440 C. JSarus and V. Strouhal — Strain-effect of sudden 



plained chemically, we inferred that the category of electrical 

 and magnetic phenomena exhibited by steel on passing from 

 hard to soft are mainty referable to changes in the character and 

 intensity of the strain which the tempered rod carries. We 

 accepted the theory of a dense shell and rare core as being the 

 most satisfactory and elegant conception of the strain in ques- 

 tion, — with the proviso* that "it must be regarded as a mere 

 diagram of the essential features of the vastly more complicated 

 structure of the glass-hard rod." These conditions premised 

 we finally interpreted all variation of strain produced by anneal- 

 ing as the combined effect changes of the viscosty of steel due 

 to temperature and of interference of thermal expansion with 

 the said strain. 



The experimental question which we are endeavoring to elu- 

 cidate may therefore be succinctly put as follows: 1. With 

 what degree of sensitiveness do the variations of the density of 

 the rod as a whole, indicate the corresponding variations of 

 strain. 2. Is it possible successively to remove layer after layer 

 of a rod without materially changing the character and intensity 

 of the strain which the rod is supposed to carry. 3. In how 

 far does the actual structure of tempered steel differ from the 

 diagrammatic distribution of density above assumed. 4. Does 

 the process of sudden cooling impart strain alike in kindf but 

 differing enormously in degree to all substances. In our paper 

 on structure we had the second and third of these points prin- 

 cipally in mind. We were unable to obtain direct and decisive 

 evidence of the occurrence of shrinkage during the removal of 

 shells from a hard steel rod. But since the structure of steel dif- 

 fers so largely with the kind of steel operated upon we do not 

 regard our experiments as made in sufficient number to be con- 

 clusive. An examination of the density of the elementary shells 

 shows that the character of the density at any point regarded 

 as a function of the distance of the point below the surface, is 

 harmonic.:}: To investigate this relation it is unfortunately 

 necessary to make the measurements for thin (|- mm ) shells. 

 Hence the mean amplitude of vibration and the limits of the 

 unavoidable errors of observation are of the same order. To 

 discriminate between the periodic effect of temper, which may 

 show regular or irregular periodicity, and the apparently peri- 

 odic distribution of mere errors of observation is difficult and 

 requires very nice and careful observation. In the Bulletin 

 (35) cited we attempt to arrive at this discrimination by making 

 minute comparisons of the structures of rods of the same kind 

 of steel, and of the same or of different diameters. This pro- 



* Bull. U. S. Geo], Survey, No. U, p. 96 ; ibid., pp. 95 to 98. 

 f We do not necessarily refer to mere volume expansion here. See Wrightson, 

 Journ. Iron and Steel J'St, ii. 1879, p. 418. % This volume, p. 386. 



