L. V. Pirsson — Rise of Petrology as a Science. 235 



and which he considered were worthy of preservation. 

 Although, naturally, from the status of analytical chem- 

 istry up to that time, most of these would now be con- 

 sidered rather crude, the publication of the work was of 

 great service and marked an epoch in geochemistry. In 

 these tables Roth lists four analyses of American igneous 

 rocks, two from the Lake Superior region by Jackson 

 and J. D. Whitney and two by European chemists, one of 

 whom was Bunsen. The material of the last two was a 

 "dolerite" and the same locality is given for each — 

 " Sierra Nevada between 38° and 41°" which was prob- 

 ably considered quite precise for western America in 

 those days. 



From these feeble beginnings the forward progress of 

 petrology on the chemical side in this country has been 

 a steady one until its development has reached the point 

 which will be indicated in what follows. 



The collection of material by the various State surveys 

 and by those initiated by the National Government led to 

 an increasing number of rocks being analyzed during the 

 petrographic period. These became also increasingly 

 good in quality, like those published by G. W. Hawes in 

 his papers. When, however, chemists were appointed to 

 definite positions on the staffs of the Government surveys 

 and especially when, after the organization of the IT. S. 

 Geological Survey in 1879, a general central laboratory 

 was founded in 1883 with F. W. Clarke in charge, 

 then a new era in the chemical investigation of rocks may 

 be said to have started. In this connection should be 

 mentioned the work of W. F. Hillebrand, who set a stand- 

 ard of accuracy and detail in rock analysis which had not 

 hitherto been attempted. As a consequence of his accu- 

 rate and thorough methods and results the mass of 

 analyses performed by him and his fellow chemists in 

 this laboratory affords us the greatest single contribu- 

 tion to chemical petrology which has been made. Up to 

 January, -1914, the report of Clarke 32 lists some 8000 

 analyses of various kinds made in this laboratory for 

 geologic purposes. Nearly everywhere also a great 

 improvement in the quality of rock-analyses is to be 

 noted, and in the manuals of Hillebrand 33 and Washing- 

 ton 34 the rock analyst has now at his command the 

 methods of a greatly perfected technique which should 

 insure him the best results. 



Roth's Tabellen have been previously mentioned; sev- 



