[ CLARK] VISCOSITY OF ETHYL ETHER 65 



for there was liquid present in the arm of the tube under observation 

 before the critical temperature was reached. But the liquid filled 

 the tube at a temperature so far below the critical temperature that 

 we may regard the tube as completely filled with a homogeneous 

 fluid, at least homogeneous in the sense that the two branches are 

 alike. Thus, the changes that occur in the branch of the tube under 

 observation are the same as those occurring in all parts of- the 

 tube. This large change shown in Figure 9 indicates that the larger 

 the amount of substance in the tube the greater is the change in the 

 decrement produced by heatiug. In order to test this point directly, 

 the tube of Figure 8 was opened and more liquid introduced. Un- 

 fortunately the tube exploded on heating and the cylinder was ruined. 

 X. When the temperature is raised to about 240° and lowered 

 quickly to near the critical temperature and then held constant, 

 there is an increase in the decrement when this temperature is 

 above the critical temperature and also when it is below. The 

 change seems to be more rapid below the critical temperature. The 

 value which it reached after the lapse of time seems to be the same 

 as it would reach when the tube is maintained at the same constant 

 temperature on the initial rise. 



DISCUSSION. 



The effects noted in II and V may be due to diffusion of the denser 

 part of the substance from the liquid side to the vapour side. It 

 was pointed out by Gouy^ and emphasized by Kuenen^ that near the 

 critical temperature where the isothermals on the Andrews' or pres- 

 sure-volume diagram are nearly horizontal that a very slight increase 

 in pressure causes a relatively great diminution in volume or a corres- 

 pondingly great increase in density. So the weight of the overlying 

 layers of the substance in the tube may cause abnormal density in 

 the lower part of the tube which will not disappear except very slowly or 

 by raising the temperature until the isothermals become steeper. Near 

 the critical temperature then, the density on the liquid side may be 

 abnormally great compared with that on the vapour side and heating 

 to a much higher temperature may equalize the density so that the 

 lower curve for falling temperatures in Figures 4, 6, 7, 8, and 9 might 

 be explained. Also maintaining at constant temperature near the 

 critical temperature may bring about this equalization and make 

 an explanation of the liquid curve of Figure 10 possible. But it seems 

 impossible that diffusion alone will account for this change, particularly 

 when we study IV and VI. 



'Gouy. Compt. rend. 115, 720, 1892. 

 - Kuenen loc. cit. 



