640 ALEXANDER MACFARLANE ON THE 
smaller diameter—1°‘5 inch. The value of a, which is ‘28433 centimetre, is greater 
than the value obtained with the large discs. But observations with the small — 
discs have not been made for as many lengths of spark as is desirable. : 
Table X. gives the results for the same conditions as in Tables III. to VII., — 
excepting that the density of the air is much less. The equation obtained by 
neglecting the higher powers of s is 
V =18-292,./{s? + 5232251, 
which gives a='26161, and b=4°7854. Thus a is increased and 6 decreased. 
This leads us to infer that when the pressure is diminished the condensed 
stratum is also diminished—a conclusion which is supported by a peculiarity in 
Table X XIII., where, while the mean value for air before any exhaustion has 
been made is 152, the mean value after exhaustion thrice given is 142. @ has 
always been found to be greater the smaller the pressure. The distance at 
which the spark was first observed to pass from the edge was 1°65 centimetre, 
compared with 1 centimetre for the ordinary pressure. 
Table XI. (fig. 4) gives the results obtained when hydrogen was substituted — 
for air. To effect this substitution I heated the brass discs thoroughly before 
the fire ; then having screwed them on, I exhausted the air and let in hydrogen, 
repeating this process twice ; but before taking any readings I left the hydrogen 
in for twenty-four hours. The curve through the readings is precisely similar 
to that for air, and the spark began to pass from the edge at very nearly the 
same distance. The equation obtained by neglecting the higher powers of s is 
V =43:190,/{s’+ 13690s', 
giving a to be ‘06845 centimetre. The ratio of the intercepts for hydrogen and 
for air is thus ‘66, which is not very different from the ratio of their electric 
strengths *—-63 given in column 6th of Table XXIII.—and agrees exactly 
with the value given in column 7th. I made only one determination of the 
absolute value of the difference of potential required to give a spark through 
hydrogen under the standard conditions. The value obtained, 12:1, is very 
nearly half of that obtained from the value for air by means of the ratio of the 
electric strengths of the two gases. I have provisionally taken the latter value. 
The necessity for leaving the hydrogen in for a length of time before taking the 
readings becomes manifest when the above table is compared with Table XIL., 
which gives the results when the observations were taken immediately after the 
gas was put in. The observations are plotted on fig. 10. The curve given by 
the equation 
V =53°940s —12:9316s", 
which represents a parabola, runs well through all the observed values. There 
* We have the high authority of Professor CLerK Maxwett for using the term the electric strength of 
the gases to denote what has sometimes less properly been called their insulating power. 

