DRAINAGE OF TEE GLACIAL PERIOD. 331 



Canaan, on the water-parting between the Merrimack and 

 the Connecticut, where there is to be found the dry bed of a 

 river which for a time flowed through a pass from the Con- 

 necticut Valley into the Merrimack, but which is five hun- 

 dred feet above the valleys. Here upon this mountain axis, 

 in central New Hampshire, nine hundred feet above the sea, 

 are numerous and large water- worn circular cavities in the 

 rock, technically known as pot-holes, such as are formed in 

 shallow rapids, wherever gravel and pebbles become lodged, 

 first, in some natural slight depression, and then, through the 

 whirling motion given them by the running water, these 

 continue to wear a symmetrical depression so long as the 

 supply of water continues, or until a channel has been cut 

 through. Pot-holes may be seen in the rapids of almost any 

 rocky stream, with the gravel and pebbles, which do the im- 

 mediate work when set in motion, still partially filling them. 

 Such pot-holes exist in the anomalous position mentioned in 

 New Hampshire, where no present stream could by any pos- 

 sibility he made to flow. One of them, measured many years 

 ago by Jackson, was eleven feet deep, four and. a half feet in 

 diameter at the top, and two feet at the bottom, and, when 

 discovered, was rilled with earth and rounded stones.* 



The explanation of this phenomenon furnished by Mr. 

 Upharn, while engaged on the New Hampshire Survey, is as 

 follows : The ice of the Connecticut Valley, being farther 

 from the glacial front, lingered considerably longer than that 

 in the Merrimack Valley to the southeast, so that for a con- 

 siderable period the drainage from the ice-front in the south- 

 eastern part of Grafton county was compelled by the ice- 

 barrier on the west to flow over this depression into the Mer- 

 rimack basin, thus furnishing exactly the conditions neces- 

 sary for the production of pot-holes and such other marks of 

 water-action as have so long puzzled geologists at this point. 



Similar pot-holes, to be accounted for in like manner, 

 have recently been described near Archibald, in Blakely 



" New Hampshire Geological Report," vol iii, pp. 63-66. 



