Obfer vat ions on Electricity. 267^ 
between the liik and culhion in weak excitations does not con- 
lift of returning eleCtricity, but of eleftricity which paftes to 
the cylinder, in confequence of its not having been fufficiently 
fupplied, during its contact with the rubbing furface. 
11. When the excitation was very ftrong in a cylinder 
newly mounted, flalhes of light were feen to fly acrofs its 
infide, from the receiving furface to the furface in contact with 
the culhion, as indicated by the brufh figure. Thefe made the 
cylinder ring as if ftruck with a bundle of frnall twigs. They 
feem to have arifen from part of the ele&ricity of the cylinder 
taking the form of a charge. This appearance was obferved in 
a nine-inch and a twelve-inch cylinder, and the property went 
off in a few weeks. Whence it appears to have been ch iefiy 
occafioned by the rarity of the internal air produced by hand- 
ling, and probably reftored by gradual leaking of the cement. 
12. With a view to determine what happens in the infide of 
the cylinder, recourfe was had to a plate machine. One 
culhion was applied with its filken flap. The plate was nine 
inches in diameter and two- tenths of an inch thick. During the 
excitation, the furface oppofite the culhion ftrongly attracted 
electricity, which it gave out when it arrived oppofite the ex- 
tremity of the flap. So that a continual ftream of electricity 
palled through an infulated metallic bow terminating in balls, 
which were oppofed, the one to the furface oppofite the extre- 
mity of the lilk, and the other oppofite the culhion ; the for- 
mer ball fhewing pofitive, and the latter negative figns. The 
knobs of two jars being fubftituted in the place of thefe balls, 
the jar, applied to the furface oppofed to the culhion, was> 
charged negatively, and the other pofitively. This dilpofition 
of the back furface feemed, by a few trials, to be weaker the 
4 ftronger.* 
