PARTULA GIBBA, GUAM. 67 
TABLE 28.—Partula gibba, Guam. Progressive regional comparisons. 
MEAN VALUE, DIFFERENCES. 

Region. Shell. Aperture. Length aper- 
ture + length 
shell, propor- 
Length. Width. Proportions. Length. Width. Proportions. tions. 
Northeast...... mm. mm, dp. ct. mm, mm. p. ct. p. ct. 
— .2303 +.0297 —.1799+.0231 | [—.0639+.0924]| [—.0405+.0168]} —.1431+4.0139| —1.0845+4.0840} +.5144+.0618 
East Central.... 
+.1262+.0275 | [+.0638+.0222]} [—.1861+.0854]} +.0808+.0157 | [+.0019+.0131]} —.5752+.0755 | [+.0670+.0554] 
South Central... 
—.1788 +.0303 —.0790+.0225 +.3100 +.0886 —.0687+.0172 | [+.0291+.0130]| +.7852+.0884 +.2846+.0660 
Coast Central... 
East Central.... 
[—.0526+.0302]] [—.0152+.0243]} [+.1239+.0906]| [—.0127+.0188]} [+.0310-+.0141]] [+.2100+.0883]} +.3516+.0694 
Coast Central... 
[—.1710+.0689]| +.4097+.0709 | +2.8421+.2657 | [+.0851+.0456]} +.1693+.0428 | +1.14314.2532| +.9082+.1895 
+.6237+.0817 | [—.0819+.0762]| —2.7660+.2864 +.1905+.0517 | [+.0454+.0471]| —1.0091+.2735 — 8364 +.2062 
West Central. . 
+.2935+.0727 | [—.0248 +.0662]]| —1.4674+.2989 | [+.0022+.0448]| [—.0391+.0408]| [—.2174+.2806]} —1.2391+.1956 
Southwest...... 

value of the summary figures need arise in view of the present precise statement as 
to what they mean. 
If every association should have exactly the same weight as any other, regard- 
less of the numbers in each, the average of the colonial averages is a different figure 
with a different biological significance. By way of concrete illustration, the char- 
acter of shell-length comes out variously according to the method of its determina- 
tion. When the shells are all combined into a single series, the value is 16.9968 mm. 
as tabulated; the regional averages, when accorded equal weight, give 17.0634 mm. 
as their mean; and if the colonial averages are treated as though they had the same 
significance, their general average is 17.0921 mm. In my own opinion, the first 
determination is nearer to the real value of this character in the species, because 
in general the weights to be assigned to the several component groups are more 
fairly estimated by the actual numbers collected than by any other method, and 
because the number of colonial groups is large enough for errors in one direction 
to be offset to some degree by errors of the opposite sign. 
The figures for the characters of the combined regional populations disclose 
some interesting relations. The areas with the largest shells are the far-separated 
Northeast and Southwest Regions, which display clearly contrasted ecological char- 
acters of soil and vegetation. Furthermore, the lowest regional averages in shell 
length are those of the Coast Centra] and East Central territories, with very different 
barometric levels. The Northeast and East Central Regions adjoin and their 
environmental characters are essentially alike, and yet in the absolute measures 
of shell length their gzbba populations are distinctly different. If the analysis 
is made on the basis of other characters of the shells, the same result is obtained, 
namely, that the ecological qualities of the surroundings as such are not responsible 
for the characters of size and shape of the shells borne by the snails which inhabit 

