fome  Electrical  'Experiments.  1 5.7 
trometer  falling  to  o,  but  not  the  leaft  explofion  is  per- 
ceived. The  cloud  then  turning  back  in  its  progreffive 
motion  in  the  frame,  is  charged  again  while  it  paffes  on 
to  b;  at  which  point,  by  means  of  its  motion  round  its 
axis,  it  revolves  over  the  conductor  b ; it  empties  itfelf, 
the  electrometer  falls,  and  no  explofion  is  perceived : the 
fame  thing  happens  in  the  paffage  over  the  houfe  c. 
The  machine  remaining  in  the  pofition  as  before,  I 
place  balls  of  a quarter  of  an  inch  diameter,  at  the  upper 
terminations  of  the  conductors  of  the  houfes  a,  b,  c,  and 
with  thefe  balls,  the  experiments  proceed  almoft  as  be- 
fore; that  is,  the  matter  pafles  off  with  a little  hilling 
noife,  and  now  and  then  it  gives  a flight  explofion,  the 
fmallnefs  of  thefe  balls  differing  little  from  points;  but 
when  I place  balls  of  three-quarters  of  an  inch  diameter 
inftead  of  the  fmall  ones,  the  cloud,  every  time  it  pafles 
over  them,  makes  one  or  more  explofions,  and  fires  the 
magazines  a,  b,  c;  and,  notwithftanding  that,  the  index 
of  the  electrometer  does  not  defcend  above  20°,  and 
ftarts  up  again  as  fuddenly  as  it  fell. 
If  balls  are  fafer  at  the  upper  ends  of  conductors  than 
points,  it  lhould  follow,  that  the  larger  the  balls  are,  the 
greater  the  fecurity;  but  from  all  thefe  experiments  I 
never  found  a fhock  with  a point,  and  not  always  with  a 
very  fmall  ball:  but  the  electrical  matter  paffes  off 
X 2 filently 
