308  SMITHSONIAN    MISCELLANEOUS    COLLECTIONS  VOL.    96 
direct  relationship.  It  is  by  no  means  improbable  that  these  Birnirk- 
like  heads  were  introduced  into  St.  Lawrence  Island  from  the  north. 
Most  of  the  harpoon  heads  which  Jenness  excavated  at  the  old 
mound  site  at  Wales  belonged  to  Mathiassen's  Thule  type  3.  On  St. 
Lawrence  Island  this  is  the  predominant  type  of  the  Punuk  period 
(open  socket  type  III  (a)  x),  of  which  161  examples  were  found 
at  Gambell.  However,  Jenness  found  that  the  lashing  arrangement  on 
this  type  of  harpoon  head  was  not  always  the  same: 
[the  modern  closed  socket  type]  was  preceded  at  Wales  and  the  Diomede 
islands  by  a  type  with  an  open  socket  into  which  the  foreshaft  of  the  harpoon 
was  lashed  through  a  series  of  holes  (either  two  or  three)  drilled  on  each  side 
of  the  socket.  Earlier  still,  a  rectangular  slot  on  each  side  fulfilled  the  same 
function  as  the  drilled  holes ;  or  else  the  edges  of  the  implement  were  trimmed 
away  for  the  lashing.    Only  the  two  last  forms  (considered  here  as  one  type) 
were   present   in  the  mound  dwellings  at  Wales The  rubbish  heap  on 
Little  Diomede  island  ....  showed  the  succession  fairly  clearly.  In  the  upper 
layers  there  were  only  closed-socketed  harpoon-heads  of  different  varieties ; 
lower  down,  nearly  all  had  open  sockets  with  drilled  holes  for  the  lashings; 
and  at  the  3-foot  level  the  open-socketed  type  with  rectangular  slots  or  no 
holes  at  all  for  the  lashing  began  to  predominate.    [Jenness,  1928  a,  p.  76.] 
In  describing  the  harpoon  heads  in  the  collection  which  Knud 
Rasinussen  purchased  from  the  Eskimos  at  East  Cape,  Siberia,  Ma- 
thiassen  (1930a,  p.  72)  remarks  on  the  fact  that  "of  the  harpoon 
heads,  the  Thule  types  are  very  much  in  the  majority."  Thule  types  i 
and  3  are  both  included  in  this  collection,  some  of  them  having  slots  and 
others  drilled  holes  for  the  lashing.  Because  these  two  lashing  ar- 
rangements occur  promiscuously  in  this  collection,  in  similar  collec- 
tions from  the  Arctic  coast  of  Alaska,  and  at  all  of  the  Thule  sites 
in  the  Central  region,  Mathiassen  (1930a,  p.  yy)  feels  that  "no 
general  rule  can  apply;  the  one  method  seems  to  be  just  as  ancient 
as  the  other.  There  is,  however,  something  that  might  indicate  that 
the  method  with  holes  is  the  one  that  has  held  out  longest."  However, 
if  we  base  our  conclusions  on  the  western  material  that  has  been  sys- 
tematically excavated,  this  question  appears  in  a  different  light.  First 
of  all,  we  see  that  these  East  Cape  harpoon  heads  can  have  no  bearing 
on  the  question  as  to  what  features  may  be  ancient  in  the  Bering 
Strait  region,  since  they  do  not  belong  to  the  older  types,  but  on  the 
contrary,  to  the  late  Punuk  types.  Thus,  the  small  bladeless  form, 
Thule  type  i,  does  not  appear  at  Gambell  until  very  late  in  the  se- 
quence— only  from  Seklowaghyaget,  the  old  section  of  Gambell,  and 
the  contemporaneous  house  ruins  which  appear  to  date  from  about  the 
eighteenth  century.  Most  of  these  Gambell  heads,  which  I  have 
described  as  open  socket  type  V  (pi.  71,  figs.  12,  13,  15,  16)  differ 
