NO.    I  ARCHEOLOGY    OK    ST.    LAWRENCE    ISLAND COLLINS  285 
are  placed  upright  there  is  nothing  to  prevent  their  taking  a  circular 
outHne  as,  in  fact,  so  many  American  Indian  houses  do,  while  the 
roof  may  according  to  custom  be  either  domed,  conical,  gabled  or 
flat.  This  has  been  pointed  out  by  Hatt  (1928,  p.  8)  who  also  cites 
the  rectangular  whale  bone  house  at  Nunligren,  northeastern  Siberia, 
described  by  Bogoras,  as  showing  that  whale  bone  houses  need  not 
be  round.  To  this  we  might  add  that  the  underground  caches  on  St. 
Lawrence  Island,  constructed  either  entirely  of  whale  bones  or  of 
whale  bones  together  with  stones  and  walrus  skulls,  are  always  quadri- 
lateral. Some  of  the  comparatively  late  St.  Lawrence  Island  houses 
had  circular  annexes  to  or  enlargements  of  the  entrance  passage,  used 
evidently  for  storage,  the  walls  of  which  were  of  whale  and  walrus 
bones  and  stones,  but  the  houses  themselves,  built  of  precisely  the 
same  materials,  were  invariably  square  or  rectangular.  When  we  see 
that  St.  Lawrence  Island  houses  from  the  time  of  the  Old  Bering  Sea 
culture  down  to  the  nineteenth  century  were  quadrilateral  in  shape 
regardless  of  whether  they  were  constructed  of  wood,  bones,  or  stone, 
we  are  led  to  the  conclusion  that  custom  alone  is  the  determining  factor 
and  that  the  nature  of  the  building  material  in  no  way  dictates  the 
shape  that  the  structure  is  to  assume. 
If  we  examine  the  various  types  of  houses  which  have  been  revealed 
on  St.  Lawrence  Island,  we  see  that  it  is  only  one  of  these — the  latest 
of  the  prehistoric  types — that  shows  any  close  agreement  to  the  general 
style  of  Eskimo  house  found  below  Bering  Strait.  This  house  (type  3, 
p.  261),  which  seems  to  have  appeared  late  in  the  Punuk  period,  con- 
tinued in  use  until  about  40  years  ago.  It  conforms  to  the  southwest 
Alaskan  type  in  having  the  roof  supported  by  uprights  rising  from 
the  floor,  in  having  the  walls  made  of  small  vertical  timbers  or  whale 
jaws,  leaning  slightly  inward,  and  in  the  presence  of  low  sleeping 
platforms  along  two  or  more  sides.  The  sudden  appearance  of  this 
type  of  house,  which  embodies  features  not  previously  known  on  St. 
Lawrence,  points  to  its  being  an  importation  from  some  other  region. 
It  would  be  premature  at  the  present  stage  of  our  knowledge  to  say 
whether  it  had  been  introduced  from  southwest  Alaska  (Bristol  Bay  to 
the  Yukon),  where  houses  of  this  general  type  prevail,  or  from  north- 
eastern Siberia.  Since  Siberia  was  the  source  from  which  many  other 
cultural  traits  were  received  during  the  Punuk  period,  the  house  might 
have  been  acquired  in  the  same  way.  That  houses  of  this  general  type 
occurred  in  Siberia  is  shown  by  Bogoras'  description  of  a  rectangular 
whale  bone  house  at  Nunligren,  but  until  we  have  more  precise  in- 
formation on  Siberian  house  types,  the  exact  relationships  of  the  St. 
Lawrence  Island  form  must  remain  uncertain. 
