202 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. IO3 



all ill zone II; the otlier, ridge complex lo, is represented by only 

 one locality in zone III, which contains only five specimens. The 

 remaining five ridge complexes show no marked variation in the 

 majority of localities between zones II and III. Ridge complex i 

 contains only i8+ shells, regardless of altitude. If the Gulick shells 

 from this general region, probably zone I and lower zone II, are 

 compared with present-day forms, size variation can be noted. The 

 lower Gulick forms are i6+ or 17+ shells. However, if all the 

 Gulick shells in institutions outside of the Bishop Museum collection 

 were measured and considered together, the mean length might be 

 greater, probably 18+ instead of 17 + . The highest locality in ridge 

 complex 7 (area 31, table 2, p. 15) has 18+ shells, while the lowest 

 locality (area 27A?) has shells with a mean length of 18.77 nim., 

 or 19+ shells. In between these locaHties, 19+ and 20+ shells occur. 



The most remarkable exceptions to altitudinal size increase are 

 ridge complexes 11 and 12, which have small 17+ shells occupying 

 the majority of localities in both zones II and III. In ridge complex 

 II, zone III, 16+ shells also occur. Localities of 18+ shells are 

 found here and there in both zones II and III in both these ridge 

 complexes. 



As with A. mustelina (Welch, 1938), no clear-cut correlation can 

 be drawn between increase of moisture and size variation. The 

 highest localities in ridge complex i near the backbone ridge range 

 in elevation from 1,650-2,000 feet, while those in ridge complex 2 

 range from 1,150-1,500 feet. However, the rainfall in upper ridge 

 complex 2 is undoubtedly greater than in upper ridge complex i, for 

 although there are no rain-gage stations in this region, south of 

 Nuuanu and Manoa the backbone ridge undoubtedly receives less 

 rainfall (compare Luakaha (upper) with Makapuu, table i, p. 10), so 

 that in this case a correlation can be drawn between increase of 

 size and increase of moisture. In a similar manner the locaHties in 

 area 8, which is a much dryer section of Nuuanu Valley than area 10, 

 have smaller shells than the upper locality even though the elevation 

 of both areas is about the same (compare Luakaha upper and 

 Luakaha lower, table i, p. 10). While a correlation can be drawn 

 between an increase of size with an increase in moisture conditions 

 in ridge complexes i and 2, it is not possible on the same basis to 

 explain the small shells of ridge complexes 11 and 12, zone III, or 

 the larger 19+ shells of ridge complexes 7, 16, and 17, zone II. 



Even though exceptions occur at random, 65 percent of the ridge 

 complexes show a size change between zone II and III. A correla- 



