90  M.  E.  Edlund's  Researches  on  the  Electromotive 
caoutchouc  tube.  I  afterwards  ascertained,  by  the  same  proce- 
dure, the  hermeticity  of  the  joints  at  the  other  extremity  of  the 
glass  tube. 
Let  us  suppose  now  that  a  wire  of  two  metals,  A  and  B,  sol- 
dered together,  has  been  introduced  into  one  of  the  cylinders, 
and  an  identical  wire  of  the  same  metals  and  of  the  same  thick- 
ness into  the  other,  that  the  glass  tube  is  fixed  hermetically  in 
its  position,  and  that  the  tubes  c  c1  are  likewise  hermetically 
closed.  If  now  one  and  the  same  galvanic  current  passes  through 
the  wires  of  both  cylinders  (from  metal  A  to  metal  B,  for  ex- 
ample), an  equal  quantity  of  heat  is  developed  in  the  cylinders. 
For  this  reason  the  index  in  the  glass  tube  remains  at  rest.  It 
is  evident  that  the  same  ought  to  be  the  case  when  the  force  of 
the  current  varies,  and  consequently  that  the  development  of 
heat  occasioned  by  the  resistance  of  the  wires  has  no  influence 
on  the  movement  of  the  index.  If,  on  the  other  hand,  the 
wires  are  united  to  one  another  in  such  a  manner  that  in  one  of 
the  cylinders  the  current  passes  from  A  to  B,  and  in  the  other 
from  B  to  A,  heat  will  be  developed  at  one  of  the  points  of  con- 
tact, and  a  cooling  will  take  place  at  the  other.  The  develop- 
ment of  heat  in  the  two  cylinders  will  no  longer  be  equal,  and 
the  index  will  move  toward  the  cylinder  presenting  the  less  de- 
velopment of  heat.  This  displacement  will  continue  until  the 
heating  of  each  cylinder  is  equal  to  the  cooling  occasioned  by 
radiation  and  by  contact  with  the  surrounding  air.  After  three 
quarters  of  an  hour  no  displacement  of  the  index  could  be  per- 
ceived; with  the  less  delicate  apparatus  which  I  used  in  my  first 
researches,  this  took  place  in  a  much  shorter  time.  Once  the 
index  is  at  rest,  the  quantity  of  heat  lost  by  the  air  of  the  cylinder 
by  radiation  and  the  contact  of  the  external  air  is  consequently 
equal  to  the  quantity  received  from  the  wires  by  the  air  in  the 
cylinder.  The  difference  between  the  loss  of  heat  in  one  cylinder 
and  in  the  other  is  therefore  found  to  be  equal  to  the  difference 
between  the  amounts  of  heat  produced  in  the  two  wires;  and,  as 
will  be  demonstrated  below,  the  former  difference  can  be  calcu- 
lated from  the  amount  of  displacement  of  the  index. 
If  any  alteration  take  place  in  the  temperature  of  the  labora- 
tory, it  can  have  no  effect  on  the  motion  of  the  index.  The  two 
copper  cylinders  are  perfectly  identical.  Surrounded  with  jackets 
of  polished  zinc,  which  are  both  of  the  same  size  and  shape  and 
contain  an  equal  quantity  of  water,  they  are  placed  at  but  a 
small  distance  one  from  the  other.  An  alteration  in  the  tempe- 
rature of  the  laboratory  acts  but  slowly  on  the  copper  cylinders ; 
for  it  has  first  to  traverse  the  considerable  mass  of  water  (5*8 
kilogrammes)  which  each  jacket  contains.  It  was  nevertheless 
shown,  in  the  experiments,  that  the  index  had  a  motion  proper 
