Classification of Rocks. — Spurr. 2 1 1 
classification was made and the rocks described he began to 
see that even this partial classification was virtually a proposed 
innovation of rock classification in general, although, as above 
stated, not designed with that exact idea. Accordingly some 
justifying statements of the reasons for classification were pre- 
pared, stating briefly where principles which are used have 
been used before, with what authorities the writer agrees and 
where he deviates from them, together with the reasons for 
the chief principles adopted in the classification. It follows 
naturally from the method in which this classification was con- 
ceived that it does not belong to any school, although it has 
some of the principles of many. 
/. The System Used in the Table. 
The table which the writer has drafted is only for feldspar 
rocks, these having been found almost exclusively during the 
season's work. Under this class of feldspar rocks the families, 
groups, species, and varieties which are defined are only such 
as are found among these rocks, and the same is true of the 
structural variations or definite types. Nevertheless, the writer 
believes that the same system elaborated and revised affords a 
good expression of the natural relations of other igneous 
rocks and he will briefly outline the larger scheme. 
In the first place, the class of feldspar rocks itself, as classi- 
fied in the table, should be so elaborated as to comprise many 
rock groups not /described in this report. The family inter- 
mediate between the alkali feldspar rock family and the sili- 
ceous soda-lime feldspar rock family, namely the monzonyte 
family, should be added; and the class of feldspar rocks would 
then comprise three families, and two intermediate families 
covering the transition stages. Within these families as many 
new groups as are necessary should be added. The number 
will be always arbitrary (since rocks form a connected series), 
and will depend upon the fineness with which distinctions are 
drawn. Good judgment must always be used in each case, 
and while discriminating as many groups as are really service- 
able to science must guard against hair-splitting. The groups 
should be based on mineralogy, and should take cognizance 
ncl only of the mineral conil)ination. l)ut of the relative 
