CHANGES OF LEVEL OP NOKTH AMRIUCAN COASTS. 157 



due to subsidence from that brouiilit alxnit by glacial erosion. A careful 

 study of numerous tril)utaries of these tiooded valleys luis, however, con- 

 vinced me tliat tlie average amount of erosion accom[)Hshed in the last 

 ice-time on the general surface of the country as far to tlie nortlnvard as 

 New Brunswick did not exceed 50 feet. It is possible, liowever, that in 

 the axes of the main streams the ice, owing to its deei)er cutting, did 

 more extensive work. The impression, however, left ui)on the observer 

 is that tliese streams owe their present depth and i)enetration of the sea 

 much more to downsinking than to the erosion of their channels by ice. 



I have endeavored to ascertain whether the channels of the streams 

 to the north of cape Cod are prolonged m depressions below the level of 

 the sea in the manner of the well known instance of the Hudson river. 

 Although the soundings give indications that such is the case, the evi- 

 dence which they afford is not as yet sufficiently clear to warrant any 

 distinct statements. It may be said, however, that the topography of 

 the bay of Maine, so far as the shape of its bottom is known, is better to 

 be explained on the supposition that it is a land surface which has re- 

 cently been depressed than in any other manner. 



Bay of Fundi/. — The bay of Fundy by its form as described by its 

 shorelines, by the normal convergence of its rivers, and by the shape of 

 its bottom, as far so that is known, appears to be an extensively flooded 

 valley, the baselevel of which was 500 feet or more below the present 

 plane of the sea. 



So, too, in the eastern extremity of the peninsula of Nova Scotia and 

 the island of cape Breton we have evidence of extensive valley-iiooding. 



The Gut of Canso seems to be the valley of two streams which headed 

 against each other, the common trough having been considerably modi- 

 fied l>y glacial action. 



The extensively digitated reentrant known as Bras d'Or, which I have 

 examined with some care, must be reckoned as a river valley, somewhat 

 modified by.'ice-work, which has been flooded to its headwaters. 



Gulf of Saint Lairrencc. — The gulf of Saint Lawrence and the narrower 

 l)art thereof, extending as far as Quebec, seems to me explicable in no 

 other satisfactory way save l)y the sui)i)osition that it is a flooded Ijasin, 

 tlie general topography of which was determined during a peri(3(l of ele- 

 vation. I lliink that this explanation is doubtless true of the wedge- 

 shaped indent from Gasi)e peninsula to the head of tidewater. The 

 field of waters commnoly termed the (Julf presents certain puzzling fea- 

 tures wliich I shall now note. 



These peculiarities consist in the numerous islands, scattered over the 

 Gulf, which have very precipitous sides. Kntry island of the Magda- 

 lenes, liird rocks, Roche Perce island and the northern coast of Aiiticosti 



.XXII — Hl-m.. Oeol. Hoc, Am., Vol. 6. 1«'J4. 



