THE PROPOSED QUANTITATIVE SYSTEM.
The distinguishing characteristic of the more rigorous system designed to meet the needs of scientific petrology is its quantitative chemical character. All igneous rocks are classified primarily according to their chemical composition and only secondarily according to their mineral constituents, texture, and other characters. The rigorous application of the system requires chemical analyses of the rocks, but as these are not available in many cases, the authors of the system have devised a method of optical mineral analysis by which the nearly exact proportions of all the constituent minerals can be determined, and by knowledge of their chemical nature the results may be converted, by computation, into chemical terms. This can only be done for holocrystalline rocks whose crystals are large enough to be measured under the microscope, but aphanitic rocks may often be approximately classified by comparison with similar rocks already accurately determined. To facilitate this method of chemical analysis by measuring the minerals, the chemical composition of certain common rock-making minerals is expressed in proportional parts and tabulated, and is used somewhat as molecular weight is in ordinary chemical analysis. Certain of these are selected as standard minerals, the selection being such that the standard minerals embrace all the essential elements that enter into the composition of rocks. All other minerals are converted into their chemical equivalents in terms of these standard minerals by the use of the tables. All the mineral constituents being thus reduced to standard minerals, the classification is built up systematically on these standard (or standardized) minerals.
A new system of names is required, and these have been very skillfully formed by selecting significant letters from the names of the leading minerals or from words signifying their preponderance, so that short terms which carry their meaning in their forms, are secured, and this has been done so that these are usually euphonious, however strange they may seem to our preoccupied senses. For example, minerals composed chiefly of silica and alumina are called salic; those of ferromagnesian minerals, femic; those of aluminous ferromagnesian minerals, alferric, etc. When in a combination of salic and femic minerals, the salic are extremely abundant, the rock is persalic; if notably dominant, dosalic; if the salic and femic minerals are nearly equal, salfemic; if the femic are dominant, dofemic; if extremely abundant, perfemic, and so on, the system being mnemonic. This method of deriving names is applicable only to a portion of the necessary divisions. For the rest, a series of roots derived from geographic names, with a system of terminations, has been employed.
All standard minerals are divided into two groups of primary importance: one of minerals characterized by alumina, as the feldspars,—orthoclase, albite, anorthite,—leucite, nephelite, sodalite, noselite, and corundum, to which are added the closely associated minerals, quartz and zircon. This is called the salic group. The second group contains minerals characterized by iron and magnesia with no alumina, as hypersthene (enstatite), acmite, olivine, magnetite, hematite, and ilmenite, to which are added the closely associated minerals, titanite, perofskite, rutile, apatite, and all other rock-making minerals except those containing alumina together with iron and magnesia. The second group is called femic.
Aluminous ferromagnesian minerals, such as hornblende, augite, mica, etc.; are called alferric, and are not classed as standard minerals, because their complexity of composition makes it better to treat them as though made up of the simpler minerals of the standard list.
The composition of all igneous rocks can be expressed in terms of the relative proportions of the two groups of the standard minerals, salic and femic. By subdividing these groups successively on a mineral and chemical basis, a series of classificatory divisions of greater and greater precision has been formed. In each stage of the series, two factors only are compared, and a simple set of ratios has been selected to limit the divisions. Assuming the possibility of a continuous range of variable mixtures of the two factors (A and B) from an extreme composed wholly of one (A), and an extreme composed wholly of the other (B), five ideal cases have been chosen as types or centerpoints about which variation in mixture may take place. These are:
A
1
A
3