NO. I ACHATINELLA APEXFULVA DIXON WELCH I93 



ridge. When it is collected on a ridge, the highest locahty is usually 

 a mile or more below the backbone ridge. An exception to this is 

 area i, which is nearer the backbone ridge. Area 115? is only an 

 approximate locality and so cannot be considered (fig. 7, p. 194). The 

 Nuuanu localities of area 10 are on subridges in the valley separated 

 from the backbone ridge by a high precipice. On the other hand, 

 although the forests are fairly continuous in zone III, migration from 

 one ridge to another would be impeded by the gulches which contain 

 water. 



Whether there is any horizontal migration in zone III is not known. 

 The forms in regions I, II, III, and IV, zone III, are certainly 

 distinct from each other but have color patterns that are more similar 

 to each other than are the color patterns in the same regions in zone II. 

 I do not believe that the more extreme horizontal variation in zone II 

 is entirely due to isolation which has come about in the last 50 or 100 

 years, because forms of A. apexfulva seem to colonize limited areas 

 and to break up into different varieties or subspecies at different eleva- 

 tions or at short distances from an adjacent colony. This is especially 

 true of zone II, which has less rainfall than zone III and consequently 

 would tend to discourage migration. Nevertheless, the present-day 

 extreme isolation and inbreeding of certain groups of genes un- 

 doubtedly accentuates the former horizontal valley-to-valley variation. 

 The characteristic oi A. a. apexfulva of breaking up into varieties in 

 limited areas differs from what Crampton found (1916, 1932) in 

 species of Parttila, which apparently migrate into various adjacent 

 areas and do not change color pattern in doing so. 



3. SIZE 



Table 2 (p. 15) gives the statistics of length range, mean length of 

 shell, and number of dextral and sinistral shells found in each locality 

 having a series of five shells or more with a length range of at least 

 three sizes such as 18.5 mm., 19.5 mm., 20.5 mm. The means of the 

 shells are grouped into length classes. For example, everything be- 

 tween 17.76 mm. and 18.25 mm. is considered 18.0 mm. and everything 

 between 18.26 mm. and 18.75 n^"!-' is taken to be 18.5 mm. This was 

 also done in the paper on A. mustelina Mighels (Welch, 1938). In the 

 following discussion on size variation the mean length of the shell 

 will be referred to without needlessly repeating the words "mean 

 length of shell." 



The various localities or areas containing forms of A. apexfulva 

 are grouped for convenience into 20 ridge complexes or ridge groups 

 in order to demonstrate size variation. A ridge complex such as 

 3 (fig. 8, p. 195) may be a single ridge which fans out near the 



