MODIFIED DRIFT ALONG MERRIMACK RIVER. 83 



made artificially in Concord. One mile south of Fisherville depot the 

 course of the river was formerly in a westerly curve, passing around 

 Goodwin's point, two thirds of a mile from its direct course. At the 

 west end of this detour it was fast undermining a long line of bluff 125 

 feet in height. When the Northern Railroad was built, in 1846, the river 

 was turned, to avoid bridging, into a new channel, by which its course 

 was made straight, being shortened fully a mile. Its old channel remains 

 filled with water, except at its south-west bend, which is nearly silted 

 across ; and the erosion of the bluff at times of freshet is greatly dimin- 

 ished. Farther south, at about three miles above the city, the river 

 flowed in two channels, of which the west one was largest, enclosing 

 Sewall's island. The railroad was built across this island, reaching and 

 leaving it by embankments instead of bridges, for which purpose the 

 west channel was dammed, when the river is said by Dr. Prescott to 

 have swept away, to widen its east channel, a width of 20 to 25 rods of 

 its bordering interval for two thirds of a mile. 



Dr. Prescott mentions that, in cutting the new channel across the base 

 of Goodwin's point, "the workmen, at the depth of about 12 feet, struck 

 upon a bed or stratum of vegetable matter, consisting of leav^es, branches, 

 and trunks of small trees, the latter from three to six inches in diameter, 

 the form of which was perfect, and the bark distinct. This vegetable 

 deposit was found embedded in a stratum of fine blue sand, which at first 

 sight was mistaken for blue clay, and was from one to three inches in 

 thickness. The trunks and large branches were recognized as belonging 

 to the natural order coniferae." He also describes, from an excavation 

 at the gas-works in Concord, supposed "fragments of the roots, trunks, 

 and branches of trees. They were found deposited in a stratum of fer- 

 ruginous sand (composed of sand and oxide of iron) ; and in some in- 

 stances the fragments of roots and branches of trees were completely 

 incased in a firm coating or crust of the oxide of iron and sand from one 

 eighth to one half an inch in thickness." This was at a depth of ten feet. 

 It appears probable that these were cylindrical concretions of oxide of 

 iron, which often show concentric rings, almost exactly imitating the 

 annual layers of wood. These were found abundantly in the excavation 

 for laying the water-works main, in 1872, near the south line of the city 

 farm, and may be occasionally met with in any alluvial sand. 



