ON THE ETHNOLOGICAL SURVEY OF CANADA. 4 I- 5 



tion, and say that probably less than a millennium ago the ancestors of 

 both these stocks dwelt together as one people and spoke a common 

 language. "Where the original home of this undivided people was, or what 

 territory they occupied before their advent here, is a question we shall 

 have to consider later. But wherever it may have been ' it is abundantly 

 clear that it was not the shores and bays of British Columbia or indeed 

 those of the adjoining States. For almost every division of these two 

 stocks have distinct names for the six different species of salmon and the 

 other varieties of tish found in these waters ; which could not conceivably 

 have been the case had they lived together here before their separation, 

 as fish, and above all salmon, is their staple food, and has been time out 

 of mind. And not only have they different names for the fish themselves, 

 but also widely differing myths to account for their origin or rather 

 presence in these waters. 



The later or more recently formed middens are easily distinguished 

 from the older kind. First by their general condition, and secondly 

 because in most instances they are known to have been old camp 

 sites of the pi-esent tribes. With very few exceptions we find the shell 

 remains in the later heaps in a good state of preservation and freer from 

 ashes and other earthy matter. So much is this the case that some of 

 the shrewder settlers in early days converted some of these shell heaps 

 into lime, for which commodity they found a ready sale. Of the relics 

 recovered from them the majority are of stone. In the old heaps 

 the reverse is the case, bone specimens preponderating. The later 

 middens, too, are comparatively small and shallow, and, as far as my own 

 investigations go, not nearly so rich in relics as the older and more 

 extensive heaps. Taking both classes of middens together, the number 

 and ubiquity of them are remarkable. The shores of the estuary and of 

 Puget Sound, as well as the coast and islands generally, are literally 

 covered with them. In the neighbourhood of Boundary Bay they stretch 

 almost continually for miles along the sound. Between Ladner's at the 

 mouth of the Fraser and Point Roberts in Washington State I found 

 them in scores, sometimes situated several miles back from the water in 

 the midst of thick bush and timber. These latter were generally speci 

 mens of the older kind ; and like those on the Lower Fraser were com- 

 posed before the forest grew there, and when the Delta was less extensive 

 than at present, and the salt water reached farther inland. 



The relative richness of these delta middens in relics is another 

 remarkable feature of them. One may dig and search for days in some 

 heaps and find scarcely anything, while others abound, or did formei'ly, 

 in bone and stone specimens of all kinds. There is one at the river-side 

 village of Hammond, on the Fraser, which has yielded an almost incredible 

 number of the most interesting relics. It extends along the bank of the 

 river for a considerable distance, and is now utilised as fruit and vegetable 

 gardens, &c., for which purpose the midden matter is admirably adapted, 

 being rich in the elements of plant life. The settlers who first cultivated 

 this ridge collected hundreds of different specimens. These, unfortunately, 

 for the most part were cast aside, or became broken or lost, or else were 



' See the writer's paper on the ' Oceanic Origin of the Kwakiutl-Nootka and Salish 

 Stocks of British Columbia,' published in the Trans. Hoy. Soc. Can., vol. iv., sect, ii., 

 1898. The views therein set forth have met with the general concurrence of Tregear 

 and other Polynesian scholars, and have been further strongly confirmed by my later 

 linguistic studies of Columbian and Oceanic stocks. 



