722 



NATURE 



[February 24, 19 16 



A convenient j-naterial to interpose is a thin sheet 

 of india-rubber, and a photograph of the colour 

 effects clearly shows the widely different char- 

 acter of the stress distribution. The stress at 

 the centre line is now the greatest, and it is, more- 

 over, artificially increased 20 per cent, or thereabouts, 

 for the same total load owing to the action of the 

 interposed film of india-rubber, as the measurements 



y/////////////////////////////// 



ELEVATION 



Fig. s. 



show. It is, moreover, not a local effect confined to a 

 small area at the ends, but exerts an effect over the 

 greater part of the block. 



The injurious effect of a lead plate, when applied 

 to give a bearing over a block of material under 

 test, is confirmed by this experiment, and the measure- 

 ments give a quantitative estimate of the increase in 

 ■stress intensity. 



NO. 2417, VOL. 96] 



To ensure uniform stress conditions, the expn 

 mental results apjjear to point to the conclusion th;i 

 the load ought to be applied through an intermedial 1 

 layer of the same material, and when this is carritd 

 out, the block shows a very nearly uniform optic, ii 

 effect, and the measurements prove that the stn - 

 intensity is nearly uniform througliout. 



Eye-Bars. 



A promising development of experimental woi 

 with polarised light, relates to the design of machin< 

 and structures, especially the detailed parts or coi: 

 ponents. It has already been pointed out that in il 

 majority of cases the stresses in even the simpU - 

 members are so complicated as to defy exact calcuhi 

 tion, and with the simplifying assumptions usually 

 adopted, it is necessary to allow for faulty methods 

 by using a large factor of safety. 



We may take a very simple example in the case of 

 a member which is merely required to transmit a pull 

 in the direction of its length by means of pins, as 

 shown in the diagram. 



If we take a member of rectangular form bored 

 out at each end to receive the pins, then it is at once 

 apparent that the material around the pin is very 

 highly stressed compared with the body of the mem- 

 ber, and, as we have already seen in the case of the 

 rivet, there is a compression stress of very large 

 intensity in one region, and a tension stress of con- 

 siderable magnitude in another, but in the main body 

 of the member, the section is excessive for the load 

 applied. We have here, in fact, the problem which 

 confronted the Deacon in Oliver Wendell Holmes's 

 story of the wonderful " One Hoss Shay," which "ran 

 a hundred years to a day." 



" Fur," said the Deacon, " 't's mighty plain 

 Thut the weakes' place mus' stan' the strain ; 

 'n' tb« way t' fix it, uz I maintain. 



Is only jesi 

 To maka that place uz strong uz the rest." 



In practice, therefore, the superfluous material of the 

 main body is cut away, and a link is obtained with 

 swelled ends, shaped to resist in the best possible 

 way the stresses which come upon them. The form 

 which these ends shall take to ensure the maximum 

 possible strength is a problem which has exercised 

 the minds of many eng|ineers, particularly thosfe( 

 engaged in the construction of large span-bridges of 

 the pin-connected type, where such members occur 

 in considerable numbers and of great size. 



Some of the forms of ends which have been 

 very generally used are shown on the dia- 

 gram, and their diversity indicates the un- 

 certainty which is felt as to the best possible 

 shape; nor is it easy to devise a method of 

 selection unless the stresses in these forms can be 

 measured. 



An optical investigation of a model does this effec- 

 tively and quickly. A very common form is that in 

 which the swelled end has a contour bounded by a 

 circle concentric with the pin, and if we use propor- 

 tions very widely adopted, it is easy to see from the 

 colour effects on the loaded model now shown, that 

 the stress distribution is unsatisfactory; in fact, at 

 the principal transverse section (Fig. 6) the normal 

 stress is that of compression at the extreme edge, 

 and undue stress is therefore thrown upon the re- 

 mainder of the section, to balance the total pull on 

 the member. This is also indicated by the lines of 

 stress which have been plotted from the observations. 

 A much better shape is one devised by Berkeley, 

 and the reason for its special excellence is, I think, 

 clear from the picture of its stressed condition which 

 is before you, from which you will see that the head 



