IMPERFECTION OF THE GEOLOGICAL RECORD 43 



coast, a large number of trees standing in the cliff and reefs, 

 or fallen to the shore, were broken up and examined, the 

 result being to discover that, with one unimportant exception, 

 the productive trees were confined to one of the beds at Coal 

 Mine Point, that from which the original specimens had been 

 obtained. Attention was accordingly concentrated on this, 

 and as many as thirty trees were at different times extracted 

 from it, of which rather more than one-half proved more or 

 less productive. By these means bones representing about 

 sixty specimens and twelve species were extracted, besides 

 numerous remains of land shells, millipedes, and scorpions. 

 In this way a very complete idea was obtained of the land life, 

 or at least of the smaller land animals, of this portion of the 

 coal formation of Nova Scotia. It is not too much to say that 

 if similar repositories could be found in the succeeding forma- 

 tions, and properly worked when found, our record of the 

 history of land quadrupeds might be made very complete. 



When in 1855 I changed my residence from Nova Scotia to 

 Montreal, and so was removed to some distance from the 

 carboniferous rocks which I had been accustomed to study, I 

 naturally felt somewhat out of place in a Cambro-Silurian dis- 

 trict, more especially as my friend Billings had already almost 

 exhausted its fossils. I found, however, a congenial field in 

 the Pleistocene shell beds; more especially as I had given 

 some attention to recent marine animals when on the sea coast. 

 The very perfect series of Pleistocene deposits in the St. 

 Lawrence valley locally contain marine shells from the bottom 

 of the till or boulder clay up to the overlying sands and gravels. 

 The assemblage was a more boreal one than that on the coast 

 of Nova Scotia, though many of the species were the same, 

 and both the climatal and bathymetrial conditions differed in 

 different parts of the Pleistocene beds themselves. The gap 

 in the record here could at that time be filled up only by col- 

 lecting recent shells. In addition to what could be obtained 



