PARTULA GIBBA, GUAM. 95 
The central problem in the field of inheritance is concerned with the free inter- 
breeding of the few or many color-classes which make up one or another representa- 
tive assemblage. If there is evidence of assortative mating on the basis of like- 
nesses in coloration, then there would be some reason to regard the several groups as 
incipient varieties in the taxonomic sense. But in the previous section on the 
measurable characters of the color-classes it was argued that the absence of any 
consistent relation between peculiar statistical features and a distinctive mode of 
coloration indicated free intercrossing of the varied members of a colony. This 
conclusion is supported by another consideration, not adduced in the discussion of 
radiolata but equally applicable, namely, the fertile condition of snails belonging 
to a color-group which is sparsely represented in a flourishing colony. Taking the 
Aniguac series as an illustration, it appears from the census (table 29) that out of 
222 adults, there were 3 castanea and 4 vespera; I of the castanea contained no 
eggs or young, but the other 2 castanea and the 4 vespera were gravid. The chances 
that the last-named had mated with others of their own respective kinds are so 
small that their fertile conditions may be taken as proof that they had mated with 
mitella or bicolor individuals. In the same way the Dededo association provides 
evidence that phea snails mate outside of their class. In the Dungcas association, 
the marginata number 6 and the mitella amount to 29; as all of the former are gravid, 
it is virtually certain that at least some of them had been fertilized by mitella 
mates. 
When the colors of the embryonic snails are compared with the colors of 
their parents, class by class, the results provide additional corroboration of the 
contention that the several color-classes freely interbreed. While the demon- 
stration is not absolute, yet the facts certainly do not favor the alternative view 
that strict selective mating takes place. By way of approach, we may recall that 
the early adolescents of unicolor and bicolor are indistinguishable because the 
ground-colors agree and the distinctive bands of the latter develop only afterwards; 
they are both whitish or very faintly corneous. The immature shells of mitella 
are similar to the foregoing in general tones and display only the faintest yellowish 
deepening on the apical whorls; the young of mitella-rubra are the same. In phea 
and castanea, however, the partly-grown individuals are deeper corneous, faint 
brown in tone, and develop their ruddy or purplish tinges only later in life. The 
vespera adolescents are likewise light corneous. Finally, the immature marginata 
are brown-corneous rather than yellowish, and they become lighter instead of darker 
with age. 
If, now, the embryonic young with well-advanced shells are examined, they 
present three distinguishable modes of coloration. In the first, type a (fig. 57, 
plate 13), the shell is whitish or a very faint whitish-yellow, and is uniform in color. 
The second mode, type 0 (fig. 58, plate 13), is more yellowish and its extreme apical 
coils are deeper in tone. In the third, type c (fig. 59, plate 13), the ground-color is 
faint corneous brown, with the apical coils still deeper in color. The numbers of 
each kind of embryonic young taken from the adults of the several color-classes 
are given in a comprehensive summary (table 48); but this table can not be taken 
