PARTULA RADIOLATA. 43 
COLOR CLASSES AND THEIR COLONIAL DISTRIBUTION. 
The shells of Partula radiolata vary considerably in coloration, although as 
compared with those of P. gibba they range within relatively narrower limits. 
Taking the species as a whole, no less than 6 color-classes are distinguished in the 
present study, of which 3 have been specifically noted in the earlier literature, while 
the other 3 are herein described for the first time. The position is taken that the 
distinctive types are not true varieties in the taxonomic sense, mainly because 
they are intermingled in the series from single localities and from the same regions, 
because their differentiation is not 
always carried very far, and because 
this differentiation is not an accom- 
paniment of that structural diversi- 
fication which was the subject of 
the foregoing section. If in the 
course of evolution the color differ- 
ences should be invariably corre- 
lated with morphological distinc- 
\ tions, then one or another might 
emerge from the highly mixed as- 
semblages of to-day to become a 
real variety, or even a recognizable 


Northeast 
1.2.3. 




East . 
/ Central 
1.2.3.5.6. 



South 
Central 
species. 
en But whatever interpretation 
1. Pallida may be preferred, the immediate 
2. Flavea ‘ : Soe 
3. Fulva task is to describe the material in 
4° Strigata the chosen terms of color-classifica- 
5. Strigata-helix : : 
Reseed tion, and to record the geographical 
distribution of the several dis- 
tinguishable types. It has been 
Fic. 2.—Partula radiolata, Guam, regional distribution of the stated earlier that no representa- 
color-classes. tive collection is exactly similar in 
pallida its color composition to that from 
| , another locality. There is variation 
flavea bicestata ‘ 
| | in the number of the color-types 
fulve, ———- strigata —______ strigata-hdlix that are present, ranging from 3 to 
Fic. 3.—Partula radiolata, Guam, probable interrelations of 6 where the material is ample and 
color-classes. . ° ; 
reliable, while the proportionate 
numbers of the classes vary from colony to colony. The larger significance of such 
facts is that, as in the case of the structural characters, we must look to the hereditary 
and not to the environmental factors for the causes of the phenomena as observed. 
The complete census of the color-types hereinafter described (table 16) gives 
the actual numbers in the collections from the several localities where adult 
radiolata were found. The graphic scheme of text-figure 2 records their geo- 
graphical distribution in another way. What are regarded as their probable 
genetic inter-relationships are shown diagrammatically in text-figure 3. 

