3 2 4 SCIENCE PROGRESS 



exerted on the metal remaining intact and apparently unaffected. 

 Deterioration is found which is not explained by mere loss in 

 weight or by change in the ordinary composition of the 

 metal." 



He describes experiments in which eight-inch lengths of 

 rail were cut up into sections and the pieces subjected to tensile 

 and other tests. On testing pieces cut from certain used rails, 

 in which cracks had been developed through defective track 

 conditions, in comparison with those from unused rails, no sign 

 of decay or deterioration could be observed in the faulty rails. 

 Under other conditions, however, rails became seriously em- 

 brittled, this being the case especially in tunnel rails, particularly 

 in those from damp tunnels. Whilst the whole of the results 

 Mr. Longmuir obtained show the embrittling effect, no two 

 samples were precisely alike ; he was not able, in fact, to trace 

 any regularity in the order of development of brittleness but 

 considers that whilst erratic in intensity, it is certain in effect. 



In many industries, especially that of wire drawing, cleaning 

 with acid is resorted to in order to remove the scale from iron ; 

 the embrittling effect induced by this treatment is more or less 

 recognised. Mr. Longmuir gives results obtained by himself 

 showing the deleterious effect such treatment produces and the 

 manner in which the original properties may be restored to the 

 metal by heating it. The brittleness is attributed to the absorp- 

 tion of hydrogen. 



Earlier in his paper Mr. Longmuir quotes analyses of rust 

 obtained from rails laid in various situations, specially calling- 

 attention to the relatively large amount of sulphur present in 

 them as sulphate, specially in that from tunnel rails ; these, 

 it is well known, are particularly subject to corrosion. It is 

 clear, therefore, that the effect acids may produce should not be 

 lost sight of. 



The Influence of Impurities on Corrosion 



As iron cannot be acted upon by an electrolyte except in 

 the presence of an electro-negative " metallic " conductor, it 

 follows that all conducting impurities which come within this 

 category must influence corrosion ; moreover, that alterations 

 in the iron itself — whether in molecular or crystalline structure, 

 which in any way affect the value either of E or of R in the 



