﻿8 Prof. L. N. G. Filon : Investigation of Stresses 



are given in Table I. below. The bar of glass used was a 

 bar of thick plate glass, made for me by the London and 

 Manchester Plate Glass Company, St. Helens, and was 

 practically free from imperfect annealing. The height, 2b, 

 was 3' 60 cm. 



Table I. 



y- 



W-if. 



2S/E. 



2S /K-(^-/-). 



Percen tage 

 difference. 



1-6 



•68 



•75 



+ •07 



10 



14 



1-28 



1-33 



+ •05 



4 



1-2 



1-80 



1-85 



+ •05 



3 



1-0 



2*24 



2-19 



-•05 



2 



•8 



260 



2-68 



+ •08 



3 



•6 



2-88 



2-85 



-•03 



1 



•4 



3-08 



2-83 



-•25 



8 



- 4 



3-08 



2-88 



-•20 



7 



- -6 



2-88 



284 



-04 



1 



- -8 



2-60 



2-49 



-•11 



4 



-1-0 



2-24 



2-20 



-•04 



2 



-1-2 



1-80 



1-79 



-•01 



1 



-1-4 



1-28 



1-25 



-03 



2 



-1-6 



•68 



•59 



-•11 



16 



Large errors occur for numerically small values of y, and 

 no reliable measurements were found to give 2JS/K for small 

 values of y. This, however, is to be expected, since when y 

 is small, tan 2/ is very great and formula (7) is liable to large 

 errors whenever a small absolute error is made in y. Large 

 percentage errors also occur near the ends, but this is prin- 

 cipally due to the smallness of the quantity on which the 

 percentage is taken. A noticeable point is that the errors 

 tend to be in a different sense on opposite sides of the neutral 

 axis. Any small error in the determination of the neutral 

 axis might easily account for this. Altogether, if the dif- 

 ficulty of measuring exactly bands of this nature be borne in 

 mind, the results give, on the whole, very good confirmation 

 of the accepted theory. 



§ 5. Experimental verification of the Bernouilli-Euler 

 Theory of Flexure for a Beam under a pure couple. 



The apparatus described in § 2 allowed of the bar being 

 strained under a pure couple. This was done by removing 

 the central load altogether and suspending weights from the 

 projecting ends. 



It seemed of some interest to verify that in this case the 

 assumptions of the Euler-Bernouilli theory are satisfied, viz. 



