NO.    I  ARCHEOLOGY    OF    ST.    LA\\RENCE    ISLAND — COLLINS  l6l 
given  a  rounded  cutting  edge  by  beveling  from  both  sides.  In  a  few 
instances  the  ends  are  pointed  ;  one  of  these,  plate  49,  figure  3,  also 
differs  in  the  manner  of  hafting :  at  the  upper  end  of  the  convex 
side  there  are  two  deep  grooves  separated  by  a  prominent  ridge,  and 
the  opposite  side  has  been  hacked  to  a  flat  surface  for  a  distance  of 
10.5  cm  in  order  to  provide  a  firm  resting  place  for  the  end  of  the 
handle.  Usually  the  upper  end  of  the  pick  was  prepared  for  the  handle 
by  scarfing  along  one  side  with  the  opposite  side  roughened  to  hold 
the  lashing  (pi.  49,  figs.  2,  5,  6).  Some,  like  plate  49.  figure  i,  have 
two  large  widely  spaced  grooves  on  one  side  with  a  single  groove 
opposite.  The  most  finished  type  is  illustrated  by  plate  42,  figure  4, 
which  has  two  deeply  cut,  rounded,  transverse  grooves.  Plate  49. 
figure  7,  was  lashed  to  the  handle  by  means  of  two  small  notches. 
Ivory  ivcdgcs. — Ivory  wedges,  mainly  for  splitting  wood,  were 
found  in  large  numbers  both  at  the  Hillside  site  and  Miyowagh  (pi. 
47,  figs.  8,  9).  They  are  usually  made  from  the  distal  end  of  the  tusk, 
which  was  beveled  to  a  rather  sharp  edge  ;  the  upper  end  usually  shows 
signs  of  hammering.  Such  wedges  were  often  used  secondarily  as 
drill  rests,  as  in  the  case  of  one  of  those  illustrated.  The  average  length 
is  around  12  cm. 
In  plate  29,  figure  17,  and  plate  47,  figures  12  and  13.  are  shown  three 
wedge-shaped  ivory  objects  from  the  Hillside  site  and  Miyowagh 
respectively,  which  are  characteristic  of  the  Old  Bering  Sea  culture, 
but  the  use  of  which  is  uncertain.  Plate  47,  figure  12,  is  6  cm  long; 
the  upper  end  is  1.4  cm  wide  by  i  cm  thick,  and  the  lower  end  2.3 
cm  wide  by  .3  cm  thick.  On  both  sides  the  right  half  has  been  cut 
down  lower  than  the  left,  thus  producing  two  opposite  offsets  which 
would  help  to  hold  the  object  firmly  in  place  if  it  were  inserted  wedge- 
like in  a  socket  shaped  in  a  similar  manner.  The  small  hole  in  the 
lower,  thin  end  was  probably  for  a  peg  or  rivet  for  fastening  to  the 
socket.  Figure  13  is  made  in  the  same  way,  but  is  somewhat  thicker 
and  less  tapering.  Plate  29,  figure  17,  from  the  Hillside  site,  shows 
the  same  characteristic  features.  It  is  possible  that  these  objects  were 
wedges  which  were  driven  into  the  end  of  the  wooden  adz  handle  to 
hold  it  tightly  in  place  in  the  socket. 
Mattock  blade. — Plate  50.  figure  2.  is  a  broken  mattock  blade,  made 
from  a  section  of  a  whale  rib.  The  upper  end  is  roughened  for  the 
lashing. 
Snoiv  shovels. — In  plate  50,  figure  6  (cut  18,  depth  49  in.)  is  an 
example  of  the  Old  Bering  Sea  type  of  snow  shovel,  made  from  a 
walrus  scapula.  The  concave  inner  surface,  shown  in  the  photograph, 
served  as  the  front  or  upper  side  of  the  shovel ;  it  was  smoothed  down 
