442 



Prof. A. W. Remold and A. W. Riicker. [May 27, 



is called the limiting sensitiveness. In any experiment the observed 

 value of dh divided by the sensitiveness gives the fractional change of 

 surface-tension. 



The numbers expressing the sensitiveness indicate the relative 

 merits of the different experimental arrangements. Some of the 

 authors' experiments were more sensitive than those of any previous 

 observer, but the possibility of improvement in this direction is 

 limited by the fact that if the arrangement is too sensitive the films 

 do not remain sufficiently steady for accurate measurement. 



The apparatus and method of observation are then described. 

 Preliminary experiments were instituted to test the observations of 

 Liidtge and van der Mensbrugghe as to the difference of surface- 

 tension between two films one of which had been formed more 

 recently than the other. These experiments showed that when one 

 of the films was kept thick by supplying liquid to its upper support 

 (Hooding), while the other was allowed to thin, a considerable 

 apparent difference of surface-tension was obtained. Before, how- 

 ever, this could be accepted as a trustworthy determination of an 

 actual difference of suriace-tension, several possible sources of 

 error had to be considered. Thus, experiment shows (1) that the 

 fact that the thicker film displays the greater surface-tension cannot 

 be attributed to any peculiarity of the apparatus or mode of thicken- 

 ing adopted ; and (2) that it is not due to the weight of the thicker 

 film. 



It was also important to determine whether the gradual disappear- 

 ance of the liquid rings by which the films were attached to their 

 solid supports could produce changes in the forms of the films which 

 would account wholly or partially for the phenomenon observed. 

 Two films of the same surface-tension attached to two rings of 

 different sizes could not both be cylinders, as the curvature and there- 

 fore the pressure exerted on the internal air would be different. A 

 difference in the forms of the films might thus be due, not to the 

 difference of surface-tension but to an inequality in the magnitude of 

 their supports. The solid rings were accurately turned to the same 

 diameter, but it was thought possible that the draining away of the 

 small liquid masses by which the thinning film was attached to the 

 solid rings might produce alterations equivalent to a change in the 

 diameters of the latter. To test this the following method was 

 adopted. 



Theory shows that provided the form of the film does not differ 

 much from a cylinder, the ratio dT/T can be readily obtained by 

 means of an expression which does not involve the diameters of the 

 supporting rings. To make use of this expression, it is necessary to 

 measure three diamjters of the film. The ratio dT/T can also be 

 calculated from measurements of the principal ordinate (i.e.> the 



