PACIFIC AND BEERING'S STRAIT. 263 



one of great interest, and will require a numerous chap. 

 and careful collection of facts before any entirely v — - r*i-> 

 satisfactory conclusion can be arrived at. I regret ^JSJ' 

 that my time did not permit me to inquire more 

 particularly into this curious matter; but having to 

 survey about fifty islands, some of which were of 

 great extent, in the space of about four months, I 

 could not accomplish more than was absolutely ne- 

 cessary to the purposes of a safe navigation of the 

 Archipelago. We were, however, not inattentive 

 to the subject, and when opportunity offered, sound- 

 ings were tried for at great depths, and the descent 

 of the islands was repeatedly ascertained as far as the 

 common lines would extend. Some of these expe- 

 riments are given in the annexed plate, representing 

 a section of a coral island from actual measurement. 



In considering the subject of these coral forma- 

 tions, my attention was drawn to the singularity of 

 the occurrence of openings in them, either opposite 

 to, or in the direction of some stream of fresh water 

 from the mountains ; and on searching several charts, 

 I find so many corroborations of the fact, that I have 

 no doubt of the truth of it : as far as my own obser- 

 vations extended, it was always so. The aversion 

 of the lithophytes to fresh water is not singular, as, 

 independent of its not being the natural element of 

 those animals, it probably supplies no materials with 

 which they can work. 



It has been suggested, that these openings being 

 opposite to valleys, the continuation of them under 

 water is the cause of the break in the reef. But 

 when we consider the narrowness of these openings, 

 compared with the width of the valleys, and that 

 the latter are already filled up to the surface and fur- 



