396 PLEISTOCENE CW? INDIANA AND MICHIGAN. 



as a sandy belt, considerably wind blown. North of Rifle River it runs northeast passing half 

 a mile south of Melita to the bank of Au Gres River in Mason Township. 



Farther northeast it runs near the crest of a morainic ridge half a mile west of the Cedar 

 Valley into Burleigh Township, Iosco County. About 2 miles east of Prescott a sandy ridge, 

 apparently corresponding in level to the Warren beach, runs south for 2 miles through south- 

 west Richland Township, Ogemaw County. 



A sandy and gravelly ridge corresponding to the Warren beach runs south a quarter of a 

 mil e east of Whittemore and west and southwest about 4 miles in moderate strength, though 

 broken and fragmentary. Four miles northeast of Whittemore, in Reno and Grant townships, 

 a faint sandy ridge, corresponding to the Warren beach, winds for 2 or 3 miles along the outer 

 edge of a great sand plain which extends northeastward about 15 miles to Au Sable River and 

 on which it finally fades away. 



In the entire stretch from where it enters Bay County to the great sand plain of the 

 Au Sable this beach appeared to mark the upper limit of submergence by lake waters. Wind- 

 blown sand and some gravels of glaciofluvial origin he at higher levels, but nothing indicating 

 wave action is present. 



North of Au Sable River in Iosco and Alcona counties a search was made by Mr. Leverett 

 and the writer for the Warren and other beaches above the Algonquin, but none were found. 



The shore of Lake Warren around the west, south, and east sides of the western peninsula 

 of Ontario was traced by Spencer about 25 years ago. 1 He did not trace it continuously 

 but identified it very accurately from place to place. He represents it, however, as extending 

 much farther to the north on both sides of the peninsula than the facts observed justify. On 

 the west side he found it a beach ridge only as far as Ripley, about 10 miles southeast of 

 Kincardine, beyond which he identifies it only as a "terrace in valley." On the east side 

 of the peninsula he found it a beach ridge as far as Brantford, describing it beyond as a "rock 

 terrace" or "terrace." 



The writer has examined this beach at several points in Ontario but has traced it con- 

 tinuously only in the region east of Sarnia. There and at some other points it shows the same 

 distinctly double phase that is so well developed on the "thumb" in Michigan and in western 

 New York. Spencer's map is on a very small scale and does not accurately represent the 

 course of the beach, which extends considerably farther southwest east of Sarnia and south- 

 west of St. Thomas than is shown. 



ALTITUDE OF WARREN BEACH. 



MICHIGAN. 



In the area of horizontality the Warren beach has an altitude of 675 to 685 feet. Its crest 

 in the Ann Arbor quadrangle is generally very close to the 680-foot contour. In the Rochester 

 quadrangle its crest is nearer 6S5 feet and rises gradually northward. One and a half miles east 

 of Troy it has an altitude of 690 feet, according to the Survey's topographic map, but this is 

 exceptional and may be due to wind-blown sand. The northward rise probably begins in this 

 quadrangle, for a mile south of Lenox it has become more distinctly double and the two ridges 

 have altitudes of about 690 and 700 feet, respectively. Troy is probably slightly north of the 

 hinge line for the Maumee beaches. The hinge for the Warren is in all probability a few miles 

 farther north. (See fig. 7.) 



At Goodells the altitude for the bay-shore ridge of the Warren is 707 feet. One-fourth mile 

 east of Abbottsf ord the altitude of the same ridge is 708 feet and that of the main barrier 709 or 

 710 feet. On the next railroad (Port Austin branch of the Pere Marquette) the altitude of the 

 beach 1J miles east of Atkins is 709 feet. The altitude at this point is probably fairly accurate. 



1 Spencer, J. W., High-level shores in the region of the Great Lakes and their deformation: Am. Jour. Sci., 3d ser., vol. 41, 1891, pp. 203-204. 



