﻿Posnjak 
  and 
  Merwin 
  — 
  Hydrated 
  Ferric 
  Oxides. 
  317 
  

  

  to 
  water 
  determination. 
  Water 
  was 
  usually 
  determined 
  

   by 
  ignition 
  loss 
  and 
  may, 
  therefore, 
  have 
  included 
  other 
  

   volatile 
  compounds. 
  As 
  the 
  proposed 
  hydrated 
  ferric 
  

   oxides 
  differ 
  only 
  slightly 
  in 
  their 
  water 
  content 
  an 
  exact 
  

   estimation 
  of 
  water 
  is 
  essential. 
  

  

  The 
  analyses 
  in 
  Table 
  I 
  were 
  made 
  on 
  carefully 
  selected 
  

   material 
  13 
  examined 
  microscopically 
  in 
  respect 
  to 
  its 
  

   homogeneity. 
  The 
  iron 
  was 
  determined 
  by 
  titration 
  with 
  

   permanganate, 
  manganese 
  was 
  determined 
  colorimetri- 
  

   cally. 
  In 
  cases 
  where 
  aluminum, 
  calcium, 
  etc., 
  were 
  to 
  

   be 
  determined, 
  the 
  iron 
  was 
  first 
  removed 
  by 
  electrolysis. 
  

   The 
  water 
  determinations 
  were 
  made 
  according 
  to 
  Pen- 
  

   field's 
  method, 
  14 
  for 
  some 
  of 
  the 
  hydrated 
  ferric 
  oxides 
  

   contained 
  considerable 
  amounts 
  of 
  carbonate 
  which 
  

   increase 
  the 
  value 
  for 
  water 
  if 
  the 
  latter 
  is 
  determined 
  by 
  

   the 
  ignition 
  loss. 
  In 
  what 
  form 
  the 
  carbonate 
  is 
  present 
  

   could 
  not 
  directly 
  be 
  ascertained, 
  for 
  microscopically 
  the 
  

   analyzed 
  hydrated 
  ferric 
  oxides 
  appeared 
  practically 
  

   homogeneous. 
  According 
  to 
  the 
  analyses 
  the 
  carbonate 
  

   might 
  either 
  be 
  present 
  as 
  ferrous 
  or 
  as 
  manganese 
  car- 
  

   bonate. 
  However, 
  the 
  corresponding 
  amounts 
  of 
  manga- 
  

   nese 
  and 
  carbon 
  dioxide 
  directly 
  determined 
  do 
  not 
  show 
  

   any 
  regularity 
  whatever, 
  though 
  small 
  amounts 
  of 
  fer- 
  

   rous 
  iron 
  could 
  be 
  detected 
  (in 
  No. 
  11 
  about 
  0-5%). 
  It 
  

   seems, 
  therefore, 
  that 
  the 
  carbonate 
  present 
  must 
  be 
  sid- 
  

   erite, 
  while 
  manganese 
  is 
  present 
  in 
  a 
  peroxidized 
  condi- 
  

   tion 
  which 
  when 
  the 
  mineral 
  dissolves 
  oxidizes 
  the 
  ferrous 
  

   iron 
  and 
  thus 
  renders 
  the 
  direct 
  determination 
  of 
  the 
  

   latter 
  impossible. 
  Manganese 
  is 
  probably 
  present 
  as 
  

   manganite 
  (Mn 
  2 
  3 
  -H 
  2 
  0) 
  which 
  is 
  isomorphous 
  with 
  goe- 
  

   thite. 
  

  

  That 
  carbon 
  dioxide 
  was 
  probably 
  derived 
  from 
  ferrous 
  

   carbonate 
  finds 
  some 
  confirmation 
  in 
  the 
  following 
  obser- 
  

   vation. 
  The 
  ignition 
  loss 
  determined 
  in 
  some 
  of 
  the 
  

   hydrated 
  ferric 
  oxides 
  gave 
  a 
  certain 
  excess 
  compared 
  

   with 
  the 
  amount 
  of 
  water 
  found 
  by 
  the 
  direct 
  method. 
  

   Assuming 
  this 
  to 
  be 
  due 
  to 
  the 
  presence 
  of 
  ferrous 
  carbo- 
  

   nate, 
  which 
  reacts 
  on 
  ignition 
  thus 
  : 
  2FeC0 
  3 
  + 
  = 
  Fe 
  2 
  3 
  

   + 
  2C0 
  2 
  , 
  the 
  amount 
  of 
  carbon 
  dioxide 
  corresponding 
  to 
  

  

  13 
  We 
  wish 
  to 
  express 
  our 
  thanks 
  to 
  Dr. 
  G. 
  P. 
  Merrill 
  of 
  the 
  National 
  

   Museum 
  for 
  the 
  privilege 
  of 
  selecting 
  these 
  specimens; 
  and 
  to 
  Dr. 
  E. 
  T. 
  

   Wherry, 
  formerly 
  of 
  the 
  National 
  Museum, 
  for 
  help 
  in 
  making 
  the 
  selections, 
  

   and 
  for 
  specimens 
  from 
  his 
  own 
  collections; 
  also 
  to 
  Professor 
  E. 
  J. 
  Holden 
  

   of 
  the 
  Virginia 
  Polytechnic 
  Institute 
  for 
  about 
  20 
  fine 
  specimens. 
  

  

  14 
  See 
  W. 
  F. 
  Hillebrand, 
  Analyses 
  of 
  silicate 
  and 
  carbonate 
  rocks. 
  

  

  