IGNEOUS ROCKS IN YELLOWSTONE PARK. 
209 
The independently uniform variation in the lime molecules 
is consistent with the fact that it enters so largely into both 
the feldspar and the ferromagnesian molecules. Their 
steady diminution from the basic to the acidic end of the 
series is in accord. with the decrease in the amount of 
augite and hornblende, and the increase of alkali feldspars, 
shown by the increase, of soda and potash and silica. 
The very noticeable reciprocal relation between the ferrous 
and ferric oxides indicates the variable oxidation of pre¬ 
existing molecules.. And it is probable that the ferrous 
molecules were the original ones; moreover, when all of the 
iron is reduced to the form of ferrous oxide, its molecules are 
found to vary in the same direction, and almost to the same 
extent as those of magnesia. Since the hornblende and 
biotite are the ferromagnesian minerals, carrying the greatest 
percentage of ferric oxide, the variation in the oxidation of 
the iron is naturally in accord with the amount of these min¬ 
erals in the rock. This is, however, significant from its 
bearing on. the question of the development of hornblende 
and biotite in the coarser grained varieties of these rocks, 
and from its possible connection with the work of mineral¬ 
izing agents. 
The chemical composition of different varieties of the 
Sepulchre Mountain rocks is shown by the accompanying 
analyses, Table IX. Nos. 219, 694, 221, 2736 and 3680 were 
analyzed by Mr. J. E. Whitfield; Nos. 214, 217 and 394, by 
Dr. T. M. Chatard; and No. 3680 by Mr. L. G. Eakins. 
The varieties represented are the two forms of breccia : 
pyroxene-andesite and hornblende-pyroxene-andesite. And 
the dike rocks including three varieties of andesite and 
dacite. They range from 55.83 to 67.49 per cent, of silica, 
very nearly the same as the group from Electric Peak. 
The chemical variation of the group is expressed graph¬ 
ically by the diagram, Table X, showing the molecular 
proportions of the essential oxides, which have been plotted 
in the same manner as in Table VIII. 
An examination of this diagram shows the same rela¬ 
tions between the oxides as those observed for the group 
