GEOMORPHY 23 



of blocks by gravity, which form a talus at the bases of the cliffs. Al- 

 though frost is absent, so easily are the fragments separated because of 

 the character of the rocks that the excavation is as effective as in colder 

 climates on the more durable ledges. In the lower reaches the streams 

 take winding courses, and thus act laterally against the sides, widening 

 the bases. 



The Koolau area is the easiest on Oahu to understand. From the de- 

 tails already presented it is seen to be elliptical, nearly 40 miles long, and 

 deeply eroded along its seaward face, with many amphitheaters, outliers, 

 and especially the long cliff opposite Kaneohe bay. There has been great 

 excavation along the western side of Koolaupoko, but comparatively 

 little on the interior side of Koolauloa. Judging from incomplete ob- 

 servations on the rainfall for the past five years, the average has been 144 

 inches two miles below the Pali (Luakaha), and about 20 inches near the 

 wharves of Honolulu ; but the rainfall is confessedly greater at the crest 

 of the ridge, probabl}^ 200 inches, and it diminishes gradually all the 

 way to the harbor. The fall along the eastern shoreline exceeds 30 

 inches, increasing to the summit ; hence it appears the water should be 

 most abundant along the crest of the range, but greater on the eastern 

 than the western slope, and whatever the fall may be on the Honolulu 

 side it came from the northeast. The erosion has been the greatest on 

 the northeastern side, as seen in the Pali, the outliers, sometimes 2,000 

 feet high, the ridges running northeasterly, and the amphitheaters. It 

 reached probably to the central axial line of elevation opposite Kaneohe 

 bay. The cliff can not very well have been eroded by the sea, since 

 there are irregular ridges and chains of hills at intervals of 2 or 3 miles 

 stretching out perpendicularly from the wall and ending in promontories. 

 Marine action would have removed these projections. The erosion 

 seems to have been most intense at the road crossing the Pali, since 

 there is a gap worn down to 1,207 feet from over 3,000 feet on either 

 side, and there are two other gaps to the north not far away. Some have 

 explained the presence of the Pali gap and the horseshoe form of the 

 land from Mokapu point to Konahuanui and thence along the main 

 range to the northeast branch, ending at Kualoa point, by assuming a 

 break or fault at the Pali gap or the existence of an enormous crater in 

 the part of the circular ridge just delineated. The best argument in 

 reply to both these volcanic theories is that the topography is in better 

 agreement with what is known elsewhere to be the results of subaerial 

 erosion. If there were one transverse fault, there must have been three, 

 quite close together, for the first cataclysmic theory ; and the theory of 

 the large crater assumes that certain cinder cones and scoria were inti- 



IV— Bull. Geol. Soc. Am., Vol. 11, 1899 



