NO.    I  ARCHEOLOGY    OF    ST.    LAWRENCE    ISLAND COLLINS  313 
and  being  for  the  most  part  undecorated.  It  is  within  this  last,  the 
dominant  Punuk  group,  that  the  Thule  type  3  harpoon  heads  belong, 
although  in  contour  they  are  closer  to  the  still  later  St.  Lawrence  form, 
open  socket  type  III  (b)  x  (pi.  yi,  figs.  5-7)  ;  the  latter,  however,  has 
the  characteristic  "  wedge-shaped  "  socket,  whereas  the  Thule  sockets 
are  all  wide  and  straight-sided. 
From  Jenness'  description  of  the  harpoon  heads  at  Wales  and  the 
Diomedes,  it  is  evident  that  the  development  of  the  open  socket  har- 
poon head  has  followed  the  same  general  course  at  Bering  Strait  as 
on  St.  Lawrence  Island,  although  probably  in  no  instance  have  the 
comparable  forms  been  identical  in  every  feature.  Everything  points 
to  the  Thule  type  3  head  being  closer  to  the  Bering  Strait  than  to  the 
Gambell  form  of  the  open  socket  type  III  (a)  x  head.  Thus,  of  the 
two  heads  of  this  type  illustrated  by  Jenness  (1928  a,  pi.  12,  b,  c)  the 
one  with  lashing  slots  is  practically  identical  to  Mathiassen,  1927, 
vol.  I,  plate  I,  figure  6;  these  later  Wales  and  Diomede  heads  have 
the  same  straight  contour  as  the  Thule  forms,  whereas  at  Gambell  a 
tendency  to  a  curvature  of  the  sides  persists  until  the  appearance  of 
the  very  latest  prehistoric  form,  that  with  the  "  wedge-shaped  "  socket. 
Furthermore,  as  we  have  already  seen,  a  close  linkage  between  the 
Thule  and  Bering  Strait  forms  is  provided  by  the  drilled  lashing  holes, 
which  at  a  rather  late  stage  supplanted  slots  at  Bering  Strait.  At 
Barrow  these  drilled  lashing  holes  are  very  rare,  and  furthermore  the 
open  socket  heads  with  end  blades  which  have  been  described  from 
there  (Mason,  1930,  pi.  5,  figs.  4-7;  Collins,  1935,  pi.  11,  figs.  5-6) 
are  not  very  close  to  the  Thule  type  3  head  in  general  shape — not 
nearly  so  close  as  are  the  Wales-Diomede  heads  already  mentioned. 
The  Thule  type  4  head  (rather  small,  with  closed  socket  and  a  blade 
at  right  angles  to  line  hole — Mathiassen,  1927,  vol.  i,  pi.  2,  figs.  1-2) 
occupies  a  subordinate  position  in  the  Thule  culture.  In  the  west,  har- 
poon heads  of  this  general  type  are  known  from  all  culture  stages, 
but  if  we  may  judge  from  the  St.  Lawrence  forms  the  older  ones 
were  almost  invariably  decorated  and  possessed  certain  minor  features 
as  of  shape,  form  of  spur,  median  ridge  between  tip  and  line  hole, 
etc.,  which  serve  to  distinguish  them  from  the  plain  Thule  type ;  later, 
the  closed  socket  heads  are  usually  undecorated  and  have  quite  simple 
contours,  but  they  have  become  larger  than  the  Thule  forms. 
It  appears,  therefore,  that  the  three  most  characteristic  forms  of 
Thule  harpoon  heads  all  have  their  counterparts  in  the  west,  and  that 
without  exception  where  systematic  excavations  have  been  made,  they 
have  been  found  to  be  later  than  the  more  elaborate  Old  Bering  Sea 
and  Birnirk  forms. 
