74  SMITHSONIAN    MISCELLANEOUS    COLLECTIONS  VOL.    96 
The  remains  of  the  rear  or  east  wall  of  house  no.  4  consisted  of 
two  whale  jaws  and  two  small  logs  laid  horizontally  behind  two  up- 
right corner  roof  supports  (pi.  8,  figs,  i,  2,  at  right  center,  and  pi.  9, 
figs.  4,  5).  The  larger  whale  jaw  was  cracked  so  that  it  was  some- 
what bowed  inward ;  it  rested  on  one  of  the  logs,  and  both  were 
propped  up  with  a  small  wooden  post  and  a  walrus  penis  bone.  The 
upper  whale  jaw  had  slipped  from  its  original  position  and  lay  about 
a  foot  to  the  rear  and  just  a  little  above  the  lower  (pi.  9,  fig.  4).  Be- 
tween the  two  lay  the  second  small  log.  Three  feet  back  and  about 
one  foot  above  the  whale  jaws  was  the  remnant  of  another  wall 
section  or  perhaps  of  the  same  wall.  It  consisted  of  four  pieces  of 
logs  and  a  fifth  fragment  superimposed,  with  a  stake  wedged  in 
between. 
The  south  wall  (pi.  9,  fig.  2,  in  background)  consisted  of  a  tier  of 
horizontal  timbers  and  a  piece  of  whale  jaw,  the  ends  extending  2 
feet  beyond  the  SE.  corner  post.  The  wall  was  braced  on  the  inside 
by  wooden  stakes  and  a  walrus  penis  bone.  Plate  9,  figure  4  (extreme 
lower  right  corner),  shows  the  outside  of  the  wall  at  the  SE.  corner, 
consisting  of  three  superimposed  logs,  just  behind  the  corner  post. 
(The  post  is  shown  wrapped  with  string  to  keep  it  from  disintegrat- 
ing.)   At  the  SW.  corner  was  another  large  upright  corner  post. 
The  north  wall  consisted  of  three  superimposed  logs,  the  back  ends 
of  which  are  visible  at  the  right  foreground,  plate  8,  figure  3,  behind 
the  cord-wrapped  NE.  corner  post.  The  end  of  one  of  the  logs  pro- 
jected 8  inches  into  house  no.  3,  at  a  level  just  below  the  floor.  Above 
the  wall  timbers  of  house  no.  4  and  partly  beneath  those  of  the  larger 
house,  no.  3,  lay  a  row  of  small  timbers  which  had  fallen  inward 
from  a  vertical  position.  They  could  hardly  have  been  part  of  a  fallen 
roof,  for  the  outer  ends  were  resting  directly  on  the  uppermost  of 
the  three  wall  timbers,  and  the  latter  alone  would  not  have  been  high 
enough  to  have  constituted  the  entire  wall.  More  likely  they  were 
supplementary  wall  timbers,  which  had  been  set  into  the  ground,  either 
upright  or  leaning  slightly  inward,  just  behind  the  horizontals  which 
formed  the  base  of  the  wall.  Twelve  of  these  fallen  timbers  remain, 
but  there  were  evidently  others,  extending  to  the  NW.  corner,  which 
had  been  broken  into  and  removed  in  the  construction  of  the  later 
house,  no.  3.  Further  and  unmistakable  evidence  of  the  superposition 
at  this  point  is  afforded  by  the  two  stakes  which  had  been  driven  into 
the  ground  to  help  support  the  rear  wall  of  the  later  house.  Plate  8, 
figure  2,  shows  the  stake  (at  the  left  center)  placed  against  the  inside 
of  the  wall  of  the  later  house,  and  just  opposite  it  the  similar  stake 
which  had  been  driven  into  the  gravel  overlying  the  fallen  walls  of 
