STRATIGRAPHIC TAXONOMY 587 



nessean boundary, which, on account of locally similar lithic characters, 

 have been referred to either side of the line. This occurred sometimes 

 through misinformation, but at other times by design. Thus the War- 

 saw formation, which, because of intervening warping, differs widely in 

 distribution from the preceding Keokuk and doubtless is separated from 

 that formation by a long emergent stage, has been occasionally placed 

 beneath the line, though more commonly associated with the formations 

 of the Meramec group. Keyes,^^ and more recently Weller,^^ refer the 

 typical Warsaw to the Osagian on the alleged ground that the Warsaw is 

 but the "superior portion of the Keokuk," and that sedimentation between 

 these two is "apparently continuous." I can not stop to discuss this 

 problem here except in a general manner. To begin with, I shall express 

 the conviction that the W^arsaw proper is everywhere separated from the 

 Keokuk by a stratigraphic hiatus. I may add that Keyes and Weller's 

 contrary belief, which was formerly shared by myself, arises chiefly from 

 the unwarranted assumption that the Keokuk is practically coextensive 

 with the Burlington horizons, and that these formations, together with 

 the Warsaw, are frequently embraced in a single stratigraphic unit. Re- 

 cent investigations tend to show — indeed, they have to a large extent 

 established — that these three formations differ widely in geographic dis- 

 tribution, diastrophic history, and age. 



The Burlington constitutes the principal part of the Osagian rocks in 

 the Mississippi Valley north of Saint Louis. This formation is well, 

 though not uniformly, developed also on the north, west, and south flanks 

 of Ozarkia, but on the east flank — that is, in the valley south of Saint 

 Louis — it is unknown save a small patch of lower Burlington on the Aux 

 "Vases River southwest of Sainte Genevieve. The Burlington, further, 

 is very commonly, if not wholly, absent in southern Illinois, in central 

 Kentucky, in central Tennessee, and in the southern Appalachian region. 

 The Keokuk, on the other hand, is usually present in the latter areas, but 

 seems to be wanting on the northwest side of Ozarkia. Further, while 

 the Burlington seems to have stretched over the greater part of the west- 

 ern two-thirds of this dome, as is indicated by remnants preserved in old 

 sink-holes, the Keokuk sea left much of this median area uncovered. 

 Evidently broad northwest-southeast tilting of the Mississippi Valley 

 occurred, the Burlington being almost confined to areas west of the Kan- 

 kakee Saint Francis axis (see map, figure 1, page 293), while the Keokuk, 

 though extending northward in the valley proper to Iowa, was chiefly 



"^2 C. R. Keyes : Missouri Geol. Survey, vol. 4. 1894, p. 69. 



'8 Stuart Weller : Illinois State Geol. Survey, Bull. 6, 1907, p. 24 ; Journal Geology, 

 vol. xvll, 1909, p. 276. 



