192 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. IO3 



similar to those of area 50 above area 43 (fig. 4, p. 53). Also, the 

 shells of area 45 on the Waimano-Manana Ridge more closely 

 resemble the shells of area 49 above area 50, although area 50 is 

 opposite area 45. In like manner the color patterns found in any 

 zone in region III will definitely dififer from those found in region 

 IV. The same holds true for regions I and 11. 



In between the regions of major dififerentiation intermediate or 

 border areas occur, such as regions la, Ila, Ilia, Illb, and areas 40, 

 49, 50. Region la contains forms which have a color pattern inter- 

 mediate between regions I and II, zone II, but closer to the color 

 pattern of the shells to the southeast, or region I. Region Ila and 

 Illb and areas 49 and 50 have shells the color pattern of which is 

 very similar to that of the shells to the southeast and entirely different 

 from those to the northwest in the same zone. Region Ilia and area 

 40, zone II, contain gray color patterns which are close to the color 

 patterns of region III, zone II, and very dififerent from those to the 

 southeast. 



As a general rule more horizontal variation occurs in zone II than 

 in zone III. This can be explained by the fact that usually the lower 

 reaches of a ridge fans out into several long ridges. Consequently, 

 there are more ridges and valleys at low elevations or in a lowland 

 zone, which increases the possible number of isolated colonies in 

 zones I and II as compared with zone III. 



Today the isolation of the lowland colonies is more complete than 

 it was originally because of the dying out of the forest and the spread 

 of the staghorn fern {Gleichenia linearis). In this manner groups 

 of trees are isolated from other groups of trees so that snails cannot 

 possibly migrate. This is more especially true of lowland forms in 

 zone II and the lower portions of zone III, In many places trees 

 exist only in the bottom of a gulch the sides of which are covered with 

 staghorn fern, while the tops of the intervening ridges are covered 

 with grass. Again the trees on the side or summit of a ridge may also 

 be surrounded by staghorn fern. 



In zone III, which includes the upper part of Hosaka's (1937) 

 Ohia Zone and all of the Cloud Zone (fig. i, p. 3), the forest is more 

 continuous, although in many regions large sections are covered with 

 staghorn fern. One would expect more horizontal migration here 

 because of the relatively larger amount of flora, greater rainfall, and 

 the connection of the backbone ridge with the main ridges of the 

 Koolau Range, which would afford a road for migration. On the 

 other hand the peaks of the backbone ridge might also serve as 

 effectual barriers. But A. apexfulva does not occur at the backbone 



