430  Mr.  R.  W.  Atkinson's  Examination  of  the 
numbers.  Hence  it  follows  that  in.  adopting  two  volumes  as  the 
unit  of  volume  for  compound  vapours,  which  depends  upon  this 
assumption,  he  also  makes  use  of  atomic  ideas ;  and  upon  this, 
again,  he  founds  his  definition  of  "  combining  numbers." 
Dr.  Wright  refers  to  the  law  of  multiple  proportions  as  one 
of  the  facts  of  chemistry.  Experiment,  however,  does  not  lead 
to  numbers  which  are  multiples  of  his  "  combining  numbers." 
Thus,  referring  again  to  Roscoe's  researches  on  the  tungsten 
compounds,  there  are  four  chlorides  described;  and  analysis 
shows  that  for  100  parts  of  the  metal  they  contain  the  following 
amounts  of  chlorine :— 
w. 
CI. 
1.  100     . 
.     .     36-1 
2.     „ 
.     .     73-8 
3.     „        , 
.     .     95-4 
4.     „        . 
.     .  114-7 
or  the  ratio  of  the  amounts  of  chlorine  is 
1  :  2-04  :  2'64  :  3-18, 
or 
2  :  4-08  :  5'28  :  6-36. 
From  this  example  it  will  be  seen  that  the  law  of  multiple 
proportions  cannot  be  termed  an  experimental  fact.  In  order  to 
adopt  the  law,  we  must  assume  that  these  numbers  are  incor- 
rect, and  replace  them  by  others  which  agree  with  the  atomic 
theory.  But  in  admitting  the  existence  of  this  law,  Dr.  Wright 
does  at  the  same  time  make  use  of  the  atomic  explanation  to 
account  for  the  differences  between  the  experimental  and  theo- 
retical numbers. 
We  have  seen  the  assumption  involved  in  Dr.  Wright's  defi- 
nition of  " combining  number;"  but  even  after  making  it,  the 
rule  does  not  always  lead  to  the  number  in  use.  Thus  the 
vapour-density  of  ferric  chloride  would  lead  to  the  "  combining 
number"  for  iron  112,  and  that  of  aluminic  chloride  to  the 
number  55.  These  examples  speak  for  themselves  as  regards 
the  value  of  the  rule  given  for  the  determination  of  the  atomic 
weights  of  those  elements  which  form  volatile  compounds. 
For  the  determination  of  the  "  combining  number "  of  non- 
volatile elements  the  rule  given  is  even  more  fallacious.  In  the 
following  Table  are  given  the  "  combining  numbers "  of  some 
elements  which  form  volatile  compounds,  their  specific  heats,  and 
the  products  of  the  two  numbers. 
