260 Proceedings of Boyal Society of Edinburgh. [sess. 
During this month I have also been making similar experiments 
with indiarubber of greater thickness. With pieces of indiarubber 
tubing, though I have found a gradual loss of weight in most cases, 
I have not had sufficiently satisfactory results to enable me to speak 
of definite numbers; but with football bladders, about 0'5 mm. 
thick in their distended condition, I find a loss of weight cor- 
responding to an escape of from 0’75 to 3 cubic cm. per sq. 
metre per day. I am sorry I did not try these thicker balloons in 
summer, when higher temperatures and larger differences in weight 
make the results more satisfactory. But, taking and averaging the 
results that are comparable, it appears that rubber twenty times 
thicker than the toy balloons allows almost ^ as much water to pass 
through it ; the loss is thus very far from inversely proportional to 
the thickness of the rubber. This suggests that the resistance to 
the passage of water through indiarubber is probably to a large 
extent a surface phenomenon, whether the obstacle be at the inner 
or the outer surface, or at both ; and that its diffusion through the 
substance of the rubber is comparatively easy. 
I have estimated, from the rough data given me by the hospital 
nurses, that in order to explain the slackening of the water-beds 
actually observed by leakage of water alone, it would be necessary 
to assume a rate of loss about as great as was found to occur from 
the thin balloons in ordinary air during summer ; and seeing that 
these beds are made of rubber at least 1 mm. thick, it seems 
hardly likely that the transudation can be as rapid as this. The 
loss must, however, be rather greater in a water-bed in actual use 
than in a laboratory experiment such as I have made at the same 
air-temperature; for the temperature, and therefore vapour-pres- 
sure, of the water inside is raised a little by the body-heat of the 
patient lying upon it. 
Some escape of water through the indiarubber there certainly 
must be ; and it gives, at all events, a partial explanation of the 
usual behaviour of water-beds. 
I thought it would be interesting to find out at the same time 
whether other liquids could also pass through the indiarubber films. 
I chose alcohol, as it has no sensible action on indiarubber. 
On filling similar balloons with absolute alcohol and methylated 
spirit, I found that the loss was enormously greater than in the case 
