MINERALOGICAL CLASSIFICATION OF IGNEOUS ROCKS 73 



There are now six dimensions in the classification, and since 

 each pigeonhole will represent not only a plutonic rock but also a 

 hypabyssal and an extrusive, we may say we have a classification 

 in seven dimensions, yet every rock may be shown by a single 

 point on a drawing in a single plane. The more detailed description 

 which follows may make this clearer. 



NUMBER AND POSITIONS OF THE VARIOUS DIVISION LINES 



Classes. — The dividing lines between the various classes, orders, 

 families, etc., were not selected at random, but an attempt was 

 made to see if they have any logical positions. For this purpose 

 the writer has been collecting data on cards for all rocks whose 

 modes in mineral percentages have been determined. The number 

 is small, less than 600 such rocks having been found. Unfortu- 

 nately this number is too small to determine definitely all points, 

 but the writer found that in most cases preliminary graphs with 

 fewer analyses showed practically the same curves as the ones 

 here given. 1 



In order to determine the positions of the dividing planes 

 between the light and the dark rocks, and to decide whether there 

 should be four or five classes (namely white, light, medium, dark, 

 and black), the rocks of the various families were plotted in Fig. 12, 

 in which the abscissae represent the proportions of light constitu- 

 ents in the rock and the ordinates the number of rocks whose 

 modes were known, the percentages being gathered by fives to make 

 a smoother and more representative curve than the individual 

 percentages would have made. The lower curve in the figure is 

 the curve of all rocks (585) of which the writer had the modes, and 

 includes the alcalic rocks as well as the families given in the upper 

 curves. All the curves except the one for gabbro, which does not 

 extend so far, show an increase at 90-95 per cent light and a 

 decrease beyond that toward 100 per cent. Consequently rocks 

 may well be called leucocratic when there are 95 per cent or more 

 of light minerals ; and there is no objection to making the melano- 

 cratic division beyond 95 per cent dark. A difficulty appears in 



'Since this paper was written, 91 additional mode-analyses have been found, 

 but the graphs remain practically as they were. 



