302 



ALEXANDER AGASSIZ 



Fig. 1 



where the so-called reef is nothing but a series of islets 

 and ledges overgrown by Serpulse and Algse, which, in 

 the long stretch from Tinker's Farm [?] to the Gibb's 

 Hill Bay on the south shore, form no end of atolls, bar- 

 rier reefs, and crescent-shaped reefs. But the rising rim 



is not formed by the growth 

 of the Serpulse. The outer 

 edge is merely protected by 

 Serpulse and Algse, and the 

 surf breaking over that pro- 

 tected edge digs out a deep 

 hole and thus are formed 

 potholes (atolls) or crescent- 

 shaped reefs, if there is an outlet broken out, or a barrier 

 reef if it is a shore ledge which is acted upon by the surf. 

 I had before seeing these little atolls always supposed 

 the vertical walls were built up by the Serpulse, but it 's 

 no such thing — four to six inches is the highest Ser- 

 pulse thickness I 've found, and everywhere where I have 

 looked into the rim of an atoll, etc., it has always been 

 made up of ceolian ? , ocks ! — the walls left by the eating 

 and scouring of the surf on blocks of seolian rocks, of 

 all shapes, after they have been reduced to the line of 

 the sea. This 

 will show you 

 my idea : — 



W h e n the 

 rock A[ Fig. 1] 

 is cut if at 

 II W M, it 



leaves a flat ledge at L WM ; the block A falls to one 

 side and is, according to its size and thickness, then acted 

 upon by the surf and breakers, and disintegrated accord- 



-H.W.M. 



L.W.M. 



Fig. 2 



