METHOD OE PLOTTING 621 



.10- of the medium plagioclase, 



.05 of the basic plagioclase, 



.21 of the nephelite, 



.21 of the leucite, 



.10 of the biotite, 



.14 of the analcite, 



.10 of the alkalic pyroxenes, et cetera. 



By the use of the microscope or analysis, closer estimates can be made 

 for the alkalies in feldspars, but the several stages of accuracy in no way 

 alter the principle on which the arrangement is made, and field obser- 

 vations may serve to show whether a particular series resembles any here 

 plotted. 



It should be noted that the plots here used are so similar in results 

 to plots by Iddihgs and others, on the basis of silica per cent and ratios 

 of alkali to silica, that no differences in the conclusions can be expected. 

 If it is argued that simple per cents of silica and alkali would be better 

 than either, it may be answered that Washington's compilation of analy- 

 ses, with accompanying norms, saves practically all calculation of 

 analyzed specimens, and for field estimates the system here adopted, is 

 simplest. 



Selection of Data 



The data used for the first plot were averages selected from Daly's 

 valuable collection. 13 Two pegmatites were added. These serve to show 

 differences in a general way, though it hap often been emphasized that 

 some particular rock families may range more widely than the average 

 of one family differs from another. The range for any family depends 

 on the strictness with which it is defined. 



The data for figure 2 and figure 3 were selected primarily on the basis 

 of their being clear cases of differentiated magmatic units, for which 

 quantative data are available. About 40 bodies have been described in 

 a way to indicate that the different rocks grade into each other and were 

 derived from a single magma. Many are so completely exposed that the 

 nature of the original instrusive magma can be fairly closely estimated — 

 that is, no large part remains deeply concealed and no large part has 

 been eroded and lost. These are naturally the most valuable in a study 

 of differentiation. In a region of disconnected bodies it must always 

 be uncertain whether all phases developed by differentiation have been 

 found. The laws of differentiation must find their surest basis in these 



13 R. A. Daly : Igneous rocks and their origin, pp. 19-36. 



