248 Report ofDr Turner's Lecture on 



forcible increase of volume. The crystallization of salts was a 

 similar pkenomenon, and gave rise to a similar effect. When a 

 stone was dipped into a saline solution, and then suspended in 

 the air to dry, the crystallization of the salt produced a certain 

 amount of injury ; and the effect due to one operation might be 

 multiplied to any extent by the repetition of the same process. 

 The experiment of a few days might thus be made to imitate 

 the effect of numerous winters, and the relative durability of 

 different materials be ascertained prior to their selection for 

 building. The salt most applicable to such substances was 

 found to be the sulphate of soda *. 



3. Chemical Action. — The affinities which principally contri- 

 bute to affect the integrity of rocks, were stated to be thoseof water 

 and carbonic acid for potash and soda, and that of oxygen for iron. 

 The changes referred to were frequent in felspathic rocks, and 

 were exemplified in a very striking manner in the formation of 

 porcelain clay from granite and other allied rocks rich in felspar. 

 All granitic regions presented examples of this nature, and in 

 none were they more remarkable than in Cornwall and Au- 

 vergne. It was probable that the long-continued action of pure 

 water might produce decomposition ; but the effect of its affinity 

 for the alkalies of the rock was materially aided by that of car- 

 bonic acid for the same bases. This was shown by the increased 

 decomposing power of water when charged with carbonic acid, 

 and by the action of moist carbonic acid gas on granite, as ex- 

 emplified in the volcanic districts of Auvergne. Basaltic rocks 

 were likewise prone to decomposition, partly in consequence of 

 containing felspar, and partly from the protoxide of iron of the 

 augite or hornblende which enters into their composition. TliiA 

 passage of the iron into a higher degree of oxidation was du^^ 

 to atmospheric oxygen applied in a liquid state to the rodk'^ 

 through the medium of water. It was probable that carboni<fc>^ 

 acid likewise co-operated ; — that, as in the rusting of iron, a ca^iX* 

 bonate of the protoxide was first generated, which subsequefitl^* 

 passed into the hydraled peroxide of iron. iiv/<>rf« 



t^^he'TOcks in which these changes occurred, underwent a totaP' 



, ^, M* I^r4Jin4|iit4'^iJhiiiVi[}frrhjs. vol. xxxviii. p. IW*!? V qfi-^ 



