CAMBRIAN AND LOWER ORDOVICIAN. 105 



from place to place that it is difficult to follow them over any considerable area. The difficulties 

 are greatly enlaanced by the apparent scarcity and imperfect preservation of the fossils and by 

 the general similarity in lithologic character exhibited by widely separated beds ia the series. 

 The occurrence of beds of sandstone at intervals in the mass of magnesian hmestones was early 

 seized upon as affording a ready means for dividing the series into formations; but experience 

 gradually taught that most of the sandstones were mere lenses, irregularly developed at unequal 

 intervals and occurring locally as many successive beds or with only two or three. In the 

 northern and central parts of the Ozark Plateau the sandstone beds seem to be more regularly 

 developed than elsewhere, and here they are of considerable use in working out the stratigraphy. 

 Along the southern border, however, only the First or St. Peter sandstone seems to have been so 

 well developed as to mark a de&iite horizon. Along the eastern border, especially in Ste. 

 Genevieve County, the upper half of the series is thicker than usual. It here contains so many 

 beds of sandstone that it is very difficult to establish the boundary line between the Gasconade 

 and Roubidoux formations and between the latter and the overlying Jefferson City limestone. 



A careful study of the cherts has yielded perhaps the most rehable of the lithologic criteria 

 employed in discrimiriating the three formations of the Potosi group. As this group covers 

 by far the greatest part of the Ozark area, the opportunities for studying and testing the value of 

 certain varieties of chert as indices of particular horizons are unusual. Having checked the 

 results in many instances by evidence afforded by fossils, which really are much commoner in 

 these rocks than is generally believed, we have gradually become convinced that the story as told 

 by the cherts is rarely at fault. The mere presence of considerable quantities of chert is at once 

 rehably indicative of the Potosi group, the underlying limestones being practically free from it, 

 as is true also of the Joachim limestone at the top of the series. It is to be remembered, however, 

 that the proportion of chert in a given bed may vary greatly in near-by exposures. This is shown 

 often very strikingly in opposite faces of a hill, the one side of which forms a bluff, the other a 

 gentle, soil-covered slope. In cases of bluff exposure, where side erosion is active and takes 

 place under conditions very different from those prevailing on gentle slopes, the limestone beds 

 may often seem to be almost entirely without chert, but on the other side of the hill the surface 

 may be thickly strewn with it. 



The classification presented in tabular form on page 104 is based upon more or less extended 

 investigations in aU parts of the field. The formations are discriminated chiefly by lithologic 

 differences and limited in most cases by at least local stratigraphic discordances. As the for- 

 mations are intended to express as much as possible Hthologic units deemed worthy of being 

 mapped separately, the evidence of the fossils found in them was accorded secondary rank. 

 However, in the determination of the age of the deposit the faunas were depended upon almost 

 exclusively. 



It wiU be observed that the division between the Ordovician and Cambrian is drawn as a 

 dotted line opposite the Roubidoux member of the Potosi. This seeming indecision concerning 

 the point at which the line should be drawn is not due to the absence of a fauna, for we have 

 succeeded in collecting a large one, but because of the confficting elements contained in it. The 

 trilobites from the Gasconade and the lower part of the Roubidoux are, so far as observed, 

 nearly all of types that hitherto have been regarded as strictly indicative of the Upper Cambrian. 

 On the other hand, the numerous gastropods and fewer bracliiopods found in the same beds 

 are nearly all of Lower Ordovician types. 



Ulrich ^^^ also gives detailed descriptions of the formations. 



The latest classification of the rocks of this series is that made by Buckley, ^°° 

 who in 1909 published a paper in which several new names were introduced for the 

 rocks lying between the Roubidoux and the Bonneterre. 



