76 VARIATION, DISTRIBUTION, AND EVOLUTION OF THE GENUS PARTULA. 
The adolescents, such as the one from Pago north (fig. 51, plate 13), are very deli- 
cately colored with the same attenuation of the colors upon the upper whorls; the 
specimen in question is entirely different from the mitella and castanea young of 
its locality. 
Apparently the Barrigarda area is the headquarters of this class, for in that 
locality the vespera-rosea individuals are absolutely abundant and they also far 
outnumber the other classes taken together. It is not strange that the same type 
should be well represented at Lolo, but the absence of vespera from the Northeast 
Region is worthy of remark. Elsewhere this class closely coincides with castanea 
in its distribution; there are only slight differences between these two classes in 
the early stages, and it is with the approach to maturity that the shells of the dis- 
tinctive adult types diverge more widely. 
marginata.—The last class is remarkable on account of its exceptional features 
of coloration and also because it was found in a single sharply restricted area of 
bush only a half mile in length at Dungcas. The shells display a corneous color 
as a basis upon which brownish streaks run transversely (figs. 52 to 56, plate 13); this 
is the only color-class in which such strigations occur. The whorls of the spire are 
margined along their lower borders with a peculiar rose-brown, and the same color 
more intensely tinges the outer half of the flaring lip; the distinctive name for the 
class is chosen on account of these features. In a decorticated individual (fig. 55, 
plate 13) the streaks and the colored band are less pronounced. The immature 
specimens (fig. 56, plate 13) prove that the differential characteristics of this class 
are developed before maturity. 
Undoubtedly this is the kind described by Pilsbry on page 316 of the Manual 
of Conchology (volume 20) and illustrated in figure 18, plate 39, of that publica- 
tion; the shell in question is an old one in which the red component has faded so as 
to make the brown more evident. Except that it is a member of the same species, 
gibba, it is in no way related to the bicolored shells which served Pease as the types 
of his variety, for the immature and the embryonic conditions are entirely different. 
The identification stated is based upon the characteristic strigations which have 
not been found in any other variety of gibba. 
This class stands by itself and its genetic derivation is problematical. If the 
immature coloration gives the clue to its origin it would be placed near vespera-rosea, 
whose adolescents are the most similar. Its close restriction to a small area of 
thicket east of Agafia and only a few score yards within the strand are the notable 
features of its geographical location. 
A brief general review of the foregoing detailed account of the color-classes 
will be profitable. The first salient fact is that mitella is the dominant form at the 
present time, for it is the most widespread and the most abundant; furthermore, 
this type figures prominently in the literature of the past century. However, 
without additional reasons it is not justifiable to regard the mitella coloration as 
the original type from which all the others have arisen by mutation; there is nothing 
in the observed facts to indicate which of the existing modes is ancestral with 
