64 THE ROYAL SOCIETY OF CANADA 



IV. Above the critical temperature the vapour side of the tube 

 shows an increase in decrement as the temperature rises and when the 

 temperature falls again the decrement decreases. The curve for 

 falling temperatures is slightly below that for rising temperatures, 

 indicating a smaller decrement or a decrease due to the heating. 

 The amount is not large but there is certainly no increase. 



V. When the tube is maintained at constant temperature above 

 and near the critical temperature, the liquid side shows a decrease 

 in the logarithmic decrement. The same is true for constant temper- 

 ature below the critical temperature. The limits for this change, 

 if such exist, have not been determined. 



VI. When the tube is maintained at constant temperature 

 above and near the critical temperature the vapour side shows at 

 first an increase in the decrement and then a decrease. Below the 

 critical temperature the initial increase is uncertain. 



VII. When the temperature is raised above the critical tempera- 

 ture and lowered again, the decrement for the liquid side decreases 

 to a certain point and then the curve turns abruptly, changes its 

 curvature and runs into the curve for rising temperature well below 

 the critical temperature. The position of the abrupt change is at 

 194° as nearly as can be determined. We may call this the critical 

 temperature. 



VIII. If the temperature be raised to a higher point than in 

 VII the curve for falling temperature is still lower and the point 

 of abrupt change is lower and better defined but at the same temper- 

 ature while the curve runs into the curve for rising temperature farther 

 from the critical temperature. 



Obviously this angle on the curve can be realized only with a tube 

 filled so that it is completely full below the critical temperature, 

 since in other cases when condensation takes place the cylinder will 

 be immersed at different depths as the temperature falls. Of course 

 this may be obviated by slightly heating one leg of the tube to cause 

 the condensation to occur only in the leg in which the cylinder is 

 hanging. But this would introduce great variation in the density 

 which would not become equallized soon enough. 



IX. Figure 8 seems to show that all of the above peculiarities dis- 

 appear when the density of filling the tube is so small that the meniscus 

 vanishes at the bottom of the tube at or near the critical tempera- 

 ture. This suggests that the changes in the decrement depend on 

 the amount of substance in the tube. Study of Figure 5 shows 

 that the change in the decrement on the vapour side of the tube is 

 considerable, and Figure 9 indicates a much larger change. Here we 

 are not dealing with the substance under exactly the same conditions, 



