38  SMITHSONIAN    MISCELLANEOUS    COLLECTIONS  VOL.    96 
of  the  slope,  and  between  the  two  points  there  were  a  great  many 
rocks,  large  and  small,  which  may  have  been  parts  of  fallen  walls, 
etc. ;  if  SO,  they  were  too  scattered  to  tell  their  story.  The  flat  stones 
of  the  floors  were  the  only  ones  that  with  certainty  could  be  said  to 
have  belonged  to  a  house.  At  the  NW.  end — down  the  slope — some 
of  the  larger  stones  seemed  to  suggest  an  entrance  but  this  again  was 
uncertain.  The  possibility  that  the  house  may  have  been  demolished 
by  a  landslide  is  seen  in  the  fact  that  at  two  places  on  the  south  side 
there  were  what  appeared  to  be  remnants  of  walls — several  medium- 
sized  stones,  one  above  the  other — supporting  two  huge  stones  which 
might  have  rolled  down  upon  them  (pi.  5,  fig.  i).  An  alternative  ex- 
planation would  be  that  the  large  stones  were  already  there  and  that 
the  small  stones  had  been  so  placed  in  order  to  keep  them  from  slipping. 
More  conclusive  evidence  of  a  rock  slide  which  descended  over  the 
spot  after  the  house  had  been  abandoned  is  seen  in  the  presence  of 
a  number  of  large,  heavy  rocks  lying  above  the  refuse  both  in  the 
paved  area  and  further  down  the  slope.  These  were  so  generally  dis- 
tributed over  the  site  that  such  an  explanation  seems  more  likely  than 
that  they  all  represented  the  remains  of  fallen  walls. 
Fortunately,  the  site  was  prolific  in  artifacts,  which  provided  the 
much  needed  information  concerning  the  implement  types  of  the  Old 
Bering  Sea  culture,  even  though  the  house  remains  were  so  disappoint- 
ing. The  excavation  which  revealed  the  house — or  paved  area — yielded 
yy  artifacts  of  bone,  ivory  and  wood,  and  62  artifacts  of  stone,  in 
addition  to  large  quantities  of  potsherds,  animal  bones,  pieces  of  stone 
implements,  and  worked  pieces  of  ivory,  bone,  and  baleen.  Twenty- 
six  artifacts  were  obtained  from  between  and  below  the  floor  stones, 
and  72  artifacts,  in  addition  to  numerous  potsherds  and  fragmentary 
objects  of  stone,  ivory,  and  bone,  were  obtained  from  patches  of 
refuse  found  at  various  places  in  the  rock  slide  adjoining  the  excava- 
tion. The  midden  between  the  paved  area  and  the  foot  of  the  slope 
yielded  188  artifacts  besides  quantities  of  fragmentary  objects.  The 
midden  ranged  in  depth  from  a  little  less  than  i  foot  at  the  bottom 
of  the  slope  to  3  feet  at  the  upper  end  and  was  taken  down  in  layers 
approximately  9  inches  in  thickness.  The  conditions  for  excavating 
at  the  Hillside  site  were  extremely  unfavorable.  The  ground  was  kept 
in  a  constant  state  of  saturation  from  the  melting  of  a  large  bank 
of  snow  high  up  the  slope,  so  that  as  the  soil  thawed  out  it  was 
promptly  converted  into  mud.  As  a  result  of  this  condition  and  also 
because  of  the  many  large  stones  which  were  encountered,  the  work 
at  the  Hillside  site  progressed  more  slowly  than  at  the  middens  situ- 
ated on  the  gravel  plain. 
