TRANSACTIONS OF SECTION B. 567 



with Swedish bar iron. The E.M.F. between the two metals was accurately 

 determined by a galvanometer. This test is really, therefore, a modified solution 

 pressure-test, and was employed by Andrews in his standard researches on 

 corrosion. 



The simple corrosion results show that an annealed steel is least liable to 

 corrosion and hardened steels most liable. A general increase in corrodibility is 

 noticed 'with an increase of carbon percentage in steel. 



The galvanic corrosion results show that when a steel is in contact with pure 

 iron the corrodibility (as indicated by strength and direction of current) is in 

 the following increasing order : Hardened, rolled, normalised, annealed. 



Solubility tests were determined in one and two per cent, sulphuric acid for 

 72 hours, each strength of acid giving similar types of curve. 



These results show to a marked degree the influence of both heat treatment 

 and composition. The relative position of the curves are, however, in exactly 

 the reverse order from that found in the case of simple corrosion which would 

 not be the case if solubility in acid were merely an accelerated corrosion test. 

 The special feature of the solubility curves is the point at 45 per cent, carbon 

 (which is also slightly marked in the corrodibility curves), and which has been 

 previously noted by Messrs. Heyn and Bauer. 



The solution pressure of each steel was separately determined in each of the 

 three following solutions : Sea-water, one per cent, sulphuric acid, and slightly 

 acidulated N/10 ferrous sulphate solution. The influence of heat treatment is in 

 each case very marked, the effect produced being practically the same in every 

 case, the annealed and rolled steels having the highest solution pressures, the 

 normalised intermediate, and the hardened steels the lowest values. 



A comparison of the corrodibility and solubility curves indicates, in the first 

 place, that the solubility results cannot be accepted as indicative of the relative 

 corrodibilities of the steels in such solutions as sea-water, whereas, for example, 

 an annealed steel is much more soluble in dilute sulphuric acid than the same 

 steel in the hardened condition; yet, when subjected to a simple corrosion test 

 in sea- water the hardened steel is found, on the contrary, to corrode much more 

 than the annealed steel. So that although a dilute acid test is of value for 

 certain purposes, yet these results show it to be unreliable as a guide to the 

 corrodibility of steel in such solutions as sea-water. 



The general influence of heat treatment on the various properties examined 

 is found to be the same in the case of solubility, solution pressure (in all three 

 solutions), and the galvanic corrosion results. In the simple corrosion results, 

 however, treatment exerts practically the opposite effect to that found in above- 

 mentioned tests. 



The results obtained suggest the probability that the solution pressure is the 

 governing factor in the solubility of steel in dilute sulphuric acid, its influence 

 being modified by auto-electrolytic actions due to constituents and stresses (set 

 up by treatment) in the steel. Consideration of simple corrosion results suggests 

 that the solution pressure only plays a secondary part in corrosion, the principal 

 factor being the auto-electrolytic action in the steel itself. 



Marked breaks in the continuity of the curves appear in every case at about 

 0'3 per cent, and 0"75 per cent, carbon; the latter break suggests the likelihood 

 of another peak having its maximum at the saturation point of the steel. We 

 cannot at present account for the break at 0'3 per cent, carbon, but further 

 work is intended on the influence of carbon and other elements on corrosion, 

 which we anticipate will elucidate this matter. 



Although heat treatment exerts considerable influence on corrosion, it 

 cannot be expected to make up for the defects due to segregation or inferior 

 material. 



9. The Crystalline Structure of Iron at High Temperatures. 

 By Walter Rosenhain, B.A., D.Sc. 



The immediate object of the experiments was to study the mode of deforma- 

 tion of iron and steel under stresses applied to the metal at high temperatures. 

 The principal questions to be decided were, whether at high temperatures the 

 mode of deformation was still of a strictly crystalline character, such as that 



