﻿Measuring 
  large 
  Molecular 
  Masses. 
  501 
  

  

  methods 
  which 
  I 
  have 
  proposed 
  and 
  applied 
  for 
  the 
  measure- 
  

   ment 
  of 
  large 
  molecular 
  masses. 
  In 
  the 
  paper 
  on 
  Ionization 
  

   in 
  Solutions 
  and 
  Two 
  New 
  Types 
  of 
  Viscosity 
  (loc. 
  cit.), 
  it 
  

   was 
  shown 
  that 
  for 
  element 
  ions 
  this 
  method 
  is 
  expressed 
  

   for 
  18° 
  C. 
  by 
  the 
  equation 
  

  

  1 
  _ 
  0-0365 
  0-0022 
  

  

  B*A 
  ~ 
  Bl 
  + 
  f(1+ 
  10-5/Bf; 
  ' 
  ' 
  * 
  * 
  W 
  

  

  in 
  which 
  B 
  is 
  the 
  volume 
  of 
  a 
  gram-atom 
  of 
  valency 
  v. 
  For 
  

   the 
  fatty 
  acid 
  ions 
  from 
  the 
  acetic 
  CH 
  3 
  COO 
  up 
  to 
  C 
  5 
  H 
  n 
  COO 
  

   the 
  same 
  formula 
  holds 
  except 
  that 
  0*0022 
  is 
  replaced 
  by 
  

   0*0097, 
  v 
  of 
  course 
  being 
  1. 
  While 
  for 
  these 
  five 
  ions 
  the 
  

   formula 
  expresses 
  the 
  experimental 
  facts 
  closely, 
  yielding 
  A 
  

   to 
  1 
  per 
  cent., 
  it 
  fails 
  for 
  formic 
  acid 
  for 
  w^hich 
  it 
  makes 
  A 
  

   too 
  large 
  by 
  18 
  per 
  cent. 
  Now 
  for 
  the 
  pyridinecarboxylic 
  

   acids 
  1 
  find 
  that 
  at 
  18° 
  C, 
  assuming 
  that 
  at 
  25° 
  A 
  is 
  1*18 
  

   times 
  its 
  value 
  at 
  18°, 
  

  

  1 
  0-00547 
  . 
  0-0097 
  

  

  B*A 
  Bf 
  ' 
  v(l 
  + 
  10-5/B*)- 
  * 
  

  

  In 
  the 
  second 
  term 
  this 
  is 
  identical 
  with 
  the 
  previous 
  result 
  

   for 
  the 
  fatty 
  acid 
  ions, 
  the 
  valency 
  v 
  appearing 
  specifically, 
  as 
  

   it 
  does 
  in 
  the 
  case 
  of 
  the 
  element 
  ions. 
  But 
  the 
  coefficient 
  

   of 
  the 
  first 
  term 
  on 
  the 
  right 
  is 
  only 
  0*15 
  times 
  that 
  for 
  the 
  

   fatty 
  acid 
  ions 
  and 
  the 
  element 
  ions. 
  I 
  take 
  this 
  striking- 
  

   result 
  to 
  be 
  due 
  to 
  the 
  fact 
  that 
  pyridine 
  is 
  a 
  base 
  giving 
  a 
  

   markedly 
  alkaline 
  reaction. 
  The 
  N 
  atom 
  in 
  it 
  carries 
  a 
  special 
  

   electron 
  pair 
  ^J, 
  and 
  the 
  positive 
  electrons 
  of 
  the 
  acids 
  

   derived 
  from 
  pyridine 
  have 
  their 
  inductive 
  effects 
  mostly 
  

   confined 
  within 
  the 
  pyridinecarboxylic 
  ions 
  by 
  this 
  pair 
  |jjp. 
  

   Now 
  the 
  first 
  term 
  on 
  the 
  right 
  originates 
  in 
  viscosity 
  

   caused 
  by 
  the 
  electron 
  of 
  the 
  ion 
  acting 
  inductively 
  on 
  the 
  

   molecules 
  of 
  the 
  solvent. 
  But 
  if 
  the 
  induction 
  is 
  confined 
  

   within 
  the 
  ion 
  mostly, 
  then 
  the 
  viscosity 
  of 
  electric 
  inductive 
  

   origin 
  must 
  become 
  relatively 
  small. 
  In 
  the 
  paper 
  just 
  cited 
  

   I 
  gave 
  reason 
  for 
  theoretically 
  expecting 
  this 
  variable 
  effect 
  

   in 
  large 
  ions, 
  and 
  was 
  surprised 
  not 
  to 
  find 
  it 
  in 
  the 
  

   fatty 
  acid 
  ions. 
  It 
  appears 
  now 
  that 
  the 
  paraffin 
  residue 
  

   C 
  n 
  H 
  2n 
  +i 
  does 
  not 
  affect 
  the 
  inductive 
  action 
  in 
  the 
  solvent 
  

   appreciably 
  until 
  in 
  the 
  formic 
  ion 
  it 
  is 
  reduced 
  to 
  H, 
  whence 
  

   the 
  exceptional 
  behaviour 
  of 
  the 
  formic 
  ion. 
  It 
  is 
  very 
  

   satisfactory 
  in 
  these 
  circumstances 
  to 
  find 
  the 
  ordinary 
  viscous 
  

   resistance 
  to 
  the 
  pyridinecarboxylic 
  ions 
  expressed 
  by 
  the 
  same 
  

   term 
  as 
  applies 
  to 
  the 
  fatty 
  acid 
  ions 
  with 
  fulfilment 
  of 
  the 
  

   additional 
  theoretical 
  condition 
  that 
  v 
  must 
  appear 
  as 
  in 
  (10) 
  

  

  