(558 
PROFESSORS A. W. REINOLD AND A. W. RTTCKER 
The ruby glass was then removed and the experiment proceeded with as follows:— 
Films counted.—Tube I., 59. II., 53. 
Position of zero.—Five readings, when the central black fringe was in the standard 
position, gave 
1° 50', 1° 58', 1° 46', 1° 45', and 2° 13' respectively. 
Films counted again.—Tube I., 58. II., 53. 
Position of zero read once more 1° 37'. 
Mean of six readings 1° 52'. 
Films broken in Tube I., then counted 
Tube I., 4. II., 53. 
New position of zero determined five times 
6° 42', G° IF, 6° 58', 7° 12', 7° 15 . 
Films counted and found unaltered. 
Position of zero read once more 7° 10'. 
Mean of six readings 6° 55'. 
Films broken in Tube II.—A slight delay occurred here, as some of the films did 
not break easily. When the operation was completed they were counted again 
Tube I., 4. II., 4. 
New position of zero determined five times 
3° 26', 3° 24', 3° 2', 3° 18', 3° 26'. 
Mean of five readings 3° 19'. 
Hence the zero shifted 6° 55'—1° 52'= 5° 3 = 303' when 54 films were broken in 
Tube I., and 6° 55' — 3° 19' = 3° 36' = 216' when 49 films were broken in Tube II. 
As therefore a =13° 39'= 819', and T = S\/an(p— 1) we get from the first tube (in 
millionths of a millimetre) 
T=303X 615/819 x54x0-4= 10-5, 
and from the second 
T=216X 615/819X49x0-4=8-3. 
Mean value T = 9"4. 
This result is less than usual. During the experiments on the first tube the zero 
was very steady, and the value obtained was almost exactly equal to the mean from 
all the experiments. Before the observations on the second tube, the zero showed a 
tendency to rise, and the slight delay which followed may have given time for a motion 
sufficient to reduce the second value to that actually found. 
The following tables give the results of the experiments. 
The films formed of plain soap solution did not last so well as those made of liquide 
ylycerique. The numbers broken in the experiments are therefore smaller. Several 
