THE EIGHT-RAYED POLYPS 



3497 



that water is constantly being thrown by the larger waves over the lower portions 

 of the reef into the canal or lagoon, and seeks either to escape as an undercurrent 

 in opposition to the flood tide, or else strengthens the ebb tide. These and other 

 similar disturbances of the water in the canals bring with them much coral detritus, 

 and render the bottom altogether unsuitable for the growth of corals. Where such 

 currents are strong, they keep the canals clean and open. The action of the oceanic 

 currents is often increased by the fresh waters coming from the central islands, and 

 harbors are therefore very often found at the mouths of valleys and of their small 

 streams. The influence of the fresh water itself on the corals is not so great as 

 is usually assumed, chiefly because it, being lighter than salt water, flows away 

 on the surface of the latter and hardly touches the animals which grow below the 

 surface. 



The form of the reef is again largely influenced by the form and constitution 

 of the sea bottom. Where deep submarine fissures occur, dipping down below the 

 level at which the corals flourish, no reef can be formed, as also in places where 

 firm ground alternates with sand and mud. All irregularities in the outline of a 

 reef or an atoll, and the formation of harbors in coral islands, can thus be simply 

 explained. 



The most important point which needs elucidation is why some reefs encircle 

 islands as a fringe extending from the shore, while others run parallel with the land, 



SECTION THROUGH A CORAL RKEK. 



no longer touching it; others, again, forming circular lagoons with no island at all in 

 the middle. This was the question which puzzled the first discoverers of reefs, and 

 at one time it was supposed that instinct guided the animals in giving their struc- 

 tures the form best suited to withstand the force ' of the waves. According to 

 another hypothesis, put forth by Steffen in 1822, the reefs represent the summits of 

 volcanic mountains, the crater being filled by the lagoon, while the channels through 

 the reef indicate the points at which the edge of the crater was destroyed by out- 

 bursts of lava. This superficially plausible view was disposed of by Darwin thirty 

 years ago. He argued that the volcanic cone thus assumed must either once have 

 stood upon dry land and then have been submerged, or else must have been formed 

 beneath the sea. In the former case, the crater would in almost all cases have been 

 destroyed during the gradual sinking; while the formation of craters by submarine 

 eruptions and their subsequent elevation is hardly conceivable. This hypothesis 

 further requires the assumption that immense numbers of volcanoes must have 

 arisen over a limited area, and a still more improbable supposition, that all these 

 volcanoes rose to almost the same level, seeing that the coral animals occur only 

 within a depth of about twenty fathoms. Craters nearly fifty miles in diameter 

 must be assumed to have existed, and others of twenty to twenty-five miles must 



