A Determination of Poissons Ratio. 887 



Searle, in his book on Experimental Elasticity, describes 

 -p 

 a method of measuring ^- based on the principle we are 



considering. He attaches metal pointers to the vertical sides 

 AC and BD (fig. 1) of the cross-section of the bar and 

 determines the transverse radius R x from the motion of one 

 pointer over a scale carried by the other. The mean value 

 of a which he obtained for a steel specimen was '285, but 

 the extreme observations from which this value was calcu- 

 lated differed from the mean by as much as 5 per cent. 



I shall now describe some experiments which I carried 

 out in order to determine this ratio for a steel bar of rect- 

 angular section. 



First Method. 



In my first experiments I attached one mirror, M^ to the 

 front face BD of the bar while the second mirror, M 2 , was 



carried by a rod CAFE which was 

 Fig. 2. bent over, as shown in fig. 2, and 



51 F soldered to AC. One mirror was 



—TVi fixed while the other could be 

 w £ rotated about horizontal and vertical 

 axes and could also be displaced 

 parallel to AB. 



The two mirrors were adjusted 

 as as to lie in the same plane, and two images of a scale 

 placed about four feet away could be obtained side by side 

 in the field of the telescope. The bar, the cross-section of 

 which was 2'49 cm. x *495 cm., was supported on two knife- 

 edges placed 40*1 cm. apart. Two equal loads were applied 

 at known equal distances beyond the knife-edges so that the 

 bar became convex upwards. 



In order to measure the longitudinal curvature a mirror 

 was mounted above one knife-edge, and the image of a scale 

 placed at a distance of 70 to 90 centimetres from the mirror 

 was observed by means of a telescope. The curvature 

 measured in this way is, of course, one-half of the total 

 curvature. These observations were checked by placing a 

 second mirror over the other knife-edge. The close agreement 

 in the observed results showed that the arrangement was quite 

 symmetrical, and that one mirror was sufficient. 



To obtain the transverse curvature the scale readings corre- 

 sponding to the mirrors Mj and M 2 were noted in the 

 telescope. The observed displacement is due partly to the 



