108 VARIATION, DISTRIBUTION, AND EVOLUTION OF THE GENUS PARTULA. 
blance to P. nodosa as indicative of a genetic relationship. The latter species 
became very narrowly restricted in its habitat, probably to the single valley of 
Punaruu, but most recently it has revived in strength and increased in numbers, so 
as to occupy the seven valleys of its present range. 
II. Partula nodosa 1s not uniform throughout its range, but varies from valley to 
valley; the local types group themselves so as to form four well-marked varieties with 
definite geographical limits. The individual peculiarities of the seven colonies have 
been described in the foregoing pages; variation has been shown (1) as regards the 
relative numbers of the species in the representative valley collections; (2) in the 
proportionate numbers of dextral and sinistral individuals; (3) in the varying fre- 
quency of the principal color-types, and (4) in the measurements and proportions 
of the shells as defined by statistical determinations. It is unnecessary at this 
point to review the details under the first three heads, but the fourth category needs 
to be re-examined. 
The statistical constants of the seven morphological characters, for all valleys, 
are brought together in table 53. To facilitate the direct comparison of the valley 
series, the differences noted in passing from Taapuna progressively to the other 
valleys are given for the mean values in table 54; the significance of any difference 
may be estimated by its proportion to its probable error. Finally, the graphic 
representation of table 55 makes it possible to appreciate at a glance the degrees of 
change in passing from one valley to the next in both the average values and in 
the ranges of variation of the several characters. 
We may take first the qualities of the shell as a whole. The dextral Taapuna 
shells are certainly longer and wider than those of Punaruu; as the excess in length 
is more considerable than the difference in width, the proportions are much more 
slender than in Punaruu. ‘The sporadic shells of Taapuna may be disregarded in 
this comparison. Passing now to Maruapoo, we may restrict our attention to the 
28 dextral shells, or we may base our comparison on the conditions of the preva- 
lent sinistral type, but the two form-classes are not certainly different from one 
another except in shell proportions. Specifically, the Maruapoo shells are about 
as long as those of Punaruu, but they are certainly narrower; hence they are more 
slender to a degree that is statistically significant. The little specimens from Atehi 
are distinctively reduced in width, but far more abbreviated in length; henee they 
are proportionately stout—by far the stoutest of all the valley colonies. Papehue 
has longer shells that are broader also, but the reduction in proportions which is 
correlated in part with the increase in length is not commensurate with the degree 
of change in the latter character; this is not statistically different from the figure for 
the Atehi shells. The Aoua colony is shorter, narrower, and more slender, proving 
that reduction in length does not always bring about a stouter shell. Finally, the 
short series from Orofere comprises greatly elongated and widened individuals of 
slender proportions, all three differences from the Aoua colony being statistically 
significant. 
