K W. Morley — Moisture in Gas. 201 



were therefore made by joints ground in a special way. Take 

 for example the connection between c and d. d is to be 

 weighed. The end of d is ground to a cone about fifteen 

 millimeters long. The end of c is ground to a hollow cone 

 about ten millimeters long, so as to leave about five millimeters 

 of the end of d free from contact with c when thrust into it. 

 If now a little fat of suitable consistency is put on the end of 

 c, d thrust into it, and both gently warmed, a joint is made 

 which is absolutely air-tight for small differences of pressure, 

 provided that the layer of fat is not disturbed after solidifica- 

 tion. It will be observed that at the joint between h and *', 

 the difference of pressure is about an atmosphere; but any 

 slight leakage here would only affect the measurement of the 

 volume of air used in the experiment, and would not affect the 

 weight of the weighed apparatus, because a slow current of 

 moist air entering here would be swept along into i by the 

 current of three liters an hour passing out of h during the 

 experiment. Further, direct experiment with such a joint 

 showed a leakage of much less than a cubic millimeter of air 

 in three months. It may be fairly assumed that the diffusion 

 of moisture through the layer of fat in each joint is negligible 

 for the present purpose. In any experiments where this as- 

 sumption ought not to be made, the joint would be surrounded 

 by air kept dry with phosphorus pentoxide, or by a vacuum. 

 It may be added that if such a layer of fat has to be disturbed 

 after solidification, as by turning a plug or stopper, the problem 

 of making an absolutely air-tight joint is very different from 

 that in the present case. 



The weighing of drying tubes with an error of less than a 

 decimilligram is difficult. Two hollow cones were ground 

 to fit each end of the tubes to be weighed ; one was used to 

 connect this tube with other parts of the apparatus, and the 

 second was fused together, making a cap, as shown between c 

 and d. Caps 1 and 2 fitted dd, caps 3 and 4 fitted efgh. All 

 these caps having a suitable quantity of fat applied, 1 and 2 

 were counterpoised against 3 and 4, and the difference of 

 weight determined by careful weighings for a few days. The 

 end of h was withdrawn from *, carefully wiped free from the 

 fat which had made this joint tight, and 1 was put in place; 

 2, 3 and 4, were successively put in place, and the barometer 

 and thermometer were noted. Each tube was carefully washed, 

 and then polished with a dry cloth, but nothing was brought 

 near the fat on the caps. The difference in weight between 

 efgh, caps and fat, and dd, caps and fat, was determined by 

 weighings for several days. The hygroscopic surface of the 

 tube efgh was so nearly equal to that of dd, and the volumes 

 of air displaced by efgh, on the one side, and by dd and the 



