March 26, 1915] 



SCIENCE 



455 



lying propositions, so interesting, so illu- 

 minating, often so amazing. 



Finally, but far from exhausting the list, 

 it remains to mention the great subjects of 

 invariants and groups. Both of them ad- 

 mit of definition perfectly intelligible to 

 disciplined laymen; both admit of endless 

 elementary illustration, of having their 

 mutual relations simply exemplified, of 

 being shown in historic perspective, and of 

 being strikingly connected, especially the 

 notion of invariance, with the dominant 

 enterprise of man : his ceaseless quest for 

 the changeless amid the turmoil and trans- 

 formation of the cosmic flux. 



Cassius J. Keyser 



Columbia TJniveesity 



PSELIMINAEY SEPOBT ON A SHALEB 

 MEMORIAL STUDY OF CORAL BEEFS 



A LIBERAL grant from the Shaler Memorial 

 Fund of Harvard University, siipplemented 

 by a generous subsidy from the British Asso- 

 ciation for the Advancement of Science with 

 an invitation to attend its meeting in Australia 

 last August as a foreign guest, enabled me to 

 spend the greater part of the year 1914 in visit- 

 ing a number of islands in the Pacific Ocean 

 with the object of testing various theories that 

 have been invented to account for coral reefs. 

 Thirty-five islands, namely, Oahu in Hawaii, 

 eighteen of the Fiji group, New Caledonia of 

 which the entire coast line was traced, the 

 three Loyalty islands, five of the New Hebrides, 

 Earotonga in the Cook group, and six of the 

 Society islands, as well as a long stretch of the 

 Queensland coast inside of the Great Barrier 

 reef of northeastern Australia, were examined 

 in greater or less detail. A brief statement of 

 my results has been published in the Pro- 

 ceedings of the National Academy of Sciences 

 for March, 1915. A full report will appear 

 later, probably in the Bulletin of the Museuin 

 of Comparative Zoology at Harvard College. 

 The general conclusions reached are here 

 briefly summarized. 



Any one of the eight or nine theories of 



coral reefs will satisfactorily account for the 

 visible features of sea-level reefs themselves, 

 provided the postulated conditions and proc- 

 esses of the invisible past are accepted: hence 

 a study of the visible features of the reefs 

 alone can not lead to any valid conclusion. 

 Some independent witnesses must be interro- 

 gated, in the hope of detecting the true theory. 

 The only witnesses, apart from sections ob- 

 tained by deep and expensive borings, available 

 for sea-level reefs are the central islands within 

 oceanic barrier reefs, or the mainland coast 

 within a continental barrier reef. The testi- 

 mony of these witnesses has been too largely 

 neglected, apparently because most investi- 

 gators of coral reefs have been zoologists, little 

 trained in the physiography of shore lines. 

 Elevated reefs afford additional testimony in 

 their structure and in the relation of their 

 mass to its foundation; but these witnesses 

 also have been insufficiently considered, perhaps 

 because most investigators of reefs have, as 

 zoologists, been little trained in structural 

 geology; hence it seemed desirable to give as 

 much time as possible on the Pacific islands to 

 questioning the independent witnesses above 

 designated, rather than to the study of the 

 reef themselves. 



The testimony of the first group of witnesses 

 — the central islands of barrier reefs — con- 

 vinced me that Darwin's theory of subsidence 

 is the only theory competent to explain not 

 only the development of barrier reefs from 

 fringing reefs, but also the shore-line features 

 of the central (volcanic) islands within such 

 reefs ; for the embayment of the central islands 

 testify emphatically to subsidence, as Dana 

 long ago pointed out: thus my results in the 

 study of this old problem of the Pacific agree 

 with those of several other recent students, 

 especially Andrews, Hedley and Taylor of Aus- 

 tralia, and Marshall of New Zealand. Darwin's 

 theory of subsidence also gives by far the 

 most probable explanation of atolls; for it is 

 unreasonable to suppose that a subsidence of 

 the ocean bottom should occur only in regions 

 where the central islands of barrier reefs are 

 present to attest it, and not in neighboring 

 regions where reefs of identical appearance. 



