214 VARIATION, DISTRIBUTION, AND EVOLUTION OF THE GENUS PARTULA. 
reason or another. Confining our attention to the five available cases, clearly each 
must be taken by itself, for the relative numbers of DD, DR, and RR adults in their 
populations must be so diverse as to preclude their union into one series. 
Tuauru VALLEY. 
The fundamental facts are that the plain adults amount to 73.2 per cent of the 
population, and the banded adults to 26.7 per cent, taken as 27 per cent; one of these 
divisions must be RR, and the other must include DD together with DR. As it 
happens, the banded pattern seems to be recessive, as in the case of Lang’s breeding 
experiments with Helix; so the analysis will first be made on that assumption for 
Partula, although the results that follow from the contrary supposition will be 
worked out, to be contrasted with the first-found figures. 
When the parents bearing young are classified according to their contents, the 
numbers are the following: 
(1) Plain adults: plain young, 175; both kinds, 10; banded young, 20=205 
(2) Banded adults: plain young, 14; both kinds, 2; banded young, 30= 46 
We begin with (2) and, assuming that the parents are themselves RR in composition, 
their mates would have been RR in 27 per cent of the 46 cases, or in 12 instances; 
this figure differs from the empirical number by 18, which is to be transferred to the 
middle class, because that excess represents the RR individuals mated with DR 
which have only a single offspring and hence can not display the two expected kinds 
of plain and banded young. Making a correction of the figure for the first group, 
to the same degree of #, we obtain: 
(3) Banded adults: plain young, 6; both kinds, 28; banded young, 12=46 
Hence the three kinds of mates of the 46 RR parents, and the three kinds of 
snails in the whole population, would be DD 6, DR 28; RR 12. 
Testing this in the case of the plain parents of (1), making a correction of the 
first figure to the empirically defined extent of %, the actual numbers are: plain 
young 70, both kinds 135, as compared with the expected numbers: plain young 
58, both kinds 147. The difference is 12 out of 205, or 5.8 per cent. 
When the contrary assumption is made, namely, that the plain coloration is 
recessive to banding, the proportions of DD:DR:RR come out as 6:144:55. 
Testing these relations in the independent class of banded snails, assumed to be 
DD+DR, the empirical figures (plain young 3, both kinds 43) depart from the 
expectation of 9.5:36 to the extent of 6.5 out of a series of 46, or 14.1 per cent. 
The former assumption seems to be more justified by the Partula data themselves, 
as well as by the analogous results of Lang for Helix. : 
FaRAPA VALLEY. 
Here the red-banded and ‘brown-banded snails are combined into one group, 
to be contrasted with the unbanded class of whatever ground-color, for the problem 
deals only with bands and their absence. The figures for the general population 
