Petrology

Petrology, the science of rocks (Gr. lrErpoS), the branch of geology which is concerned with the investigation of the composition, structure and history of the rock masses which make up the accessible portions of the earth's crust. Rocks have been defined as " aggregates of minerals." They are the units with which the geologist deals in investigating the structure of a district. Some varieties cover enormous areas and are among the commonest and most familiar objects of nature. Granite, sandstone, clay, limestone, slate often form whole provinces and build up lofty mountains. Such unconsolidated materials as sand, gravel, clay, soil are justly included among rocks as being mineral masses which play an important role in field geology. Other rock species are of rare occurrence and may be known in only one or two localities in distant parts of the earth's surface. Nearly all rocks consist of minerals, whether in a crystalline or non-crystalline state, but the insoluble and imperishable parts of the skeletons of animals and plants may constitute a considerable portion of rocks, as for example, coral limestone, lignite beds and chalk.

Treatment of the Subject

In this paragraph the subject matter of the science of petrology is briefly surveyed; the object is to point out the headings under which particular subjects are treated (there is a separate article on the terms printed in italics). General questions as to the nature, origin and classification of rocks and the methods of examination are discussed in the present article; mineralogy comprises similar matter respecting the component minerals; metamorphism, metasomatism, pneumatolysis and the formation of concretions are agencies which effect rocks and modify them. Three classes of rocks are recognized: the igneous, sedimentary and metamorphic. The plutonic, or deep-seated rocks, which cooled far below the surface, and occur as batholites, bosses, laccolites, and veins, include the great classes granite, syenite, diorite, gabbro and peridotite; related to the granites are aplite, greisen, pegmatite, schorl rock and micro pegmatite; to the syenites, borolanite, monzonite, nepheline-syenite and ijolite; to the diorites, aphanite, napoleonite and tonalite; to the gabbros, pyroxenite and theralite, and to peridotites, picrite and serpentine. The hypabyssal intrusive rocks, occurring as sills, veins, dikes, necks, &c., are represented by porphyry and porphyrite (including bostonite, felsite and quartz-porphyry), diabase and lamprophyre; some pitchstones belong to this group and contain crystallites and spherulites. The volcanic rocks, found typically as lava flows, include rhyolite and obsidian (with sometimes perlite), trachyte and phonolite (and leucitophyre which is treated under leucite), andesite and dacite, basalt (with the related dolerite, variolite and tachylyte), nephelinite and tephrite. Among sedimentary rocks we recognize a volcanic group (including tuff, agglomerate and some kinds of pumice); an arenaceous series such as sand (some with glauconite), sandstone, quartzite, greywacke and gravel; an argillaceous group including clay, firebrick, phyllite, laterite, shale and slate; a calcareous series with chalk, limestone (often forming stalactites and stalagmites), dolomite and marls or argillaceous limestones (flint occurs as nodules in chalk); the natural phosphates may be mentioned here. The metamorphic rocks are commonly gneisses and schists (including mica-schist); other types are amphibolite, charnockite, eclogite, epidiorite, epidosite, granulite, itacolumite, hornfels, mylonite and the scapolite rocks.

Composition

Only the commonest minerals are of importance as rock formers. Their number is small, not exceeding a hundred in all, and much less than this if we do not reckon the subdivisions into which the commoner species are broken up. The vast majority of the rocks which we see around us every day consist of quartz, felspar, mica, chlorite, kaolin, calcite, epidote, olivine, augite, hornblende, magnetite, haematite, limonite and a few other minerals. Each of these has a recognized position in the economy of nature. A main determining factor is the chemical composition of the mass, for a certain mineral can be formed only when the necessary elements are present in the rock. Calcite is commonest in limestones, as these consist essentially of carbonate of lime; quartz in sandstones and in certain igneous rocks which contain a high percentage of silica. Other factors are of equal importance in determining the natural association or paragenesis of rock-making minerals, principally the mode of origin of the rock and the stages through which it has passed in attaining its present condition. Two rock masses may have very much the same bulk composition and yet consist of entirely different assemblages of minerals. The tendency is always for those compounds to be formed which are stable under the conditions under which the rock mass originated. A granite arises by the consolidation of a molten magma (a fused rock mass; Gr. ih yua, from pacraav, to knead) at high temperatures and great pressures and its component minerals are such as are formed in such circumstances. Exposed to moisture, carbonic acid and other subaerial agents at the ordinary temperatures of the earth's surface, some of these original minerals, such as quartz and white mica are permanent and remain unaffected; others " weather " or decay and are replaced by new combinations. The felspar passes into kaolin, muscovite and quartz, and if any black mica (biotite) has been present it yields chlorite, epidote, rutile and other substances. These changes are accompanied by disintegration, and the rock falls into a loose, incoherent, earthy mass which may be regarded as a sand or soil. The materials thus formed may be washed away and deposited as a sandstone or grit. The structure of the original rock is now replaced by a new one; the mineralogical constitution is profoundly altered; but the bulk chemical composition may not be very different. The sedimentary rock may again undergo a metamorphosis. If penetrated by igneous rocks it may be recrystallized or, if subjected to enormous pressures with heat and movement, such as attend the building of folded mountain chains, it may be converted into a gneiss not very different in mineralogical composition though radically different in structure to the granite which was its original state.

Structure

The two factors above enumerated, namely the chemical and mineral composition of rocks, are scarcely of greater importance than their structure, or the relations of the parts of which they consist to one another. Regarded from this standpoint rocks may be divided into the crystalline and the fragmental. Inorganic matter, if free to take that physical state in which it is most stable, always tends to crystallize. Crystalline rock masses have consolidated from solution or from fusion. The vast majority of igneous rocks belong to this group and the degree of perfection in which they have attained the crystalline state depends primarily on the conditions under which they solidified. Such rocks as granite, which have cooled very slowly and under great pressures, have completely crystallized, but many lavas were poured out at the surface and cooled very rapidly; in this latter group a small amount of non-crystalline or glassy matter is frequent. Other crystalline rocks such as rock-salt, gypsum and anhydrite have been deposited from solution in water, mostly owing to evaporation on exposure to the air. Still another group, which includes the marbles, mica-schists and quartzites, are recrystallized, that is to say, they were at first fragmental rocks, like limestone, clay and sandstone and have never been in a molten condition nor entirely in solution. Certain agencies however, acting on them, have effaced their primitive structures, and induced crystallization. This is a kind of metamorphism.

The fragmental structure needs little explanation; wherever rocks disintegrate fragments are produced which are suitable for the formation of new rocks of this group. The Fragmental . or i g i nal materials may be organic (shells, corals, Rocks. crystalline or " massive "; others are fragmental. The massive igneous rocks include a few which are nearly completely vitreous, and still more which contain a small amount of amorphous matter, but the majority are completely crystallized. Among the best known examples are obsidian, pumice, basalt, trachyte, granite, diorite. The fragmental igneous rocks consist of volcanic ashes more or less firmly compacted. The sedimentary rocks form a second group; they surface subject to the conditions of temperature, moisture and pressure which obtain there. They include fragmental and crystalline varieties. The former consist of the debris of pre-existing rocks, accumulated in seas, lakes or dry land and more or less indurated by pressure and cementing substances. Gravel, sand and clay, conglomerate, sandstone, shale are well-known examples. Many of them are fossiliferous as they contain fragments of organisms. Some are very largely made up of remains of animals or plants, more or less altered by mineralization. These are sometimes placed into a special group as rocks of organic origin; limestone, peat and coal are typical of this class. The crystalline sediments are such as rock-salt and gypsum, deposits of saline lakes or isolated portions of the sea. They were formed under conditions unfavourable to life and hence rarely contain fossils. The metamorphic rocks are known to be almost entirely altered igneous or sedimentary masses. Metamorphism consists in the destruction of the original structures M etamor- p hic Rock s. and the development of new minerals. The chemical composition of the rocks however suffers little change. The rock becomes as a rule more crystalline; but all stages in the process may be found and in a metamorphosed sediment, e.g. a sandstone, remains of the original sand grains and primary fragmental structure may be observed, although extensive recrystallization has taken place. The agencies which produce metamorphism are high temperatures, pressure, interstitial moisture and in many cases movement. The effects of high temperatures are seen best in the rocks surrounding great outcrops of intrusive granite, for they have been baked and crystallized by the heat of the igneous rock (thermo-metamorphism). In folded mountain chains where the strata have been greatly compressed and their particles have been forced to move over one another a different type of metamorphism prevails (regional or dynamic metamorphism).

Methods of Investigation

The macroscopic (Gr. yarcpos, large) characters of rocks, those visible in hand-specimens without the aid of the microscope, are very varied and Macroscopic difficult to describe accurately and fully. The Characters. geologist in the field depends principally on them and on a few rough chemical and physical tests; and to the practical engineer, architect and quarry-master they are allimportant. Although frequently insufficient in themselves to determine the true nature of a rock, they usually serve for a preliminary classification and often give all the information which is really needed. With a small bottle of acid to test for carbonate of lime, a knife to ascertain the hardness of rocks and minerals, and a pocket lens to magnify their structure, the field geologist is rarely at a loss to what group a rock belongs. The fine grained species are often indeterminable in this way, and the minute mineral components of all rocks can usually be ascertained only by microscopic examination. But it is easy to see that a sandstone or grit consists of more or less rounded, waterworn sand-grains and if it contains dull, weathered particles of felspar, shining scales of mica or small crystals of calcite these also rarely escape observation. Shales and clay rocks generally are soft, fine grained, often laminated and not infrequently contain minute organisms or fragments of plants. Limestones are easily marked with a knife-blade, effervesce readily with weak cold acid and often contain entire or broken shells or other fossils. The crystalline nature of a granite or basalt is obvious at a glance, and while the former contains white or pink felspar, clear vitreous quartz and glancing flakes of mica, the other will show yellow-green olivine, black augite and grey striated plagioclase.

But when dealing with unfamiliar types or with rocks so fine grained that their component minerals cannot be determined with the aid of a lens, the geologist is obliged to have recourse to more delicate and searching methods of Microscopic g Characters. investigation. With the aid of the blowpipe (to test the fusibility of detached crystals), the goniometer, the magnet, the magnifying glass and the specific gravity balance, the earlier travellers attained surprisingly accurate results. Examples of these may be found in the works of von Buch, Scrope, Darwin and many others. About the end of the 18th century, Dolomieu examined crushed rock powders under the microscope and Cordier in 1815 crushed, levigated and investigated the finer ground-mass of igneous rocks. His researches are models of scrupulous accuracy, and he was able to announce that they consisted essentially of such minerals as felspar, augite, iron ores and volcanic glass, and did not differ in nature from the coarser grained rocks. Nicol, whose name is associated with the discovery of the Nicol's prism, seems to have been the first to prepare thin slices of mineral substances, and his methods were applied by Witham (1831) to the study of plant petrifactions. This method, of such far-reaching importance in petrology, was not at once made use of for the systematic Crystalline Rocks. Igneous dated from a state of fusion. Some of them are Rocks. Sedi ocks mentary have all been laid 'down as deposits on the earth's plants) or vitreous (volcanic glasses) or crystalline (granite, marble, &c.); the pulverizing agent may be frost, rain, running water, or the steam explosions which shatter the lava within a volcanic crater and produce the fragmental rocks known as volcanic ash, tuffs and agglomerates. The materials may be loose and incoherent (sand, clay, gravel) or compacted by pressure and the deposit of cementing substances by percolating water (sandstone, shale, conglomerate). The grains of which fragmental rocks are composed may be coarse or fine, fresh or decayed, uniform or diverse in their composition; the one feature which gives unity to the class is the fact that they are all derived from pre-existing rocks or organisms. Because they are made up of broken pieces these rocks are often said to be " clastic." Origin of Rocks. - The study of the structure of rocks evidently leads us to another method of regarding them, which is more fundamental than those enumerated above, as the structure depends on the mode of origin. Rocks are divided into three great classes, the Igneous, the Sedimentary and the Metamorphic. The igneous (Lat. ignis, fire) rocks have all consoli investigation of rocks, and it was not till 1858 that Sorby pointed out its value. Meanwhile the optical study of sections of crystals had been advanced by Sir David Brewster and other physicists and mineralogists and it only remained to apply their methods to the minerals visible in rock sections. Very rapid progress was made and the names of Zirkel, Allport, Vogelsang, Schuster, Rosenbusch, Bertrand, Fouque and Levy are among those of the most active pioneers in the new field of research. To such importance have microscopical methods attained that textbooks of petrology at the present time are very largely devoted to a description of the appearances presented by the minerals of rocks as studied in transparent micro-sections. A good rock-section should be about one-thousandth of an inch in thickness, and is by no means very difficult to make. A thin. splinter of the rock, about as large as a halfpenny may be taken; it should be as fresh as possible and free from obvious cracks. By grinding on a plate of planed steel or cast iron with a little fine carborundum it is soon rendered flat on one side and is then transferred to a sheet of plate glass and smoothed with the very finest emery till all minute pits and roughnesses are removed and the surface is a uniform plane. The rock-chip is then washed, and placed on a copper or iron plate which is heated by a spirit or gas lamp. A microscopic glass slip is also warmed on this plate with a drop of viscous natural Canada balsam on its surface. The more volatile ingredients of the balsam are dispelled by the heat, and when that is accomplished the smooth, dry, warm rock is pressed firmly into contact with the glass plate so that the film of balsam intervening may be as thin as possible and free from air-bubbles. The preparation is allowed to cool and then the rock chip is again ground down as before, first with carborundum and, when it becomes transparent, with fine emery till the desired thickness is obtained. It is then cleaned, again heated with a little more balsam, and covered with a cover glass. The labour of grinding the first surface may be avoided by cutting off a smooth slice with an iron disk armed with crushed diamond powder. A second application of the slitter after the first face is smoothed and cemented to the glass will in expert hands leave a rock-section so thin as to be already transparent. In this way the preparation of a section may require only twenty minutes.

The microscope employed is usually one which is provided with a rotating stage beneath which there is a polarizer, while above the objective or the eyepiece an analyser is mounted; alternatively the stage may be fixed and the polarizing and analysing prisms may be capable of simultaneous rotation by means of toothed wheels and a connecting-rod. If ordinary light and not polarized light is desired, both prisms may be withdrawn from the axis of the instrument; if the polarizer only is inserted the light transmitted is plane polarized; with both prisms in position the slide is viewed between " crossed nicols." A microscopic rocksection in ordinary light if a suitable magnification (say 30) be employed is seen to consist of grains or crystals varying in colour, size and shape. Some minerals are colourless and transparent (quartz, calcite, felspar, muscovite, &c.), others are yellow or brown (rutile, tourmaline, biotite), green (diopside, hornblende, chlorite), blue (glaucophane), pink (garnet), &c. The same mineral may present a variety of colours, in the same or different rocks, and these colours may be arranged in zones parallel to the surfaces of the crystals. Thus tourmaline may be brown, yellow, pink, blue, green, violet, grey or colourless, but every mineral has one or more characteristic, because most common tints. The shapes of the crystals determine in a general way the outlines of the sections of them presented on the slides. If the mineral has one or more good cleavages they will be indicated by systems of cracks (see Pl. III.). The refractive index is also clearly shown by the appearance of the sections, which are rough, with well-defined borders if they have a much stronger refraction than the medium in which they are mounted. Some minerals decompose readily and become turbid and semitransparent (e.g. felspar); others remain always perfectly fresh and clear (e.g. quartz), others yield characteristic secondary products (such as green chlorite after biotite). The inclusions in the crystals are of great interest; one mineral may enclose another, or may contain spaces occupied by glass, by fluids or by gases.

Lastly the structure of the rock, that is to say, the relation of its components to one another, is usually clearly indicated, whether it Macro- be fragmental or massive; the presence of glassy matter. in contradistinction to a completely crystalline or holo-crystalline condition; the nature and origin of organic fragments; banding, foliation or lamination; the pumiceous or porous structure of many lavas; these and many other characters, though often not visible in the hand specimens of a rock, are rendered obvious by the examination of a microscopic section. Many refined methods of observation may be introduced, such as the measurement of the size of the elements of the rock by the help of micrometers; their relative proportions by means of a glass plate ruled in small squares; the angles between cleavages or faces seen in section by the use of the rotating graduated stage, and the estimation of the refractive index of the mineral by comparison with those of different mounting media.

Further information is obtained by inserting the polarizer and rotating the section. The light vibrates now only in one plane, and in passing through doubly refracting crystals in the h slide is, speaking generally, broken up into two rays, sm. which vibrate at right angles to one another. In many coloured minerals such as biotite, hornblende, tourmaline, chlorite, these two rays have different colours, and when a section containing any of these minerals is rotated the change of colour is often very striking. This property, known as " pleochroism " (Gr. irXeicov, more; X pcos, colour), is of great value in the determination of rock-making minerals. It is often especially intense in small spots which surround minute enclosures of other minerals, such as zircon and epidote; these are known as " pleochroic halos." If the analyser be.now inserted in such a position that it is crossed relatively to the polarizer the field of view will be dark where there are no minerals, or where the light passes through isotropie substances such as glass, liquids and cubic crystals. All other crystalline bodies, being doubly refracting, will appear bright in some position as the stage is rotated. The only exception to this rule is provided by sections which are perpendicular to the optic axes of birefringent crystals; these remain dark or nearly dark during a whole rotation, and as will be seen later, their investigation is of special importance. The doubly refracting mineral sections, however, will in all cases appear black in certain positions as the stage is rotated. They are said to be " extinguished " when this takes place. If we note these positions we may measure the angle between them and any cleavages, faces or other structures of the crystal by means of the rotating stage. These angles are characteristic of the system to which the mineral belongs and often of the mineral species itself (see Crystallography). To facilitate measurement of extinction angles various kinds of eyepieces have been devised, some having a stauroscopic calcite plate, others with two or four plates of quartz cemented together; these are often found to give more exact results than are obtained by observing merely the position in which the mineral section is most completely dark between crossed nicols.

The mineral sections when not extinguished are not only bright but are coloured and the colours they show depend on several factors, the most important of which is the strength of the double refraction. If all the sections are of the same thickness as is nearly true of wellmade slides, the minerals with strongest double refraction yield the highest polarization colours. The order in which the colours are arranged is that known as Newton's scale, the lowest being dark grey, then grey, white, yellow, orange, red, purple, blue and so on. The difference between the refractive indexes of the ordinary and the extraordinary ray in quartz is. 009, and in a rock-section about - 5 - 1 00 - of an inch thick this mineral gives grey and white polarization tints; nepheline with weaker double refraction gives dark grey; augite on the other hand will give red and blue, while calcite with still stronger double refraction will appear pinkish or greenish white. All sections of the same mineral, however, will not have the same colour; it was stated above that sections perpendicular to an optic axis will be nearly black, and, in general, the more nearly any section approaches this direction the lower its polarization colours will be. By taking the average, or the highest colour given by any mineral, the relative value of its double refraction can be estimated; or if the thickness of the section be precisely known the difference between the two refractive indexes can be ascertained. If the slides be thick the colours will be on the whole higher than in thin slides.

It is often important to find out whether of the two axes of elasticity (or vibration traces) in the section is that of greater elasticity (or lesser refractive index). The quartz wedge or selenite plate enables us to do this. Suppose a doubly refracting mineral section so placed that it is " extinguished "; if now it is rotated through 45° it will be brightly illuminated. If the quartz wedge be passed across it so that the long axis of the wedge is parallel to the axis of elasticity in the section the polarization colours will rise or fall. If they rise the axes of greater elasticity in the two minerals are parallel; if they sink the axis of greater elasticity in the one is parallel to that of lesser elasticity in the other. In the latter case by pushing the wedge sufficiently far complete darkness or compensation will result. Selenite wedges, selenite plates, mica wedges and mica plates are also used for this purpose. A quartz wedge also may be calibrated by determining the amount of double refraction in all parts of its length. if now it be used to produce compensation or complete extinction in any doubly refracting mineral section, we can ascertain what is the strength of the double refraction of the section because it is obviously equal and opposite to that of a known part of the quartz wedge.

A further refinement of microscopic methods consists of the use of strongly convergent polarized light (konoscopic methods). This is obtained by a wide angled achromatic condenser above the polarizer, and a high power microscopic objective. Those sections are most useful which are perpendicular to an optic axis, and consequently remain dark on rotation. If they belong to uniaxial crystals they show a dark cross or convergent light between crossed nicols, the bars of which remain parallel to the wires in the field of the eyepiece. Sections perpendicular to an optic axis of a biaxial mineral under the same conditions show a dark bar which on rotation becomes curved to a hyperbolic shape. If the section is perpendicular to a " bisectrix " (see Crystallography) a black cross is seen which on rotation opens ova to form two hyperbolas, the apices of which are turned towards one another. The optic axes emerge at the apices of the hyperbolas and may be surrounded b y coloured rings, though owing to the thinness of minerals in rock sections these are only seen when the double refraction of the mineral is strong. The distance between the axes as seen in the field of the microscope depends partly on the axial angle of the crystal and partly on the numerical aperture of the objective. If it is measured by means of an eye-piece micrometer, the optic axial angle of the mineral can be found by a simple calculatio-l. The quartz wedge, quarter mica plate or selenite plate permit the determination of the positive or negative character of the crystal by the changes in the colour or shape of the figures observed in the field. These operations are precisely similar to those employed by the mineralogist in the examination of plates cut from crystals. It is sufficient to point out that the petrological microscope in its modern development is an optical instrument of great precision, enabling us to determine physical constants of crystallized substances as well as serving to produce magnified images like the ordinary microscope. A great variety of accessory apparatus has been devised to fit it for these special uses.

The separation of the ingredients of a crushed rock powder from one to another in order to obtain pure samples suitable for analysis is also extensively practised. It may be effected by means of a powerful electro-magnet the strength of which can be regulated as desired. A weak magnetic field will attract magnetite, then haematite and other ores of iron. Silicates containing iron will follow in definite order and biotite, enstatite, augite, hornblende, garnet and similar ferro-magnesian minerals may be successively abstracted; last only the colourless, non-magnetic compounds, such as muscovite, calcite, quartz and felspar, will remain. Chemical methods also are useful. A weak acid will dissolve calcite from a crushed limestone, leaving only dolomite, silicates or quartz. Hydrofluoric acid will attack felspar before quartz, and if employed with great caution will dissolve these and any glassy material in a rock powder before dissolving augite or hypersthene. Methods of separation by specific gravity have a still wider application. The simplest of these is levigation (Lat. levigare, to make smooth, levis) or treatment by a current of water; it is extensively employed in the mechanical analysis of soils and in the treatment of ores, but is not so successful with rocks, as their components do not as a rule differ very greatly in specific gravity.

Fluids are used which do not attack the majority of the rockmaking minerals and at the same time have a high specific gravity. 'Solutions of potassium mercuric iodide (sp.r. 3.196), cadmium borotungstate (sp. gr. 3.30), methlyene iodide (sp. gr. 3.32), bromoform (sp. gr. 2.86), or acetylene bromide (sp. gr. 3.00) are the prin- .cipal media employed. They may be diluted (with water, benzene, &c.) to any desired extent and again concentrated by evaporation. If the rock be a granite consisting of biotite (sp. gr. 3.1), muscovite (sp. gr. 2.85), quartz (sp. gr. 2.65), oligoclase (sp. gr. 2.64) and orthoclase (sp. gr. 2.56) the crushed minerals will all float in methylene iodide; on gradual dilution with benzene they will be precipitated in the order given above. Although simple in theory these methods are tedious in practice, especially as it is common for one rock-making mineral to enclose another. But expert handling of fresh and suitable rocks yields excellent results and much purer powders may be obtained by this means than by any other.

Although rocks are now studied principally in microscopic sections the investigation of fine crushed rock powders, which was the first branch of microscopic petrology to receive attention, is by no means discontinued. The modern optical methods are perfectly applicable to transparent mineral fragments of any kind. Minerals are almost as easily determined in powder as in section, but it is otherwise with rocks, as the structure or relation of the components to one another, which is an element of great importance in the study of the history and classification of rocks, is almost completely destroyed by grinding them to powder.

In addition to naked-eye and microscopic investigations chemical methods of research are of the greatest practical utility to the petrographer. The crushed and separated powders, obtained by the processes described above, may be analysed and thus the chemical composition of the minerals in the rock determined qualitatively or quantitatively. The chemical testing of microscopic sections and minute grains by the help of the microscope is a very g Y P P elegant and valuable means of discriminating between the mineral components of fine-grained rocks. Thus the presence of apatite in rock-sections is established by covering a bare rock-section with solution of ammonium molybdate; a turbid yellow precipitate forms over the crystals of the mineral in question (indicating the presence of phosphates). Many silicates are insoluble in acids and cannot be tested in this way, but others are partly dissolved, leaving a film of gelatinous silica which can be stained with colouring matters such as the aniline dyes (nepheline, analcite, zeolites, &c.).

Complete chemical analyses of rocks are also widely made use of and are of the first importance, especially when new species are under description. Rock analysis has of late years (largely under the influence of the chemical laboratory of the United States Geological Survey) reached a high pitch of refinement and complexity. As many as twenty or twenty-five components may be determined, but for practical purposes a knowledge of the relative proportions of silica, alumina, ferrous and ferric oxides, magnesia, lime, potash, soda and water will carry us a long way in determining the position to which a rock is to be assigned in any of the conventional classifications. A chemical analysis is in itself usually sufficient to indicate whether a rock is igneous or sedimentary and in either case to show with considerable accuracy to what subdivision of these classes it belongs. In the case of metamorphic rocks it often establishes whether the original mass was a sediment or of volcanic origin.

The specific gravity of rocks is determined in the usual way by means of the balance and the pycnometer. It is greatest in those rocks which contain most magnesia, iron and heavy metals; least in rocks rich in alkalis, silica and water. It diminishes with weathering, and generally those rocks which are highly crystalline have higher specific gravities than those which are wholly or partly vitreous when both have the same chemical composition. The specific gravity of the commoner rocks ranges from about 2.5 to 3.2.

The above methods of investigation, naked eye, physical, microscopical, chemical, may be grouped together as analytical in contradistinction to the synthetic investigation of rocks, which proceeds by experimental work to reproduce different rock types and in this way to elucidate their origin and explain their structures. In many cases no experiment is necessary. Every stage in the origin of clays, sands and gravels can be seen in process around us, but where these have been converted into coherent shales, sandstones and conglomerates, and still more where they have experienced some degree of metamorphism, there are many obscure points about their history upon which experiment may yet throw light. Up to the present time these investigations have been almost entirely confined to the attempt to reproduce igneous rocks by fusion of mixtures of crushed minerals or of chemicals in specially contrived furnaces. The earliest researches of this sort are of those of Faujas St Fond and of de Saussure, but Sir James Hall really laid the foundations of this branch of petrology. He showed (1798) that the whinstones (diabases) of Edinburgh were fusible and if rapidly cooled yielded black vitreous masses closely resembling natural pitchstones and obsidians; if cooled more slowly they consolidated as crystalline rocks not unlike the whinstones themselves and containing olivine, augite and felspar (the essential minerals of these rocks). Many years later Daubree, Delesse and others carried on similar experiments, but the first notable advance was made in 1878, when Fouque and Levy began their researches.

They succeeded in producing such rocks as porphyrite, leucitetephrite, basalt and dolerite, and obtained also various structural modifications well known in igneous rocks, e.g. the porphyritic and the ophitic (Gr. dcpcs, serpent). Incidentally they showed that while many basic rocks (basalts, &c.) could be perfectly imitated in the laboratory, the acid rocks could not, and advanced the explanation that for the crystallization of the latter the gases never absent in natural rock magmas were indispensable mineralizing agents. It has subsequently been proved that steam, or such volatile substances as certain borates, molybdates, chlorides, fluorides, assist in the formation of orthoclase, quartz and mica (the minerals of granite). Sir James Hall also made the first contribution to the experimental study of metamorphic rocks by converting chalk FIG. I. - Banded Obsidian, Kirghiz (x 25).

The rock consists of alternate bands of brown and colourless glass which have been arranged in stripes by the fluxion movement of the viscous mass before solidification. The glass is rendered granular by very minute crystals.

FIG. 2. - Fluidal Rhyolite, Hungary (x 15).

In the centre are crystals of felspar, rather turbid through weathering. The matrix is partly glassy, partly felsitic, and shows the effects of streaming movements with eddies behind the felspar crystals.

FIG. 3. - Obsidian, Mexico (x 15).

This rock has a damascened pattern owing to the irregular mingling of streams of brown and of colourless glass. It is nearly quite free from minute crystals.

FIG. 4. - Perlitic Obsidian, Tokai, Hungary (x 15).

The clear glassy rock is traversed by a large number of cracks, some long and straight, while others are nearly circular. These are rendered more distinct by the deposit of thin films of secondary limonite in them. The cracks are due to contraction on cooling.

FIG. 5. - Perlitic Pitchstone, Meissen, Germany (x 15).

The perlitic, rounded cracks are very clearly seen, because the rock is decomposing and becoming slightly opaque along them. At the top there is a corroded crystal of felspar, showing cleavage, with large circular enclosures of brownish glass.

FIG. 6. - Obsidian, Iceland (x 17).

In the clear glassy base there are rounded yellow spots (spherulites) arranged in fluxion streams.

FIG. 7. - Spherulitic Rhyolite, Hlinik, Hungary (x io).

The white, angular patches are crystals of quartz and of sanidine felspar. Between them there is a yellowish glass showing circular areas with a well-defined radiate fibrous structure (spherulites).

XXI. 326.

FIG. 8. - Spherulitic Felsite, Arran, Scotland (x Io).

The round spherulites of this rock are large and sometimes composite; their radiate structure is obvious. This is a devitrified pitchstone, no longer glassy but finely crystalline, and at the centres of the spherulites there are spaces occupied by a secondary deposit of quartz.

FIG. 9. - Porphyritic And Fluidal Rhyolite, Hungary (x 12).

The ground mass is partly glassy, partly felsitic, and shows fluxion-banding. The large quartz is a double hexagonal pyramid, but its edges and corners are rounded by corrosion and large irregular areas of glass penetrate to its centre.

FIG. I. - Porphyritic Pitchstone, Scuir Of Eigg, Scotland (x Io). A large porphyritic felspar crystal is seen lying in a pale-brown glassy base and containing many glass inclusions of irregular shape. The felspar, in one margin especially, shows corroded outlines.

FIG. 2. - Trachyte, Orotava, Teneriffe (x 12).

There are larger porphyritic felspars of the first generation, and smaller ones of later origin composing part of the ground-mass, which also contains a considerable amount of yellow vitreous material.

FIG. 3. - Trachyte, Perlenhardt, Germany (x 10).

In this rock there are porphyritic crystals of felspar and of dark brown biotite (nearly black in the photograph), with a few of green augite and magnetite. The ground-mass is finely crystalline.

FIG. 4. - Granite Rubislaw, Aberdeen (x io).

This is a non-porphyritic, holocrystanine rock. Among its components the crystals of dark mica are conspicuous, and with them occur also a few plates of white mica, with perfect cleavage. The slightly turbid or granular substance is felspar, a little decomposed, and the large clear spaces are crystals of quartz.

FIG. 5. - Hornblende-Granite, Dalbeattie, Scotland (x 15).

The dark crystal with fine parallel lines of cleavage is biotite; the others, with two less perfect cleavages, are hornblende. At the top there is a long rod-shaped grain of sphene. The granular-looking substance is felspar, and the quartz, as usual, is clear and transparent.

FIG. 6. - Graphic Granite, Bodenmais, Bavaria (x Io).

This rock consists of angular patches of clear quartz scattered through a striated dull matrix of felspar. The different quartz areas have all the same optical orientation, as if they were parts of a single crystal.

FIG. 7. - Luxullianite, Luxulyan, Cornwall (x 15).

In this variety of tourmaline-granite there are many blue needles of tourmaline, grouped in stellate clusters which are embedded in a matrix of clear quartz. These pointed needles diverge from the surfaces of larger grains of tourmaline.

FIG. 8. - Granophyre, Braemar, Scotland (x 37).

This photograph is taken between crossed nicols to show the graphic structure of the ground-mass, similar to that of Fig. 6, but on a much finer scale. The quartz towards the centre of the field appears as white, angular areas, embedded in a grey matrix of felspar, and each mineral reacts in a uniform fashion to polarized light over an area of moderate extent.

FIG. 9. - Diorite, Hodritch, Hungary (x 10).

The dark crystals are green hornblende, and show the outlines which are characteristic of that mineral. The cloudy grey substance between them is felspar in a somewhat weathered state.

into marble by heating it in a closed gun-barrel, which prevented the escape of the carbonic acid at high temperatures. Adams and Nicholson have carried this a stage farther by subjecting marble to great pressure in hydraulic presses and have shown how the foliated structures, frequent in natural marbles, may be produced artificially.

Roth Classification

The three great classes of rocks above enumerated - the igneous, the sedimentary and the metamorphic - are subdivided into many groups which to a small extent resemble the genera and species under which the naturalist classifies the members of the animal kingdom. There are, however, no hard and fast boundaries between allied rocks. By increase or diminution in the proportions of their constituent minerals they pass by every gradation into one another; the distinctive structures also of one kind of rock may often be traced gradually merging into those of another. Hence the definitions adopted in establishing rock nomenclature merely correspond to selected points (more or less arbitrary) in a continuously graduated series. This is frequently urged as a reason for reducing rock classification to its simplest possible terms, and using only a few generalized rock designations. But it is clear that many apparently trivial differences tend regularly to recur, and have a real significance, and so long as any variation can be shown to be of this nature it deserves recognition.

The igneous rocks (crystalline and fragmental) form a well-defined group, differing in origin from all others. The crystalline or massive varieties may occur in two different ways; the lavas have Igneous been poured out at the surface and have consolidated Rocks. after ejection, under: conditions which are fairly well understood, seeing that they may be examined at active volcanoes in many parts of the world; the intrusive rocks, on the other hand, have been injected from below into cracks and fissures in the strata and have cooled there beneath masses which conceal them from view till exposed by denudation at a subsequent period. The members of these two groups differ in many respects from one another, so that it is often possible to assign a rock to one or other of them on mere superficial inspection. The lavas (or effusive rocks), having cooled rapidly in contact with the air, are mostly finely crystalline or have at least fine-grained ground-mass representing Lavas or that part of the viscous semi-crystalline lava flow which Effusive was still liquid at the moment of eruption. At this Types. time they were exposed only to atmospheric pressure, and the steam and other gases, which they contained in great quantity, were free to escape; many important modifications arise from this, the most striking being the frequent presence of numerous steam cavities (vesicular structure) often drawn out to elongated shapes subsequently filled up with minerals by infiltration (amygdaloidal structure). As crystallization was going on while the mass was still creeping forward over the surface of the earth, the latest formed minerals (in the ground-mass) are commonly arranged in subparallel winding lines following the direction of movement (fluxion or fluidal structure) (see Pl. I. figs. 2 and 9, Pl. II. fig. 2), and the larger early minerals which had previously crystallized may show the same arrangement. Most lavas have fallen considerably below their original temperatures before they are emitted. In their behaviour they present a close analogy to hot solutions of salts in water, which, when they approach the saturation temperature, first deposit a crop of large, well-formed crystals (labile stage) and subsequently precipitate clouds of smaller less perfect crystalline particles (metastable stage). In igneous rocks the first generation of crystals generally forms before the lava has emerged to the surface, that is to say, during the ascent from the subterranean depths to the crater of the volcano. It has frequently been verified by observation that freshly emitted lavas contain large crystals borne along in a molten, liquid mass. The large, well-formed, early crystals are said to be porphyritic (P1.III. figs. 1, 2, 3); the smaller crystals of the surrounding matrix or ground-mass belong to the post-effusion stage. More rarely lavas are completely fused at the moment of ejection; they may then cool to form a non-porphyritic, finely crystalline rock, or if more rapidly chilled may in large part be non-crystalline or glassy (vitreous rocks such as obsidian, tachylyte, pitchstone (Pl. I. figs. I, 4, 5). A common feature of glassy rocks is the presence of rounded bodies (spherulites: Gr. ball), consisting of fine divergent fibres radiating from a centre (Pl. I. figs. 7, 8); they consist of imperfect crystals of felspar, mixed with quartz or tridymite; similar bodies are often produced artificially in glasses which are allowed to cool slowly. Rarely these spherulites are hollow or consist of concentric shells with spaces between (lithophysae: Gr. stone; 4uim, bellows). Perlitic structure, also common in glasses, consists in the presence of concentric rounded cracks owing to contraction on cooling (see Perlite).

The phenocrysts (Gr. to show; crystal) or por phyritic minerals are not only larger than those of the groundmass. As the matrix was still liquid when they formed they were free to take perfect crystalline shapes, not being interfered with by the pressure of adjacent crystals. They seem to have grown rapidly, as they are often filled with enclosures of glassy or finely crystalline material like that of the ground-mass (Pl. II. fig. I). Microscopic examination of the phenocrysts often reveals that they have had a complex history. Very frequently they show successive layers of different composition, indicated by variations in colour or other optical properties; thus augite may be green at the centre and various shades of brown outside this; or may be pale green centrally and darker green with strong pleochroism (aegirine) at the periphery. In the felspars the centre is usually more basic and richer in lime than the surrounding faces, and successive zones may often be noted, each less basic than those which lie within it. Phenocrysts of quartz (and of other minerals), instead of sharp, perfect crystalline faces, may show rounded corroded surfaces (Pl. I. fig. 9), with the points blunted and irregular tongue-like projections of the matrix into the substance of the crystal. It is clear that after the mineral had crystallized it was partly again dissolved or corroded at some period before the matrix solidified. Corroded phenocrysts of biotite and hornblende are very common in some lavas; they are surrounded by black rims of magnetite mixed with pale green augite. The hornblende or biotite substance has proved unstable at a certain stage of consolidation and has been replaced by a paramorph of augite and magnetite which may be partially or completely substituted for the original crystal but still retains its characteristic outlines.

Let us now consider the characteristics of a typical deep-seated rock like granite or diorite (Pl. II. figs. 4, 5, 9). That these are igneous is proved by the manner in which they have burst through the superincumbent strata, filling the Plutonic or cracks with ramifying veins; that they were at a very T yssal high temperature is equally clear from the changes which yp. they have induced in the rocks in contact with them. But as their heat could dissipate only very slowly, because of the masses which covered them, complete crystallization has taken place and ma vitreous rapidly chilled matter is present. As they have had time to come to rest before crystallizing they are not fluidal. Their contained gases have not been able to escape through the thick layer of strata beneath which they were injected, and may often be observed occupying cavities in the minerals, or have occasioned many important modifications in the crystallization of the rock. Because their crystals are of approximately equal size these rocks are said to be granular; there is typically no distinction between a first generation of large well-shaped crystals and a fine-grained ground-mass. Their minerals have formed, however, in a definite order, and each has had a period of crystallization which maybe very distinct or may have coincided with or overlapped the period of formation of some of the other ingredients. The earlier have originated at a time when most of the rock was still liquid and are more or less perfect; the later are less regular in shape because they were compelled to occupy the interspaces left between the already formed crystals (Pl. II. figs. 5, 9). The former are said to be idiomorphic (or automorphic), the latter are anidiomorphic (allotriomorphic, xenomorphic).1 There are also many other characteristics which serve to distinguish the members of these two groups. Orthoclase, for example, is the typical felspar of granite, while its modification sanidine occurs in lavas of similar composition. The same distinction holds between elaeolite and nepheline. Leucite is common in lavas, very rare in plutonic rocks. Muscovite is confined to the intrusives. These differences show the influence of the physical conditions under which consolidation takes place.

There is a certain class of intrusive rocks which have risen upwards towards the surface, but have failed to reach it, and have solidified in fissures as dikes and intrusive sills at no great depth. To this type the name intrusive (or hypIntrusive or abyssal) is often given in distinction to the plutonic (or Ilypabyssal abyssal) which formed at greater depths. As might Types. be expected, they show structures intermediate between those of the effusive and the plutonic rocks. They are very commonly porphyritic, not rarely vitreous, and sometimes even vesicular. In fact many of them are indistinguishable petrologically from lavas of similar composition.

The attempt to form a special group of hypabyssal (intrusive and dike) rocks has met with much criticism and opposition. Such a group certainly cannot rank as equally important and equally well characterized with the plutonic and the effusive. But there are many kinds of rock which are not found to occur normally in any other manner. As examples we may cite the lamprophyres, the aplites and the porphyrites. These never occur as lava flows or as great plutonic bosses; if magmas of the same composition as these rocks occur in either of these ways they consolidate with different assemblages of minerals and different structures.

In subdividing the plutonic, the hypabyssal and the effusive rocks, the principle is followed of grouping those together which resemble one another in mineral con- Subdivisions stitution and in chemical composition. In a broad of Igneous sense these two properties are interdependent. Rock Class. 1 Idiomorphic, having its own characteristic form, Gr. tikos, belonging to one's self, µopen (form); allotriomorphic, from Gr. &XXorpcos, belonging to another (etXkos), a stranger (Eivos) The commoner rock constituents are nearly all oxides; chlorine, sulphur and fluorine are the only important exceptions to this and their total amount in any rock is usually much less than 1 %. F. W. Clarke has calculated that a little more Characters than 47%° of the earth's crust consists of oxygen. It occurs principally in combination as oxides, of which the chief are silica, alumina, iron oxides, lime, magnesia, potash and soda. The silica functions principally as an acid, forming silicates, and tall the commonest minerals of igneous rocks are of this nature. From a computation based on 1672 analyses of all kinds of rocks Clarke arrived at the following as the average percentage composition: S10 2 = 59.71, Al 2 0 3 = 15.41, Fe203 = 2.63, FeO =3.52, MgO = 4.36, CaO = 4.90, Na 2 0 =3.55, K 2 0 =2.80, H 2 O =1.52, TiO 2 = o 60, P20 5 = 0.22, total 99.22%. All the other constituents occur only in very small quantities, usually much less than i %.

These oxides do not combine in a haphazard way. The potash and soda, for example, with a sufficient amount of alumina and silica, combine to produce felspars. In some cases they may take other forms, such as nepheline, leucite and muscovite, but in the great majority of instances they are found as felspar. The phosphoric acid with lime forms apatite. The titanium dioxide with ferrous oxide gives rise to ilmenite. Part of the lime forms lime felspar. Magnesia and iron oxides with silica crystallize as olivine or enstatite, or with alumina and lime form the complex ferro-magnesian silicates of which the pyroxenes, amphiboles and biotites are the chief. Any excess of silica above what is required to neutralize the bases will separate out as quartz; excess of alumina crystallizes as corundum. These must be regarded only as general tendencies, which are modified by physical conditions in a manner not as yet understood. It is possible by inspection of a rock analysis to say approximately what minerals the rock will contain, but there are numerous exceptions to any rule which can be laid down.

Hence we may say that except in acid or siliceous rocks containing 66% of silica and over, quartz will not be abundant. In basic rocks (containing 60% silica or less) it is rare and accidental. If magnesia and iron be above the average while silica is low olivine may be expected; where silica is present in greater quantity other ferro-magnesian minerals, such as augite, hornblende, enstatite or biotite, occur rather than olivine. Unless potash is high and silica relatively low leucite will not be present, for leucite does not occur with free quartz. Nepheline, likewise, is usually found in rocks with much soda and comparatively little silica. With high alkalis soda-bearing pyroxenes and amphiboles may be present. The lower the percentage of silica and the alkalis the greater is the prevalence of lime felspar as contracted with soda or potash felspar. Clarke has calculated the relative abundance of the principal rock-forming minerals with the following results: Apatite = o6, titanium minerals = 15, quartz = 12.0, felspars = 59.5, biotite = 3.8, hornblende and pyroxene = 16.8, total = 94.2%. This, however, can only be a rough approximation. The other determining factor, namely the physical conditions attending consolidation, plays on the whole a smaller part, yet is by no means negligible, as a few instances will prove. There are certain minerals which are practically confined to deep-seated intrusive rocks, e.g. microcline, muscovite, diallage. Leucite is very rare in plutonic masses; many minerals have special peculiarities in microscopic character according to whether they crystallized in depth or near the surface, e.g. hypersthene, orthoclase, quartz. There are some curious instances of rocks having the same chemical composition but consisting of entirely different minerals, e.g. the hornblendite of Gran, in Norway, containing only hornblende, has the same composition as some of the camptonites of the same locality which contain felspar and hornblende of a different variety. In this connexion we may repeat what has been said above about the corrosion of porphyritic minerals in igneous rocks. In rhyolites and trachytes early crystals of hornblende and biotite may be found in great numbers partially converted into augite and magnetite. The hornblende and biotite were stable under the pressures and other conditions which obtained below the surface, but unstable at higher levels. In the ground-mass of these rocks augite is almost universally present. But the plutonic representatives of the same magma, granite and syenite contain biotite and hornblende far more commonly than augite.

Those rocks which contain most silica and on crystallizing yield free quartz are erected into a group generally designated the " acid ", rocks. Those again which contain least silica and most Acid Intermediatemagnesia and iron, so that quartz is absent while olivine and Bas i c is usually abundant, form the " basic " group. The i ntermediate " rocks include those which are character ized by the general absence of both quartz and olivine. R An important subdivision of these contains a very high percentage of alkalis, especially soda, and consequently has minerals such as nepheline and leucite not common in other rocks. It is often separated from the others as the " alkali " or " soda " rocks, and there is a corresponding series of basic rocks. Lastly a small sub-group rich in olivine and without felspar has been called the " ultrabasic " rocks. They have very low percentages of silica but much iron and magnesia.

Except these last practically all rocks contain felspars or felspathoid minerals. In the acid rocks the common felspars are orthoclase, with perthite; microcline, oligoclase, all having much silica and alkalis. In the basic rocks labradorite, anorthite and bytownite prevail, being rich in lime and poor in silica, potash and soda. Augite is the commonest ferro-magnesian of the basic rocks, but biotite and hornblende are on the whole more frequent in the acid.

The rocks which contain leucite or nepheline, either partly or wholly replacing felspar are not included in this table. They are essentially of intermediate or of basic character. We might in consequence regard them as varieties of syenite, diorite, gabbro, &c., in which felspathoid minerals occur, and indeed there are many transitions between syenites of ordinary type and nepheline - or leucite - syenite, and between gabbro or dolerite and theralite or essexite. But as many minerals develop in these " alkali " rocks which are uncommon elsewhere, it is convenient in a purely formal classification like that which is outlined here to treat the whole assemblage as a distinct series.

	Alkali Felspar,	Soda Lime Felspar	Nepheline or
Commonest	Nepheline or Leu-	Nepheline or Leu-	Leucite, Augite,
Minerals.	cite,Augite,Horn- blende, Biotite.	cite,Augite,Horn- blende (Olivine).	Hornblende, Olivine.
Plutonic	Nepheline-syenite.	Essexite and	Ijolite and
type.	Leucite-syenite.	Theralite.	Missourite.
Intrusive 1 type.	Nepheline- porphyry
Effusive)		and	Nepheline-
type or l Lavas. S	Phonolite, Leucitophyre.	Tephrite Basanite.	basalt. Leucite-basalt

Nepheline and Leucite-bearing Rocks. Phis classification is based essentially on the mineralogical constitution of the igneous rocks. Any chemical distinctions between the different groups, though implied, are relegated to a subordinate position. It is admittedly artificial but it has grown up with the growth of the science and is still adopted as the basis on which more minute subdivisions are erected. The subdivisions are by no means of equal value. The syenites, for example, and the peridotites, are far less important than the granites, diorites and gabbros. Moreover, the effusive andesites do not always correspond to the plutonic diorites but partly also to the gabbros. As the different kinds of rock, regarded as aggregates of minerals, pass gradually into one another, transitional types are very common and are often so important as to receive special names. The quartz-syenites and nordmarkites may be interposed between granite and syenite, the tonalites and adamellites between granite and diorite, the monzonites between syenite and diorite, norites and hyperites between diorite and gabbro, and so on.

	Acid.	Intermediate.		Basic.	Ultrabasic.
	Quartz	Little or no Quartz.
Commonest	Orthoclase			No Quartz	No Felspar
Minerals.	(and Oligo- clase), Mica, Hornblende, Augite.	Orthoclase Hornblende, Augite, Biotite.	Plagioclase Hornblende, Augite, Biotite.	Plagioclase Augite, Olivine.	Augite, Hornblende, Olivine.
Plutonic or
A b y s s a l type.	Granite.	Syenite.	Diorite.	Gabbro.	Peridotite.
Intrusive or
H y p a b y s- sal type.	Quartz- porphyry.	Orthoclase- porphyry.	Porphyrite.	Dolerite.	Picrite.
Lavas o E ff u s i v e r type. J	Rhyolite, Obsidian.	Trachyte.	Andesite.	Basalt.	Limburgite.

There is of course a large number of recognized rock species not included in the tables given. These are of two kinds, either belonging to groups which are subdivisions of those enumerated (bearing the same relation to them that species do to genera) or rare and exceptional rocks that do not fall within any of the main subdivisions proposed. The question may be asked - When is a rock entitled to be recognized as belonging to a distinct species or variety and deserving a name for itself? It must, first of all, be proved to occur in considerable quantity at some locality, or better still at a series of localities or to have been produced from different magmas at more than one period of the earth's history. In other words, it must not be a mere anomaly. Moreover, it should have a distinctive mineral constitution, differing from other rocks, or something individual in the characters of its minerals or of its structures. It is often surprising how peculiar types of rock, believed at first Mineral Constitu- to be unique, turn up with identical features in widely scattered regions, alniiite, for example, occurs in Norway, Scotland, Montreal, British Columbia, New York and Brazil, tinguaite in Scotland, Norway, Brazil, Montana, Portugal, &c. This indicates that underlying all the variations in mineralogical, structural and chemical properties there are definite relationships which tend to repeat themselves, producing the same types whenever the same conditions are present.

Although in former years the view was widely current, especially in Germany, that igneous rocks belonging to different geological epochs should receive different names, it is now admitted on all sides that this cannot be upheld.

In 1902 a group of American petrographers brought forward a proposal to discard all existing classifications of igneous rocks and to substitute for them a " quantitative " classification based on chemical analysis. They showed how vague and often unscientific was much of the existing terminology and argued that as the chemical composition of an igneous rock was its most fundamental characteristic it should be elevated to prime position. Geological occurrence, structure, mineralogical constitution, the hitherto accepted criteria for the discrimination of rock species were relegated to the background. The completed rock analysis is first to be interpreted in terms of the rock-forming minerals which might be expected to be formed when the magma crystallizes, e.g. quartz felspars of various kinds, olivine, akermannite, felspathoids, magnetite, corundum and so on, and the rocks are divided into groups strictly according to the relative proportion of these minerals to one another. There is no need here to describe the minutia of the process adopted as the authors have stated them very clearly in their treatise (Quantitative Classification of Igneous Rocks, Chicago, 1902), and there is no indication that even in the United States it will ever displace the older classifications.

We can often observe in a series of eruptives belonging to one period and a restricted area certain features which distinguish them as a whole more or less completely from other. similar assemblages. Such groups are often said to sanguinity be consanguineous, and to characterize a definite " petrological province. " Excellent examples of this are furnished by the Devonian igneous rocks of southern Norway as described by BrSgger, the Tertiary rocks of the Hebrides (Harker), the Italian lavas studied by H. S. Washington. On a larger scale the volcanoes which girdle the Pacific (Andes, Cordillera, Japan, &c.), and those which occur on the volcanic islands of the Atlantic, show the same phenomena. Each of these groups has been formed presumably from a single deep-seated magma or source of supply and during a period which while necessarily prolonged was not of vast duration in a geological sense.

On the other hand, each of the great suites of eruptive rocks which constitute such a petrological province embraces a great Differentia- range of types. Prolonged eruptions have in a few cases a somewhat monotonous character, owing to the predominance of one kind of rock. Thus the lavas of the Hawaiian Islands are mostly basaltic, as are those of Oregon, Washington and the Deccan, all of which form geological masses of enormous magnitude. But it is more usual to find basalts, andesites, trachytes, dacites and many other rocks occurring in a single eruptive complex. The process by which a magma splits up into a variety of partial products is known as " differentiation." Its importance from the standpoint of theoretical petrology is very great, but as yet no adequate explanation of it has been offered. Differentiation may show itself in two ways. In the first type the successive emissions from a volcanic focus may differ considerably from one another. Thus in the Pentland Hills, near Edinburgh, the lavas which are of lower Devonian age, were first basaltic, then andesitic, trachytic and dacitic, and finally rhyolitic, and this succession was repeated a second time. Yet they all must have come from the same focus, or at any rate from a group of foci very closely connected with one another. Occasionally it is found that the earlier lavas are of intermediate character and that basic alternate with acid during the later stages of the volcanic history.

Not less interesting are those cases in which a single body of rock has in consolidation yielded a variety of petrographical types often widely divergent. This is best shown by great plutonic bosses which may be regarded as having once been vast subterranean spaces filled with a nearly homogeneous liquid magma. Cooling took place gradually from the outer surfaces where the igneous rock was in contact with the surrounding strata. The resultant laccolite (Gr. Mims, pit, crater, XiDos, stone), stock or boss, may be a few hundred yards or many miles in diameter and often contains a great diversity of crystalline rocks. Thus peridotite, gabbro, diorite, tonalite and granite, are often associated, usually in such a way that the more basic are the first-formed and lie nearest the external surfaces of the mass. The reverse sequence occurs occasionally, the edges being highly acid while the central parts consist of more basic rocks. Sometimes the later phases penetrate into and vein the earlier; evidently there has been some movement due to temporary increase of pressure when part of the laccolite was solid and part still in a liquid state. This links these phenomena with those above described where successive emissions of different character have proceeded outwards from the focus.

According to modern views two explanations of these facts are possible. Some geologists hold that the different rock facies found in association are often due to local absorption of surrounding rocks by the molten magma (" assimilation "). Effects of this kind are to be expected, and have been clearly proved in many places. There is, however, a general reluctance to admit that they are of great importance. The nature and succession of the rock species do not as a rule show any relation to the sedimentary or other materials which may be supposed to have been dissolved; and where solution is known to have gone on the products are usually of abnormal character and easily distinguishable from the common rock types.

Hence it is generally supposed that differentiation is to be ascribed to some physical or chemical processes which lead to the splitting up of a magma into dissimilar portions, each of which consolidates as a distinct kind of rock. Two factors can be selected as probably most potent. One important factor is cooling and another is crystallization. According to physico-chemical laws the least soluble substances will tend to diffuse towards the cooling surfaces (Ludwig - Sorets's principle). This is in accordance with the majority of the observed facts and is probably a vera causa of differentiation, though what its potency may be is uncertain. As a rock solidifies the minerals which crystallize follow one another in a more or less well-defined order, the most basic (according to Rosenbusch's law) being first to separate out. That in a general way the peripheral portions of a laccolite consist mainly of those early basic minerals suggests that the sequence of crystallization helps largely in determining the succession (and consequently the distribution of rock species in a plutonic complex). Gravity also may play a part, for it is proved that in a solution at rest the heaviest components will be concentrated towards the base. This must, however, be of secondary importance as in laccolites the top portions often consist of more basic and heavier varieties of rock than the centres. It has also been argued that the earliest minerals being heaviest and in any case denser than the fused magma around them, will tend to sink by their own weight and to be congregated near the bottom of the mass. Electric currents, magnetic attraction and convection currents have also been called in to account for the phenomena observed. Magmas have also been compared to liquids which, when they cool, split up into portions no longer completely soluble in one another (liquation hypothesis). Each of these partial magmas may dissolve a portion of the others and as the temperature falls and the conditions change a range of liquids differing in composition may be supposed to arise.

All igneous magmas contain dissolved gases (steam, carbonic acid, sulphuretted hydrogen, chlorine, fluorine, boric acid, &c.). Of these water is the principal, and was formerly believed to have percolated downwards from the earth's surface to the heated rocks below, but is now generally admitted to be an integral part of the magma. Many peculiarities of the structure of the plutonic rocks as contrasted with the lavas may reasonably be accounted for by the operation of these gases, which were unable to escape as the deep-seated masses slowly cooled, while they were promptly given up by the superficial effusions. The acid plutonic or intrusive rocks have never been reproduced by laboratory experiments, and the only successful attempts to obtain their minerals artificially have been those in which special provision was made for the retention of the " mineralizing " gases in the crucibles or sealed tubes employed. These gases often do not enter into the composition of the rockforming minerals, for most of these are free from water, carbonic acid, &c. Hence as crystallization goes on the residual liquor must contain an ever-increasing proportion of volatile constituents. It is conceivable that in the final stages the still uncrystallized part of the magma has more resemblance to a solution of mineral matter in superheated steam than to a dry igneous fusion. Quartz, for example, is the last mineral to form in a granite. It bears much of the stamp of the quartz which we know has been deposited from aqueous solution in veins, &c. It is at the same time the most infusible of all the common minerals of rocks. Its late formation shows that in this case it arose at comparatively low temperatures and points clearly to the special importance of the gases of the magma as determining the sequence of crystallization.

When solidification is nearly complete the gases can no longer be retained in the rock and make their escape through fissures towards the surface. They are powerful agents in attacking the minerals of the rocks which they traverse, and instances of their operation are found in the kaolinization of granites, tourmalinization and formation of greisen, deposit of quartz veins, stanniferous and auriferous veins, apatite veins, and the group of changes known as propylitization. 1 These " pneumatolytic " (Gr. 7rv€i aa, spirit, vapour, May, to loose, dissolve) processes are of the first importance in the genesis of many ore deposits. They are a real part of the history of the magma itself and constitute the terminal phases of the volcanic sequence.

The complicated succession from basic (or ultrabasic) to acid types exemplified in the history of many magmas is reflected with 1 The term " propylite " (Gr. irp6 ruXov, a gateway) was given by Richthofen to a volcanic rock which is supposed to have marked a new epoch in volcanic geology (see Andesite).

astonishing completeness in the history of individual products. In each class of rock crystallization follows a definite course. The first minerals to separate belong to a group known Sequence of as the minor accessories; this includes zircon, apatite, Crystalliza- sp h ene. iron oxides; then follow in order olivine, augite, tion. hornblende, biotite, plagioclase, felspar (beginning with > P g ? ? P (g ? g the varieties most rich in lime and ending with those which contain most soda), orthoclase, microcline and quartz (with micropegmatite). Many exceptions to this rule are known; the same mineral may crystallize at two different periods; two or more minerals may crystallize simultaneously or the stages in which they form may overlap. But the succession above given holds in the vast majority of cases. Expressed in this way: the more basic minerals precede the less basic; it is known as Rosenbusch's law.

Types of Structure

In some rocks there seems to be little tendency for the minerals to envelop one another. This is true of many gabbros, aplites and granites (Pl. III, fig. 7). The grains then lie side by side, with the faces of the latter moulded on or adapted to the more perfect crystalline outlines of the earlier. More commonly some closer relationship exists between them. When the smaller idiomorphic crystals of the first-formed are scattered irregularly through the larger and less perfect crystals of later origin, the Poililltic. structure is said to be poikilitic (Gr. 7roueXos, many coloured, mottled). A variety of this, known as ophitic (Pl. III, fig. 6), is very characteristic of many dolerites and diabases, in which large plates of augite enclose many small laths of plagioclase felspar. Biotite and hornblende frequently enclose felspar ophitically; less commonly iron oxides and sphene do so. In peridotites the " lustre-mottled " structure arises from pyroxene or hornblende enveloping olivine in the same manner (Pl. III, fig. 8). In these cases no crystallographic relation exists between the two minerals (enclosing and enclosed).

But often the surrounding mineral has been laid down on the surface of the other in such a way that they have certain crystalline faces or axes parallel to one another. This is known Parallel . as " parallel growth." It is best seen in zoned crystals Growths of plagioclase felspar, which may range in composition from anorthite to oligoclase, the more acid layers being deposited regularly on the surfaces of the more basic. Biotite and muscovite, hornblende and augite, enstatite and diallage, epidote and orthite, very frequently are associated in this way.

When two minerals crystallize simultaneously they may be intergrown in " graphic " fashion. The best example is quartz lc and orthoclase occurring together as micropegmatite &aph Inter- (Pl. II, figs. 6 and 8). The quartz forms angular growths. patches in the felspar, which though separated have gro the same crystalline orientation and one position of extinction, while the felspar on its part behaves in the same way. Two porous crystals thus interpenetrate but the scattered parts of each mineral maintain their connexion with the others. There may be also a definite relation between the crystalline axes of the two crystals, though this is not known in all cases. Augite also occurs in graphic intergrowth with hornblende, olivine and felspar; and hornblende, cordierite, epidote and biotite in graphic intergrowth with quartz.

Physical Chemistry of Igneous Rocks

The great advances that have been made in recent years in our knowledge of physical chemistry have very important bearings on petrological investigations. Especially in the study of the genesis of igneous rocks we anticipate that by this means much light will be thrown on problems which are now very obscure and a complete revolution in our ideas of the conditions which affect crystallization may yet be the consequence. Already many important results have been gleaned. As yet little work of an exact and quantitative nature has been done on actual rocks or on mixtures resembling them in composition, but at the Carnegie Institution in Washington, an elaborate series of experiments in the synthesis of minerals and the properties of mixtures of these is being carried on, with all the refinements which modern science can suggest. The work of Doelter and of Vogt may also be mentioned in this connexion. At the same time the mathematical theory of the physical processes involved has received much attention, and serves both to direct and to elucidate the experimental work.

A fused mixture of two minerals may be regarded as a solution of one on the other. If such a solution be cooled down, crystalliza- Magmas as tion will generally set in and if the two components be Solutions independent (or do not form mixed crystals) one of them may be expected to start crystallizing. On further cooling, more of this mineral will separate out till at last a residue is left which contains the two components in definite proportions. This mixture, which is known as the eutectic mixture, has the lowest melting-point of any which can be formed from these minerals. If heat be still abstracted the eutectic will consolidate as a whole; its two mineral components will crystallize simultaneously. At any given pressure the composition of the eutectic mixture in such a case is always the same.

Similarly, if there be three independent components (none of which forms mixed crystals with the others), according to their relative amounts and to the composition of the eutectic mixture one will begin to crystallize; then another will make its appearance in solid form, and when the excess of these has been removed, the ternary eutectic (that mixture of the three which has the lowest melting-point) will be produced and crystallization of all three components will go on simultaneously.

These processes have without doubt a very close analogy to the formation of igneous rocks. Thus in certain felsites or porphyries which may be considered as being essentially mixtures of quartz and felspar, a certain amount of quartz has crystallized out at an early period in the form of well-shaped porphyritic crystals, and thereafter the remainder of the rock has solidified as a very fine-grained, cryptocrystalline or sometimes micrographic groundmass which consists of quartz and felspar in intimate intermixture. The latter closely resembles a eutectic, and chemical studies have proved that within somewhat narrow limits the composition of these felsitic ground-masses is constant.

But the comparison must not be pushed too far, as there are always other components than quartz and felspar (apatite, zircon, biotite and iron oxides being the most common), and in rocks of this type the gases dissolved in the magma play a very important part. As crystallization goes on, these gases are set free and their pressure must increase to some extent. Moreover, the felspar is not one mineral but two or perhaps three, there being always soda felspar and potash felspar and usually also a small amount of lime felspar in these porphyries.

In a typical basic rock the conditions are even more complex. A dolerite, for example, usually contains, as its last products of crystallization, pyroxene and felspar. Of these the latter consists of three distinct species, the former of an unknown number; and in each case they can form mixed crystals, to a greater or less extent with one another. From these considerations it will be clear that the properties of solutions of two or three independent components, do not necessarily explain the process of crystallization in any igneous rock.

Very frequently in porphyries not only quartz but felspar] also is present in large well-formed early crystals. Similarly in basalts, augite and felspar may appear both as phenocrysts and as components of the ground-mass. As an explanation of this it has been suggested that supersaturation has taken place. We may suppose that the augite which was in excess of the proportion necessary to form the felspar-augite, eutectic mixture, first separated out. When the remaining solution reached the eutectic composition the felspar did not at once start crystallizing, perhaps because nuclei are necessary to initiate crystal-growth and these were not at hand; augite went on crystallizing while felspar lagged behind. Then felspar began and as the mixture was now supersaturated with that mineral a considerable amount of it was rapidly thrown out of the solution. At the same time there would be a tendency for part of the augite, already crystallized, to be dissolved and its crystals would be corroded, losing their sharp and perfect edges, as is often observed in rocks of this group. When the necessary adjustments had been made the eutectic mixture would be established and thereafter the two minerals would consolidate simultaneously (or nearly so) till crystallization was complete.

There is a good deal of evidence to show that supersaturation is not unimportant in igneous magmas. The frequency with which they form glasses proves that under certain conditions the molten rocks are highly viscous. Much will depend also on the presence, accidental or otherwise, of nuclei on which a mineral substance can be deposited. It is known that minerals differ in their tendency to crystallize, some doing so very readily while others are slow and backward. The rate at which crystallization goes on depends on many factors, and there are remarkable differences in this respect between minerals.

On the other hand, there is plenty of evidence to show that supersaturation, though probably one of the causes, is not the principal cause of the appearance of more than one mineral in two generations of crystals. In some of the quartz-porphyries, for example, there are phenocrysts not only of quartz and felspar but also of micropegmatite. These prove that quartz and felspar were not crystallizing successively or alternately but simultaneously.

The great majority of the minerals found in igneous rocks are not of simple composition, but are mixtures of various elementary minerals in very different proportions. This enormously complicates the theoretical problems of consolidation. It has been found, for example, that in the case of three minerals - one of which is independent, while the two others can form mixed crystals - there is a large number of possible sequences; and, what is very important, one mineral may separate out entirely at an early stage, or its crystallization may be interrupted and not continuous. The ternary eutectic, which is produced by a mixture of three independent minerals, may not in such a case be the last substance to crystallize, and may not be present at all. This is very much in accordance with the observed facts of petrology; for usually in a rock there is one mineral which indubitably was the last of all to finish crystallizing and contained no appreciable quantity of the others.

As yet we know little about such important questions as the composition of the eutectic mixtures of rock-minerals, their latent heat of fusion, specific heats, mutual solubilities, inversion temperatures, &c. Until we are in possession of a large body of accurate information on such points as these the theoretical treatment of the processes involved in the formation of igneous rocks cannot be successfully handled. But every day sees an increase in the amount of data available, and encourages us to believe that sooner or later some of the simpler igneous rocks at any rate will be completely explicable on physico-chemical principles.

Rock masses of igneous origin have no sooner consolidated than they begin to change. The gases with which the magma is charged Post are slowly dissipated, lava-flows often remain hot and steaming for many years. These gases attack the cornponents of the rock and deposit new minerals in cavities and fissures. The beautiful zeolites, so well known to collectors of minerals, are largely of this origin. Even before these " post-volcanic " processes have ceased atmospheric decomposition begins. Rain, frost, carbonic acid, oxygen and other agents operate continuously, and do not cease till the whole mass has crumbled down and most of its ingredients have been resolved into new products. In the classification of rocks these secondary changes are generally considered unessential; rocks are classified and described as if they were ideally fresh, though this is rarely the case in nature.

Epigenitic change (secondary processes) may be arranged under a number of headings, each of which is typical of a group of rocks or rock-forming minerals, though usually more than Changes. - hanges. one of these alterations will be found in progress in the same rock. Silicification, the replacement of the minerals by crystalline or crypto-crystalline silica, is most common in acid rocks, such as rhyolite, but is also found in serpentine, &c. Kaolinization is the decomposition of the felspars, which are the commonest minerals of igneous rocks, into kaolin (along with quartz, muscovite, &c.); it is best shown by granites and syenites. Serpentinization is the alteration of olivine to serpentine (with magnetite); it is typical of peridotites, but occurs in most of the basic rocks. In uralitization secondary hornblende replaces augite; this occurs very generally in diabases; chloritization is the alteration of augite (biotite or hornblende) to chlorite, and is seen in many diabases, diorites and greenstones. Epidotization occurs also in rocks of this group, and consists in the development of epidote from biotite, hornblende, augite or plagioclase felspar.

The sedimentary rocks, which constitute the second great group, have many points in common that distinguish them from the igneous and the metamorphic. They have all originated on the surface of the earth, and at the period of their formation were exposed only to the temperature of the air and to atmospheric pressure (or the pressures which exist at the bottoms of seas and lakes). Their minerals are in most cases not susceptible to change when exposed to moist air or sea, and many of them are hydrated (chlorite, micas, &c.), or oxidized (iron ores), or contain carbonic acid (calcite, dolomite). The extent, however, to which this is the case depends largely on the rapidity with which they have accumulated; coarse rocks quickly piled up often consist of materials only partly weathered. When crystalline, the sedimentary rocks are usually soluble at low temperatures. The members of this group occur in beds or strata, hence they are often known as the stratified rocks; the upper beds are always of later formation than those which underlie them, except (as may happen when great disturbance has taken place) the whole series is inverted or overturned. Many of the stratified rocks have been formed by the agency of moving water (rivers, currents, &c.) and are grouped together as " aqueous " rocks; others have been deposited by the wind in deserts, on sandy beaches, &c. (these are " aeolian "). Others are the remains of animals or of plants, modified by the action of time, pressure and percolating water. Lastly, we find beds of crystalline nature, such as rock-salt and gypsum, which have been formed by the desiccation of saline waters; other crystalline stratified rocks, such as dolomite and many bedded iron-stones, are replacement products due to the introduction of mineral matter in solution, which replaced the original rock mass partially or wholly.

When the rocks exposed at the earth's surface give way before the attack of the agencies of denudation, they crumble down and are resolved into two parts. One of these consists of solid material (sand, clay and angular debris) insoluble in carbonated waters; the other part is dissolved and washed away. The undissolved residues, when they finally come to rest, form clastic sedimentary rocks (sandstone, conglomerate, shale, &c.). The dissolved portions are partly transferred to the sea, where they help to increase its store of salts, and may again be precipitated as crystalline sedimentary rocks; but they are also made use of by plants and by animals to form their skeletal and vital tissues. From this latter portion the rocks of organic origin are built up. These may also contain certain ingredients derived from the atmosphere (nitrogen, carbon in coals, &c.).

We have thus three types of sediments of distinct origin, which may be named the clastic (or fragmental), the crystalline and the organic.

The clastic materials may accumulate in situ, and then differ chiefly in their disintegrated and weathered state from the parent rock masses on which they rest. The best example of these are the soils, but in elevated regions angular broken rock often covers large areas. More usually they are transported by wind or water, and become sorted out according to their size and density. The coarsest debris comes first to rest and is least worn and weathered; it includes screes, gravels, coarse sands, &c., and consolidates as conglomerates, breccias and pebbly grits. The bedding of these rocks is rudimentary and imperfect, and as each bed is traced along its outcrop it frequently changes its character with the strata on which it rests. The most finely divided sediment travels farthest, and is laid down in thin uniform sheets of wide extent. It is known as mud and clay; around the shores of our continents, at distances of a hundred miles and more from land, great sheets of mud are spread over the ocean floors. This mud contains minute particles of quartz and of felspar, but consists essentially of finely divided scaly minerals, which by their small size and flat shape tend to remain suspended in water for a very long time. Chlorite, white micas and kaolin are the best examples of this class of substances. Wind action is even more effective than water in separating and removing these fine particles. They to a very large extent escape mechanical attrition, because they are transported in suspension and are not swept along the ground or the bottom of the sea; hence they are mostly angular. Fragments of intermediate magnitudes (from 1 o of an inch to a of an inch) are classed as sands. They consist largely of quartz, because it does not weather into scaly minerals like felspar, and having but a poor cleavage does not split up into flakes like mica or chlorite. These quartz grains have been rolled along and are Usually rounded and worn (Pl. IV., fig. i). More or less of garnet felspar, tourmaline, zircon, rutile, &c., are mixed with the quartz, because these are hard minerals not readily decomposed.

The mechanical sorting by the transporting agencies is usually somewhat incomplete, and mixed types of sediment result, such as gravels containing sand, or clays with coarser arenaceous particles.. Moreover, successive layers of deposit may not always be entirely similar, and alternations of varying composition may follow one another in thin laminae: e.g. laminae of arenaceous material in beds of clay and shale. Organic matter is frequently mingled with the finer-grained sediments.

These three types have been named the psephitic (or pebbly; Gr. 1,q(pos, pebble); psammitic (or sandy, Gr. ifiaµµos, sand), and pelitic (or muddy: Gr. mud).

Two groups of clastic sediments deserve special treatment. The pyroclastic (Gr. and KXavrO, broken) rocks of volcanic origin, consist mostly of broken pieces of lava (bombs, ash, &c.) (Pl. IV. fig. 2), and only accidentally contain other rocks or fossils. They are stratified, and may be coarse or fine, but are usually much less perfectly sorted out, according to their fineness, than ordinary aqueous or aeolian deposits. The glacial clays (boulder clays), representing the ground moraines of ancient glaciers and ice sheets, are characterized by the very variable size of their ingredients and the striated, blunted sub-angular form of the larger rock fragments. In them stratification is exceptional and fossils are very rare.

The crystalline sedimentary rocks have been deposited from solution in water. The commonest types, such as rock-salt, gypsum, anhydrite, carnallite, are known to have arisen by the ? evaporation of enclosed saline lakes exposed to a dry atmosphere. They occur usually in beds with layers of red clay and marl; some limestones have been formed by calcareous waters containing carbonate of lime dissolved in an excess of carbonic acid; with the escape of the volatile gas the mineral matter is precipitated (sinters, Sprudelstein, &c.). Heated waters on cooling may yield up part of their dissolved mineral substances; thus siliceous sinters are produced around geysers and hot springs in many parts of the world. There seems no reason to separate from these the veinstones which fill the fissures by which these waters rise to the surface. They differ from those above enumerated in being more perfectly crystallized and in having no definite stratification,. but only a banding parallel to the more or less vertical walls of the fissure. Another subdivision of this class of rocks is due to recrystallization or crystalline replacement of pre-existing sediments. Thus limestones are dolomitized or converted into ironstones, flints and cherts, by percolating waters which remove the lime salts and substitute for them compounds of iron, magnesia, silicon, and so on. This may be considered a kind of metamorphism; it is generally known as metasomatism (q.v.).

The rocks of. organic origin may be due to animals or plants.. They are of great importance, as limestones and coals belong to, this group. They are the most fossiliferous of all rocks; but clastic sediments are often rich in fossils though crystalline sediments rarely are. They may be subdivided, according to their dominant components, into calcareous, carbonaceous, siliceous, ferruginous, and so on. The calcareous organic rocks may consist principally of foraminifera, crinoids, corals, brachiopoda, mollusca, polyzoa, &c. Most of them, however, contain a mixture of organisms. By crystallization and metasomatic changes they often lose their organic structures; metamorphism of any kind has the same effect. The carbonaceous rocks are essentially plant deposits; they include peat, lignite and coal. The siliceous organic rocks include radiolarian and diatom oozes; in the older formations they occur as radiolarian cherts. Flint nodules owe their silica to disseminated fossils of this nature which have been dissolved and redeposited by concretionary action. Some kinds of siliceous sinter may be produced by organisms inhabiting hot silicated waters. Calcareous oolites in the same way may have arisen through the agency of minute plants. Bog iron ores also may be of organic rather than of merely chemical origin. The phosphatic rocks so extensively sought after as sources of fertilizing agents for use in agriculture are for the most part of organic origin, since they owe their substance to the remains of certain varieties of animals which secrete a phosphatic skeleton; but most of them no longer show organic structures but have been converted into nodular or concretionary forms.

All sediments are at first in an incoherent condition (e.g. sands, clays and gravels, beds of shells, &c.), and in this state they may remain for an indefinite period. Millions of years have. elapsed since some of the early Tertiary strata gathered on the ocean floor, yet they are quite friable (e.g. the London Clay) and differ little from many recent accumulations. There are few exceptions, however, to the rule that with increasing age sedimentary rocks become more and more indurated, and the older they are the more likely it is that they will have the firm consistency generally implied in the term " rock." The pressure of newer sediments on underlying masses is apparently one cause of this change, though not in itself a very powerful one. More efficiency is generally ascribed to the action of percolating water, which takes up certain soluble materials and redeposits them in pores and cavities. This operation is probably accelerated by the increased pressure produced by superincumbent masses, and to some extent also by the rise of temperature which inevitably takes place in rocks buried to some depth beneath the surface. The rise of temperature, however, is never very great; we know more than one instance of sedimentary deposits which have been buried beneath four or five miles of similar strata (e.g. parts of the Old Red Sandstone), yet no perceptible difference in condition can be made out between beds of similar composition at the top of the series and near its base. The redeposited cementing material is most commonly calcareous or siliceous. Limestones, which were originally a loose accumulation of shells, corals, &c., become compacted into firm rock in this manner; and the process often takes place with surprising ease, as for example in the deeper parts of coral reefs, or even in wind-blown masses of shelly sand exposed merely to the action of rain. The cementing substance may be regularly deposited in crystalline continuity on the original grains, where these were crystalline; and even in sandstones (such as Kentish Rag) a crystalline matrix of calcite often envelopes the sand grains. The change of aragonite to calcite and of calcite to dolomite, by forming new crystalline masses in the interior of the rock, usually also accelerates consolidation. Silica is less easily soluble in ordinary waters, but even this ingredient of rocks is dissolved and redeposited with great frequency. Many sandstones are held together by an infinitesimal amount of colloid or cryptocrystalline silica; when freshly dug from the quarry they are soft and easily trimmed, but after exposure to the air for some time they become much harder, as their siliceous cement sets and passes into a rigid condition. Others contain fine scales of kaolin or of mica. Argillaceous materials may be compacted by mere pressure, like graphite and other scaly minerals. Oxides and carbonates of iron play a large part in many sedimentary rocks and are especially important as colouring matters. The red sands and limestones, for example, which are so abundant, contain small amounts of ferric oxide (haematite), which in a finely divided state gives a red hue of all rocks in which it is present. Limonite, on the other hand, makes rocks yellow or brown; oxides of manganese, asphalt and other carbonaceous substances are the cause of the black colour of many sediments. Bluish tints result sometimes from the presence of phosphates or of fluorspar; while green is most frequently seen in rocks which contain glauconite or chlorite.

Metamorphic Rocks

The metamorphic rocks, which form the third great subdivision, are even more varied than the igneous and the sedimentary. They include representatives of nearly all kinds of the other two classes, their common characteristic being that they have all undergone considerable alterations in structure or in mineral composition. The agencies of metamorphism (q.v.) are of two kinds - thermal and regional. In the former case contact with intrusive igneous masses, such as granite, laccolites or dikes, have indurated and recrystallized the original rock. In the second case the actions are more complex and less clearly understood; it is evident that pressure and interstitial movement have had a powerful influence, possibly assisted by rise of temperature. In thermal or contact alteration the rocks are baked, indurated, and often in large measure recrystallized. In regional metamorphism recrystallization also goes on, but the final products are usually schists and gneisses. It is as a rule not difficult to distinguish the two classes of metamorphic rocks at a glance, and they may conveniently be considered separately.

When a rock is contact altered by an igneous intrusion it very frequently becomes harder, more crystalline and more lustrous, owing to the development of many small crystals in its mass. Many altered rocks of this type were formerly called hornstones, and the term hornfelses (Ger.

or- Hornfels) is often used by geologists to signify those fine grained, compact, crystalline products of thermal metamorphism. A shale becomes a dark argillaceous hornfels, full of tiny plates of brownish biotite; a marl or impure limestone changes to a grey, yellow or greenish lime-silicate-hornfels, tough and splintery, with abundance of augite, garnet, wollastonite and other minerals in which lime is an important component. A diabase or andesite becomes a diabase hornfels or andesite hornfels with a large development of new hornblende and biotite and a partial recrystallization of the original felspar. A chert or flint becomes a finely crystalline quartz rock; sandstones lose their clastic structure and are converted into a mosaic of small close-fitting grains of quartz.

If the rock was originally banded or foliated (as, for example, a laminated sandstone or a foliated calc-schist) this character may not be obliterated, and a banded hornfels is the product; fossils even may have their shapes preserved, though entirely recrystallized, and in many contact altered lavas the steam cavities are still visible, though their contents have usually entered into new combinations to form minerals which were not originally present. The minute structures, however, disappear, often completely, if the thermal alteration is very profound; thus small grains of quartz in a shale are lost or blend with the surrounding particles of clay, and the fine ground-mass of lavas is entirely reconstructed.

By recrystallization in this manner peculiar rocks of very distinct types are often produced. Thus shales may pass into cordierite rocks, or may show large crystals of andalusite (and chiastolite, Pl. IV., fig. 9), staurolite, garnet, kyanite and sillimanite. A considerable amount of mica (both muscovite and biotite) is simultaneously formed, and the resulting product has a close resemblance to many kinds of schist. Limestones, if pure, are often turned into coarsely crystalline marbles (Pl. IV., fig. 4); but if there was an admixture of clay or sand in the original rock such minerals as garnet, epidote, idocrase, wollastonite, will be present. Sandstones when greatly heated may change into coarse quartzites composed of large clear grains of quartz. These more intense stages of alteration are not so commonly seen in igneous rocks, possibly because their minerals, being formed at high temperatures, are not so easily transformed or recrystallized.

In a few cases rocks are fused and in the dark glassy product minute crystals of spinel, sillimanite and cordierite may separate out. Shales are occasionally thus altered by basalt dikes, and felspathic sandstones may be completely vitrified. Similar changes may be induced in shales by the burning of coal seams or even by an ordinary furnace.

There is also a tendency for interfusion of the igneous with the sedimentary rock. Granites may absorb fragments of shale or pieces of basalt. In that case hybrid rocks arise which have not the characters of normal igneous or sedimentary rocks. Such effects are scarce and are usually easily recognized. Sometimes an invading granite magma permeates the rocks around, filling their joints and planes of bedding, &c., with threads of quartz and felspar. This is very exceptional, but instances of it are known and it may take place on a large scale.