42 VARIATION, DISTRIBUTION, AND EVOLUTION OF THE GENUS PARTULA. 
to the East Central Region, although the two areas are virtually identical in nature. 
The increase of the shells in length in passing from the Apra Region to the West 
Central, and further to the Southwest Region, is even more clearly indicated, although 
the second step involves no change in surroundings. Clearly the innate constitutional 
factors are responsible for the qualities displayed by the regional groups as well as 
for the varying colonial characters of the individual associations. 

Region. 
Northeast 
(Tarague).... 
East Central.... 
Coast Central... 
South Central... 
West Central 
(Salifan)...... 
Southwest...... 
Southeast 
(Inarajan).... 

TABLE 15.—Partula radiolata, Guam. Progressive regional comparisons. 
Length. 
mm. 
— 0.6768 +.0746 
+0.4880 +.0599 
[ —0.0894 -+.0509] 
—1.4322 +.0330 
+0.8184+.1587 
+1.0090 +.1601 
—1.3240+.0872 
+0.9288 +.0796 
Shell. 
Width. 
mm. 
—0.2192 +.0416 
+0.2547 +.0402 
— 0.3593 +.0388 
—0.7623 +.0228 
+0.9435 +.0839 
+0.2025 +.0848 
—0.8643 +.0574 
+0.4806+.0551 

Mean value 
Proportions. 
p. ct. 
+1.0332 +.1940 
[—0.2247 +.2104] 
—1.7818 +.1960 
+0.6098 +.1034 
+2.4669+.3512 
—2.3161+.3457 
[—0.2891 +.2250] 
—0.4715 4.2194 

differences. 
Length. 
Aperture. 
Width. 
Length aper- 
ture + length 
shell, propor- 
Proportions. tions. 
mm. 
—0.2817 +.0420 
+0.2639 +.0365 
—0.3068 +.0344 
—0.8027 +.0205 
+0.7377 +.0790 
+0.3832 +.0804 
—0.6779 +.0451 
+0.3597 +.0415 
mm. 
40.2259 -£.0293 |[ 0.0919 +.1897]| +-0.6709 +..1759 
+0.1176+.0296 
—0.2436+.0281 
—0.5000 +.0154 
+0.6977 +.0478 
{[+0.0340 +.0497] 
—0.5101 +.0345 |[ —0.2300 +.2026]| —0.5497 +.2110 


db. ct. dD. ct 
— 0.7436 +.1900 | —0.2200 +.1430] 
—0.7658+.1791 | —1.4046+.1282 
+1.4023 +.1008 |{ —0.1155 +.0949] 
+1.5768 4.2896 | +1.9642 +.4142 
—2.4590 4.2882 | —1.3235 4.4143 
South Central.. J 

The final figures of the tables (tables 12 to 14) precisely define the whole collec- 
tion of radiolata, combined into a single inclusive series; they give an ultimate 
quantitative description of the species, so far as the actual material in hand is 
concerned. Inevitably the relative weights of the local collections are various, in 
correspondence with the lesser or larger numbers taken in the several representa- 
tive areas; thus the fuller series from Dededo has a much greater effect upon the 
summary figures than the smaller collection from Barrigarda. In general, however, 
the irregularities balance one another in sign, and in addition a large collection at a 
given place usually means a greater local abundance of the species. Undoubtedly 
the statistical results would differ from those here recorded if a uniform representa- 
tion were taken in all places, when, for example, a greater number of short shells from 
Inarajan would reduce the general average, while more specimens from the Umatac 
area would raise the figures. But the biological significance would then be different, 
for the summary statistics would be the general averages of the colonial averages. 
The actual discrepancies are not large when the second method is followed; when 
the 23 colonial averages of shell length are themselves averaged, the result is 16.6215 
mm., as compared with the tabulated summary figure of 16.6218 mm. 
No misconception can arise as to the significance of the data here presented, in 
view of the explicit statements above as to the material from which they are derived 
and which they empirically describe. 
+0.2784 +.0323 |[—0.2901 +.1993]}[+0.0245 +.2048] 
