86  M.  E.  Edlund's  Researches  on  the  Electromotive 
§2. 
When  a  galvanic  current  passes  through  a  wire  composed  of 
two  different  metals,  there  result  two  sources  of  heat,  one  inde- 
pendent of  the  other :  in  the  first  place,  in  consequence  of  the 
galvanic  resistance,  a  quantity  of  heat  proportional  to  that  resist- 
ance and  to  the  square  of  the  intensity  of  the  current  is  developed 
in  the  wire ;  in  the  second  place,  there  ensues  at  the  point  of 
contact  a  production  or  absorption  of  heat  proportional  both  to 
the  electromotive  force  of  that  point  and  to  the  intensity  of  the 
current.  The  first  source  of  heat  is  in  general,  but  above  all 
with  metallic  wires  which  present  great  resistance,  infinitely 
greater  than  the  second,  and,  besides,  increases  more  rapidly 
with  the  intensity  of  the  current.  If,  then,  in  an  experiment  the 
intensity  varies  a  little,  it  is  not  impossible  that  the  variations  in 
the  quantity  of  heat  produced  by  the  resistance  may  be  greater 
than  the  total  quantity  of  heat  to  be  measured,  and  consequently 
the  result  become  doubtful.  For  the  accuracy  of  the  measure- 
ments, therefore,  it  is  evidently  of  great  importance  to  give  to 
the  apparatus  such  a  construction  that  its  indications  shall  be 
perfectly  independent  of  the  first  source  of  heat.  Whatever  the 
principle  on  which  the  apparatus  is  constructed,  and  however 
in  other  respects  it  be  arranged,  it  is  further  necessary  that  those 
indications  depend  as  little  as  possible  on  foreign  influences  of 
temperature  (as,  for  example,  variations  in  the  temperature  of 
the  place  in  which  the  experiments  are  made,  &c.) — a  condition 
difficult  to  fulfil,  seeing  that  the  sensitiveness  of  the  apparatus 
ought  to  be  so  great  that  it  will  indicate  the  slightest  variations 
of  temperature.  It  is  therefore  necessary  to  regulate  the  expe- 
riments in  such  a  manner  that  those  influences  of  foreign  tem- 
perature can  be  eliminated,  if  we  cannot  render  ourselves  per- 
fectly independent  of  them.     The  question  is,  to  measure  the 
as  to  pass  from  the  copper  to  the  bismuth  in  one  of  the  calorimeters,  and 
vice  versa  in  the  other.  The  water  of  one  of  the  calorimeters,  therefore, 
became  hotter  than  the  water  of  the  other ;  and  the  difference  between 
their  temperatures,  measured  by  means  of  an  ordinary  thermometer,  gave 
the  measure  of  the  quantities  of  heat  sought.  This  method  has  the  advan- 
tage of  giving,  in  ordinary  units  of  heat,  the  measure  of  the  quantities  in 
question.  On  the  other  hand,  it  is  not  so  delicate,  by  far,  as  that  employed 
by  me.  M.  le  Roux  obtained,  for  the  bismuth-copper  combination,  a  dif- 
ference of  temperature  of  1'7  degree.  With  a  combination  for  which  the 
absorption  or  production  of  heat  is  a  fiftieth  or  a  hundredth  part  of  the 
above,  the  total  difference  of  temperature  to  be  measured  by  the  mercurial 
thermometer  would  only  amount  to  some  hundredths  of  a  degree ;  the 
measurement  of  such  a  difference  by  means  of  thermometers  of  which  the 
divisions  are  tenths  of  a  degree  cannot  be  particularly  exact.  Doubtless, 
also,  this  is  in  part  the  cause  of  the  differences  between  M.  le  Roux's  re- 
sults and  mine. 
