312  SMITHSONIAN    MISCELLANEOUS    COLLECTIONS  VOL,    96 
48;  1935,  p.  462).  The  oldest  site  at  which  Ford  excavated  was 
Birnirk.  Here  the  bulk  of  the  harpoon  heads  were  of  the  type  that 
had  been  previously  designated  by  that  name — the  Birnirk  type — being 
made  usually  of  bone,  with  an  open  socket,  a  side  blade  of  chipped 
flint  with  an  opposite  barb  (sometimes  with  another  blade  in  place  of 
the  barb),  and  with  two  or  more  asymmetrical  spurs  at  the  base.  Only 
one  harpoon  head  of  the  Thule  no.  2  type  (with  two  barbs)  was  found 
at  Birnirk.  However,  this  type  was  common  at  the  more  recent  site 
of  Utkiavik,  where  somewhat  later  it  developed  into  forms  charac- 
teristic of  the  period  just  preceding  the  historic,  as  shown  by  their 
association  with  metal  and  late  types  of  implements.  There  is  no 
certain  evidence  here  as  to  the  origin  of  the  Thule  type  2  head,  al- 
though on  typological  grounds,  it  might  be  considered  a  descendant  of 
the  Birnirk  type  (Collins,  1935,  pi.  n,  fig.  i)  in  which  the  single  side 
blade  had  been  replaced  by  another  barb  and  the  bifurcated  spur  by 
a  plain  spur.  But  whether  or  not  the  Thule  head  developed  in  this 
particular  manner,  the  significant  fact  revealed  by  Ford's  excavations 
is  that  it  is  not  one  of  the  earliest  forms  of  the  Barrow  sequence,  but 
a  form  that  appeared  after  the  typical  Birnirk  head  and  that  eventually 
gave  rise  to  the  dominant  protohistoric  closed  socket  type,  which  was 
flat  on  one  side,  convex  on  the  other,  and  with  one  or  two  pairs  of 
opposed  barbs  (Collins,  1935,  pi.  n,  fig.  4). 
With  regard  to  the  Thule  type  3  head  (Mathiassen,  1927,  vol.  i, 
pi.  I,  figs.  6-12),  we  now  see  that  this  belongs  to  that  variable  group 
of  harpoon  heads  which  in  one  form  or  another  has  held  an  important 
place  in  the  West  from  the  Old  Bering  Sea  period  down  to  modern 
times.  At  Gambell  the  earliest  forms  ( from  the  Hillside  site  and  the 
older  section  of  Miyowagh)  described  as  open  socket  type  HI  x  (pi. 
23,  fig.  4;  pi.  24,  figs.  11-18)  were  large,  had  prominent  asymmetrical 
spurs  that  were  tri furcated,  bifurcated,  or  otherwise  irregular,  and 
were  decorated  in  Old  Bering  Sea  style.  From  this  old  form  there 
developed  through  gradual  modification  a  simpler  form,  marked  by  a 
reduction  in  size,  by  a  simpler  though  still  irregular  form  of  spur, 
and  usually  an  early  Punuk  ornamentation  (pi.  24,  figs.  19-23).  This 
relatively  simple  form  in  turn  gave  way  to  a  still  simpler  form  (open 
socket  type  HI  (a)  x,  pi.  28,  figs.  8-14),  which  had  a  straight  plain 
spur,  plain  edges,  and  a  Punuk  decoration.  This  form,  which  at 
Miyowagh  was  practically  restricted  to  the  later  northwestern  sec- 
tion of  the  midden,  continued  in  use  at  levoghiyoq,  where  it  was 
represented  by  98  examples  (pi.  70,  figs.  8-15).  Most  of  these,  how- 
ever, had  undergone  additional  slight  modifications,  and  again  in  the 
direction  of  simplification,  becoming  somewhat  straighter  and  heavier 
