J. D. Dana on Denudation in the Pacific. 59 
As the gorge increases in steepness, the excavations above 
deepen rapidly,—the more rapid descent more than compensa- 
ting, it may be, for any difference in the amount of water. 
Moreover, as the rains are generally most frequent at the very 
summits, the rills in this part are kept in almost constant action 
through the year, while a few miles nearer the sea they are 
often dried up or absorbed among the cavernous rocks. The 
denudation is consequently at all times great about the higher 
parts of the gorge, (especially after the slopes have become steep 
by previous degradation ;) thus finally a steep precipice forms the 
head of the valley. 
t. The waters descending the ridges either side of the valley 
or gorge, are also removing these barriers between adjacent val- 
leys, and are producing as a first effect. a thinning of the ridge at 
summit to a mere edge; and as a second, its partial or entire re- 
moval, so that the two valleys may at last be separated only by a 
low wall, or even terminate in a common head,—a wide amphi- 
theatre enclosed by the lofty mountains. In one case, the ridge 
between the two valleys, which towards the shores of the islaud 
has rather a broad back, high up in the region of mists and fre- 
quent rains becomes a narrow wall, and thus connects with the 
central summit. Jn the second, the ridge finally terminates ab- 
ruptly, and a deep valley separates it from the main mountain. 
The following sketch may assist the mind in conceiving of the 
action upon the Pacific mountains. It represents one o - 
leys of Tahiti from the centre to the shore, excepting its irregu- 
larities of direction and descent, and the uneven character of its 
walls, arising from lateral valleys and minor denudations. The 
height of Tahiti is about eight thousand feet; its radius ¢s is 
ten geographical miles. ‘The head of the valley at a@ is three 
thousand feet below the summit peak p. The descent along the 
ascertained, ) it would still give four hundred feet to the mile. 
This subject is beautifully illustrated in some of the tufa cones 
of Odhu, where, on a smaller scale, we have the same kind o 
sorge and valley; and in this case, there is no doubt that de- 
Hudation was the cause by which they were preduced. ‘The 
Valleys have the direction of the slopes, and are similar in form 
and winding character to those of the mountains. The inter- 
Vening ridges are also similar. Many of them become very 
@ 
