1918] on Studies on Liquid Films 891 



lating a separate lighfc glass ring, on which small weighed pieces of 

 steel wire could be hung. A graduated pull could be exerted by a 

 small solenoid below. Tiny horizontal magnets were attached to the 

 columns, and by their tendency to set in the meridi m were able to 

 retard irregular oscillations. For this purpose they were fixed to 

 hght cotton rings by being soaked in nitro-cotton solution and dried. 



A coUimn of six segments each of 150 c.c. was elongated by one- 

 third of its length with a weight of 0* 186 gram ; the original length, 

 free of superfluous liquid, being 81 '0 cm., and the maximum elonga- 

 tion, 1 1 cm. ; the smallest additional weight then broke the column. 

 The same weight, hung on a column of five bubbles 23 cm. long, 

 caused an elongation of 8 • 3 cm., or 35 per cent. On the removal of 

 a drop of liquid weighing six centigrams a contraction of 2 '3 cm. 

 followed. 



A column of three segments gave the following result : with a 

 weight of '075 gm. its initial length was 14*7 cm., and the elonga- 

 tion produced was 1*85 cm. 



The following values were obtained from the gradual increase of 

 load when a loss of buoyancy occurred l)y diffusion of hydrogen from 

 a column of eight equal segments of 1 50 c.c. blown with 7 per cent, 

 hydrogen. The initial length of 85 cm. was increased to 89 * 5 cm. 

 when a weight of 7*5 centigrams was hung on. "When the hydrogen 

 had diffused out the total length became 40 • 3 cm., or an extension 

 of 0'<s cm. from an additional weight of 8*4 milligrams. These 

 results are recorded in the following table (I.). 



Table I. 



Where n is the number of segments in the column ; / is its length 

 in cms. ; iu is the attached weight in gms. ; e is the elongation in 

 cms. ; X is the elongation per unit length of column ; and E is the 

 corresponding value of Young's modulus. 



It will be noticed, on comparing the fourth set of these ob- 

 servations with the first set, that (roughly) halving the weight and 

 length reduced the elongation per unit length to one-fourth, as might 



be expected. For, a; = ^ and W = Ey, hence for the first case 

 YOL. XXII. (Xo. 112) 2 E 



