February 6, 19 13] 



NATURE 



substance on the front face of each block. But the 

 surface of the second block shows a feature which 

 Darwin did not notice, although it is quite as essen- 

 tial a consequence of his theory as any other. This 

 is the invasion of the previously eroded valleys of the 

 subsiding island by the sea, so that the relatively 

 simple shore line of the first stage Is in 

 the second stage transformed into an em- 

 bayed shore line, possessing several of "those 

 deep arms of the sea . . . which," as Darwin said, 

 "penetrate nearly to the heart of some encircled 

 islands" ("Coral" Reefs," 1842, p. 49). So long as 

 subsidence continues the bays cannot be filled with 

 deltas and the ridges cannot be cut back in cliffs. 

 Darwin recognised that the central island must 

 diminish in size as it subsides, but he did not also 

 perceive the necessary modification of its outline. 



It requires but a brief e.xamination of large-scale 

 charts of the Pacific island-groups to discover that the 

 central islands of barrier reefs are repeatedly char- 

 acterised bv an embayed shore line, that the bays 



1839. Several months earlier, during " the ascent of 

 Mt. Aorai on Tahiti, in September of 1839," he had 

 conceived the production of an embayed shore line 

 as a necessary result of the subsidence of a dissected 

 land mass. Let it be noted in passing that he was 

 the first clearly to announce this important principle, 

 which then had no place in geology or geography, 

 although it is hinted at in De la Beciie's "Researches 

 in Theoretical Geology " (London, 1834, p. 193). In 

 Dana's first report he says, when following Darwin 

 in explaining barrier reefs and atolls by subsidence : — 

 " The very features of the land, the deep indentations, 

 are sufficient evidence of subsidence to one who has 

 studied the character of the Pacific islands " 

 ("Geology," U.S. E.xpl. Exped., 1849, p. 131); and 

 on a later page a more explicit statement is made 

 under the general heading, " Evidence of Subsidence," 

 and the special heading, "Deep Bay-indentations in 

 Coasts as the Terminations of Valleys " : — " In the 

 remarks upon the valleys of the Pacific islands, it 

 has been shown that they were in general formed by 



-Stage's 



are not filled with delta-plains, and that the ridge ends 

 are not truncated in cliffs. Kandavu and Mbengha 

 in the Fiji group (.'\dmiralty chart 167), the western 

 members of the Society group (chart 1060), and 

 especially Bora Bora (1428), Gambler Island (1112), 

 a western member of the Paumotus, and Rossel 

 Island (1473), in the Louisiade archipelago of British 

 New Guinea, may be instanced among many others 

 as affording good illustration of at least some of these 

 features, particularly of embayed shore lines. 



In view of the remarkable accordance that is thus 

 found between the inferred consequences of the theory 

 of subsidence and the observed features of central 

 islands of barrier reefs, and in memory of the failure 

 of corresponding consequences of other theories to 

 match the facts, an open-minded inquirer cannot hesi- 

 tate long in making choice among the several ex- 

 planations that have been suggested for barrier reefs, 

 and with them, of atolls. 



Dana first learned of Darwin's theory when the 

 Wilkes expedition reached Sydney near the end of 



NO. 2258, VOL. 90] 



\X.in^ 



the waters of the land, unaided by the sea; that the 

 sea tends only to level off the coast, or give it an even 

 outline. When, therefore, we find the several valleys 

 continued on beneath the sea, and their enclosing 

 ridges standing out in long narrow points, there is 

 reason to suspect that the island has subsided after 

 the formation of the valleys. For such an island as 

 Tahiti could not subside even a few scores of feet 

 without changing the even outline into one of deep 

 coves or bays, the ridges projecting out to sea on 

 every side. . . . The absence of such coves, on the 

 contrary, is evidence that any subsidence which has 

 taken place has been comparatively small in amount " 

 (P- 393)- Similar statements are made in Dana's first 

 book on this subject, " On Coral Reefs and Islands " 

 (New York, 1853, pp. 118-119), and in all the editions 

 of his larger book, "Corals and Coral Islands" (New- 

 York, first edition, 1872, pp. 319-320; third edition, 

 i8qo, pp. 273-274). _ 



It is remarkable how rarely the value of this capital 

 point has been recognised. It is referred to in 



