PARTULA GIBBA, GUAM. 73 
extension of the characteristic color of the spire, the two red areas are not directly 
continuous, for the younger portion of the last whorl as viewed in oral aspect shows 
only the general corneous-yellow ground-color (fig. 48, plate 12), The inner sur- 
face in this order is deeper in color than in the ordinary mitella in almost all of the 
examples; but the lip is usually white, and only in a small minority is it stained 
with yellow or yellow-brown. 
The shells of the Lonfit colony are quite unique within this group, inasmuch as 
the younger whorls are lighter in color on the sides toward the apex (figs. 49, 50, 
plate 12), and hence they simulate the bicolor shells in pattern, although their 
tints are entirely different. 
Returning to Tarague, the representatives of this order are much more darkly 
colored (figs. 51, 52, plate 12). Still more extreme shells were found at Dededo 
(figs. 53 to 56, plate 12), in which the spire grades from red-brown to purple-brown 
and the body-whorl is almost entirely suffused with deep colors that usually cor- 
respond with those of the upper coils. Yet even in such specimens there are 
always some areas which lack the overtints. The illustrations show a degree of 
correspondence between the inner and outer surfaces as regards the depth of color. 
The outer border of the lip is more frequently stained with brown or orange- 
brown in the deeply colored shells like those of Dededo. In mitella-rubra, as in 
the other orders and classes, the white border below the suture varies considerably 
in width and distinctness. 
The adolescent individuals of this order are exactly like those of mitella. 
This is evident from the fact that no differences can be observed among the mem- 
bers of associations which comprise numerous mitella and mitella-rubra adults, 
as at Dededo and Tarague. The case of Asados is still more demonstrative, 
because no adult mitella were taken there, but among the immature individuals are 
many which are indistinguishable from the adolescents found at the Ylig, where 
all of the full-grown specimens were mitella. Obviously the two groups in ques- 
tion are intimately related, and are separable only on the basis of differences which 
arise with the approach to maturity. 
The order is represented in the collections from 15 localities, and it is next 
in total abundance to the related mitella group. Only in the Asados association 
does it occur without any of its near relatives. 
phea.—Next to the foregoing in abundance is the brown-shelled class to 
which the distinctive name of phea is given; it comprises two color-orders, rubra 
and purpurea, whose differences are indicated by their names. 
The Lolo examples of phza-rubra may be taken as the representatives of 
their order, because the absence of phza-purpurea and of mitella-rubra causes them 
to stand out collectively in sharp contrast with all of their associates. Typically 
the shell is a rich brown in color (figs. 1 to 3, plate 13), with a certain amount of red- 
dish admixture which produces a ruddy effect when it is abundant (figs. 4, 5, plate 
13); the red tints are more pronounced when decortication occurs (figs. 6 to 8, 
plate 13). Another distinctive feature of this and of the purpurea order of the phea 
class is the attenuation of the color on the apical whorls, shown by all specimens 
