NO.    I  ARCHEOLOGY    OF    ST.    LAWRENCE    ISLAND COLLINS  69 
in  place  and  one  large  upright  timber  that  had  evidently  been  a  roof 
support. 
Cut  2/. — Excavated  by  M.  B.  Chambers  in  1931.  Near  the  center 
of  the  midden,  overlapping  the  SE.  corner  of  cut  19;  4  feet  4  inches 
deep ;  taken  down  in  1 1  levels ;  bottom  not  reached.  This  cut  revealed 
the  remains  of  a  rectangular  structure,  either  a  cache  or  a  small  house 
with  four  corner  posts  and  walls  of  horizontal  timbers.  Plate  7,  figure 
I,  looking  northwest,  with  cut  19  in  the  background,  shows  the  pro- 
jecting ends  of  three  of  the  corner  posts,  that  in  the  upper  right  corner 
being  a  whale  jaw,  the  others  logs ;  at  the  center  may  be  seen  a  hollow 
whale  vertebra,  probably  a  ventilator.  Plate  7,  figure  2,  is  a  view  from 
the  same  point  taken  later  when  the  wall  timbers  were  beginning  to 
appear.  Plate  7,  figure  3,  is  a  view  of  the  cut  looking  in  the  opposite 
direction,  southeast,  with  cut  19  in  the  foreground.  The  four  corner 
posts  are  all  visible,  one  of  them,  in  the  foreground,  with  a  section  of 
wall  abutting  (the  post  and  wall  timbers  are  lashed  together  to  pre- 
vent them  from  falling  as  the  earth  was  removed  from  around  them). 
The  timbers  behind  the  man  may  also  have  been  part  of  the  wall. 
From  this  cut  came  270  artifacts.  In  contrast  to  cut  19,  there  were 
more  Old  Bering  Sea  than  Punuk  pieces.  The  latter  is  represented 
by  a  fragmentary  wrist  guard  from  a  depth  of  9  inches  (pi.  22, 
fig.  10)  and  a  harpoon  head,  16  inches  deep  (pi.  28,  fig.  20)  ;  the 
Old  Bering  Sea  pieces  come  from  the  following  depths:  29  inches 
(pi.  24,  fig.  12)  ;  32  inches;  38  inches. 
HOUSES    AT    MIVOWAGTT 
The  house  remains  mentioned  above,  which  were  found  at  or  near 
the  base  of  the  midden  at  a  number  of  places,  were  too  poorly  pre- 
served to  afford  any  clear  idea  as  to  their  exact  size,  shape,  and  manner 
of  construction.  fJowever,  fragmentary  though  the  evidence  was,  it 
indicated  that  the  earliest  houses  at  Miyowagh,  like  those  at  the  Hill- 
side site,  had  stone  floors,  a  narrow  entrance  passage  at  a  lower  level 
than  the  house  floor,  and  walls  built  of  horizontal,  superimposed 
timbers.  Fortunately,  two  other  house  ruins  were  found,  which  pro- 
vide a  much  better  although  by  no  means  a  wholly  satisfactory  basis 
for  reconstructing  the  house  types  of  the  Old  Bering  Sea  and  early 
Punuk  periods.  These  are  the  two  houses,  nos.  3  and  4,  which,  as 
mentioned  above,  were  found  in  cut  9  and  in  the  area  immediately 
to  the  south.  Although  the  excavation  of  cut  9  was  begun  on  August 
I  and  continued  into  October,  the  major  part  of  the  work  on  the  two 
houses,  nos.  3  and  4,  was  conducted  by  J.  A.  Ford  during  my  absence. 
6 
