Prof.  R.  Clausius  on  the  Mechanical  Theory  of  Heat.      445 
in  question  is  one  in  which  the  intensity  of  the  current  is  not 
continually  increased  by  greater  heating  of  one  junction,  but 
from  a  certain  temperature  upward  the  current  again  diminishes 
in  intensity,  and  with  a  still  further  rise  of  temperature  may 
even  change  its  direction. 
This  phenomenon  I  have  likewise  already  discussed,  in  my 
above-mentioned  memoir.  I  have  endeavoured  to  explain  it  by  the 
assumption  that,  in  one  of  the  two  metals  of  which  such  a  circuit 
consists  (or  even  in  both),  the  change  of  temperature  gives  rise 
to  a  change  of  molecular  state,  the  effect  of  which  is  that  the 
altered  and  unaltered  portions  of  the  metal  have  the  same  elec- 
trical relation  to  each  other  as  two  different  metals.  As  soon  as 
a  change  of  this  kind  occurs,  electromotive  forces  act  not  only  at 
the  places  of  contact  of  different  metals,  but  also  wherever  differ- 
ently constituted  portions  of  the  same  metal  are  in  contact. 
Accordingly  heat  will  be  generated  or  expended  not  merely  at 
the  junctions,  but  also  in  other  parts,  in  the  interior  of  the  in- 
dividual metals ;  and  hence,  in  order  to  determine  all  the  pas- 
sages of  heat  that  occur,  we  must  consider  not  merely  the  tem- 
peratures of  the  junctions,  but  also  the  temperatures  of  those 
other  parts. 
Of  course  the  thing  becomes  thereby  more  complicated. 
Besides,  of  the  alterations  mentioned,  although  their  existence 
in  individual  cases  has  already  been  shown,  we  have  yet  too  little 
special  knowledge  to  be  able  to  trace  in  detail  all  that  take  place 
in  such  a  thermoelectric  circuit.  Meanwhile  it  will  not  be  dis- 
puted that  in  the  assumption  I  have  made  we  have  at  least  a 
possible  explanation ;  and  at  any  rate  it  will  be  admitted  that  a 
phenomenon  in  which  circumstances  as  yet  unknown  cooperate 
is  ill  adapted  to  be  used  as  a  proof  for  or  against  a  theorem 
advanced. 
Finally  Mr.  Tait  says  further  that  by  the  introduction  of  the 
ideas  of  internal  work  and  disc/relation  I  have  done  harm  to 
science. 
What  Mr.  Tait  has  against  the  notion  of  internal  work  is  to 
me  incomprehensible.  Since,  in  my  first  memoir  on  the  mecha- 
nical theory  of  heat,  I  distinguished  the  work  accomplished  by 
heat,  in  the  change  of  state  of  a  body,  into  external  and  internal 
work,  and  then  showed  that  these  two  quantities  of  work  essen- 
tially differ  in  their  behaviour,  this  distinction  has  been  in  like 
manner  employed  by  all  the  authors  who,  to  my  knowledge, 
have  written  on  the  mechanical  theory  of  heat. 
With  respect  to  the  notion  of  disgregation,  investigations  by 
Boltzmann  and  myself  have  just  been  published  by  which  it  ac- 
quires a  universal  mechanical  significance;  and  although  the 
investigations  relative  to  this  subject  are  not  yet  concluded,  I 
