»rcous smid, iiiterrupled only ar intervals, Lij shelves composed (probably) 

 »of dead coral rock. To c.irry on the analogy , flie blades of grass grew 

 Dtbiiiner and ihiiiner, till at last, the soil was so sterile , lliat nothing 

 Tjsprung from it. 



All corals are efiiciwit agents in forming a reef, the la- 

 mclliform perhaps the least so of any. Mr. Dauwin himself, ob- 

 scrvcd in the cold waters of the seas of Terra deiruego, that, 

 //even at the depth of forty or fifty fathoms, small strong co- 

 //rallines, were abundant on the sea bottom," and, captain Sir 

 James Clarke Eoss (who doubtless had seen), seems to have 

 correctly estimated these statements of Mr. Dauwin, as he very 

 quietiy observes, respecting a bank in the Pacific , on which he had 

 sounded. // This coral bank is thus growing up, from a depth of 

 // over three hundred fathoms."" The whereabouts of those soundings 

 taken by captain riTziiOY is a desideratum, because the reef 

 on which the chain of islets is situated, slopes down from the 

 breakers on its brim to some distance , the slope from that brim 

 outwards being composed of patches of sand and stools of coral 

 havinf>- from forty to fifty fathoms at about a cable's length 

 from the reef. In fact live coral in abundance and in large mas- 

 ses are to be seen in a calm day , in more than fifteen fathoms 

 depth , and pieces of the same have been brought up (on an an- 

 chor) from over fifty fathoms in depth. Dead or rotten coral , 

 could not have broken all capt. I'iïsr.OY's // small anchors^ 

 //hooks, p-rappling irons and chains, one after another as soon as 

 // a strain v/as hove upon them by , or i]i his largest boats." I'itz-, 

 roy, Volume I page 034. 



The shelves are therefore most probably of basaltic, if not 

 of live corai rock formation. 



II. » As long as no facls bcyond these relating to the siruclurc of lagoon 

 » isiands were known , so as to establish sonie more comprehensivc iheory , 

 » the belief that corals constructed iheir habitations, or speaking more cor- 

 .■.)reclly, their skelelous on the circular crests of submarine cralers, was 

 ïbolb ingenious »ud very plausiblc. Yet , the sinuous margin of some , as 

 »iii the Radack Islands of Kolzebue , one of which is fifly Iwo miles long 

 j) by Iwenly broad , and the narrowness of others, as in T'ow Island , must 

 »ha\e slartled cvcry one who considercd this subject." 



