In a preliminary article (1) dealing with the British Quaternary strand-line oscillations it was shown that unexpected correlations had emerged as a result of the methods employed. Of the possible assumptions the only one to prove workable was that there had been, during Quaternary time, a temporary lowering of ocean level of 1,000 metres or more. The graphs left no reasonable doubt that if the ocean level had moved in the way assumed the Mindel-Kiss interval of Penck's Alpine sequence must be equated with the major glaciation of North-West Europe and the British Isles. In order to test this, the only assumption made, it was applied to the coral reef problem (2) and there it was successful in offering what should be an acceptable solution. It was surprising to find that Darwin had actually used an apparent rise of ocean level, of the same order of magnitude, to illustrate his own theory of subsiding islands. The success of the postulate when faced with the coral reef problem was held to have invested it with a high degree of probability. It was pointed out, however, that changes in ocean level of that magnitude could best be explained by the sinking of the ocean floor, and the North Pacific Ocean was indicated as the probable position. The cause of the changes of ocean level will now be investigated and then its consequences discussed in the light of ascertained facts.