522 E. O. ULRICH REVISION OF THE PALEOZOIC SYSTEMS 



Weisner quartzite in Georgia and Alabama, which probably includes beds 

 that are strictly contemporaneous with similar beds in the Chilhowee 

 series in Tennessee, and in the lower Cambrian quartzites in Virginia, 

 Pennsylvania, and Vermont. But their exact determination seems as 

 yet impossible. In fact, exact correlations of individual beds of lower 

 Cambrian deposits in the Appalachian Valley seem possible only to a 

 very limited extent. Combining all the methods we can go but little fur- 

 ther than to say that the quartzites containing Olenellus in the Appala- 

 chian region are of lower Cambrian age. 



But there are some sandstones that are of much greater practical value 

 in correlation. Such are the well defined beds that occur at intervals in 

 shale formations, as the one (Kiefer sandstone) at the base of the Cayugan 

 series in Maryland and Pennsylvania, and the red Bloomsburg sand- 

 stone, which is found higher up in the series and separates the lower 

 (McKenzie) formation from the shaly middle (Wills Creek) formation 

 of the Cayugan series; also the conglomerate sandstone, which marks 

 the lower limit of the typical Chemung fauna in the Jennings formation. 

 In going both eastward and westward from the middle part of the Ap- 

 palachian Valley these sandstones usually become thinner and may 

 finally be lost entirely ; but so far as they are recognizable they afford the 

 most ready, and I believe also the most reliable, of the several means em- 

 ployed in correlating the local subdivisions of the late Silurian and late 

 Devonian deposits. Similar, though perhaps less accurate, results are 

 attained by using the sandstone beds in the Ozarkian system in Missouri. 

 The value of these as datum planes was recognized long ago by Swallow 

 and Shumard, and though less regularly distributed than they thought 

 and therefore often misidentified, the present revised classification of the 

 Cambrian and Ozarkian rocks in Missouri would, I fear, have been im- 

 possible without the aid of these sandstones. 



(3) Similarity in lithologic character and stratigraphic position does 

 not establish contemppraneity. — Correlation of similar lithologic units 

 based otherwise only on apparent likeness of stratigraphic position is 

 always unsafe and never conclusive. The practice of this simple method 

 seemed safe enough under the old idea of universal seas, but with the 

 proof of frequent and diverse sea oscillation growing stronger every day 

 it can no longer be relied on. Geological literature is full of errors due 

 to such insufficiently grounded correlations, and the chances for error 

 are not much greater when the correlated units are dissimilar lithically. 

 Contemporaneity is to be assumed in cases of lithologic similarity of 

 beds not continuously traceable only when the organic criteria prove that 



