either an island, or connected by a narrow isthmus at the lower canon 

 of the Ruby with the Tobacco Root range to the east. 



East of this great lake across the Gallatin range was another lake 

 in what is now the upper Yellowstone valley. * 



To the west of the continental divide another lake began in the 

 vicinity of Butte, extending westward and then northward, occupy- 

 ing the Deer Lodge valley, and perhaps down the Hell Gate to the 

 vicinity of Drummoud. There was another smaller one in the vicinity 

 of Phillipsburg on Flint creek a tributary of the Hell Gate. There was 

 also a large lake occupying the Bitter Root and Missoula valleys. 



Near White Sulphur Springs in the valley of Smith or Deep river 

 was a comparatively small lake, belonging, according to Di". W. B. 

 Scott, to the later .Tolin Day ** and early Loup Fork divisions; while 

 the other beds l)elong to the White River and later Ijoup F^ork 

 horizons. As the Deep River beds have been studied, and their verte- 

 brate remains described by Cope and Scott, I will confine myself, prin- 

 cipally, to the more extensive deposits in the otlier parts of the state. 



By far the best exposure is in the lower Madison valley. This 

 valley extends about twenty miles south from the Three Forks, where 

 the Jefferson, Madison, and Gallatin rivers unite to form the Missouri. 



On the west side of the river extending eiglit or ten miles in a north 

 and south direction is a series of nearly perpendicular cliffs composed 

 principally of clay, fine sand, volcanic ash and conglomerate, gray in 

 color, but often stained with iron. They are usually soft, yet there are 

 in places some very hard layers. The river has not cut through the 

 T)ed so as to expose the rock beneath, but tlie thickness of the ex- 

 posure must be at least two hundred feet. 



Above these beds, on the east side of the valley, are softer, marly 

 beds, almost white, weathering into conical hills and bluffs with even- 

 ly sloping sides which have an angle of twenty-five to thirty-five de- 

 grees. I have observed no distinct line of separation between these 

 beds and the ones just de'seribed. 



Between the marly beds and those above, there is a sharp dividing 

 line and the formations can l>e distinguished as far as they can be 

 seen. Thre is a sudden change from the white slopes of marl to the 

 cream colored clays and darker sands and conglomerates of the upper 

 beds whose sloping sides are broken by cliffs of liarder material. 

 There is also unconformity of ei'osion. The material of the lower beds 



* See Livingston Atlas Sheet (Montana). 



**W. B. Scott informs me that Matthew believes the uppermost 

 White River horizon which he has found in Colorado corresponds in 

 age to the lower beds in the Smith River Valley. 



