THAMES EIVER TEKRACES IN CONNECTICUT 493 



of material was used, thus nearly obliterating the little esker that played a most 

 important part in the history of the deposits now to be discussed. 



Another difficulty with the hypothesis of a flooded river is that these terraces 

 have numerous depressions, 10 to 15 feet in depth, which are very similar in form 

 to kettleholes seen in sandplains in other parts of New England. These kettle- 

 holes would suggest that ice was probably present when the deposits were being 

 built up, the sand having been washed around ice blocks. 



The cuts made at the navy yard will now be considered. In order that there 

 should be no grade crossings at this point, the road leading from the highwaj' to- 

 ward the river was cut below the level of the track. At this point, therefore, the 

 sections were exposed much deeper into the terrace than at any other place, and 

 the writer was so fortunate as to obtain a number of photographs of these cuts be- 

 fore the banks were sloped back to make the sides of the road. Four of these are 

 here reproduced. They all show delta structure. 



Figure 1 of plate 53 is a general view looking south across the cut for the road. 

 The granite blocks lie on the level where the railroad now runs, the cut for which 

 through the overlying deposits may be seen to the right of the cart. These over- 

 lying deposits are some 10 feet thick and made up of coarse gravel and sand in 

 nearly horizontal layers, and are typical top-set beds of a glacial delta. At the 

 place where the view is taken they have been removed almost entii'ely, leaving the 

 fore set beds beneath them. This photograph is taken where the fore-set beds 

 frijm one lobe merge into those of another lobe of the delta, the upi^er portions of 

 all the layers being cut off by the succeeding top-set beds, a small portion of which 

 had fortunately not been removed when the view was taken. 



Figure 2 of plate 53 is taken in the cut looking north at southward-dipping fore- 

 set beds, the water-laid character of which is evident from a glance at the structure 

 as shown in the photograph. A few inches of the top-set beds are shown here also 

 just beneath the small house. 



Figure 1 of plate 54 is also taken in the cut, but looking eastward, and therefore 

 at right angles to the general direction of water and ice flow down the valley. The 

 connection of the toi>set with the fore-set beds is here very clearly shown, in jjlaces 

 continuous with them and in places cutting otf the tops of the previously laid fore- 

 sets. The delta character of this dejjosit is here unmistakable, and this is in a re- 

 gion where the flat-topped dej^osit laps up against the till-covered slopes of the sides 

 of the valley in a manner strongly to suggest the terrace origin. 



Figure 2, plate 54, is taken at the extreme eastern end of the cut and shows the 

 better stratified deposits overlying the less stratified deposits at the edge of the 

 delta-terrace, where the stratified drift ends and the unstratified drift begins. Less 

 than 20 feet east of where this i^hotograph is taken, there is unmistakable till. 

 Along the margin of the terrace at this point the sections show, mixed with the less 

 stratified water-worn material, some more angular rock waste that is morainal in 

 character. This point would thus appear to have been an ice-margin at a time 

 immediately preceding that of delta-building. This deposit would indicate that 

 there was a tongue of ice extending down the valley with its margin at this point 

 when the water from the meltingglaciercameoutfartherdown the valley, perhaps 

 at New London; and that a little later, when the tongue did not quite fill the val- 

 ley, a glacial delta was built at this point, filling in completely the space between 

 the ice and the.valley side. The stream which supplied this waste was some 2 miles 

 above the navy yard, as is indicated by the esker mentioned above. For the method 



