428 THE CAMBRIAN. [bull. 81. 



In this connection the student should read Mr. J. E. Marr's paper on 

 Hotnotaxis. 1 



Percentage of species. — This method of paleontologic correlation was 

 employed by Sir Charles Lyell. He found a certain percentage of 

 species identical with living forms in one horizon, the Eocene; a larger 

 percentage in the Miocene, and a still larger one in the Pliocene. He 

 then concluded the Eocene represented the oldest and the Pliocene the 

 latest deposit. He thus classified the formations and gave an indirect 

 method for their correlation in different areas. It is not capable of 

 application where all the members of a fauna are extinct. 



III. LITHOLOGIC CHARACTER. 



Lithologic character has very little value in establishing correlations, 

 when comparing formations outside of limited areas. The constituents 

 of a rock may differ in their chemical and mechanical condition in a 

 stratum that is proved by other evidences to be continuous. In some 

 instances lithologic characters may be of value, especially in areas 

 where the stratigraphic sequence of the beds is known and paleonto- 

 logic evidence is lacking. Dr. R. D. Irving used the method success- 

 fully among the pre-Oambrian rocks of the Lake Superior region, and 

 says of it: 



Its value in tracing formations from point to point can hardly be overestimated, 

 being as great as that of paleoatological evidence and, in my judgment, of much the 

 same nature. As we trace formations and their minor subdivisions from place to 

 place we must of course constantly check lithological evidence by stratigraphy. As 

 we pass from one extremity of the field, to the other, changes in lithological charac- 

 ters of course come in, and these changes might lead to unsafe conclusions were we to 

 compare strata! successions too distantly removed from one another ; but when wo 

 work from point to point such changes are detected as they gradually appear, and 

 are provided for. 2 



An illustration of the value of the method is shown bv the correla- 

 tions made by the brothers Roger's 3 in Virginia and Pennsylvania, where 

 they correlated the quartzites, shales, and limestones by their litho- 

 logic characters along an extent of several hundred miles. When, 

 however, they came to carry this correlation without that area of sedi- 

 mentation, the inadequacy of the method is shown by the fact they 

 correlated the Lower Cambrian quartzite with the Upper Cambrian 

 quartzite about the Adirondacks of New York. A still more striking 

 illustration of erroneous correlation by this method is that made by 

 Prof. N. H. Winchell 4 in correlating, as members of one formation, the 



1 Proe. Cambridge Phil. Soc, vol. 6, 1887, pp. 74-82. 



2 Seventh Ann. Rep. TJ. S. Geo!. Survey, 1885-'86, 1888, pp. 378, 379. 



3 A system of classification and nomenclature of the Paleozoic rocks of the United States, with an 

 account of their distiihution, more particular! y in the Appalachian mountain chain. Am. Jour. Sci., 

 vol. 47, 1844, p. 112. 



4 The crystalline rocks of Minnesota. General report of progress made in the study of their field 

 relations. Statement of problems yet to be solved. Geol. and Nat. Hist. Survey of Minn., 17th Ann. 

 Rep. for 1888, 1889, p. 49. 



