260       Dr.  C.  R.  A.  Wrisrht  on  the  Relations  between  the 
enters,  this  group  of  atoms  is  indivisible,  and  to  it  therefore 
the  term  atom  may  be  extended,  although  of  coarse  in  strictness 
of  speech  the  term  thus  applied  is  a  misnomer.  Thus  water, 
acetic  acid,  and  caustic  soda  may  be  regarded  as  produced  by 
the  union  of  an  atom  of  hydroxyl  (OH)  with  one  of  hydrogen, 
acetyl  (C2  H3  0),  and  sodium  respectively.  The  term  replace- 
ment thus  acquires  a  grammaticallyexact  application,  one  or  more 
atoms  of  one  kind  being  removed  from  a  molecule  and  their 
places  filled  up  by  others.  The  atomic  weight  or  combining 
number  of  a  radical  (compound  atom)  is  therefore  the  sum  of  the 
weights  of  its  component  elementary  atoms ;  and  its  equivalent 
is  the  atomic  weight  divided  by  the  number  of  unsaturated  affi- 
nities which  it  exhibits  in  the  reaction  in  question. 
32.  Changes  in  valency  are  thus  readily  explainable  by  as- 
suming that  one  or  more  pairs  of  affinities  either  mutually 
saturate  each  other  or  cease  to  do  so.  Thus  the  atom  of  free 
mercury  (identical  with  the  molecule)  exhibits  no  valency ;  but 
when  combined  with  chlorine,  bivalency ;  the  two  affinities  ex- 
hibited in  the  chloride  have  therefore  saturated  one  another  in 
the  free  metal.  Carbon  is  quadrivalent ;  two  carbon  atoms  may 
unite  by  mutually  saturating  1,  2,  or  3  pairs  of  affinities,  thus 
giving  rise  to  a  radical  C2,  which  mav  be  sexi-,  quadri-,  or  bi- 
valent accordingly:  thus,  C2H6,  C2H4,  C2  H2.  Hence  the 
observed  fact  that  changes  in  valency  almost  invariably  proceed 
by  even  numbers  is  readily  explainable ;  so  also  Gerhardt's  law  of 
even  numbers. 
Those  few  cases  where  valency  does  not  differ  by  even  num- 
bers (e.  g.  the  oxides  of  nitrogen  not  considered  to  be  dissociated, 
the  chlorides,  oxychlorides,  &c.  of  vanadium  and  tungsten,  &c.) 
require  the  assumption  that  single  affinities  can  sometimes  re- 
main, as  it  were,  latent :  the  precise  meaning  of  this  phrase 
depends  on  the  view  taken  of  the  nature  of  these  affinities;  but 
as  it  can  hardly  be  doubted  that  they  represent  the  action  of  a 
force  of  some  kind,  their  exertion  or  transference  being  always 
connected  with  the  manifestation  or  absorption  of  force  of  some 
kind,  the  anomalous  result  is  arrived  at  that  an  atom  is  capable 
of  exerting  ordinarily  certain  forces  which  at  other  times  it 
does  not  exert, — a  proposition  somewhat  of  the  same  nature  as 
this,  that  a  given  substance  should  sometimes  weigh  5  lbs.  and 
sometimes  only  4  lbs. 
33.  The  dissected  formula  of  a  compound  hence  represents, 
on  the  atomic  hypothesis,  not  merely  the  existence  of  certain 
reactions,  but  also  certain  connexions  and  relationships  between 
particular  groups  of  the  constituent  atoms.  Of  course  the  for- 
mula CH3 — CO — OH  does  not  represent  in  space  the  relative 
positions  of  the  compound  atom3 ;  it  bears  to  the  true  position 
