43 
continued the work for the National Museum of Canada, he made a recon- 
naissance near the mouth of Skeena river in 1915, and some intensive 
excavations in 1919 on the northeastern part of Graham island, Queen 
Charlotte group — the most western large island of Canada. It is worth 
mentioning here that during his first visit to British Columbia he saw 
large quantities of clam and cockle shells thrown out of the houses of some 
Kwakiutl Indians on the north end of Vancouver island. Again, in 1919, 
he saw a small kitchen-midden and shell-deposit on the eastern side of 
Massett inlet, Graham island, so recent that the grass upon which it was 
built had not yet decayed. Thus certain middens in their entirety, and 
at least the upper layers of others, have been formed under the observation 
of Europeans within the past fifty years. Nevertheless investigations 
leave little doubt that the majority of middens along this coast date back 
many centuries. 
What methods are available for determining their antiquity? Else- 
where certain writers have attempted to calculate the age of middens by 
estimating the population at a given site, and the number of shells an 
Indian would throw out each day. Such a method seems impossible in 
British Columbia, where there is no information concerning the population, 
the duration of their occupation of the various sites, or the proportion of 
shell-fish to other foods in their diet. An abundance or scarcity of other 
food would decrease or increase the number of shell-fish consumed, and the 
number of shells thrown away. A large population, other things being 
constant, would discard more shells than a smaller population, but would 
probably scatter them over a larger area; and even a small population 
might build up a very deep heap if it occupied a site for a long period. 
Again, a site might be inhabited continuously, or only intermittently at 
certain periods of the year. Even if everything except population is 
discounted, the antiquity of a site, calculated from the number of shells 
contained in the kitchen-midden, would increase or decrease in inverse 
ratio to the number of people assumed to have inhabited it. 
The presence in a shell-heap of European or Asiatic objects of known 
antiquity might give some clue as to its date; but such objects have seldom 
been found in British Columbia middens. It would appear, therefore, 
that many of the heaps were abandoned before or soon after European 
contact. The absence of perishable materials, such as wood, horn, and 
feathers, confirms this conclusion. 
We may consider now the culture revealed by the shell-heaps, and 
inquire; (1) whether it differs from the culture of the present-day natives; 
(2) whether there are differences in the cultures of different shell-heaps; 
and (3) whether there are differences in culture in different layers of any 
one heap. Considering the last point first, the writer has been unable to 
find any differences between the upper and lower layers in any shell-heap 
that he has investigated. There remain then only two points to be con- 
sidered, the relation of the shell-heap culture to that of the modern Indians, 
and the possible variations of culture in the shell-heaps themselves. 
In both ancient and modern times the material culture of the natives 
must have been greatly influenced by the physical environment. Living 
is easy here. The climate is moist, the weather neither very warm nor 
very cold, and the vegetation luxuriant. Trees of red cedar abound; the 
wood is easily split for boards or hollowed for canoes, and the bark lends 
itself to weaving. Berries and roots are abundant. The numerous shell- 
