DISRUPTIVE DISCHARGE OF ELECTRICITY. 637 
the observations made. The mode of treating the observations was as fol- 
lows :—I plotted the arguments and entries of each series of observations on 
section paper, and drew the curve which best satisfied these points of observa- 
tion. After comparing all the curves so obtained, which related to one question, 
I found an equation, 
¥ =S(2); 
which satisfied a curve to a first approximation, and calculated out the 
values of 7’ for all the points of observation ; then since 
y = yee), 
where y denotes the most probable observed value, 
y 
$(z) was obtained by plotting : on an enlarged scale, drawing a curve by the 
graphic method, and finding an equation for that curve. | 
Measurement of the Difference of Potential required to pass a Spark through 
Air at the Atmospheric Pressure between Parallel Metal Plates at Dif- 
Jerent Distances. 
This problem has been investigated for small distances by Sir W1ILLIAM 
THOMSON ; and it was from a perusal of his paper on the subject, in 
“Papers on Electrostatics and Magnetism,” and the discussion of these results 
in CLERK MaxweEtv's “ Electricity,” that I was led to take the above subject for 
an experimental inquiry. Tables IIL, IV., V., VI., VIL, VIIL, and figs. 1, 2, 
3, give the observations and inferences on this problem. The large discs men- 
tioned as forming the electrodes were each of brass, 4 inches diameter, well 
rounded at the edge, the one having its face flat, the other slightly convex. 
A spherometer, the side of the base of which is 2°5 inches, gave the height of the 
convex segment through its feet to be ‘05 centimetre. The convex face 
belonged to the lower electrode. The diameter of the cylindrical part of the 
receiver used is 19 centimetres. The constants were the same for the five 
curves (Tables III. to VII.), excepting that the pressure and the temperature 
may have been slightly different. The maximum deflection could always be 
observed with precision, and the differences in the values for the same length 
of spark were probably due in great part to the spark not always passing 
exactly in the axis of the discs. The zero was always read after the second 
system of conductors had been completely discharged, which was always done 
after reading the maximum deflection. The values in the column headed Mean 
Difference of Potential are not always the mean of the values in the preceding 
_ column; but when not such, they are values got by concluding from inde- 
| pendent evidence that greater weight ought to be attached to some of the 
entries than to the others. The entries to which greater weight has been 
VOL. XXVIII. PART It. 8 C 
