482 BOUNDARIES OF THE SEA 



phorite, glauconite, or pyrites and other iron compounds. Nor- 

 mally these are formed in situ or at least in the environment where 

 they become buried. The behavior of silica is still little understood. 

 It migrates in the recent sediment and part of the flint nodules of 

 ancient limestones were evidently formed at an early stage. The 

 precipitation of lime as calcite or aragonite in the shape of oolites 

 or mud is a common phenomenon, but still insufficiently known. 

 The change from calcium carbonate to dolomite has involved 

 mighty building stones of the earth's crust. No wonder geologists 

 are keenly interested in learning under what conditions these 

 chemical processes take place. Precipitation appears to be limited 

 to warm waters, but curiously enough the same can be said of 

 solution in surface waters. This solution of limestone, especially 

 of reef rock, has played a part in the coral reef problem. We still 

 only suspect what the chemical and biological explanation is. \\'e 

 are still less certain as to the rate at which this solution takes 

 place. Perhaps it can only model the surface, but the possibility 

 exists that it acts swiftly and that all coral reefs were beheaded 

 during the low levels of the Ice Age. 



I have had occasion to point out that if this proves to be true 

 a much more satisfactory explanation can be given of atolls and 

 barriers. Darwin's upgrowth during subsidence evidently built 

 the main part of the structure. But the present configuration 

 probably results dominantly from postglacial growth on plat- 

 forms produced by low-level solution in rain water and the surface 

 film of tropical waters. For Darwin's theory fails to explain the 

 usual combination of barrier with fringing reefs. Neither can it 

 account for the almost total absence of passages through the reefs 

 at depths greater than 100 m. These and other difficulties are 

 accounted for by low-le^'el beheading of the preglacial limestone 

 mass. 



This hypothesis is a modified form of Daly's glacial control 

 that I call "decapitation." The advantages over Daly's views are 

 that the very large atolls need not represent excessively wide 

 erosion platforms and that coral growth need not have been sup- 

 pressed to allow wave erosion. Hence, postglacial growth could 

 start immediately on the outer rims and also in localities that 



