NO.    I  ARCHEOLOGY    OF    ST.    LAWRENCE    ISLAND COLLINS  5/ 
of  less  than  5  feet.  However,  this  is  due  to  the  encroachment  of  the 
gravel,  for  the  bottom  of  the  midden  was  found  to  extend  several  feet 
below  the  level  of  the  surrounding  plain,  at  one  place  reaching  a  total 
depth  of  8  feet.  In  spring  and  early  summer  the  small  stream  which 
flows  down  among  the  rocks  of  the  Hillside  site  passes  along  the 
southeastern  edge  of  the  Miyowagh  midden  and  empties  into  the  lake. 
House  pits  are  represented  by  shallow  depressions,  more  or  less  cir- 
cular in  outline.  Some  18  or  20  of  these  pits  were  visible,  but  the 
surface  of  the  midden  had  been  so  disturbed  by  the  Eskimos  digging 
for  ivory  that  it  was  not  always  possible  to  recognize  a  house  pit 
with  certainty.  The  surface  of  the  midden  was  practically  free  of 
whale  bones  and  timbers,  such  as  are  found  at  some  of  the  more 
recently  abandoned  sites.  No  remains  of  roof  or  wall  structures  were 
visible,  most  of  them  no  doubt  having  been  utilized  in  the  construc- 
tion of  later  houses. 
The  Miyowagh  midden,  as  well  as  the  three  other  middens  situated 
on  the  gravel  plain,  ditTered  in  several  respects  from  that  at  the  Hill- 
side site.  Being  on  the  level  ground,  it  had  not  been  subjected  to 
erosion  or  to  disturbances  from  the  shifting  or  sliding  of  rocks  as 
had  the  Hillside  midden.  As  a  result  the  Miyowagh  midden  represents 
a  full  record  (although  one  which  may  have  been  altered  here  and 
there  through  sloughing  or  from  the  excavation  of  house  pits), 
whereas  at  the  Hillside  site  the  present  configuration  of  the  surface 
indicates  that  an  unknown  amount  of  refuse  has  been  washed  down 
the  slope.  There  was  also  a  difference  in  the  nature  of  the  deposits : 
at  the  Hillside  site,  as  in  kitchen  middens  generally,  the  debris  of 
human  origin  was  mixed  with  a  certain  amount  of  natural  soil,  for 
the  gently  sloping  lower  part  of  the  hillside  is  covered  with  a  fairly 
deep  soil  layer.  At  Miyowagh,  on  the  other  hand,  the  first  houses 
were  built  on  a  barren  gravel  plain,  and  the  extensive  pile  of  debris 
which  slowly  accumulated  around,  and  eventually  over  them,  was  al- 
most entirely  artificial,  the  greater  part  of  it  resulting  from  the 
decomposition  of  animal  remains  and  other  organic  materials.  The 
midden  therefore  consists  of  a  heavy,  rich,  black  matrix  containing 
great  quantities  of  animal  bones  and  baleen  in  addition  to  artifacts 
and  rejectage  of  stone,  bone,  ivory,  wood,  and  baleen.  The  midden 
was  found  to  be  compact  and,  in  general,  unstratified.  Occasionally  a 
thin  layer  of  mussel  shells  or  a  mass  of  bones,  etc.,  would  stand  out 
distinctly,  and  along  the  western  periphery  lenses  of  midden  material 
were  sometimes  separated  by  masses  of  gravel.  As  a  rule,  however, 
there  was  no  stratification  in  the  usual  sense,  the  bones,  artifacts,  and 
rejectage  being  a  relatively  homogeneous  mass,  held  together  in  a 
