DECOMPOSING SHELLS. 135 



has been analyzed for me by Mr. T. Reeks : it 

 consists of sulphates and. muriates both of lime and 

 soda, with very little carbonate of lime. It is 

 known that common salt and carbonate of lime, left 

 in a mass for some time together, partly decom- 

 pose each other ; though this does not happen 

 with small quantities in solution. As the half-de- 

 composed shells in the lower parts are associated 

 with much common salt, together with some of the 

 saline substances composing the upper saline layer, 

 and as these shells are corroded and decayed in a 

 remarkable manner, I strongly suspect that this 

 double decomposition has here taken place. The 

 resultant salts, however, ought to be carbonate of 

 soda and muriate of lime ; the latter is present, but 

 not the carbonate of soda. Hence I am led to 

 imagine that by some unexplained means the car- 

 bonate of soda becomes changed into the sulphate. 

 It is obvious that the saline layer could not have 

 been preserved in any country in which abundant 

 rain occasionally fell : on the other hand, this very 

 circumstance, which at first sight appears so highly 

 favourable to the long preservation of exposed 

 shells, has probably been the indirect means, 

 through the coinmon salt not having been washed 

 away, of their decomposition and early decay. 



I was much interested by finding on the terrace, 

 at the height of eighty-five feet, embedded amidst 

 the shells and much sea-drifted rubbish, some bits 

 of cotton thread, plaited rush, and the head of a 

 stalk of Indian corn. I compared these relics with 

 similar ones taken out of the Huacas, or old Peru- 

 vian tombs, and found them identical in appear- 

 ance. On the mainland in front of San Lorenzo, 

 near Bellavista, there is an extensive and level 

 plain about a hundred feet high, of which the low- 

 er part is formed of alternating layers of sand and 



