304 proceedings: chemical society 



Experiments on corrosion in soils from various sources show a wide 

 variation in the coefficient of corrosion. The rates of electrolytic cor- 

 rosion were not materially different for Ingot iron, wrought iron, cast 

 iron, and machine steel. 



A series of tests in which nitrates, carbonates, sulphates, chlorides, 

 and chromates were added to the soil showed that only the chromates 

 retarded the electrolytic corrosion. The character of the corroded 

 surface depends upon the chemical in the soil. Pitting is not due en- 

 tirely to the non-homogeneity of the anode. 



The coefficient of corrosion in normal concrete is usually about 1 

 per cent. High current density, temperatures above 50°C, the addi- 

 tion of salt to the cement, and the exposure of small green specimens to 

 C0 2 greatly accelerate the corrosion. 



Electrolytic corrosions with alternating currents show low coefficients 

 of corrosion, unless the length of the cycle or the character of the soil is 

 such as to prevent the corroded iron from plating back on the reversal 

 of the current. The coefficient of corrosion is usually low, even when 

 the period of a complete cycle is several hours. 



On account of the reversing polarity of the underground structures 

 throughout a large part of the areas affected by stray currents from 

 street railways, electrolytic damage will usually be much less than 

 might be expected. 



Coefficients of corrosion greater than 100 per cent are only found 

 when the current density is low, and are probably due to accelerated 

 self-corrosion. (Author's abstract.) 



G. K. Burgess and P. D. Merica, Bureau of Standards (given by 

 Mr. Merica) : Some examples of metal failures. The talk was illustrated 

 by many slides of great interest. 



The question of metal failures is a very comprehensive one and may 

 indeed be said to embrace all cases in which a metal does not fulfill, as 

 well as may be, the use or uses to which it is put. Whether or not a 

 metal is a failure under given conditions may therefore be a relative 

 matter and there cannot usually be an ideal standard of service. 



Metal failures may evidently be classified as to type or cause, whether 

 due to inherent chemical or physical imperfections or to some incor- 

 rect treatment (thermal, mechanical or chemical) which it may have 

 received either in manufacture or subsequently. 



It is often necessary, in considering metal failures, to fix the responsi- 

 bility for failure, and it may be necessary at times to decide whether 

 the fault lay in the metal or in the specifications which may have been 

 so incorrectly or inadequately drawn as to be entirely unsuited to the 

 metal in question. 



Examples of metal failures and of imperfections in metals were given 

 with illustrations, the same being taken from the experience at the 

 Bureau of Standards. 



The effect of pipe, segregation, and blowholes in ingots of steel can be 

 readily traced in the properties of articles manufactured from these 

 ingots. Steel rails, for instance, will contain longitudinal seams, due 



