508         Dr.  C.  R.  A.  Wright's  Reply  to  "An  Examination 
Mr.  Atkinson  says  that  the  definition  given  of  a  combining 
number  does  not  always  lead  to  the  number  in  use,  instancing 
ferric  and  aluminic  chlorides  as  cases  in  point.  It  is  almost 
superfluous  to  point  out  that  when  the  combining  number  is 
defined  as  the  minimum  number  of  grams  of  an  element  con- 
tained in  two  volumes  of  the  homogeneous  vapour  of  any  of  its 
compounds,  the  examination  of  only  one  vapour  is  not  sufficient 
to  point  out  a  minimum ;  thus,  if  benzene  only  were  examined, 
the  combining  number  of  carbon  would  be  found  to  be  72 ;  if 
several  compounds  cannot  be  examined,  then  the  combining 
number  deduced  from  only  one  or  two  is  liable  to  be  diminished 
by  arguments  drawn  from  chemical  analogy  &c.  In  the  case 
of  aluminium,  Mr.  Atkinson  is  rather  unfortunate  in  his  ex- 
ample ;  for  the  experiments  of  Buckton  and  Odling  tend  to  show 
that  the  formulas  of  aluminic  methide  and  ethide  are  respec- 
tively AlMe3  and  AlEt3,  and  hence  that  the  combining  num- 
ber is  27*5  instead  of  55  as  deduced  from  the  chloride  alone. 
Mr.  Atkinson  says  that  the  rule  given  for  the  determination 
from  its  specific  heat  of  the  combining  number  of  an  element 
that  does  not  yield  a  sufficient  number  of  gasefiable  compounds 
to  allow  of  the  application  of  the  above  rule  is  "  even  more 
fallacious ; "  but  he  does  not  apply  it  correctly  in  the  exam- 
ples he  gives  :  the  rule  is  deduced  from  the  relations  between 
the  specific  heat  in  the  solid  state  of  the  elements  "  mercury, 
sulphur,  selenium,  tellurium,  bromine,  iodine,  phosphorus, 
arsenic,  antimony,  bismuth,  tin,  osmium,  zinc,  and  probably 
others/'  Mr.  Atkinson,  however,  appplies  it  to  oxygen  and  chlo- 
rine, which  are  not  usually  regarded  as  solids* — and  instances, 
as  exceptional  cases  which  show  that  the  data  on  which  the  rule 
is  founded  are  too  variable  to  justify  reliance,  boron,  carbon,  iron, 
and  aluminium.  As  regards  the  first  two,  they  are  not  men- 
tioned in  connexion  with  the  rule ;  and  the  fact  of  their  not 
agreeing  with  it  no  more  vitiates  it  than  the  irregular  beha- 
viour of  carbon  dioxide  when  near  its  condensing-point  vitiates 
Marriotte's  law ;  while  aluminium  is  perfectly  regular,  the  com- 
bining number  deduced  from  the  methide  being  taken.  As  to 
iron,  it  has  been  pointed  out  in  the  last  paragraph  that  the 
combining  number  cannot  be  fixed  from  one  or  two  vapours 
only  ;  and  hence  this  metal  cannot  be  cited  as  an  exception. 
Mr.  Atkinson  considers  that  the  extension  of  the  term  com- 
fi.fi 
bining  number  to  relative  numbers  approximately  equal  to  -^- 
is  an  admission  of  the  "  worthlessness  of  the  rule."     If  the  cir- 
*  Mr.  Atkinson  not  only  does  not  take  the  best  determinations  of  the 
specific  heats  of  oxygen  and  chlorine,  but  he  takes  the  specific  heats  under 
constant  volume  instead  of  under  constant  pressure. 
