NO.    I  ARCHEOLOGY   OF    ST.    LAWRENCE    ISLAND — COLLINS  283 
Still  in  use  at  East  Cape  were  all  to  be  regarded  as  Thule  houses.  In 
one  important  respect,  however,  the  East  Cape  houses  do  not  conform 
to  the  Thule  type,  for  instead  of  a  platform  they  have  an  inner  sleep- 
ing compartment  or  polog  like  other  Siberian  Eskimo  and  Chukchee 
houses.  The  rounded  shape  of  the  stone  foundation  and  the  skin  roof 
covering  are  also  features  which  apparently  are  to  be  explained  on 
the  basis  of  a  relationship  with  the  other  Chukchee  and  Eskimo  forms. 
For  this  reason  I  should  regard  these  East  Cape  houses  as  only  vari- 
ants of  the  prevailing  type  of  modern  house  which  is  octagonal  in 
shape,  with  a  domed  skin-covered  roof  and  a  skin  sleeping  room  on 
the  inside. 
Except  for  these  modern  Siberian  houses,  which  are  obviously 
related  to  the  circular  dome-shaped  houses  of  the  nomadic  tribes  to 
the  westward,  and  which  have  only  in  relatively  recent  times  sup- 
planted the  underground  wood  and  whale  bone  house  in  northeastern 
Siberia,  there  seems  to  be  no  evidence  of  circular  houses  anywhere 
in  this  region.  The  only  examples  cited  by  Mathiassen  as  showing 
that  round  houses  have  occurred  in  Alaska  are  some  ruins  described 
by  Nelson  on  Kotzebue  Sound : 
On  Elephant  point,  at  the  head  of  Kotzebue  Sound,  I  saw  the  site  of  an  old 
village,  with  about  fifteen  pits  marking  the  locations  of  the  houses.  The  pits 
sloped  toward  the  center  and  showed  by  their  outlines  that  the  houses  had  been 
small  and  roughly  circular,  with  a  short  passageway  leading  into  them,  the  entire 
structure  having  been  partly  underground.    [Nelson,  1899,  pp.  264,  265.] 
But  aside  from  the  fact  that  the  outward  appearance  of  a  house 
pit,  after  it  has  been  long  abandoned,  gives  little  indication  of  its 
original  shape,  there  seems  to  be  direct  evidence  that  the  particular 
houses  referred  to  by  Nelson  were  originally  square.  In  Beechey's 
account  of  his  exploration  of  Kotzebue  Sound  in  1826  we  find  the 
following  reference  to  a  deserted  village,  apparently  the  same  one  seen 
by  Nelson  in  1881 : 
We  sailed  up  the  bay,  which  was  extremely  shallow,  and  landed  at  a  deserted 
village  on  a  low  sandy  point,  where  Kotzebue  bivouacked  when  he  visited  the 
place,  and  to  which  I  afterwards  gave  the  name  of  Elephant  Point,  from  the 
bones  of  that  animal  being  found  near  it The  deserted  village  upon  the 
low  point  consisted  of  a  row  of  huts,  rudely  formed  with  drift-wood  and  turf, 
about  six  feet  square  and  four  feet  in  height.    [Beechey,  1831,  p.  222.] 
Both  Waterman  and  Birket-Smith  have  attempted  to  show  that 
the  round  house  was  the  original  Eskimo  form.  In  speaking  of  Alaskan 
houses.  Waterman  says  (1927,  p.  211)  : 
Houses  were  built  in  a  circular  style  when  the  first  European  explorers  arrived, 
and  nobody  as  far  as  I  know  has  ever  suggested  the  existence  of  any  type  that 
