278 VARIATION, DISTRIBUTION, AND EVOLUTION OF THE GENUS PARTULA. 
Testing the correspondence of the expected numbers with the empirical obser- 
vations for the class of red adults of (1), it appears that the DD members amount to 
30, whose young would be red in color whatever their mates. The remaining 52 
DR individuals would have DD mates in 18 instances, and would also bear red 
young only, while the remaining 42 would theoretically produce both kinds of young. 
The expected figures would thus be: 
(4) Red adults: red young, 40; red and yellow young, 42 =82 
Empirically the numbers are, respectively, 51: 31 =82; but when the second and third 
classes of (1) are combined, and the figure for the first class is reduced by the # to 
be transferred to the second group, the figures become 36:46, which differ from 
expectation by 4 out of 82 cases, or 4.8 per cent. 
Proceeding on the alternative assumption, namely, that the yellow is the domi- 
nant color, we classify gravid adults with young according to their contents, each 
parent taken once, with the following results: 
(5) Yellow adults: yellow young only, 60; yellow and red young, 5; red young only, 25 =9o0 
(6) Red adults: yellow young only, 19; yellow and red young, 12; red young only, 51 =82 
Beginning with the red series (6), we would expect their mates to have been 
RR in 53 per cent of the cases, or 43.46, with red young only as their products. 
Empirically, however, there are 51 such gravid parents, and hence the surplus over 
43.5, or 7.5, are individuals mated with DR yellow snails, where yellow young do not 
appear among the few young produced. Therefore, 7.5 is to be subtracted from 
the third class of (6) and added to the middle one. A further correction is necessary 
on account of the presence in the first class of individuals, RR in composition, that 
have mated with DR, but whose potential red young do not appear; as a measure 
for the amount of reduction we have the relation of 7.5 out of 51 in the analysis of 
the third class, or a proportion of 14.7 per cent, equal to about 3 individuals. 
Making this adjustment, (6) becomes: 
(7) Red adults: yellow young, 16; yellow and red young, 23; red young, 43 =82 
These are the proportionate numbers of RR adults which have mated with DD, 
DR, and RR, respectively, and at the same time the relative numbers of the three 
gametic classes. 
Now we may proceed to analyze the yellow adults of (5) to test the probability 
of the results obtained from the contrasted class. The second and third groups 
are to be combined because yellow as well as red young are to be expected if yellow 
is a dominant character. Out of the total of 90 cases, DD yellows amount to 37, 
if the proportions of (7) are correct; whether mated with DD, DR and RR, their 
young would be yellow only. Of the remaining 53 which are DR, we would have 
DD mates and yellow young only in to cases, while in 43 cases the mates would be 
DR and RR, with both kinds of young among the offspring. Combining these 
figures, expectations would be: 
(8) Yellow adults: yellow young, 47; yellow and red young, 43 =90 
