176 



thickness the values of the surface tension calculated by equation (1) are 

 quite concordant. In the article already referred to the author gives re- 

 sults for frames ranging in thickness from .0013 cm. to .02067 cm., the 

 greatest variation being less than six-tenths per cent. Equation (2) 

 gave results with a maximum range of four per cent, the difference be- 

 ing greatest for thick frames. But in practice it is not necessary to use 

 thick frames. In the case of the variation of the surface tension with 

 temperature all the measurements may be made with a single frame. In 

 this experiment the frame was .0102 cui. thick and 6.642 cm. long. 



k\.\\\\V\\\\V<\\\\\<\V\\\^ 



TiQ.Z. 



Fig. 2 shows the arrangement of the apparatus for measuring the 

 temperature coefficient of the surface tension of water between room 

 temperature and 80°. A mica frame F was hung on a wire W attached 

 to one arm of a balance — sensitive, at this load, to .002 mg.. The balance 

 itself rested on a wooden box shown in section. This box had a door in 

 front (practically air tight) with a double glass window through which 

 the apparatus inside could be seen and tlie thermometers read. The 

 wooden box enclosed a double walled tin box or tank T^ with walls about 



