58 James D. Dana, Esq., on the Structure, 



These facts, it may be said, are only negative, as the 

 sounding lead, especially in the manner it is thrown in sur- 

 veys, would fail of giving decisive results. The character of 

 a growing coral bed is so strongly marked in its uneven 

 surface, its deep holes and many entangling stems, to the 

 vexation of the surveyor, that in general the danger of 

 mistake is small. But allowing uncertainty as great as sup- 

 posed, there can be little doubt after so numerous observations 

 over so extended regions of reefs. 



The depth of the water in harbours and about shores where 

 there is no coral confirms the view here presented. At 

 Upolu the depth of the harbours varies generally from 12 to 

 20 fathoms. On the south side of this island Lieutenant 

 Perry, off Falealili, one hundred yards from the rocky shores, 

 found bare rocks in 18 and 19 fathoms, with no evidence 

 of coral. There is no cause here which will explain the 

 absence of coral, except the depth of water ; for corals and 

 coral reefs abound on most other parts of Upolu. Below 

 Falelatai, off the same island, an equal depth was found, with 

 no coral. Off the east cape of Falifa harbour, on the north 

 side of Upolu, Lieutenant Emmons found no coral, although 

 the depth was but 18 fathoms. About the outer capes of 

 Fungasar harbour, Tutuila, there was no coral, with a depth 

 of 15 to 20 fathoms ; and a line of soundings across from 

 cape to cape, afforded a bottom of sand and shells, in 15 

 to 21^ fathoms. About the capes of Oafonu harbour, on 

 the same island, there was no coral, with a depth of 15 

 fathoms. 



Similar results were obtained about all the islands sur- 

 veyed, as the charts satisfactorily shew. There is hence 

 little room to doubt that 20 fathoms may be received as near 

 the range in depth for reef corals ; and probably the limit 

 lies between 15 and 20 fathoms, or not far from 100 feet. 



It may be here remarked, that soundings with reference 

 to this subject are liable to be incorrectly reported by persons 

 who have not particularly studied living zoophytes. It is of 

 the utmost importance, in order that an observation supposed 

 to prove the occurrence of living coral should be of any 

 value, that it be unequivocally determined whether the frag- 

 ments, which a lead may bring up, are alive or not when 



