424 On the Mechanical Stretching of Liquid*. [Feb. 4, 



centrifugal method devised by Osborne Reynolds, (iii) the method of 

 cooling discovered in 1850 by Berthelot, and pointing out that the 

 fi -st two afford means of measuring stress but not strain, while the 

 third gives a measure of strain but not stress, the author proceeds to 

 describe the manner in which he had used the method of Berthelot in 

 combination with a new mode of determining the stress, and had 

 succeeded in obtaining simultaneous measures of tensile stress and 

 strain for ethyl alcohol up to a tension of more than 17 atmospheres, 

 or 255 Ibs. per square inch. 



The liquid, deprived of air by prolonged boiling, is sealed in a strong 

 glass vessel, which it almost fills at a particular temperature, the 

 residual space being occupied only by vapour. On raising the tem- 

 perature, the liquid expands and fills the whole. On now lowering 

 the temperature, the liquid is prevented from contracting by its 

 adhesion to the walls of the vessels, and remains distended, still 

 filling the whole and exerting an inward pull on the walls of the 

 vessel. The tension exerted is measured by means of the change in 

 capacity of the ellipsoidal bulb of a thermometer sealed into the 

 vessel and called the ** tonometer." This bulb becomes slightly more 

 spherical, and therefore more capacious, under the pull of the liquid, 

 and the mercury in the tonometer stem falls. The tension correspond- 

 ing to the fall is previously determined from observation of the rise 

 produced by an equal pressure applied over the same surface. 



The liquid is caused at any desired instant to let go its hold and 

 spring back to the unstretched volume corresponding to its tempera- 

 ture and to its saturated vapour-pressure by heating for a moment, by 

 means of an electric current, a fine platinum wire passing transversely 

 through the capillary tube that forms part of the vessel. The space 

 left vacant in the tube represents the apparent extension uncorrected 

 for the yielding of the glass vessel. 



The measures obtained show that, within the limits of observa- 

 tional error, the stress and this appparent strain are proportional up 

 to the highest tension reached (17 atmospheres) ; but, since the small 

 yielding of the nearly rigid glass vessel must itself be proportional to 

 the stress, it follows that the stress and absolute strain are propor- 

 tional. 



By subjecting the liquid to a pressure of 12 atmospheres in the same 

 vessel, it was found that the apparent compressibility was the same as 

 the apparent extensibility, whence it is deduced that between pres- 

 sures of +12 and 17 atmospheres the absolute coefficient of elasti- 

 city is, within the limits of observational errdr, constant. Its actual 

 value is best obtained by observations of compressibility. 



The paper concludes with a description and explanation of a 

 peculiar phenomenon of adhesion between two solids in contact when 

 immersed in a liquid that is subjected to tension. 



