Vicforidii. Igneous Bocks. 265 



As tliLi iinalyses sIkiwlmI a considerable range in the amount of 

 water, and as tho totals varied somewhat, it was felt that the 

 analyses as rliry stood wvw nol (luitr comiiaraljle with oik- another, 

 so that the analyses wi'rr recaUulatL'tl to lUU per eent. with the 

 water onutted. Portion of the ferrous oxide originally present in 

 the rocks had been converted tu ferric oxide, and as the amount 

 of oxidation varied it was thought l)ctter to calctdatc all the iron as 

 ferrous oxiile. and iiianganous oxide was added to the ferrous oxide. 



A second variation diagram based on these recalculated analyses 

 was nuide, and is given in Fig. 2. The difference between these two 

 diagrams is very maiked, as Fig. 1 shows a series of sigmoidal 

 curves, wliereas in Fig. 2 the variation is expressed by straight 

 lines. As this second diagram Avas considered to bring out the 

 relationship of the vaiious types to one another better than was 

 the case in the first diagram, other diagrams used in the paper have 

 been plotted from recalculated analyses. 



(a) 77(e Jlacedon Gvaiiodiorite-Dacite Series. 



Seven analyses of rocks belonging to this series have been made 

 in tlie Mines Department Laboratory, and these are given in Table 

 I. 



TABLE I. 



I. 11. III. IV. V. Vi. VII. 



Sp. Gr. - 7.789 2.781 2.773 2.722 2.708 2.688 2.630 



