82 WHITMAN CROSS 



ses. Other classifications may use theoretical criteria and they 

 will often serve useful purposes, nay, they are indeed distinctly 

 necessary to the progress of petrology, but such arrangements 

 must always be considered as subject to revision. 



The criteria available for the systematic classification of 

 rocks, fall into two groups, viz., the properties of the rocks them- 

 selves as objects, and their relationships to each other and to the 

 earth, which is made up of them. There has always been con- 

 flict of views as to the use of these criteria in establishing a 

 systematic classification. In the early years of this century 

 there were two opposing schools, one represented by the German 

 geologist, Werner, who classified all objects in the mineral king- 

 dom as geological bodies, the other best represented by the 

 French mineralogist, Hauy, to whom rocks were purely mineral 

 aggregates. For present purposes it is not essential, however 

 interesting, to trace the development of systematic petrography, 

 but it is worthy of note that the geological classification of rocks 

 is still most strongly advocated in Germany, and the mineralogical 

 classification is still most nearly realized in France. 



But the early systems of petrography failed necessarily 

 because based upon ignorance. The material constitution of 

 rocks was but very imperfectly known, and their geological 

 relations were in many respects matters of crude hypothesis. 

 All systems to the present time have failed for these reasons, 

 and the systems of today are not free from the weaknesses due 

 to the application of theoretical criteria. 



If we review the situation as regards our present knowledge 

 of the properties of rocks as objects it does not seem too much 

 to say that the development of chemistry and mineralogy, and 

 the application of the microscope to the study of rocks, have 

 given us an accurate insight into their chemical and mineral- 

 ogical composition, their structure and texture, which cannot be 

 essentially modified by future discoveries. These are the prop- 

 erties universally recognized as most applicable for subclassi- 

 fication. 



With respect to the geological relationships of rocks the case 



