W. M. Davis — Shale?' Memorial Study of Coral Beefs. 227 



depressions now occupied by embayments originated as valleys 

 of subaerial erosion, which can be excavated only above sea 

 level, and not as sunken fault blocks, which are not definitely 

 related to sea level ; but on this point the form of the valleys 

 and their branches, upstream from the bayheads, leaves no 

 doubt whatever. 



I believe that the same conclusion applies to all barrier reefs, 

 first for the reason that those which I visited were not selected 

 because they were thought to be in any way unlike other mem- 

 bers of their kind, but because they were easily accessible ; and 

 second, because all the charts of other barrier reefs that I have 

 examined show that they also have central islands with embayed 

 shorelines. For none of these islands can any one of the still- 

 stand theories hold good. 



Contrast of Cuban and Pacific Reefs. — It is, however, 

 worth noting that the embayments here considered have a 

 quite different relation to the adjacent coral reefs from that 

 found, according to Hayes, Vaughan and Spencer, in the 

 pouched-reef harbors of Cuba : all the embayments that I saw 

 inside of sea-level barrier reefs in the Pacific islands occupy 

 valleys that are older than the reefs; but in Cuba the valleys, 

 and still more the subsidence which drowned them in produc- 

 ing the pouched harbors, are described by the above-named 

 authors as younger than the elevated reefs which enclose them ; 

 and such valleys do not bear on the origin of the reefs, as 

 appears from the following extract : — " The depressions occu- 

 pied by the water forming these harbors appear to be due 

 entirely to erosion by streams flowing into the sea during a 

 recent geological period when the land stood somewhat higher 

 than now. In other words, they are drowned drainage basins. 

 Their peculiar shape, a narrow seaward channel and a broad 

 landward expansion, is due to the relation of hard and soft 

 rocks which generally prevail along the coast. Wherever the 

 conditions are favorable for the growth of corals a fringing 

 reef is built upon whatever rocks happen to be at sea level, 

 and as the land rises or sinks this reef rock forms a veneer of 

 varying thickness upon the seaward land surface. The rocks 

 on which this veneer rests are generally limestones and marls 

 much softer and more easily eroded than the coral rock. 

 Hence several small streams, instead of each flowing directly 

 to the sea by its own channel, are diverted to a single narrow 

 channel through the hard coral rock, while they excavate a 

 basin of greater or less extent in the softer rocks back from 

 the coast." It is briefly explained on a later page that the ele- 

 vated reefs now stand at an altitude of 30 or 40 feet ; that 

 their formation was associated with a subsidence of from 80 to 

 100 feet, and one may infer that they were then formed dur- 



