484 REPORTS ON THE STJtTE OF SCIENCE. — 1919. 



the corrected E.L. stresses are much greater than those at the corresponding 

 yields. 



There is no doubt that inaccurate loading lowers the mean elastic 

 limit stress in the case of annealed mild steels, and that its true value 

 varies somewhat. The two samples SC. 4 show approximately the same 

 stress distribution, but the E.L. stresses are 8'55 and 9-32 tons per sq. in. 

 before correction. BA. 1, 6, and 7 afford an even better example because 

 the higher E.L. accompanies the lower Y.P. stress. 



Evidently the measured E.L. stresses need some correction to give 

 true values, but it seems that that indicated by the strain distribution 

 is too drastic if the eccentricity of loading be large, e.g., SC steel. Of 

 the BA specimens 7 and 8 agree fairly well after correction with each other 

 and with the yield stresses, but for 6 the value is still low. The S2 steel 

 shows similar inequalities, 3 and 5 are over-corrected, but for 6 and 7 

 the results are fairly good. 



It is desirable to remark that too close agreement cannot be expected 

 between values for stresses at the elastic limit, nor at yield when the stress 

 is unequally distributed. In general, the stress-strain curve does not 

 diverge greatly from a straight line at first, and, since small errors of 

 observation arise, it is not easy exactly to locate the required point. 



Cast steel as supplied in the bar gives results which are not improved 

 by the suggested correction, but the annealed specimens need it. 



The values for aluminium and the bronzes are in no better agreement 

 after the adjustment. 



Only two specimens of brass rod are included, but these show the best 

 agreement of all between the corrected E.L. stresses. 



Unfortunately, it is not possible to draw many general conclusions 

 from the data available. Only in the cases of annealed steels and brass is 

 the proposed correction of the elastic limit stress, in the ratio of maximum 

 to mean strain, at all justified, and then the factor is too large if the in- 

 accuracy of loading is considerable. 



The degree of eccentric loading which is likely to arise in tension- 

 testing has less effect on the yield point stress than the differences which 

 are found along the same bar, and, as was expected, there is no evidence of 

 the maximum stress being affected. 



Should the elastic limit of a steel or the yield point of other material 

 be required, it is necessary to take precautions to ensure that the loading 

 is accurate, and thus minimise the necessity for correction, which latter 

 cannot be applied in any case with certainty. 



Experiments on the Effect of Alterations of Tensile Stress at Low 

 Frequencies on the Elastic Properties of Mild Steel. By Angus R. 

 Fulton, M.B.E., B.Sc, A.M.Inst.C.E., Temp. Majw Royal Air 

 Force. 



The work was carried out in the Engineering Laboratory of University 

 College, Dundee. It was directed to several points raised in the investiga- 

 tions into the theory of failure due to repetitions of stress by Reynolds 

 and Smith, Stanton and Bairstow, and others, which appeared to require 

 further examination. 



