ARTICLES 443 



where shown, to reject the assumption that the lowered tem- 

 perature of the Glacial ocean sufficed to extinguish reef growth 

 and permit the abrasion of Pre-glacial islands ; and the mobility 

 that is found to characterise the ocean floor beneath all archi- 

 pelagoes of decipherable history leads me to question very 

 seriously the long-continued stability of the ocean floor in the 

 regions of atolls. 



The most that can, to my belief, be said for the Glacial- 

 control theory is that the fluctuations of ocean level, which 

 it properly postulates, have combined with the elevations and 

 subsidences of reef foundations in such a way as sometimes 

 to favour reef submergence, sometimes reef upgrowth, some- 

 times reef outgrowth, sometimes reef emergence ; and more 

 particularly that the Glacial rise of ocean level has co-operated 

 with widely prevailing movements of subsidence in such a way 

 as to cause an alnlost universal submergence of reef foundations 

 in recent geological time, as a result of which nearly all present- 

 day reefs are of narrow or immature form, and very few are so 

 far broadened as to deserve the name of mature reef plains. 

 It is possible, however, that the theory has fuller application 

 along the border of the coral zone. 



The upshot of all this is that, wherever opportunity is given 

 to make a critical test of the conditions under which coral reefs 

 have been formed, a local subsidence of their foundation, as 

 testified by the embayment of the adjoining coasts and the 

 unconformity of the reef limestones on the underlying rocks, 

 appears to have been a determining condition of their con- 

 struction. It is for this reason that I have come to regard 

 Darwin's theory of coral reefs as much more competent than 

 any other. 



Molengraaff's Theory of Subsiding Volcanic Islands. — Darwin 

 was led by his theory of the subsidence of reef foundations to 

 assume that subsidence has recently prevailed over large areas 

 of the Pacific and Indian Oceans. It has been urged that, 

 if such wide-spread recent subsidence had prevailed, the conti- 

 nental coasts should show features of emergence, whereas they 

 frequently show features of submergence. A way out of this 

 difficulty has lately been opened by Molengraaff, who suggests 

 that the subsidence, in response to which reef upgrowth has 

 taken place, need not be a wide-spread movement of the ocean 

 floor, but a local sinking of volcanic islands by reason of their 



