NO.    I  ARCHEOLOGY    OF    ST.    LAWRENCE    ISLAND COLLINS  315 
necessary  for  the  production  of  these  complicated  later  forms?  Un- 
doubtedly some  developments  of  this  nature  have  occurred  ;  the  barbed 
Thule  type  2  head,  for  instance,  is  no  doubt  the  ancestral  form  that 
gave  rise  to  the  various  barbed  types  of  the  Hudson  Bay  region  and 
Greenland,  just  as  it  is  also  ancestral  to  the  late  barbed  forms  of 
Arctic  Alaska.  But  the  most  characteristic  features  of  many  of  the 
eastern  harpoon  heads  are  a  "  flat  "  shape,  a  bifurcated  symmetrical 
spur,  and  a  curving  line  hole  with  both  openings  on  the  same  side. 
These  are  features  which  none  of  the  four  Thule  types  possesses, 
although  some  of  the  less  extreme  forms  of  these  "  flat  "  heads  are 
already  present  at  the  Thule  sites.  If  the  Thule  heads  are  not  adapted 
structurally  to  have  served  as  the  prototype  of  these  flat  forms,  where 
are  we  to  look  for  their  origin  ?  It  seems  to  me  that  the  Dorset  harpoon 
heads  are  to  be  considered  in  this  connection.  That  there  is  a  rela- 
tionship of  some  kind  can  hardly  be  doubted,  for  the  curious  arrange- 
ment of  the  line  hole  and  the  prominent  divided  spur  of  the  modern 
central  type  are  also  found  on  some  of  the  larger  forms  of  the  Dorset 
heads  (c.  f .,  Jenness,  1925,  fig.  5,  a,  b,  both  modern  types,  with  i,  j,  k, 
Dorset  types).  The  only  real  difference  between  them  is  that  in  the 
Dorset  heads  the  closed  socket  is  narrow  and  rectangular  whereas 
in  the  later  heads  it  is  round.  Assuming  these  larger  and  heavier 
Dorset  heads  to  represent  the  later  forms,  their  prototypes  would  be 
the  smaller,  thinner  forms  with  the  same  kind  of  rectangular,  enclosed 
sockets  and  bifurcated  spurs,  but  with  the  double  line  holes  extending 
through  from  side  to  side. 
In  accordance  with  his  view  that  the  Dorset  is  nothing  more  than  a 
locally  stamped  phase  of  the  Thule  culture,  Mathiassen  (1927,  vol.  2, 
p.  30)  has  sought  to  include  the  Dorset  heads  in  the  Thule  group : 
With  the  occurrence  of  these  harpoon  heads  in  the  Thule  finds,  with  their  wide 
spread  from  King  WiUiam's  Land  to  North  Greenland  and  with  the  occurrence 
of  their  prototype  by  the   Bering   Strait we   can  no  longer  look  upon 
these  harpoon  heads  as  a  local  phenomenon ;  they  must  be  regarded  as  a  link 
in  the  Thule  culture  and  must,  like  many  of  its  other  elements,  be  presumed  to 
have  originated  in  the  west. 
It  is  difficult  to  see  why  the  inclusion  of  a  few  Dorset  harpoon 
heads  among  the  Thule  finds  should  invalidate  them  as  a  type  and 
place  them  in  a  group  to  which  they  are  structurally  unrelated;  for 
with  the  exception  of  the  one  Dorset  form  which  has  an  open  socket 
and  single  spur  like  the  Thule  type  i,  the  others  stand  entirely  apart 
from  the  Thule  heads.  Furthermore,  as  Jenness  has  already  observed 
(1933-  P-  394).  the  single  harpoon  head  from  Siberia  which  Mathias- 
sen regards  as  the  prototype  of  the  Dorset  heads  is  not  at  all  closely 
