NO.    I  ARCHEOLOGY    OF    ST.    LAWRENCE    ISLAND — COLLINS  99 
tion  by  letters  or  numbers,  such  as  Mathiassen  has  used  in  describing 
the  harpoon  heads  of  the  Thule  culture ;  or  second,  merely  to  describe 
and  give  an  appropriate  name  or  number  tO'  each  type.  As  stated 
above,  Alaskan  harpoon  heads  are  complex,  possessing  at  least  six 
variable  features,  the  inclusion  of  which  alone  would  require  cum- 
bersome formulae  for  the  designation  of  the  types.  At  that,  the 
features  included  would  be  only  those  which  would  correspond  in 
a  biological  sense  with  the  more  fundamental  divisions — class,  subclass, 
order,  family,  subfamily,  and  genus ;  a  still  further  division  would 
sometimes  be  necessary  to  include  the  specific,  subspecific,  and  varietal 
features  on  which  a  demonstration  of  close  genetic  relationship  would 
depend.  In  view  of  this  condition,  I  have  therefore,  after  fully  testing 
both  methods,  chosen  the  second,  that  of  merely  describing  and  nam- 
ing the  various  types,  with  numerals  and  letters  used  as  sparingly  as 
possible.  Eventually,  when  systematic  excavations  have  revealed  the 
sequence  of  harpoon  types  at  more  places  and  when  homologies  can 
be  clearly  distinguished  from  analogies,  it  should  be  possible  to  estab- 
lish a  comprehensive  classification  in  which  the  various  regional  forms 
can  be  arranged  in  the  order  of  their  development.  At  the  present 
time  this  cannot  be  done  for  the  entire  Eskimo  area,  not  only  because 
we  have  too  few  series  for  which  chronological  relationships  have  been 
definitely  established,  but  also  because  the  original  forms  remain  to 
be  determined.  Thus  we  have  no  knowledge  whatsoever  of  the  pre- 
sumably simpler  forms  antecedent  to  the  complicated  harpoon  heads 
of  the  Old  Bering  Sea  culture. 
The  excavations  at  Gambell  produced  abundant  evidence  of  a  num- 
ber of  stages  of  development  through  which  harpoon  heads  passed, 
beginning  with  the  complicated  and  numerous  forms  of  the  Old  Bering 
Sea  culture  and  ending  with  the  one  simple  form  now  in  use  on  St. 
Lawrence  Island  (text  fig.  24,  opposite  p.  216). 
At  the  Hillside  site  i6  harpoon  heads  were  found  which  were  suffi- 
ciently complete  for  type  identification;  14  of  these  are  shown  in 
plate  23.  Among  them  were  examples  of  the  several  Old  Bering  Sea 
types  already  known,  in  addition  to  other  types  previously  unknown. 
From  Miyowagh  there  are  164  harpoon  heads,  more  or  less  complete, 
in  addition  to  an  even  larger  number  of  fragments.  We  have  at  our 
disposal,  therefore,  a  large  body  of  material,  fully  documented  as  to 
depth  and  locality,  upon  which  to  base  our  study  of  the  harpoon  heads 
of  the  Old  Bering  Sea  and  early  Punuk  periods. 
The  harpoon  heads  are  divided  below  into  two  groups,  those  with 
open  sockets  and  those  with  closed  sockets.  The  types  are  indicated 
by  roman  numerals;  subtypes  by  letters  (a,  b,  c,  etc.)  in  parentheses; 
