364  MESSES.  A.  HAEKER  AND  J.  E.  MAEE  ON  THE  [Aug.  I  893, 
similar  appearance,  which  have  no  action  on  polarized  light,  or  the 
interior  of  a  crystal  is  seen  to  be  birefringent,  the  marginal  portion 
isotropic.  The  converse  relation,  described  by  Wichmann 1  as 
occurring  in  the  rocks  of  Berggiesshiibel  and  Schwarzenberg  in 
Saxony,  we  have  not  observed,  but  a  shell  of  doubly-refracting 
sometimes  occurs  within  a  shell  of  singly-refracting  garnet,  and 
encloses  a  kernel  which  is  very  nearly  isotropic.  The  absorption- 
colour,  presumably  connected  with  the  relative  proportions  of  lime 
and  ferrous  oxide  in  the  mineral,  does  not  seem  to  stand  in  any  very 
evident  relation  with  the  optical  anomalies,  but  it  may  be  noted 
that  we  have  not  found  birefringence  in  the  more  deeply  coloured, 
and  so  more  ferriferous,  variety. 
As  a  curious  illustration  of  the  facility  with  which  lime-garnets 
are  produced  in  connexion  with  thermal  metamorphism,  we  may  be 
permitted  to  cite  a  literally  far-fetched  example.  It  occurs  among 
the  rocks  collected  by  Mr.  J.  J.  Lister  from  the  Tonga  Islands. 
The  island  of  Mango  2  consists,  above  sea-level,  entirely  of  volcanic 
tuffs  and  breccias,  but  these  are  presumed  to  rest  upon  an  old  coral- 
reef,  and  fragments  of  coral,  some  6  inches  in  diameter,  are  found 
mingled  with  blocks  of  lava,  etc.,  among  the  volcanic  ejectamenta. 
Although  these  corals  preserve  perfectly  their  characteristic  struc¬ 
tures,  the  interstices  are  completely  occupied,  in  the  specimens 
examined,  by  crystalline  garnet,  doubtless  grossularite.  The 
mineral  is  colourless  in  thin  slices,  and  though,  from  the  nature  of 
the  case,  there  are  no  external  crystal-faces,  the  birefringence,  the 
poly  synthetic  structure,  and  the  concentric  zones  are  beautifully 
exhibited  [1274,  1331]. 
In  addition  to  the  minerals  already  named,  these  metamorphosed 
vesicles  show  under  the  microscope  a  colourless  augite,  granules  of 
felspar  (?),  patches  of  magnetite,  and  occasional  crystals  of  sphene.  An 
examination  of  the  manner  in  which  the  several  minerals  mould  or 
enclose  one  another  proves  that  theyhave  generally  followed  a  definite 
order  of  crystallization,  which  is  not  often  departed  from.  The  order 
is — iron  ores,  garnet,  sphene,  augite,  green  hornblende  and  actino- 
lite,  epidote,  quartz,  calc-ite.  So  far  as  it  goes,  this  bears  a  close 
resemblance  to  the  normal  order  of  consolidation  of  minerals  from 
igneous  fusion,  as  expressed  in  Rosenbusch’s  ‘  law  of  decreasing 
basicity.’  The  iron  ore  in  the  centre  of  the  vesicles  is  often  pyrites, 
that  near  the  edge  magnetite  or  some  other  black  ore.  The  little 
patches  of  green  hornblende  are  most  plentiful  near  the  margin,  where 
they  are  associated  with  a  finely  granular  mass  which  seems  to  be 
felspar.  The  epidote  is  most  abundant,  as  a  rule,  in  the  marginal 
portion  of  the  large  vesicles,  while  the  calcite  and  the  best-developed 
garnets  are  found  in  the  central  portion.  Many  of  the  minerals, 
however,  show  a  rather  intricate  commingling.  The  augite  forms 
either  crowds  of  little  imperfect  crystals  or  larger  crystal-plates,  in 
the  latter  case  after  enclosing  a  quantity  of  green  actinolite  or 
actinolitic  hornblende  in  needles  or  shreds,  intergrown  with  the 
1  Zeitschr.  Deutsch.  geol.  Gesellsch.  vol.  xxvii.  (1875)  pp.  749-751. 
2  Quart.  Journ.  Geol.  Soc.  vol.  xlvii.  (1891)  p.  596. 
