400  DR.  C.  CALLAWAY  ON  THE  ORIGIN  OF  THE  [Aug.  3? 
white  and  yellow  muscovite  which  undergoes  no  further  chemical 
change.”  Hornblende,  according  to  the  same  authority,  loses  only 
a  trace  of  magnesia  under  similar  circumstances  ;  but  since  horn¬ 
blende  is  easily  decomposed  into  chlorite,  and  chlorite  may  be 
changed  into  biotite,  it  is  plain  that  hornblende  may  indirectly  be 
a  source  of  muscovite. 
III.  Stages  of  Metamorphism. 
(1)  Decomposition  and  Corrosion. — Ordinary  decomposition,  such 
as  the  passage  of  hornblende  into  chlorite,  epidote,  and  iron  oxide, 
or  the  conversion  of  plagioclase  into  calcite  and  quartz,  appears  to 
be  associated  with  the  access  of  mineral  solutions.  Where  the 
rigid  rock  is  crushed,  there  is  always  more  or  less  decomposition, 
and  the  presence  of  water  is  evinced  by  the  nature  of  the  deposits 
which  fill  the  cracks.  Generally  speaking,  the  crushing  accelerates 
the  process.  AYhen  the  pressure  has  increased  so  as  to  produce 
planes  of  discontinuity,  decomposition  is  greatest  at  these  planes. 
Thus,  in  a  diorite  the  planes  of  cleavage  or  shearing  are  often 
marked  by  secondary  products,  while  the  intervening  seams  are 
comparatively  unchanged. 
Corrosion  requires  special  mention.  Crystals  of  hornblende, 
felspar,  and  some  other  minerals  are  very  commonly  corroded  in 
an  excessive  degree  in  and  near  granite-diorite  complexes.  Idio- 
morphic  individuals  thus  appear  perforated  with  holes,  or  their 
margins  are  cut  like  a  dentate  or  crenate  leaf,  and  frequently  the 
original  crystalline  form  is  nearly  or  entirely  destroyed.  The 
change  is  most  common  in  the  medium-black  diorite. 
The  hornblende  of  this  diorite  sometimes  shows  curvilinear  out¬ 
lines  at  considerable  distances  from  a  sheared  plexus  of  veins,  and 
I  must  not  be  understood  to  deny  that  in  these  cases  the  mineral 
may  have  originally  crystallized  in  this  form ;  but  the  characters 
become  more  marked  towards  the  zone  of  shearing,  the  crystals 
being  more  and  more  deeply  eroded  until  only  shreds  of  them  are 
left.  The  mineral  which  remains  in  contact  with  the  eroded  part  is 
generally  transparent  and  water-clear.  It  is  sometimes  quartz, 
sometimes  felspar.  The  latter  occasionally  displays  polysynthetic 
twinning,  but  usually  it  is  recognized  as  felspar  only  in  con¬ 
vergent  light.  Hornblende  can  rarely  be  followed  into  a  granite- 
diorite  shear-zone,  since  it  is  normally  converted  through  chlorite 
into  black  mica.  This  mica  also  frequently  suffers  corrosion,  is 
reduced  to  shreds,  and  sometimes  destroyed. 
The  corrosion  of  the  felspar  is  very  similar.  The  crystals  are 
invaded  by  quartz  or  water-clear  felspar,  and  sometimes  the 
original  substance  entirely  disappears.  Frequently,  simple  felspar- 
crystals  sheathed  in  biotite  are  replaced  by  quartz,  the  mica 
remaining  intact.  More  commonly,  the  corrosion  has  acted  in 
patches,  so  that  a  large  part  of  a  slide  of  modified  diorite  may  be 
occupied  with  water-clear  granules.  Corrosion  in  felspar  and  biotite 
is  shown  in  fig.  3  of  pi.  xvi.  in  paper  No.  II. 
