NO.    I  ARCHEOLOGY    OF    ST.    LAWRENCE    LSLAND — COLLINS  279 
on  the  Northwest  Coast  that  Birket-Sniith  seeks  the  origin  of  the 
rectangular  house  of  the  Western  Eskimo,  while  denying  a  relation- 
ship between  the  latter  and  the  houses  of  northern  Asia. 
The  rectangular  houses  among  the  Western  Eskimos  are  in  direct  geographical 
connection  with  the  rectangular  houses  which  extend  right  from  Alaska  south- 
wards to  northern  California.  All  these  dwellings  are,  like  the  Eskimo  dwellings, 
partly  dug  down  in  the  ground  and  inside  are  furnished  with  platforms.  With 
regard  to  construction  there  is,  it  is  true,  a  great  difference  between  the  Eskimo 
type  which  we  find  in  its  most  primitive  form  at  Point  Barrow,  and  the  big 
plank  buildings  of  the  nearest  Indian  peoples,  the  Tlingit,  Haida  and  Tsimshian; 
but  from  central  Washington  southwards  we  find  smaller  wooden  houses  which, 
with  their  vertical  wall-boards  and  lack  of  interior  timbering,  are  surprisingly 
like  the  Point  Barrow  type.  The  principal  difference  between  them  is  that  the 
Indian  forms  have  no  entrance  passageway ;  this,  however,  conforms  with  the 
fact  ....  that  the  entrance  passage  in  the  Alaskan  house  has  apparently  been 
taken  from  earlier  types  for  practical  reasons.  Just  this  divided  diffusion  of 
the  simple,  rectangular  houses,  interrupted  partly  by  the  peculiar  "  shedlike " 
type  at  Puget  Sound  and  Juan  de  Fuca  Strait,  which  are  perhaps  connected 
with  the  advance  of  the  Salish  to  the  coast,  and  partly  by  the  more  developed 
forms  among  the  three  principal  peoples,  the  Tlingit,  Haida  and  Tsimshian, 
argue  an  inner  connection  which  points  in  the  direction  of  the  North  Pacific 
coast  as  the  home  of  the  rectangular  house  of  the  Western  Eskimos,  as  it  has 
been  for  many  of  their  other  culture  elements  (their  rich  wood-carving,  .the 
grotesque  masks,  the  potlatch  feasts,  slavery,  the  raven  myths,  head  trophies, 
the  finer  development  of  basket  work,  etc.  etc.).  [Birket-Smith,  1929,  vol.  2, 
pp.  47,  48.] 
I  must  confess  that  I  am  unable  to  see  the  cogency  of  this  argu- 
ment unless  structure  is  to  be  wholly  subordinated  to  considerations 
of  geographical  position  and  of  cultural  relations  which  are  of  doubt- 
ful value  in  this  connection  because  of  the  strong  probability  that 
they  are  relatively  recent  manifestations.  If  we  leave  out  of  considera- 
tion the  Point  Barrow  house,  which,  as  the  most  divergent  of  the 
Alaskan  Eskimo  houses,  is  also  the  farthest  removed  from  the  Asiatic 
forms,  we  see  that  there  are  clear  and  unmistakable  resemblances  of 
a  fundamental  nature  between  the  Alaskan  Eskimo  houses  south  of 
Bering  Strait  (and  also  the  Mackenzie  house)  and  those  of  the  peoples 
of  northeastern  Asia.  The  Northwest  Coast  houses,  on  the  other 
hand,  dififer  fundamentally  from  the  Siberian  and  Eskimo  forms  in 
such  features  as  shape,  roof  structure  and  support,  sleeping  platforms, 
wall  structure,  and  the  form  of  entrance.  It  may  well  be,  as  Birket- 
Smith  suggests,  that  the  houses  of  the  modern  Northwest  Coast  tribes 
may  not  represent  the  original  type  there ;  but  they  diiTer  in  so  many 
important  respects  from  the  houses  both  to  the  north  and  south  that 
it  is  difficult  to  postulate  just  what  the  ancestral  form  may  have 
been.  Birket-Smith  finds  support  for  the  view  that  the  western  Eskimo 
