NO.    I  ARCHEOLOGY    OF    ST.    LAWRENCE    ISLAND COLLINS  37 
In  the  course  of  years  this  has  no  doubt  washed  away  most  of  the 
refuse  from  between  the  rocks.  At  several  places  refuse  was  found 
to  extend  beneath  the  rocks,  in  tilted  strata  resulting  from  water  ac- 
tion, which  no  doubt  had  produced  a  certain  amount  of  shifting  in  the 
positions  of  some  of  the  rocks  as  the  earth  binding  between  them  was 
washed  away. 
After  finding  refuse  among  the  rocks,  a  systematic  search  of  the 
adjoining  hillside  was  made.  Immediately  to  the  south  of  the  rock  slide, 
at  the  base,  a  small  test  pit  revealed  refuse  held  in  the  black  moist 
earth  beneath  the  heavy  sod.  A  trench  was  then  begun  along  the  bot- 
tom of  the  slope  and  continued  upward  (pi.  4,  fig.  5).  About  a  foot 
beneath  the  sod  refuse  became  abundant.  It  consisted  mainly  of  walrus 
and  seal  bones ;  there  were  also  bird,  dog,  and  a  few  fish  bones,  to- 
gether with  occasional  potsherds  and  artifacts  of  stone,  bone,  and  ivory. 
Realizing  that  the  refuse  at  the  bottom  must  have  rolled  down,  we 
next  started  an  excavation  25  feet  up  the  slope  to  see  if  a  house  could 
be  found  (pi.  4.  fig.  5,  small  excavation  in  background). 
Immediately  beneath  the  sod  potsherds  and  stone  chips  were  found 
in  considerable  numbers.  All  other  artifacts,  even  those  of  such  rela- 
tively durable  materials  as  bone  or  ivory,  seemed  to  have  undergone 
complete  disintegration  in  the  surface  layer,  although  they  were  found 
in  abundance  beginning  at  about  the  12-inch  level.  It  was  only  among 
the  rocks,  where  there  had  been  considerable  disturbance  of  original 
conditions,  that  artifacts  of  bone  and  ivory  were  found  nearer  the 
surface. 
At  a  depth  of  2  feet,  flat  stones  were  found,  covering  an  area  of 
about  10  feet  square;  these  were  closely  spaced  and  apparently  rei)re- 
sented  the  remains  of  an  old  floor  (pi.  4,  figs.  6,  7).  Three  small  and 
poorly  preserved  timbers  were  found  on  or  just  above  the  stones. 
The  pit  was  then  enlarged  so  as  to  afiford  a  better  view  of  the  struc- 
ture, which  will  be  designated  house  no.  i.  Immediately  to  the  north- 
ward more  refuse  was  found  but  the  stone  flooring  did  not  continue. 
Instead,  there  was  a  jumble  of  stones,  none  of  them  very  large,  which 
gave  the  impression  of  being  part  of  a  fallen  wall  (pi.  4,  fig.  8).  A 
similar  arrangement  of  fallen  stones  was  found  at  the  rear,  to  the 
eastward  (pi.  4,  figs.  6,  7).  Immediately  in  front  of  these  was  a  second 
layer  of  flat  stones,  about  a  foot  above  the  other,  which  at  first  we 
thought  to  be  a  house  platform.  However,  it  was  soon  found  that 
the  stones  rested  on  refuse  which  had  accumulated  on  the  lower  floor, 
showing  that  they  had  been  laid  down  at  a  later  period.  Unfortunately, 
no  further  details  of  the  structure  could  be  determined.  Midden  ma- 
terial was  found  all  the  way  from  the  paved  area  down  to  the  bottom 
4 
