FORMATION OF LEUCITE IN IGNEOUS ROCKS 375 
about 2 per cent. of leucite, and its alkali-feldspar is Or,Ab,; both 
rocks being chemically closely alike." 
It should, of course, be possible to test this theoretical withdrawal 
of silica by the crystallization of soda-lime feldspar and soda-ortho- 
clase by calculation based on the modes of actual rocks. The data 
necessary for this would be an exact knowledge of the order of crys- 
tallization, both qualitatively and quantitatively, and the compositions 
of the several modal minerals. It is clear that the former would be 
practically almost unattainable, owing to the difficulty of deciding the 
relative periods in many cases, and especially on account of the 
simultaneous crystallization of many of the minerals. 
An attempt was made, however, on several of the Italian leucitic 
rocks, the modes of which have been quite exactly estimated. It was 
assumed that all the femic and alferric minerals, the ores, augite, 
and olivine, as well as all the soda-lime feldspar, had crystallized 
before that of the soda-orthoclase, leucite, and nephelite had begun. 
This, of course, may or may not be in accordance with the facts, but 
no definite, quantitative estimate of the relations was possible in any 
case. Unfortunately the composition of the alkali-feldspar was also 
unknown, and would have been almost impossible to determine in 
any case, as an analysis of the orthoclase phenocrysts would not neces- 
sarily indicate that of the more numerous groundmass feldspar laths, 
which cannot be separated mechanically and satisfactorily from the 
other constituents, owing to their small size and intimate juxtaposition. 
Even though, on this basis, the calculations resolved themselves 
largely into working out backward the norm from the estimated mode, 
yet the fact that the amounts of some of the minerals, especially the 
leucite, had been estimated physically, and the composition of the 
soda-lime feldspar had been determined optically, gives them some 
weight. It would take up too much space to give the details here, 
but it may be stated that the results were fully confirmatory of the 
view that the early crystallization of soda-lime feldspars and of soda- 
t The Roman Comagmatic Region (1906), pp. 75, 77- The determination of the 
composition of the Arso feldspar here, based partly on Rosiwal estimates of the amounts 
of leucite, augite, and olivine, and partly by readjustments of the other constituents, 
agrees well with that determined chemically by Fuchs (7. M. P. M. [1872], p. 233), 
who gives figures yielding Or;,.Ab;. The Poggio Cavaliere feldspar has not yet been 
analyzed. 
