[CLARK] VISCOSITY OF ETHYL ETHER 51 



the liquid is heated in a water bath and by an auxiliary heating coil 

 inside the flask and a steady stream of ether vapour is allowed to flow 

 through the apparatus. Then the tube is cooled at E so that the first 

 distillate runs over the joint into the tube F, washing out any deposit 

 that may have been introduced. Then the tube is cooled below F 

 and the liquid condenses and runs down into the experimental tube. 

 By alternately closing and opening the aspirator cock to reduce and 

 raise the pressure inside the tube condensation is facilitated. The 

 distillate is boiled away and collected again repeatedly by alternately 

 immersing the tube in beakers of warm and cold water until it is im- 

 possible to boil the ether with ebulition, showing that the air is nearly 

 all if not quite removed, — indeed, it becomes very difficult to remove 

 the liquid from the tube toward the end of the filling except very 

 slowly. A slight excess over the required amount is collected in the 

 tube and the still cock closed. Further heating removes the excess 

 and the tube is sealed off first at B and then at A. The final sealing is 

 accomplished under reduced pressure so there is no difficulty in 

 accomplishing this. 



The ether used was Kahlbaum's purest ether kept over sodium 

 for several months and distilled into the tube. After the tube is com- 

 plete it is placed in a testing heater and heated to about 240°. If 

 it sustains this test without explosion it is judged good enough for use. 

 Of course if the tube explodes the cylinder is ruined and has to be 

 repolished if not entirely remade. 



PROCEDURE. 



Two lines of work were carried out, either of which supplies 

 sufficient data for the discussion of the problem. But the two lines 

 supplement each other and render the results more trustworthy. 



I. Observations of the logarithmic decrement were made with 

 the cylinder immersed in the liquid for a number of temperatures up 

 to and beyond the critical temperature, and then for falling tempera- 

 tures down to the critical temperature again. The same thing was 

 done with the cylinder immersed in the vapour. If the curves for 

 rising and falling temperatures do not coincide, comparison of these 

 for liquid and vapour may enable us to interpret the changes in the 

 logarithmic decrement shown by them. 



II. The temperature was maintained constant within a few 

 hundredths of a degree with the cylinder in the liquid at different 

 temperatures near the critical temperature. The same work was car- 

 ried out with the tube in the vapour. In this way changes with the 

 time can be studied. By comparing we may be able to discover 



