NO.    I  ARCHEOLOGY   OF    ST.    LAWRENCE    ISLAND— COLLINS  277 
The  following  is  their  method  of  building:  they  dig,  in  the  ground,  an  oblong 
pit,  which  rarely  exceeds  fifty  feet  in  length,  and  twenty  in  breadth;  but  the 
dimensions  are  in  general  smaller.  Over  this  excavation  they  form  the  roof  of 
wood,  which  they  cover  first  with  grass,  and  then  with  earth,  so  that  the 
external  appearance  resembles  a  dung-hill.  Near  each  end  of  the  roof  is  left 
a  square  opening,  which  admits  the  light ;  one  of  these  openings  being  intended 
only  for  this  purpose,  and  the  other  being  also  used  to  go  in  and  out  by,  with 
the  assistance  of  a  ladder ;  or  rather  a  post,  in  which  steps  are  cut.  In  some 
of  the  houses  there  is  another  entrance  below,  but  this  is  rather  uncommon. 
Round  the  sides  and  ends  of  the  habitations,  the  families,  several  of  which 
dwell  together,  have  their  separate  apartments,  where  they  sleep,  and  sit  at 
work;  not  on  benches,  but  in  a  sort  of  concave  trench,  dug  entirely  round  the 
inside   of  the   house,   and   covered  with   mats,   so  that  this  part  is   kept  pretty 
clean  and  decent We  did  not  observe  a  fire-place  in  any  one  of  their 
habitations.   They  are  lighted,  as  well  as  heated,  by  lamps ;  .  .  .  . 
Cook's  statement  that  the  Aleut  house  had  "  a  sort  of  concave 
trench,  dug  entirely  round  the  inside  of  the  house  ",  which  took  the 
place  of  the  customary  platform,  is  somewhat  surprising  in  the  light 
of  our  knowledge  of  Eskimo  and  other  northern  houses  generally. 
The  possibility  suggests  itself  that  what  was  observed  might  not  have 
been  an  excavated  trench  but  a  narrow  space  along  the  walls  enclosed 
by  logs  laid  end  to  end.  These  logs  (which  are  visible  in  Webber's 
drawing,  Cook,  vol.  2,  opposite  p.  no)  served  to  set  off  the  sleeping 
places  from  the  rest  of  the  room ;  and  the  enclosed  space  when  viewed 
from  a  distance  might  have  a  somewhat  sunken  appearance.  A  similar 
arrangement  of  logs  bordering  the  sleeping  spaces  along  the  side 
and  rear  walls  was  reported  from  Kodiak  Island  by  Lisiansky  (1814, 
P-  213). 
This  early  historic  form  of  the  Aleutian  house  appears  so  far 
reinoved  from  Eskimo  houses  generally,  and  the  Point  Barrow  house 
particularly,  that  it  is  impossible  to  regard  it,  as  Mathiassen  does,  as 
a  variant  of  the  Point  Barrow  house  (1927,  vol.  2,  p.  153);  and 
Birket-Smith's  statement  (1929,  vol.  2,  p.  51)  that  it  is  "a  local 
alteration  of  the  usual  type,  obviously  arising  out  of  the  building 
together  of  several  smaller  houses  "  does  not  explain  it  at  all.  In  two 
important  features — the  absence  of  an  entrance  room  and  the  presence 
of  a  ladder  and  roof  entrance  through  the  smoke  hole — it  differs  from 
other  Alaskan  houses.  Conversely,  it  is  these  very  features  that  link 
the  Aleutian  house  with  that  of  the  Kamchadal. 
From  Holmberg's  description  (1856,  pp.  376-378)  we  see  that  the 
house  of  the  Kodiak  Eskimo  was  a  semisubterranean  wooden  struc- 
ture, consisting  of  a  central   inner  room  to  which  were  connected 
several  smaller  sleeping  rooms.    The  walls  were  formed  of  vertical 
19 
