1897-98.] The Passage of Water through Indiarubber Films. 261 
of water. Here, however, it did not remain constant, as with the 
water balloons, but showed a steady and somewhat rapid diminution 
at each weighing. It was influenced in the same direction, but to 
a very much less extent, by exposure to dry and to moist air ; that 
is to say, it was increased a little in dry air, and diminished in moist 
air (see PI. fig. III., and Tables K. and M.). The clue to this is 
probably given by an interesting and unexpected result. When a 
water and spirit balloon were placed together in the moist chamber, 
the water balloon not only did not lose weight, but actually gained 
slightly; and on one occasion, when one of the spirit balloons 
happened to burst in that chamber, the gain of the water balloon 
beside it was very considerable (fig. I. B., and Table B.). Similarly, 
no doubt, when a balloon containing alcohol is placed in an atmos- 
phere containing water-vapour, water-vapour passes through the 
film in consequence of the lower pressure of water-vapour within it. 
In a moist atmosphere this effect must be increased, and in a dry 
one abolished or reversed. 
Taking the first observations in the case of each balloon, the 
figures, reckoned in the same way as for water, are as follows : — 
Extreme rates of loss. 
For alcohol, .... 369/x and 516/x 
For methylated spirit, . . . 339/x and 497 g 
— that is to say, from 15 to 30 times as great as the loss of water 
under similar conditions. 
It was noticed in the course of the experiments that while the 
outside of the balloons was free from pigment at the beginning of 
the observation, stains of the same colour as that with which the 
balloons were dyed frequently appeared on the paper on which 
they were standing ; and sometimes there was sufficient colouring- 
matter on them to soil the fingers. I therefore determined to try 
whether any diffusion of salts in solution in water took place. I 
filled bladders with strong solutions of magnesium sulphate and 
ammonium chloride. The former speedily began to leak, appar- 
ently from some chemical action of the sulphate on the indiarubber ; 
but the latter showed no leak, and in ordinary air lost weight just 
like the water-bladders. They were then immersed in jars of 
water, and there gained weight slightly but steadily; while the 
