GEOLOGY: W. M. DAVIS 
279 
where the rocks are hard, while no cliffs are developed on the other side of the 
island, even though the cycle of erosion on the weaker rocks that there prevail 
advanced in the same measure of absolute time to a late stage of development, 
when reef extinction and retrogressive abrasion are expectable occurrences. 
The prevention of cliff development on the southwestern coast is favored by 
assuming that the submergence which lowered block 5 to block 6 was caused 
by progressive subsidence. In any case, the general submergence by which 
block 5 is changed to block 6 is easily demonstrated. It is therefore in view 
of some such geologically modern sequence of moderate deformation and pro- 
longed erosion as is here sketched, uncertain and shadowy in its earlier stages, 
better certified in its later stages, that the development of the ore-bearing 
laterites must be explained. 
The enrichment of the present ore deposits could not have begun on the 
serpentine areas in the immature stages of the earlier cycle of erosion in which 
block 2 was worn down to block 3, for at that time the valley-side slopes, 
profiles 1, 2, figure 2, were steep enough, just as they are in the immature 
stage, profiles 1', 2' ', of the present cycle in the serpentine areas, to allow the 
FIG. 2. RELATION OF ORE-BEARING LATE RITE TO TWO CYCLES OF EROSION 
removal of disintegrated rock about as fast as it was weathered. Even during 
the mature stage, profile 3, of the earlier cycle, removal rather than accumula- 
tion must have prevailed; but as the later stages of the cycle were reached, 
soil removal from the subdued hills, profiles 4, 5, between the wide-open valleys 
must have been much slackened; the thickness of the disintegrated materials 
there occurring and with it the surface enrichment of the metallic ores by down- 
ward concentration must have thenceforward increased as the subdued hills 
were worn down to the gentler and gentler gradients of old age, as in profile 6; 
and the area on which concentration was important must have been on the low 
swells between the broad valleys. If these inferences are correct, it follows 
that the nickel and cobalt content of the greater part, MNO, of the primeval 
rock mass has been carried away and deposited on the adjoining sea floor, 
and that the deposits now worked contain chiefly the concentrated savings from 
only a quarter or a sixth of the primeval total, shaded in the upper half of 
figure 2. This explanation traverses Glasser's supposition' 2 that the serpentine 
masses have not been much eroded; for in view of their form alone, apart from 
the evidence of erosion given by ore concentration, that supposition seems 
untenable. 
