Velocity of Sound in Air. 449 
also was not registered. The method is beautifully adapted 
for comparative rather than for absolute results. 
Table I. 
Velocity of Sound in Tubes, in metres per second, at 0° C. 
(Kundt.) 
Diameter 
of tube. 
Velocity. 
Difference 
millim. 
metre. 
metre. 
55 
332-801 
332-73J 
•07 
26 
329-47 1 
323-00 1 
2-26 
13 
6-47 
6-5 
305-42 ) 
7-58 
3-5 
Le Roux's method consisted in employing a U-shaped tube, 
- 07 metre (70 millim.) diameter, closed by a membrane at each 
end. One membrane was tapped with a small beater, and the 
time occupied by the resulting wave in travelling between the 
two membranes, as indicated by the disturbance of the second 
one, was registered. It appears to me, however, that the em- 
ployment of membranes may introduce a source of error in this 
way: — Let A B be a tube closed by a rigid material at one 
end, and of a length to give the 
maximum resonance to a quarter- 
wave. Now, instead of the rigid 
end, let the tube be closed by a 
membrane : this will require to be in the position C, i. e. nearer 
to A than B is, the exact position depending upon the tension 
of the membrane. In Le Roux's experiments, unless the two 
membranes were of exactly the same tension, a source of 
error would be introduced. 
We may now turn to RegnauhVs experiments, a summary 
of which is here given. 
Table II. 
ity of Sound in 
Tubes, in metres 
(Regnault.) 
per second, at 0° C 
Diameter 
Velocity 
Velocity 
of tube. 
(mean). 
(limiting). 
millim. 
metres. 
metres. 
1100 
330-5 
330-3 
300 
330-3 
329-25 
108 
327-52 
324-25 
Free air. 
330-71 
330-60 
Phil. Mag. S. 5. Vol. 16. No. 102. Dec. 1883. 2 I 
