390 



Mr. T. Proctor Hall on New Methods of 



None of the methods in use seem to exceed in general 

 accuracy or convenience the ordinary plan with capillary 

 tubes, the limitations of -which have been already referred to. 

 The direct dependence of surface-tension upon molecular 

 cohesion, and its consequent importance as a reliable indication 

 of changes of condition in liquids, makes it very desirable that 

 some more accurate plan of measurement should be devised. 



II. New Methods. 



Two years ago, at the suggestion of Professor Michelson 

 and under his direction, I undertook the investigation of a 

 direct method which, from some preliminary experiments, he 

 expected to give results ten times as accurate as those found 

 by any method now in general use. 



This method consists in finding the weight of a thin hori- 

 zontal bar, suspended a few millimetres above the liquid sur- 

 face, whose extremities are bent at a right angle so as to dip 

 into the liquid. The weight is taken first when a film of the 

 liquid is extended between the bar and the level surface, and, 

 secondly, in exactly the same position after the film is broken. 

 The difference between the two weights divided by twice the 

 width of the film gives the surface-tension in grams per centi- 

 metre, which is reduced to dynes per centimetre when 



Fiff. 1. 



multiplied by the constant g, whose value at Clark University 

 is 980-3. 



For nearly all ordinary liquids a glass rod is suitable, but 



