NO.    I  ARCHEOLOGY    OF    ST.    LAWRENCE    ISLAND — COLLINS  317 
though  under  a  broad  definition  they  would  be  included  with  the 
Thule  heads.  To  understand  the  relation  of  these  old  south  Alaskan 
heads  to  those  of  the  Arctic  we  would  need  to  know  something  of 
the  archeology  of  the  extensive  stretch  of  coast  intervening.  In  view 
of  the  fact  that  this  south  Alaskan  culture  has  in  so  many  respects 
developed  along  special  lines  and  that  it  shows  so  few  significant 
resemblances  to  northern  Eskimo  culture,  particularly  in  its  Arctic 
phases,  it  seems  rather  hazardous  to  postulate  an  underlying  cul- 
tural connection  on  the  basis  of  the  harpoon  heads  alone.  It  is  true 
that  the  excavations  in  northern  Alaska  have  thrown  little  light  on 
the  origin  of  the  Thule  type  i  head ;  we  only  know  that  it  appears 
suddenly,  and  very  late,  at  Gambell.  However,  the  conditions  are 
somewhat  clearer  with  regard  to  the  still  more  important  Thule 
types  2  and  3,  for,  as  pointed  out  above,  these  are  found  at  Gambell, 
at  Bering  Strait,  and  at  Barrow,  but  in  no  case  have  they  appeared 
at  the  lowest  levels ;  on  the  contrary,  the  available  evidence  points 
rather  to  their  having  developed  from  the  preceding  Old  Bering  Sea 
and  Birnirk  forms.  In  view  of  these  conditions,  it  seems  highly  im- 
probable that  these  harpoon  heads  could  also  have  belonged  to  a 
"  Thule  "  phase  of  culture  in  Alaska  which  antedated  the  Old  Bering 
Sea  and  Birnirk  phases. 
The  composite  toggle  harpoon  heads  found  along  the  Northwest 
Coast  from  southeast  Alaska  to  California  are  so  fundamentally  differ- 
ent from  the  Eskimo  forms  that  an  immediate  relationship  seems  out 
of  the  question.  In  fact  these  southern  composite  heads  are  struc- 
turally further  removed  from  the  Eskimo  types  than  are  a  number 
of  Old  World  forms  including  some  from  as  far  away  as  Scandi- 
navia or  even  New  Zealand  (Edge-Partington  and  Heape,  "  Ethno- 
graphical Album  of  the  Pacific  Islands  ",  3d  series,  pi.  179,  figs.  3-5). 
Although  systematic  excavations  have  not  been  made  in  north- 
eastern Siberia,  a  number  of  old  type  harpoon  heads  have  been  re- 
ported. Those  in  Rasmussen's  collection  from  East  Cape,  figured 
by  Mathiassen  (1927,  vol.  2,  fig.  11,  1-6)  belong  for  the  most  part 
to  the  late  Punuk  stage ;  except  for  6,  which  appears  to  be  a  local 
form,  the  others  conform  to  types  which  at  Gambell  were  found  only 
at  the  three  later  sites.  Bogoras  (1904-09,  vol.  i,  p.  117,  fig.  33) 
illustrates  several  harpoon  heads  excavated  from  old  Eskimo  and 
Chukchee  villages  (probably  only  Eskimo)  which  appear  to  be  older; 
two  of  these,  c  and  d,  are  probably  early  Punuk.  Sverdrup  obtained 
several  Old  Bering  Sea  harpoon  heads  from  sites  on  the  Siberian  side 
of  Bering  Strait  and  the  writer  has  described  another  from  Plover 
Bay  (Collins,  1929,  pp.  4,  5,  fig.  i).   The  indications  are,  therefore, 
