[ 33 6 ] . 
turn of the fire might be rendered more complete 
hv increafmg the thicknefs of the filk, or by ap 
p ying to it fome nonconduding fubftance, that 
nlhf confine the excited fluid more perfectly to 
the° furface of the revolving cylinder. 
B^es-wax being a nonconduflmg fubftance ea- 
Jffty procured, I rubbed the filken flap with it, and 
found as I expeaed, that the return of the fire to 
the cufliion at 'the anterior part of the machine was 
by that means much dimmilhed, and consequently 
th» excitation of the glafs was apparently incteafed. 
The addition, however, of more filk was ftill more 
effeaual in confining the fire to the glals ; and when 
it was employed ten or twelve times doub ed > '* 
feemed to deny any paflage from the g a.s o 
CU AU thus difcovered the method of remedying the 
-common defedt in the conftruaion ofthe anterior 
part of the cuihion, I next attended to that part 
which correfponds with the defending fide of the 
cylinder. Being convinced that this part of the 
rubber was alone concerned in the excitation, I ima- 
gined that the reverfe of what was neceffary ante- 
riorly fliould be adopted in the ftruftute of the po- 
fterior part; that, inftead of placing nonconduaing 
fubftances between the glafs and cufliion, we ou 
here make the afflux of the eledtac matter as gmat 
as poffible, by the application of the moll pertet y 
conducting bodies. Confining therefore the amal- 
gam to that place where the glafs firft comes in 
contact with the rubber I placed feme tinfoil clofe 
to the amalgam, and, bending it back, fecuted t 
-to the metallic plate below the cuihion. By th 
