[CLARK] VISCOSITY OF ETHYL ETHER 45 



impure liquid which will behave differently in different parts of the 

 tube. This may explain other abnormalities. So it may be that 

 these molecular theories are unnecessary; but it may be also that 

 there are abnormalities which are not accounted for in either of these 

 two explanations. It occurred to the writer that a different line of 

 experiment might contribute data for the settlement of the question 

 and the following paper is the account of an attempt to gain such data. 

 So far as the writer is aware there have been no direct attempts to 

 demonstrate the existence or non-existence of the two kinds of mole- 

 cules. 



The most obvious lines of investigation are experiments in which 

 molecular actions are brought into play, as in viscosity determinations 

 and investigations of optical properties. The work of Galitzine and 

 Wilip' deals with the latter subject but not in a way to give much 

 help in the present instance. Young^ has investigated the question 

 of opalescence but there is not sufficient data accumulated to decide 

 the question, although the evidence given is in favour of the classical 

 theory. It seemed to the writer that since viscosity is essentially a 

 molecular phenomenon and that the value of the coefficient depends 

 on the mass of the molecules, the determination of the viscosity 

 coefficient under proper conditions might lead to valuable information, 

 so the following investigation was undertaken. The course of the 

 work has been subject to annual interruptions with the beginning 

 of the teaching session; but the apparatus and technique have been 

 perfected gradually. Of all the methods^ available for the determina- 

 tion oTthe viscosity coefficient that of measuring logarithmic decrement 

 of an oscillating cylinder suspended inside a concentric cylinder 

 seemed most practicable, and the results have justified the choice 

 of this arrangement. 



If such a suspended cylinder be set in torsional vibration, the 

 logarithmic decrement can be determined and from the change in the 

 decrement the changes in the viscosity and therefore in the molecular con- 

 dition can be studied. In these experiments, the outer cylinder of the 

 system is the glass tube containing the fluid, and the oscillating cylinder 

 is described below. In choosing a substance for investigation it is 

 necessary to select one with a moderately low critical pressure, one 

 which is fairly easy to prepare, and one whose critical temperature 

 is neither extremely high nor extremely low. Ethyl ether was chosen 



1 Galitzine and Wilip-. Rapports du Congrès International de Physique, de 

 1900, t. I, p. 675. 



2 Young, F. B. Phil. Mag. (6), 20, 793, 1901. 



' See Brillouin, La Viscosité des Liquides et des Gaz for an excellent discussion 

 of the different methods of determining the coefficient of viscosity and the results 

 obtained. 



