28 CENOZOIC MAMMAL HORIZONS OF WESTERN NORTH AMERICA. 



river-channel beds of coarse materials, from 700 feet to a mile in 

 width, with an easterly direction. The most tenable theory at pres- 

 ent seems to be that of periodic overflow deposition in very shallow 

 sheets of water, too transitory or seasonal to support any of the 

 aquatic animals — such deposition as is left by the annual overflow of 

 the Nile, for example. The nilometer at Roda shows an annual accu- 

 mulation of silt of 0.12 centimeter, equivalent to 12 meters in ten 

 thousand years, as cited by Lyons and by Beadnell.^ 



Summary. — The sum of the present opinion appears to be this : The 

 topography of the Plains Region was in Oligocene to lower Pleisto- 

 cene time, as now, level or gently undulating, not mountainous. 

 On the gentle eastward slopes of the Rocky Mountains and the Black 

 Hills were borne broad streams with varying channels, backwaters, 

 and lagoons, sometimes spreading into shallow lakes but never into 

 vast fresh-water sheets. Savannas were interspersed with grass- 

 covered pampas, traversed by broad, meandering rivers which fre- 

 quently changed their channels. 



This accounts for the presence of true conglomerates, true sand- 

 stones, calcareous grits, gypsum, fijie clays, fuller's earth, fine loess, 

 eolian sands, and even, far out on the plains of Nebraska ^ and Kansas, 

 widespread deposits of volcanic dust, wind borne from distant craters 

 in the mountains to the west and southwest. In the early Oligocene 

 and Miocene the deposits were chiefly fluviatile or river sandstones 

 and conglomerates interspersed with fine flood-plain or overflow 

 deposits, perhaps locally lacustrine, partly of volcanic ashes. This 

 interpretation is presented in PI. Ill, which has. been prepared to 

 show the actual relations of the unstratified stream-channel deposits 

 to the finer and partly stratified surrounding deposits. These rocks 

 still await petrographic analysis. 



As the desiccation or aridit}^ of the country increased, the moun- 

 tain-fed rivers became smaller and narrower, while the eolian or loess 

 deposits apparently became more common, beginning in the middle 

 Miocene. The deposits also became more and more restricted in 

 extent as the Miocene advanced. The newer river channels cut down 

 into the older series, thus using the erosion materials a second time. 



Thus geology and petrography unaided fail to complete the picture. 

 Paleontology goes hand in hand with these sciences to restore the true 

 picture of former conditions on the Great Plains; but far more exten- 

 sive petrographic and paleontologic investigation than has as 3^et been 

 made is necessar}^ to establish a final geologic theor3\ 



a Lyons, H. G., The physiography of the River Nile and its basin: Survey Dept. Egypt, Cairo, 1906, 

 pp. 313, 317, 334. 



b Beadnell, H. J. L., The topography and geology of the Faylim province of Egypt: Survey Dept. Egypt, 

 Cairo, 1905, p. 80. 



c See Barbour, E. H., The deposits of volcanic ash in Nebraska: Proc. Nebraska Acad. Sci., 1894-95. 

 The heaviest beds and the coarsest ash occur in the southwestern counties. Even as far east as Missouri 

 River (Cuming Countyj there are beds 7 feet in thickness. 



