384 



NATURE 



[December 5, 1912 



ot points in the body corresponding to definite 

 inclinations of the principal axes of stress. If, 

 for example, we take a transparent model of some 

 stressed object, such as a cement briquette, Fig-. 2, 

 of the form used by engineers for testing the 

 tensile strength of cements, we can observe the 

 movements of the bands shown on this model as 

 the Nicols are rotated, and can mark the 

 positions of the axes of principal stress at every 

 point in the specimen. 



A series of positions of the central lines of 

 these isoclinic bands is shown in Fig. ^i'^. 

 for this case, and from these curves we 

 can readily obtain, by graphical or other 

 processes, a map of the lines of principal stress 

 (Fig. 3b) throughout the body. The isoclinic 

 lines are especially valuable for verifying the 

 results of mathematical calculations, 1 as only small 

 loads need to be employed, thereby avoiding the 

 fracture of costly glass specimens and the possible 

 variation of the physical properties of the material 



suggested,- but in many cases this may be accom- 

 plished fairly accurately by approximate 

 methods, especially where one principal stress is 

 very great compared with the other. In the 

 present instance, the chief interest lies in the 

 distribution of stress at the minimum section, 

 where fracture is intended to take place. The 

 minor principal stress at this section is small 

 everywhere, and vanishes at the ends. Hence, the 

 experimental curve of values of the difference of 

 principal stresses at the section, Fig. 3c, also 

 shows the tensions at the ends accurately, and 

 very nearly so at other points. 



If this stress curve is integrated and compared 

 with the stress applied, a mean value of the minor 

 principal stress may be determined, and an 

 approximation to the minor principal stress 

 distribution obtained. Even without this we can 

 see that the stress across the section of a cement 

 briquette probably varies very greatly, and that 

 the universal method of reckoning: the stress 



at high stresses. In technical problems, how- 

 ever, and in cases where a mathematical solution 

 is not available, it is generally advisable, and it 

 may be necessary, to measure the intensity of the 

 double refraction produced by load. 



The colour 'fringes indicating the stress are 

 observed more accurately with circularly polarised 

 light as the isoclinic bands are then absent, and 

 the difference of the principal stresses at a point 

 can be obtained from a colour or a wave-length 

 scale. A direct measure can also be obtained by 

 stressing a simple tension member, set along one 

 direction of principal stress, until the field is 

 reduced to blackness at th.e point desired. This 

 has the advantage of being a zero method, and 

 is simple to carry out with ordinary mechanical 

 appliances. 



The problem of determining accurately the 

 principal stresses separately is, in general, one of 

 some difficulty, and a combined method depending 

 on optical and thermo-elastic properties has been 



' " The Investigation of Stresses in .•> Rpctangul.ir liar by means of 

 Polarised Light." By 1,. N. G. Filon, /'////. Mag., January, 1912. 



NO. 2249, VOL. 90] 



intensity, by dividing the total applied load by 

 the cross sectional area, is inacc:urate and mis- 

 leading. Experiment shows also that models of 

 the standard briquettes of Continental Europe, 

 America and England differ appreciably in their 

 stress distribution curves, and have, in fact, no 

 common basis for the comparison of results. ^ 



For examining models of structures and parts 

 of machines it is usually essential to obtain a 

 field of view in circularly polarised light far 

 beyond the scope of the largest Nicol prisms 

 and quarter wave plates hitherto constructed. In 

 collaboration with Prof. Silvanus P. Thompson 

 these difficulties have been overcome by the con- 

 struction of polariscopes and quarter-wave plates 

 of a size beyond any immediate requirements.* 



One of these instruments is shown in cross- 



r^S!^es in Springs and other Bodies by 

 By E. G. Coker, Brit. Assoc, iqiz, 



2 " The Determination of the 

 Optical and Electrical Methods 

 and EitgineeriH!^, Septeniher 20, 



^ " The Distribution of Stres; 

 Briquette." By E. G. Coker, 

 M.lterials, New York, 1912. 



4 "The Design and Construction of Large Polariscopi 

 E. G. Coker and S. P. Thompson, Optical Convention, Lond 



le Minimum Section of a Cement 

 nternational Congress for Testing 



