174 DR. E. P. PERM AN, DR. W. RAMSAY, AND MR. J. ROSE-INNES, 
the experimental tube which was inside the pressure apparatus, and the result was 
that the ether was almost invariably increased somewhat in amount at the higher 
temperature. Thus, in the above case, calculation of weight at 130° G., 140° C., 
and 150° C., gave the following numbers as means :—0'03246, 0'03387, 0'03174, 
the individual results agreeing as closely as those in the preceding table. Each 
temperature was therefore treated on its merits, and unless no alteration in weight 
had been proved to occur on changing temperatui'e, the calcadations were made 
on the basis of measurements at different volumes and pressures at the same 
temperature. 
III. Determination of the Frequency of the Yibrating Rod. 
Ten sets of experiments were made in order to ascertain accurately the wave¬ 
length in air of the note given bj^ the glass rod. The mean result of each is 
here given ; the temperature was 15'5° C. 
L IT III. IV. V. VI. VII. VIII. IX. X. 
19-4 19-5 19-2 19-2 19-3 19-44 19-22 19-71 20*05 19*33 
Mean 19*46 millims. 
From these results the number of vibrations per second of the rod may be calcu¬ 
lated by the formula 
n = V/\. 
The velocity of sound in air was taken as 33253 centims. per second at 0° C. 
(‘ Encyc. Brit.,’ Article Sound). The wave-length at 0° is 18'93 millims. Hence 
n — 17,566 per second. The note was a shrill squeak. 
The close agreement of the values for the wave-length quoted above seems to 
show that the accidental sources of error are here only trifling; and they may 
be considered as practically eliminated from the final mean value. But it is still 
possible that there may be some constant source of error affecting all the experiments 
equally ; thus it is well known that the wave-length in gases is affected by the 
diameter of the tube, and this effect may become quite noticeable if the tube is 
narrow. It is not likely that such an effect really existed in the above experiments, 
for though the tube was rather narrow, the pitch of the note was exceedingly high. 
But it was thouglit as well to make quite cei'tain, and a set of experiments was 
carried out to test the accuracy of the value of n obtained. 
(1.) Confirmatory Experiments until Hydrogen. 
The velocity of sound in hydrogen has been made the subject of investigation by a 
large number of experimenters, some of vdiom used tubes sufficiently wide to obviate 
