Yol.  49.]  PROF.  J.  W.  JUDD  ON  COMPOSITE  DYKES  IN  ARRAN.  539 
district  and  is  over  30  feet  wide  ;  its  centre  is  a  4  mica-syenite- 
porphyry  ’  with  63  per  cent,  of  silica  ;  but  it  passes  at  its  margins 
into  a  fine-grained  kersantite  or  mica-diorite  containing  only  47 
per  cent,  of  silica.  Not  only  magnetite,  but  pyrites  and  apatite 
occur  in  much  greater  abundance  at  the  sides  of  the  dyke  than  in 
its  centre. 
These  examples  may  be  taken  as  sufficiently  illustrating  the  first 
class  of  composite  dykes ;  but  it  is  to  the  second  class — in  which  it  is 
impossible  to  explain  the  differences  between  the  rocks  in  the  dyke 
by  segregation  going  on  during  its  consolidation — that  I  desire 
to  call  especial  attention  in  this  paper.  It  is  a  fortunate  circum¬ 
stance  that  we  have  in  the  Island  of  Arran  some  very  beautiful,  if 
not  indeed  actually  unique,  examples  of  this  class  of  composite 
dykes,  and  some  of  the  most  remarkable  of  them  will  be  described 
in  the  following  pages. 
II.  The  Latest  Yolcanic  Eocks  of  the  British  Isles. 
In  describing  the  succession  of  volcanic  ejections  in  the  Western 
Isles  of  Scotland,  and  comparing  them  with  those  of  Hungary  and 
the  Lipari  Islands,  I  pointed  out  many  years  ago  that  the  nature  of 
the  latest  erupted  lavas  suggests  the  conclusion  that  a  differentiation 
of  the  materials  poured  out  during  earlier  eruptions  must  have  taken 
place.1  Mr.  Iddings  has  recently  shown  how  very  constantly  the 
phenomenon  of  the  ejection  of  lavas  of  intermediate  composition, 
followed,  after  an  interval,  by  the  extrusion  of  markedly  acid  and 
basic  types  respectively,  is  exhibited  at  volcanic  centres,  in  widely 
distant  districts  and  belonging  to  every  geological  period.2  The 
same  conclusion  was  strongly  insisted  upon  as  early  as  1868,  though 
perhaps  without  certain  very  necessary  qualifications,  by  Baron  von 
Eichthofen  in  his  4  Natural  System  of  Yolcanic  Eocks.’ 
The  latest  erupted  of  the  igneous  masses  of  the  Western  Isles  of 
Scotland  are  of  peculiar  interest  to  geologists,  not  only  from  the 
circumstance  that  they  afford  evidence  concerning  the  most  recent 
volcanic  outbursts  in  our  area,  but  on  account  of  the  illustrations 
which  they  supply  concerning  the  process  of  differentiation  in  igneous 
magmas. 
There  are  only  two  localities  known  at  present  at  which  the  surface- 
ejections  of  these  most  recent  British  volcanoes  can  be  studied 
— namely,  Beinn  Hiant  in  Ardnamurchan  and  the  Sgur  of  Eigg.3 
In  the  case  of  the  former  I  have  shown  how  the  differentiation  of 
1  Geol.  Mag.  for  1875,  p.  60;  Quart.  Journ.  Geol.  Soc.  vol.  xxxii.  (1876) 
pp.  315-316  ;  ‘  Volcanoes,  what  they  are  and  what  they  teach,’  1881,  pp.  199-205. 
2  ‘  The  Origin  of  Igneous  Bocks,’  Bull.  Phil.  Soc.  Washington,  vol.  xii.  (18921 
pp.  89-214. 
3  Quart.  Journ.  Geol.  Soc.  vol.  xlvi.  (1890)  pp.  371-382.  The  fact  that  the 
Sgiir  of  Eigg  is  not  a  dyke,  as  supposed  by  Hay  Cunningham,  but  a  lava- 
stream,  overlying  the  deposits  with  Pinites  eiggensis,  was  recognized  by  Hugh 
Miller  (‘  Cruise  of  the  Betsy,’  a  work  which  was  published  in  book-form  m 
1857,  the  substance  of  it  having  appeared  as  a  series  of  newspaper  articles 
some  years  earlier).  Sir  Archibald  Geikie  made  this  conclusion  perfectly  clear, 
and  added  new  details  in  1871  (Quart.  Journ.  Geol.  Soc.  vol.  xxvii  dd  303- 
309).  -  *  * 
