I CLARK] VISCOSITY OF ETHYL ETHER 67 



If we grant that the change in the decrement above the critical tem- 

 perature is due in part to a change in the substance, to what shall we 

 attribute it? The theories of Traube, de Heen, etc., seem to apply 

 here even though their validity has been attacked by such weighty 

 authorities. Apparently the objections raised are, that the anomalies 

 observed by all of the experimenters whose work shows them are due 

 either to the effects of gravity or the presence of impurities. We 

 may grant the first for the liquid side of the tube at least; but we cannot 

 do so for the vapour side, as there is no permanent increase in the 

 viscosity in that side . Apparently the effect of impurities is to empha- 

 size the effect of gravity, that is, to increase the time for the equaliza- 

 tion of the densities. Also the value of the critical temperature would 

 be changed. But when heated to 240° and then lowered again until 

 condensation occurs we find the same amount of liquid in both legs 

 of the tube, showing that we had identical substances in both legs. 

 Also, we find a decrease in decrement occurring for the vapour side 

 where there should have been an increase if the material had not been 

 identical in the two legs of the tube. 



If chemical change is responsible for the change in the viscosity 

 we should expect to find it permanent and then we would have a 

 different value if the cycle of readings around the critcial point be 

 repeated; but we find no such difference. If chemical dissociation 

 were the cause of the changes in viscosity observed, they should appear 

 in Figure 8, where they are noticeably absent. Thus the objections 

 to thej:lassical theory seem to be sustained by the foregoing investi- 

 gation, at least in its present stage. Further study is contemplated. 

 Certainly the liquidogenique and gazogenique -theory furnishes the 

 most reasonable explanation of all of these experimental results. 



Apparently the change below the critical temperature is confined 

 to a limited region near that temperature. Its lower limit seems to 

 depend on the highest temperature reached before cooling, and is very 

 near the critical temperature unless the substance has been heated 

 well above the critical temperature. Thus we must conclude that the 

 amount of this change at any distance from the critical temperature is 

 v^ery small indeed, so that at ordinary temperature the process of 

 evaporation is not attended with this physical dissociation of the 

 molecules. 



In conclusion I have to express my thanks to the Rumford 

 Committee of the American Academy of Arts and Sciences for financial 

 assistance in carrying on this investigation. 



Physical Laboratory, 

 Queen's University, 



Kingston, Ont. 



