﻿822 
  Prof. 
  J. 
  Clerk 
  Maxwell 
  on 
  the 
  

  

  The 
  stability 
  (K) 
  o£ 
  any 
  given 
  phase 
  (A) 
  with 
  respect 
  to- 
  

   any 
  other 
  phase 
  (B), 
  is 
  expressed 
  in 
  the 
  following 
  form 
  : 
  — 
  

  

  K 
  = 
  e 
  — 
  up 
  -\-rjt 
  — 
  m 
  l 
  /uL 
  1 
  — 
  &c 
  — 
  m 
  >t 
  fi 
  n 
  , 
  

  

  where 
  e 
  is 
  the 
  energy, 
  v 
  the 
  volume, 
  rj 
  the 
  entropy, 
  and 
  

   ???!, 
  m 
  a 
  , 
  &c. 
  the 
  components 
  corresponding 
  to 
  the 
  second 
  

   phase 
  (B), 
  while 
  p 
  is 
  the 
  pressure, 
  t 
  the 
  temperature, 
  and 
  

   /*!, 
  /x 
  2 
  , 
  &c. 
  the 
  potentials 
  corresponding 
  to 
  the 
  given 
  phase 
  

   (A). 
  The 
  intensities 
  therefore 
  are 
  those 
  belonging 
  to 
  the 
  

   given 
  phase 
  (A) 
  , 
  while 
  the 
  magnitudes 
  are 
  those 
  corresponding 
  

   to 
  the 
  other 
  phase 
  (B). 
  

  

  We 
  may 
  interpret 
  this 
  expression 
  for 
  the 
  stability 
  by 
  

   saying 
  that 
  it 
  is 
  measured 
  by 
  the 
  excess 
  of 
  the 
  energy 
  in 
  the 
  

   phase 
  (B), 
  above 
  what 
  it 
  would 
  have 
  been 
  if 
  the 
  magnitudes 
  

   had 
  increased 
  from 
  zero 
  to 
  the 
  values 
  corresponding 
  to 
  the 
  

   phase 
  B, 
  while 
  the 
  values 
  of 
  the 
  intensities 
  were 
  those 
  

   belonging 
  to 
  the 
  phase 
  (A). 
  

  

  If 
  the 
  phase 
  (B) 
  is 
  in 
  all 
  respects 
  except 
  that 
  of 
  absolute 
  

   quantity 
  of 
  matter 
  the 
  same 
  as 
  the 
  phase 
  (A), 
  K 
  is 
  zero 
  ; 
  

   but 
  when 
  the 
  phase 
  (B) 
  differs 
  from 
  the 
  phase 
  (A), 
  a 
  portion 
  

   of 
  the 
  matter 
  in 
  the 
  phase 
  (A; 
  will 
  tend 
  to 
  pass 
  into 
  the 
  

   phase 
  (B) 
  if 
  K 
  is 
  negative, 
  but 
  not 
  if 
  it 
  is 
  zero 
  or 
  positive. 
  

  

  If 
  the 
  given 
  phase 
  (A) 
  of 
  the 
  mass 
  is 
  such 
  that 
  the 
  value 
  

   of 
  K 
  is 
  positive 
  or 
  zero 
  with 
  respect 
  to 
  every 
  other 
  phase 
  

   (B), 
  then 
  the 
  phase 
  (A) 
  is 
  absolutely 
  stable, 
  and 
  will 
  not 
  of 
  

   itself 
  pass 
  into 
  any 
  other 
  phase. 
  

  

  If, 
  however, 
  K 
  is 
  positive 
  with 
  respect 
  to 
  all 
  phases 
  which 
  

   differ 
  from 
  the 
  phase 
  (A) 
  only 
  by 
  infinitesimal 
  variations 
  of 
  

   the 
  magnitudes, 
  while 
  for 
  a 
  certain 
  other 
  phase, 
  B, 
  in 
  which 
  

   the 
  magnitudes 
  differ 
  by 
  finite 
  quantities 
  from 
  those 
  of 
  the 
  

   phase 
  (A), 
  K 
  is 
  negative, 
  then 
  the 
  question 
  whether 
  the 
  mass 
  

   will 
  pass 
  from 
  the 
  phase 
  (A) 
  to 
  the 
  phase 
  (Bj 
  will 
  depend 
  on 
  

   whether 
  it 
  can 
  do 
  so 
  without 
  any 
  transportation 
  of 
  matter 
  

   through 
  a 
  finite 
  distance, 
  or, 
  in 
  other 
  words, 
  on 
  whether 
  

   matter 
  in 
  the 
  phase 
  B 
  is 
  or 
  is 
  not 
  in 
  contact 
  with 
  the 
  

   mass. 
  

  

  In 
  this 
  case 
  the 
  phase 
  (A) 
  is 
  stable 
  in 
  itself, 
  but 
  is 
  liable 
  

   to 
  have 
  its 
  stability 
  destroyed 
  by 
  contact 
  with 
  the 
  smallest 
  

   portion 
  of 
  matter 
  in 
  certain 
  other 
  phases. 
  

  

  Finally, 
  if 
  K 
  can 
  be 
  made 
  negative 
  by 
  any 
  infinitesimal 
  

   variations 
  of 
  the 
  magnitudes 
  of 
  the 
  system 
  (A), 
  the 
  mass 
  

   will 
  be 
  in 
  unstable 
  equilibrium, 
  and 
  will 
  of 
  itself 
  pass 
  into 
  

   some 
  other 
  phase. 
  

  

  As 
  no 
  such 
  unstable 
  phase 
  can 
  continue 
  in 
  any 
  finite 
  mass 
  

   for 
  any 
  finite 
  time, 
  it 
  can 
  never 
  become 
  the 
  subject 
  of 
  expe- 
  

   riment 
  ; 
  but 
  it 
  is 
  of 
  great 
  importance 
  in 
  the 
  theory 
  of 
  

  

  