549 
Yol.  49.]  PROF.  J.  W.  JUDD  ON  COMPOSITE  DYKES  IN  ARRAN. 
In  the  vitreous  base  we  at  once  recognize  the  green  glass,  so 
familiar  to  all  who  have  studied  the  Arran  pitchstone  or  the  drawings 
of  it  given  by  Zirkel,  Allport,  and  Teall.  Disseminated  through  the 
colourless  glass  basis  we  find  the  dust  of  fine  microlites,  and  among 
this — each  group  surrounded  by  its  clear  space  ( ‘  court  of  crystal¬ 
lization  ’) — are  seen  the  exquisitely  beautiful,  fern-like  tufts  of 
skeleton-crystals  of  hornblende.  That  the  reference  of  these 
crystals  to  hornblende  is  correct  is  proved  in  this  particular  rock 
by  the  cleavage  and  extinction-angles  of  the  larger  shafts. 
The  ‘pitchstone’  of  the  Cir  Mhor  dyke  sometimes  exhibits,  as 
pointed  out  by  Allport,  a  distinct  banded  or  fluidal  appearance  and, 
occasionally,  an  approach  to  perlitic  structure. 
There  is  one  feature,  however,  presented  by  this  pitchstone  which 
is  of  especial  interest,  and  which  has  not  hitherto,  I  believe,  been 
observed  in  any  rock  of  the  same  kind.  Each  of  the  quartz-crystals 
and  the  angles  of  many  of  the  felspar-crystals  are  found  to  be 
enveloped  in  globular  aggregations  of  a  perfectly  clear  and  glassy 
material,  and  similar  globules  are  found  scattered  through  the  mass 
(see  PL  XIX.  fig.  1).  Under  polarized  light,  these  globules  exhibit 
the  black  cross  resolving  itself  into  hyperbolas,  evidently  due  to 
strain,  the  double  refraction  being  negative  (see  PI.  XIX.  fig.  1  a). 
When  very  thin  sections  of  the  rock  are  examined  with  high 
powers,  additional  details  of  much  interest  are  revealed  by  the  study 
of  these  globules.  They  are  seen  to  be  made  up  of  two  or  three 
concentric  coats,  the  surfaces  of  the  inner  ones  being  ragged,  while 
those  of  the  outside  layers  are  usually  smooth  and  mammillated. 
Careful  study  with  properly  adjusted  illumination  shows  that  the 
ragged  surfaces  are  due  to  a  number  of  overlapping  plates  like 
those  so  generally  recognized  by  petrographers  as  tridymite,  being 
attached  to  the  surfaces  of  the  glassy  coats.  In  the  case  of  the 
isolated  globules,  not  formed  round  large  crystals  in  the  rock,  a'group 
of  these  overlapping  scales  of  tridymite  may  usually  be  detected 
forming  a  nucleus  to  the  globules  ;  and  scattered  plates  of  the  same 
substance  may  often  be  detected  diffused  through  them.  The  white 
glassy  substance  usually  exhibits  a  distinctly  radial  appearance ;  in 
some  cases  high  powers  show  this  to  be  due  to  needles  of  hornblende 
radiating  from  the  crystals,  and  there  is  such  a  perfect  gradation 
between  these  radiating  rods,  of  which  the  character  can  be  ascer¬ 
tained,  and  those  which  are  too  minute  for  determination,  as  to 
point  to  the  conclusion  that  the  radiated  structure  is  due  to  similar 
hornblende-microlites  of  ultra-microscopic  dimensions.  (See  PL 
XIX.  fig.  1  b.) 
The  optical  characters  of  these  globules  lead  one  to  conclude 
that  they  are  hyalite ;  but  as  this  is  a  perfectly  new  mode  of 
occurrence  of  that  mineral,  I  sent  a  specimen  of  the  rock  to  my 
friend  Prof.  Dosenbusch,  who  has  very  kindly  given  me  the  benefit 
of  his  wide  knowledge  and  experience.  While  fully  confirming 
the  optical  and  other  characters  which  I  have  described,  he 
suggests  the  possibility  that  globules  of  very  clear  glass  might 
separate  out  from  the  matrix,  and  under  the  influence  of  strain 
