306 THWAITES AND TWENHOFEL THE WINDROW FORMATION 



type and the source of the iron was proliably the residuum of the older 

 rocks. Only locally has replacement of limestone by iron oxide been 

 observed. 



SEDIMENTARY STRUCTURES 



These consist of bedding and lamination and ripple-marks. Such bed- 

 ding as has been observed is poorly defined. Cross-lamination was ob- 

 served in a few places. The iron-bearing portions show essentially noth- 

 ing in the way of bedding, but this is possibly due to the development of 

 concretionary structure subsequent to deposition, which would have 

 brought about the elimination of bedding. A few current rlpple-marks 

 were observed at Tunnel Number 3. None of the structures suggest 

 marine origin and all are in harmony Avith the view of stream deposition. 



DISTRIBUTION OF THE DEPOSITS 



In its existing occurrences the AVindrow formation consists of isolated 

 patches scattered over a wide area. (See map, figure 1.) It might be 

 assumed that the knoAvn deposits of the Windrow formation are remnants 

 of a one-time continuous sheet, but the irregular local occurrence of 

 patches of gravel where conditions for preservation are apparently sim- 

 ilar — their occurrence in a locality at a certain elevation, while only a 

 short distance away no pebbles occur at that elevation — suggests that in 

 their former distribution the gravels were never continuous. The theory 

 of a once continuous sheet is a necessary consequence to the view that the 

 gravels are marine conglomerates ; but, if they are stream deposits, it is 

 very improbable that the patches were ever continuous and the present 

 distribution is readily explained. 



It has been suggested that these gravels were deposited on a peneplain 

 (13, 24, 26, 27, 28). To the writers this does not seem a necessary con- 

 clusion. The surface beneath the pebbles cuts across all formations from 

 the Huronian to the Carboniferous, but that is what occurs in the valley 

 of any stream; so that this fact alone does not prove the existence of a 

 peneplain. The size of the pebbles, the distance some of them must have 

 been transported, the pot-holes at Devils Lake, and the deep crevice filled 

 with gravel at Seneca do not agree with the usual conception of the slow- 

 flowing streams of a peneplain. These characteristics are more in har- 

 mony with valleys wdth a considerable extent of floodplain, with consid- 

 erable gradient to the stream bed, an occasional fall or rapid, and divides 

 of considerable height. 



NATURE OF THE UNDERLYING SURFACE 



The nature of the surface underlying the Windrow formation is known 

 at only four localities. At Devils Lake it is a water-worn quartzite in 



