conditions, for these are the same in the eastern half 
of the southern territory where the species flourishes. 
Cabras Island bears only radiolata, which, however, is not actually more 
abundant than in places on the mainland, for as a matter of fact the numbers 
secured here in a given time were smaller than elsewhere. The ecological conditions 
of Cabras are not especially favorable for radiolata and adverse for other species 
like gibba; the truth of the matter is that gibba is absent owing to its inability to 
reach the detached area by migration from the larger land-mass, or to its complete 
disappearance from Cabras, assuming that it was present in earlier times. 
