Gabbroid Rocks of Minnesota. — VVinckell. 387 
Therefore, the first six rocks are wholly comparable, show- 
ing- only minor dififerences ; from these the other three all dif- 
fer markedly, while among themselves the orthoclase gabbro 
and cordierite noryte are comparable especially if the potash 
be disregarded, while the quartz gabbro stands alone. These 
dififerences and irregularities are well shown in the plate. In 
the lower figure the characteristics of the silicoferrolyte are 
strikingly evident : enormous per cent of iron ; silica high for 
so basic a rock on account of the presence of free quartz; low 
magnesia, and a notable excess of alumina. 
The various ratios are conspicuously variable. In both 
figures the plagioclasyte is a marked rock with its high alumina 
and calcium oxide, and practical absence of all ferromag- 
nesian elements. The cordierite noryte is ajso conspicuous by 
its relatively low content of feldspathic constituents, high fer- 
romagnesian silica and extraordinary excess of alumina. 
In conclusion, it is seen that the comparison of these vari- 
ous methods of graphic representation shows that the method 
of Iddings, while excellent for its purpose, is not at all suited 
for a single rock analysis, nor well adapted to general use; 
Ihe method of Becke gives too much prominence to the re- 
lation K:Na:Ca, which is not characteristic for the alkaline- 
earth constituents, except when no excess of CaO exists. It 
is too complicated for general use, and many important rela- 
tions are not brought out conspicuously. 
The object of Brogger is simplicity, but the choice of the 
relative number of molecules (quotientzahl) introduces an un- 
necessary complication, since the relations of the relative 
weights are fully as characteristic. He avoided the assump- 
tion of M'ichel Levy that all the alumina is contained in the 
feldspars, which is admittedly only an approximation, but in 
avoiding this Brogger loses some of the important advantages 
of the method of Michel Levy. The latter method then is 
evidently the best hitherto devised, for a single analysis, as 
well as for general use. Its weak point, of course, is the as- 
sumption just mentioned, but this loses its importance if it is 
remembered that it is not supposed to be necessarily an ex- 
pression of fact, but only a possibility the verification of which 
is to be accomplished by the mineralogical study. 
