﻿Fenner 
  — 
  Stability 
  Relations 
  of 
  Silica 
  Minerals. 
  341 
  

  

  words, 
  the 
  formation 
  and 
  destruction 
  of 
  each 
  of 
  two 
  phases 
  

   proceed 
  at 
  the 
  same 
  rate. 
  Below 
  the 
  inversion 
  -point 
  the 
  

   formation 
  of 
  one 
  of 
  the 
  phases 
  exceeds 
  its 
  destruction 
  ; 
  above 
  

   it 
  the 
  other 
  phase 
  is 
  thus 
  characterized. 
  

  

  Let 
  us 
  apply 
  this 
  conception 
  to 
  the 
  specific 
  case 
  of 
  the 
  

   simultaneous 
  formation 
  of 
  crystals 
  of 
  quartz, 
  tridymite, 
  and 
  

   cristobalite 
  in 
  a 
  sodic 
  tungstate 
  melt. 
  In 
  this 
  experiment 
  there 
  

   is 
  some 
  evidence 
  that 
  a 
  reversible 
  reaction 
  of 
  this 
  kind 
  proceeds 
  : 
  

  

  Si0 
  o 
  + 
  Na 
  WO, 
  :zi±Na 
  SiC> 
  + 
  WO 
  

  

  3' 
  

  

  going 
  from 
  left 
  to 
  right 
  at 
  higher 
  temperatures 
  and 
  from 
  right 
  

   to 
  left 
  at 
  lower. 
  In 
  this 
  manner 
  a 
  multitude 
  of 
  Si0 
  2 
  molecules 
  

   are 
  constantly 
  contributed 
  to 
  the 
  liquid, 
  in 
  which 
  they 
  are 
  in 
  

   active 
  movement. 
  By 
  chance 
  collisions 
  numbers 
  of 
  these 
  

   meet 
  in 
  such 
  manner 
  as 
  to 
  form 
  groups 
  of 
  definite 
  patterns, 
  

   which 
  tend 
  to 
  cohere. 
  According 
  to 
  their 
  arrangement, 
  they 
  

   form 
  the 
  quartz, 
  tridymite, 
  or 
  cristobalite 
  molecules, 
  each 
  of 
  

   which 
  appears 
  and 
  disappears 
  in 
  countless 
  numbers 
  at 
  each 
  

   instant. 
  Under 
  the 
  conditions 
  of 
  the 
  experiment, 
  a 
  stream 
  

   of 
  ungrouped 
  molecules 
  is 
  constantly 
  added 
  to 
  the 
  cooler 
  portions 
  

   of 
  the 
  melt 
  by 
  the 
  breaking-up 
  of 
  Na 
  2 
  Si0 
  3 
  (or 
  by 
  con- 
  

   vection 
  currents 
  from 
  below), 
  yielding 
  such 
  an 
  over- 
  

   whelming 
  excess 
  over 
  what 
  would 
  occur 
  under 
  more 
  uniform 
  

   conditions 
  of 
  temperature 
  that 
  all 
  possible 
  configurations 
  of 
  

   grouping 
  (which 
  seems 
  to 
  mean 
  those 
  appropriate 
  to 
  quartz, 
  

   tridymite, 
  and 
  cristobalite) 
  are 
  formed, 
  and 
  the 
  number 
  of 
  each 
  

   kind 
  formed 
  is 
  in 
  great 
  excess 
  over 
  those 
  destroyed. 
  Each 
  sort 
  

   of 
  crystal 
  is, 
  therefore, 
  deposited 
  from 
  the 
  melt. 
  If 
  none 
  of 
  

   the 
  crystals 
  settled 
  into 
  the 
  hotter 
  regions 
  below, 
  we 
  should 
  

   ultimately 
  find 
  that 
  after 
  the 
  exhaustion 
  of 
  the 
  supply 
  of 
  

   ungrouped 
  molecules, 
  two 
  of 
  the 
  kinds 
  of 
  crystals 
  would 
  pass 
  

   into 
  the 
  third, 
  as 
  for 
  this 
  the 
  excess 
  of 
  production 
  over 
  

   destruction 
  would 
  be 
  greater 
  than 
  for 
  either 
  of 
  the 
  other 
  two, 
  

   and 
  this 
  would 
  be 
  the 
  stable 
  form 
  at 
  this 
  temperature; 
  but 
  

   under 
  the 
  influence 
  of 
  convection 
  currents, 
  the 
  process 
  con- 
  

   tinues 
  around 
  and 
  around 
  in 
  a 
  circle 
  indefinitely. 
  

  

  In 
  the 
  process 
  thus 
  presented, 
  I 
  have 
  conceived 
  that 
  by 
  regular 
  

   arrangements 
  of 
  the 
  simple 
  Si0 
  2 
  molecule 
  more 
  complex 
  groups 
  

   corresponding 
  to 
  quartz, 
  tridymite, 
  and 
  cristobalite 
  molecules 
  

   are 
  first 
  formed 
  in 
  the 
  liquid, 
  and 
  that 
  these 
  again 
  group 
  

   themselves 
  in 
  appropriate 
  patterns 
  to 
  form 
  the 
  respective 
  

   crystals. 
  The 
  question 
  may 
  arise 
  whether 
  such 
  primary 
  groupings 
  

   occur 
  apart 
  from 
  the 
  crystal 
  groupings. 
  There 
  is 
  abundant 
  

   evidence, 
  however, 
  from 
  various 
  sources, 
  that 
  simple 
  molecules 
  

   do 
  form 
  groups 
  of 
  this 
  nature 
  in 
  a 
  liquid, 
  so-called 
  associated 
  

   molecules. 
  Moreover, 
  evidence 
  will 
  be 
  presented 
  later 
  that 
  

   cristobalite 
  crystals 
  at 
  least 
  give 
  phenomena 
  which 
  are 
  best 
  

  

  