Obfervations cn Electricity* 2^3 
of the cylinder, without touching it. The fparks at the elec- 
trometer were four times as numerous when this metallic piece 
was thus projected; but no electricity was obferved to pafs 
between it and the cylinder. The metallic piece was then 
held in the hand to regulate its diftance from the glafs ; and 
it was found, that the fparks at the electrometer increafed in 
number as it was brought nearer, until light appeared between 
the metal and the cylinder, at which time they became fewer 
the nearer it was brought, and at laft ceafed when it was in 
contaCt. 
The following conclufions appear to be deducible from thefe 
experiments. 1. The line of light on a cylinder departing 
from a fimple cufhion coniifts of returning eleCtricity ; ,2. 
the projecting part of the cufhion compenfates the eleCtricity 
upon the cylinder, and by diminifhing its intenfity prevents its 
ftriking back in fuch large quantities as it would otherwife do ; 
3. that if there were no fuch compenfation, very little of the 
excited eleCtricity would be carried off; and, 4. that the com- 
penfation is diminifhed, or the intenfity increafed, in an higher 
ratio than that of the diftance of the compenfating fubftance ; 
becaufe if it were not, the eleCtricity which has been carried 
off from an indefinitely fmall diftance, would never fly back 
from a greater diftance and form the edge of light. 
22. I hope the confiderable intenfity I fhall fpeak of will be 
an apology for deferibing the manner in which I produce it. 
I wifh the theory of this very obfeure procefs were better 
known; but no conjecture of mine is worth mentioning. The 
method is as follows: 
Clean the cylinder, and wipe the filk. 
Greafe the cylinder by turning it againft a greafed leather till 
it is uniformly obfeured. I ufe the tallow of a candle. 
Turn 
