602 
PROFESSOR GRAHAM ON THE MOTION OF GASES. 
The velocity with which air passed into a vacuum was, by C ! inch, 1 cubic inch 
in 21 '26 seconds ; by C 2 inches, 1 cubic inch in 35'43 seconds ; and by C 4 inches, 
1 cubic inch in 70 seconds. 
C 1 inch exhibits (Table XIV.) great variation in its rate of transpiration at differ- 
ent pressures, as was to be expected. At the head of the scale, the rate of air is 
0'9138, of hydrogen 0'3967, and of carbonic acid 0*8725 ; which all indicate a great 
interference of effusion. At the bottom of the scale, on the other hand, where the 
pressure is small and the resistance to passage consequently great, the rate of air is 
0*8905, of hydrogen 0*4374, and of carbonic acid 0*7529, or nearly normal. 
With C 2 inches (Table XV.), the rates of air are pretty uniform at different 
parts of the scale, and sufficiently normal ; the deviation from 8 to 4 inches appears 
to be an accidental anomaly. The rate of hydrogen also is never distant from 0*4375, 
except at the very bottom of the scale. This appears to be the length of a tube of so 
small a bore which gives the most uniform results at different pressures : 
For with C 4 inches (Table XVI.), where the resistance is excessive, air has a rate 
corresponding sufficiently with its specific gravity at the head of the scale, but 
diverging rapidly in the lower part of the scale, the time of transpiration becoming 
rapid as compared to that of oxygen. With a tube then like the present, the relation 
between the times of transpiration is only to be looked for within a limited range, 
and that at a high degree of pressure approaching to a whole atmosphere. 
4. Transpiration by Capillary Tube H. 
As the existence of any simple numerical relation among the gases, such as the 
preceding experiments on transpiration render probable, would be a point of funda- 
mental importance in their history, I was induced to try new capillary tubes and to 
multiply experiments, varying the circumstances in which they were made, and 
taking additional precautions against the interference of disturbing causes. The 
arbitrary nature of the coefficients of transpiration indeed produces a more than 
usual necessity for strong evidence and numerous confirmations, as the numbers 
themselves have no a priori probability in their favour. 
It appeared desirable to try capillary tubes of larger diameter than those already 
employed, and of great length; partly to vary the conditions of the experiment, and 
partly because extreme shortness of tube appears, in the experiments made with C, 
to be unfavourable to uniformity of rate at different pressures. 
Several portions of capillary tube, of which the bore was as nearly equal as could 
be judged of by the eye, were accordingly selected and cemented together at the 
blowpipe, so as to form a continuous tube 22 feet in length, which was bent up into 
coils for convenience in using it, as represented in figure 4, Plate XXXIII. The 
extremities of this capillary were connected with the block-tin tubes proceeding 
from the drying tube and air-pump jar respectively, by means of thick caoutchouc 
adopters, which diminished the rigidity of the arrangement and protected the glass 
tube from the effect of the shocks, which are unavoidable in working the air- 
