278 EIGHTH REPORT — 1838. 



clear water, strongly saline, or merely brackish, calcareous, al- 

 kaline, or chalybeate, holding much or little cai-bonic acid, acts 

 most powerfully in producing rust, that is, produces the largest 

 quantity in a given time from any one specimen, or in what re- 

 lative degree. This statement is made exclusively, of course, of 

 the better-understood cases of mine-waters, such as solutions of 

 sulphate of copper, &c., whose action on iron belongs to another 

 branch of our subject. 



64. There are some cases of local action on wrought iron, 

 however, which appear remarkable, and need investigation. 

 The very purest and finest specimens of wrought iron, when ex- 

 posed, with turned or boi'ed surfaces, to fresh water, frequently 

 corrode entirely locally, by deep and destructive jntting. A 

 portion of a turned wrought iron valve was presented, acted on 

 in this way, in about two months, by a remarkably pure stream 

 of fresh water, holding nothing but air and a trace of carbonate 

 of lime and iron in solution. This sort of reaction appears the 

 converse of M. Payen's tubercles, and its explanation would 

 seem to lie in the iron so affected possessing a damasked 

 structure, that is, in fact, being composed of two sorts of iron 

 chemically different, and united by welding. This is unavoidably 

 the case with all "scrapped" wrought iron, or that forged or 

 rolled from scraps of various sorts. Now these differently- con- 

 stituted irons, being in different electrical relations, give rise to 

 a voltaic circuit, in which the most positive is corroded fastest ; 

 but as all the surface is in some degree affected, the oxides 

 formed cannot adhere, and hence, while pitting goes on in some 

 places, tubercles are formed in none. This view suggests the 

 importance of having large bars, which require to be formed 

 by rolling several smaller ones together, formed from iron all 

 of the same " make," if for any important purpose as regards 

 durability. 



It shows that rolled bars, as being more uniform, are pre- 

 ferable, for the same reason, to "scrapped" or hammered iron; 

 and it points out that the practice adopted in some cases of 

 making railway bars by rolling together two bars of different 

 sorts of iron, the one hard and rigid, to give a durable upper or 

 working face to the bar, and the other tough and soft, to resist 

 extension, is highly objectionable as regards the duration of the 

 rail, under the influence of air and moisture — eminently so, be- 

 cause the lower segment of the rail, viz. the extended part, will 

 corrode the fastest, thus losing strength where it is most wanted. 



65. There were also presented the two extremities of a bar of 

 wrought iron found at the bottom of a very deep well in a 



