T. Crook—Titaniferous Volcanic Rocks. 163 
Again, we may call attention to the fact that Grubenmann found as 
much as 3 per cent. of chromium sesquioxide in the highly titaniferous- 
melilite basalts of Baden. It seems that this constituent has not been 
determined in the Western Mediterranean basalts, in spite of the fact 
that they also might be expected to contain chromium. 
Chlorine, too, finds no place in the table of analyses, even among 
the undetermined constituents, although the norms for the fourteen 
specimens analysed exhibit an average of about 2 per cent. of apatite, 
and one norm shows the unusual amount of 5°04 per cent. of that 
mineral. Dr. Washington informs us that chlorine is present in some 
of the Linosa rocks, but its amount has not been estimated, and it is 
regarded as being derived from sea-water. One might have thought, 
however, that such quantities of apatite would have rendered it 
necessary to provide a place for chlorine (or fluorine) in stating the 
results of a first-rate analysis; and there are other reasons why the 
estimation of chlorine in such rocks should not be disregarded. In 
point of quantity, at any rate, it is likely to be of as much importance 
as nickel oxide, for which a place is provided. 
On the whole, these Western Mediterranean basalts are very similar 
to one another, and such important chemical differences as they 
exhibit are abundantly manifest in their mineral composition. Indeed, 
their essential similarity is so striking that one is surprised to 
find these rocks described under names so various as Monchiquose, 
Camptonose, Limburgose, Andose, Akerose, and Andose-Camptonose. 
It may be necessary, as a matter of convenience, to give names to 
rocks, but such a multiplication of names is neither necessary nor 
useful. Such names only conceal the truth, and lead people to 
believe that the evidence at present available for rock classification 
is of a higher order of exactness than it really is. The plain fact of 
the matter is that, just as in biology the process of evolution to some 
extent defies the wisdom of the systematist, so in petrology the processes 
of differentiation and admixture confound his artifices. Dr. Evans 
has recently shown that the American system has not escaped from 
confusion.' 
Such being the case, it is interesting to note the new turn that has 
been given to petrology by the study of the physico-chemical laws 
which underlie the process of differentiation. A fuller knowledge of 
these laws will probably emphasise the significance of the mineral 
composition of rocks as an index to chemical conditions. When these 
laws have been worked out, it is not unlikely that mineralogical methods 
(including chemical analysis as a subsidiary feature) will meet the full 
demand which a natural-history study of rock relationships makes 
upon them. Such an issue would be welcome, as it would simplify 
rather than complicate the problem of classification; at any rate, 
it would render unnecessary the unnatural system and unwieldy 
terminology which the Americans have introduced. 
In conclusion, we may be excused for emphasising the fact that 
many defects attend an exclusively chemical and too narrow view of 
rock analysis. In this respect, it i8 pleasing to be able to call in 
Cordier’s work as an illustration of a liberal mode of enquiry, for the 
1 «The Quantitative Classification of Igneous Rocks’’: Science Progress, No. 2, 
1906. 
