Atomic  Hypothesis  and  Dissected  {Structural)  Formula.     259 
given  quantity  of  heat  in  raising  the  temperature  of  a  solid  mass 
containing  a  given  number  of  elementary  atoms  all  of  the  same 
kind  is  independent  of  the  nature  of  the  atoms.  Apart  from  the 
circumstance  that  this  assumption  is  not  true  (e.  g.,  silicon,  car- 
bon, and  boron),  it  affords  no  "raison  d'etre"  for  the  fact,  it 
gives  no  clue  to  the  cause. 
29.  The  atomic  hypothesis  gives  no  clue  to  the  explanation  of 
the  remarkable  approximate  relationships  existing  between  the 
numerical  values  of  the  atomic  weights,  nor  to  the  remarkable 
sequence  in  which  those  elements  at  present  known  follow  each 
other.  Exact  experiment  (Stas)  disproves  the  possibility  of  all 
elementary  matter  being  the  same,  and  of  our  so-called  elements 
being  simply  variable  numbers  of  primary  atoms  united  together 
(i.  e.  allotropic  modifications)  ;  for  if  so,  a  common  divisor  of  all 
the  atomic  weights  would  exist.  This  does  not  seem  to  be  the 
case  even  in  members  of  the  same  family. 
30.  The  facts  referred  to  by  the  terms  equivalencj',  valency, 
radical,  replacement,  are  in  perfect  harmony  with  the  atomic  hy- 
pothesis, if  this  be  slightly  extended  by  the  definition  that  the 
power  of  union  together  of  atoms  to  form  a  molecule  is  said  to  be 
due  to  their  possession  of  combining  affinities  (Verwandtschaftsein- 
heiten)  which  mutually  saturate  one  another,  univalent  atoms 
being  such  as  ordinarily  exhibit  but  one  such  affinity,  bivalent 
ones  two,  trivalent  three,  quadrivalent  four,  and  so  on. 
A  consequence  of  this  definition  might  be  that  the  n  affinities 
of  a  multivalent  atom  might  not  be  all  of  equal  character,  and 
thus  that  two  or  more  isomeric  ternary  compounds  might  be 
produced  by  the  union  of  two  atoms  of  lower  valency  with  an 
atom  of  higher  valency,  according  to  the  affinities  which  were 
saturated  by  the  first  two  respectively.  Thus  the  atom  of  oxygen 
possesses  two  affinities,  each  saturated  by  one  affinity  of  an  atom 
of  univalent  hydrogen  in  the  molecule  water.  If,  however,  one 
affinity  were  saturated  with  hydrogen  and  the  other  with  chlo- 
rine, two  isomeric  hypochlorous  acids  might  exist  differing  in 
that  the  affinity  which  is  saturated  by  hydrogen  in  the  one  is 
saturated  by  chlorine  in  the  other,  and  vice  versa.  Similarly 
four  methyl  chlorides  might  exist,  according  as  the  chlorine 
atom  saturated  the  1st,  2nd,  3rd,  or  4th  affinity  of  the  quadri- 
valent carbon  atoms.  No  satisfactory  evidence  of  such  differ- 
ence of  affinity-value  has  yet  been  obtained,  several  such  sup- 
posed cases  having  disappeared  on  closer  examination. 
31.  A  radical,  then,  is  a  portion  of  a  molecule,  a  group  of 
atoms,  the  affinities  of  which  do  not  wholly  saturate  one  ano- 
ther, the  radical  being  uni-,  bi-,  tri-,  quadri-,  &c.  valent,  accord- 
ing as  1,  2,  3,  4,  ...  .  affinities  are  left  unsaturated.  Quoad 
the  series  of  bodies  into  the  formulae  of  each  of  wrhich  the  radical 
S2 
