49 2 WHITMAN CROSS 



The system of Walther seems specially intended to express 

 the changes rocks undergo rather than their characters as now 

 seen, and it is not apparent that the'author had in mind the apt 

 and logical analysis of the broad science of rocks which we owe 

 to Naumann. 



That the general treatment proposed by Walther for igneous 

 rocks, in naming them volcanic, and making the primary division 

 into unstratified, stratified and dike rocks, has many objections 

 will be sufficiently clear from the preceding discussions of this 

 review. The same is true of the assumption that there exists a 

 satisfactory system for the classification of igneous rocks. The 

 definition of tuffs as composed of magma consolidated in small 

 particles certainly applies to but a small part of the pyroclastic 

 deposits. 



SUMMARY. 



The science of petrography, the systematic and descriptive 

 science of rocks, was first fairly outlined by von Leonhard (1823) 

 and Brongniart (1827) through the distinction between the rock 

 and the geological terrane, and the setting up of logical classifi- 

 cations for the former. Neither of these masters gave the 

 science a name. 



The systems of von Leonhard and Brongniart necessarily used 

 the condition of ignorance concerningthe character of many rocks 

 as a ground for classification. With the increase of knowledge of 

 rocks there have been many attempts to apply new information to 

 systematic purposes. Since both the geological relations and the 

 properties of rocks are highly varied many unlike systems have 

 been proposed during the century, expressing individual opinions 

 as to the relative importance or adaptability of principles for 

 the end in view. Up to the present time, however, no compre- 

 hensive classification of rocks has been proposed which even pre- 

 tends to be natural or logically consistent in all its parts. 



When we view past petrographic systems, to judge as to how 

 far they possess natural or artificial features, it is first of all to 

 be noted that the system of Cordier is practically the only one 

 starting from the conception that rock species are natural 



