-f S? 1 ^L , /x? % <-^/ ! & PETKOLOGY FOE STUDENTS. EonDon: C. J. CLAY AND SONS, CAMBEIDGE UNIVERSITY PEESS WAREHOUSE, AVE MAEIA LANE, AND H. K. LEWIS, 136, GOWER STREET, W.C. laggoto: 263, ARGYLE STREET. 2.eiv?is: F. A. BROCKHAUS. Hork: THE MACMILLAN COMPANY. Cambrftgr Natural J^n'mtr JHamtals. PETROLOGY FOE STUDENTS: AN INTRODUCTION TO THE STUDY OF ROCKS UNDER THE MICROSCOPE. BY ALFRED BARKER, M.A., F.G.S. FELLOW OP ST JOHN'S COLLEGE, AND DEMONSTRATOR IN GEOLOGY (PETROLOGY) IN THE UNIVERSITY OF CAMBRIDGE. SECOND EDITION, REVISED. CAMBRIDGE: AT THE UNIVERSITY PRESS. 1897 [All Rights reserved.} First Edition, 1895. Second Edition, 1897. Geol. Lib. PREFACE TO FIRST EDITION. THE greatest difficulty that I have experienced as a teacher of petrology has been in recommending a text-book suitable for English students. As an attempt to meet this difficulty the following pages have been written. They are intended as a guide to the study of rocks in thin slices, and are of course assumed to be supplemented throughout by demonstrations on actual specimens. For this reason the examples are chosen, so far as possible, from British rocks. No systematic account has been given here of the crystal lographic and optical properties of minerals : Pro- fessor Iddings' translation of Rosenbusch's text-book and Dr Hatch's translation of the same author's tables have rendered this unnecessary. In particular I have made no reference to methods depending on the use of con- vergent light. I am indebted, as every writer on this subject must be, to the works of Zirkel, Rosenbusch, Fouque and Levy, and other authorities, as well as to Mr Teall's "British Petrography " ; but, so far as was possible, all descriptions of rocks have been written directly from specimens. Numerous references to original sources have been given in foot-notes, but I have thought it advisable to restrict these references to easily accessible English works. VI PREFACE. I have often cited also the coloured plates in some standard works of reference, to which most students will have access. In view of the difficulty of adequately representing rock-sections by means of process-blocks, the figures in this book are selected chiefly to illustrate simple structural characters, and some of them are necessarily rather diagrammatic. A. H. ST JOHN'S COLLEGE, CAMBRIDGE. May, 1895. PREFACE TO SECOND EDITION. IN this edition, which is revised throughout and in part re-written, I have endeavoured to profit by the criticisms of reviewers and of private friends. In particular, I have, at the instance of more than one correspondent, devoted somewhat more attention to American examples, at least among the igneous rocks. For figure 57, illustrating Girvanella, I am indebted to the kindness of Mr E. Wethered, F.G.S. A. H. ST JOHN'S COLLEGE, CAMBRIDGE. April, 1897. CONTENTS. CHAPTER PAGE I. INTRODUCTION 1 A. PLUTONIC ROCKS . 22 II. GRANITES 27 III. SYENITES (INCLUDING NEPHELINE-SYENITES) 42 IV. DIORITES 54 V. GABBROS AND NORITES . . . . . 67 VI. PERIDOTITES (INCLUDING SERPENTINE-ROCKS) 84 B. HYPABYSSAL ROCKS . . . '.. 98 VII. ACID INTRUSIVES . . . . . .99 VIII. PORPHYRIES AND PORPHYRITKS ... 112 IX. DIABASES . . . . . : *vV' . . 122 X. LAMPROPHYRES . . . . . . 134 C. VOLCANIC ROCKS . . ... . .144 XL RHYOLITES . 147 XII. TRACHYTES AND PHONOLITES .... 163 XIII. ANDESITES . ... . ... 174 'XIV. BASALTS 188 XV. LEUCITE- AND NEPHELINE-BASALTS, ETC. . 204 D. SEDIMENTARY ROCKS 215 XVI. ARENACEOUS ROCKS . .... . 216 XVII. ARGILLACEOUS ROCKS . . . . 230 XVIII. CALCAREOUS ROCKS ... . . 240 XIX. PYROCLASTIC ROCKS . . . . . . 264 APPENDIX TO SEDIMENTARY ROCKS . . 275 viii CONTENTS. CHAPTER PAGE E. METAMORPHISM . . .... . . .279 XX. THERMAL METAMORPHISM .... 282 XXI. DYNAMIC METAMORPHISM 304 XXII. VARIOUS CRYSTALLINE ROCKS . . . 318 INDEX . . 331 REFERENCES. Berwerth, Mikroskopische Structurbilder der Massengesteine (chromo- lith.), Stuttgart, 1895- . Cohen, Sammlung von Mikrophotographien...von Mineralien und Gesteinen, Stuttgart, 1881-2. Rosenbusch-Iddings, Microscopical Physiography of the Rock-forming Minerals (with photographic plates), 1888. Fouque and Levy, Mine'ralogie micrographique (with atlas of coloured plates), 1879. Teall, British Petrography (with numerous coloured plates), 1888. Rosenbusch-Hatch, Petrographical Tables. Cole, Studies in Microscopical Science, 1882-3. Watts, Guide = Guide to the Colle"tions of Rocks and Fossils be- longing to the Geological Survey of Ireland, 1895, Dublin. ABBREVIATIONS. G. M. = Geological Magazine. M. M. = Mineralogical Magazine. Q.J. G. S. = Quarterly Journal of the Geological Society. A. J. S. = American Journal of Science. CHAPTER I. INTRODUCTION. IN this section will be included such notes on the optical properties of minerals as may be of use to a novice ; but there will be no attempt to supersede the use of books dealing systematically with the subject. Microscope 1 . We shall assume the use of a microscope specially adapted for petrological work, and therefore fitted with polarizing and analysing prisms, rotating stage with graduated circle and index, and ' cross-wires ' of spider's web properly adjusted in the focus of the eye-piece. The sub-stage mirrors attached to such instruments usually have a flat and a concave face. With day-light the flat face should be used ; with artificial light things should be so arranged that the mirror, used with the concave face, gives as nearly parallel rays as possible. A double nose-piece, to carry two objectives, is very useful, although it usually gives very imperfect centring for high powers. The most useful objectives are a 1 inch or 1^ inch and a ^ inch, but for many purposes a |- inch is also very de- sirable. For minute objects, such as the 'crystallites' in glassy rocks and the fluid-pores in crystals, a high power is indispens- able, and for very fine-textured sedimentary rocks an immer- sion-lens offers great advantages. A. selenite-plate, a quartz-wedge, and other special pieces of 1 For a brief historical sketch of the application of the microscope to petrology see G. H. Williams' pamphlet Modern Petrography (Monographs on Education), Boston, 1886. 2 FORM OF SECTION : MEASUREMENT OF ANGLES. apparatus will be of use for various purposes. The methods involving their use may be found in the mineralogical text- books ; where too the student will find guidance as to the examination of crystal-slices by convergent light. Form of section of a crystal and cleavage- traces. A well-formed crystal gives in a thin slice a poly- gonal section, the nature of which depends not only upon the forms present on the crystal, but also on the direction of the section and on its position in the crystal, as, e.g., whether it cuts through the centre or only truncates an edge or corner. A cube cut parallel to one pair of faces gives a square, by merely truncating one solid angle of the crystal we get a triangle, and a parallel section through the centre gives a hexagon. Again, the same shape of section may be obtained from very different crystals. Thus we may get a regular hexagon not only from a cube, as remarked, but from an octa- hedron or a rhombic dodecahedron cut in the same crystallo- graphic direction, or again from a hexagonal prism or pyramid cut perpendicularly to its vertical axis. Nevertheless, if several crystals of one mineral are present in a rock-slice, we can by comparison of the several polygonal sections obtain a good idea of the kind of crystal which they represent. Further, if by optical or other means we can determine approximately the crystallographic direction in which a particular crystal is cut, we can usually ascertain what faces are represented by the several sides of the polygon. For this purpose we may require to measure the angle at which two sides meet, and this is easily done with a microscope provided with a graduated circle. Bring the angle to the intersection of the cross-wires, adjust one of the two sides to coincide with one of the cross-wires, and read the figure at the index of the circle. Then rotate until the other side is brought to coincide with the same cross-wire, and read the new figure. The angle turned through is the angle between the two sides of the section. This angle is the same as that between the corresponding faces of the crystal only provided the plane of section cuts these two faces perpendicularly. For a section nearly perpen- dicular to the two faces, however, the error will not be great. CLEAVAGE-TRACES. 3 In consequence of the mechanical forces which affect rock- masses, and also as a result of the process of grinding rock- slices, the minerals often become more or less fractured or even shattered. In a strictly homogeneous substance the resulting cracks are irregular, but if there be directions of minimum cohesion in crystals (cleavage), the cracks will tend to follow such directions, and will appear in a thin slice as fine parallel lines representing the traces of the cleavage-planes on the plane of section. The regularity and continuity of the cracks give an indication of the degree of perfection of the cleavage- structure, but it must also be borne in mind that a cleavage making only a small angle with the plane of section will, as a rule, not be shewn in a slice. In the case of a mineral like augite or hornblende, with two directions of perfect cleavage, the angle which the two sets of planes make with one another is, of course, a specific character of the mineral, or at least characteristic of a group of minerals, such as the pyroxenes or the amphiboles. In a slice perpendicular to both the cleavages the traces will shew the true angle ; for any other direction of section the angle between the cleavage-traces will be different, but it will not vary greatly for slices nearly perpendicular to both the cleav- ages, and will often suffice for discrimination, as for instance between the 87 of the pyroxenes and the 55^ of the amphi- boles. In a slice parallel to the intersection of the two clea- vages the two sets of cleavage-traces reduce to one, and a slice of a mineral such as augite or hornblende which exhibits but one set of cleavage-traces may be assumed to be nearly parallel to the intersection of the cleavages. It may be remarked that minerals having two good cleavages tend to develope in such a way that their greatest elongation is parallel to the intersection of the two cleavages, as in the minerals named. In some cases this is only true of the smaller crystals, e.g. in the felspars. Similarly a mineral like the micas, with one strongly marked cleavage, usually presents a tabular habit with broad faces parallel to the cleavage-direction. A mineral not possessing any good cleavage often shews irregular cracks in rock-slices (e.g. quartz and usually olivine). This is especially the case in brittle minerals. 4 ABSORPTION-COLOURS. Sometimes the cracks, though not regular enough to indicate a good cleavage, may have a tendency to follow a particular direction, as in the ' cross-jointing ' of apatite and the ' trans- verse fissures ' of tourmaline. Again, a system of ' gliding- planes ' may closely imitate the effect of a true cleavage, as in the transverse ' parting ' of cyanite, of some augites, etc. Transparency, colours, and refractive indices of minerals. Only a few rock-forming minerals remain opaque even in the thinnest slices : such are graphite, magnetite, pyrites, and pyrrhotite ; usually haematite, ilmenite, limonite. and kaolin ; sometimes chromite or picotite. These should always be examined in reflected light ; the lustre and colour, combined with the forms of the sections and sometimes the evidence of cleavage, will usually suffice to identify any of these minerals. The great majority of rock-forming minerals become transparent in thin slices. Those which seen in hand-specimens of rocks appear opaque, are often strongly coloured in slices, while those which in hand-specimens shew colours, are frequently colourless in thin slices. In the case of many minerals these ' absorption-tints ' are thoroughly charac- teristic, but still more so are the differences of colour (pleo- chroism) in one and the same crystal according to the direction of the slice and the direction of vibration of a polarized beam traversing it, as noticed below. The colours ascribed to minerals in the following pages and the epithet colourless apply to thin slices of the minerals. Apart from colour, the aspect of a mineral as seen in thin slices by natural light varies greatly according to its refractive index 1 , and it is of great importance for the student to learn to appreciate at a glance the effects due to a high or a low refractive index. If a thin slice of a single crystal be mounted by itself in some medium of the same colour and refractive index as 1 By this must be understood its mean refractive index. A crystal of any system other than the regular has in any section two refractive indices, the magnitudes of which depend further upon the direction of the section ; but these differences in any one mineral are usually small as compared with the differences between the mean indices in different minerals. REFRACTIVE INDEX. 5 the crystal, its boundaries and surface-characters will be in- visible, while its internal structure may be studied to the best advantage. Quartz mounted in Canada balsam (both colour- less and of very nearly the same refractive index) is almost invisible. If olivine, a colourless mineral of much higher refractive index, be mounted in balsam, its boundaries and the slight roughness of its polished surface will be very apparent. In ordinary rock-slices, mounted in balsam, a roughened or ' shagreened ' appearance may be taken as the mark of a mineral having a refractive index considerably higher than that of the medium used. Again, a highly refringent mineral surrounded in the slice by others less highly refringent is seen to be more strongly FIG. 1. VARIOUS MICROSCOPIC INCLUSIONS, HIGHLY MAGNIFIED. a. Gas-pores; in obsidian. b. Fluid-pores with bubbles; in quartz. c. Fluid-pore with bubble and cube of salt ; in quartz. d. Fluid-cavity in form of negative crystal, containing two fluids and bubble ; in quartz. e. Fluid-cavities in form of negative crystals, with bubbles; in quartz. /. Glass-inclusions in form of negative crystals, with bubbles ; in quartz, g. Schiller-inclusions consisting of three sets of flat negative crystals filled with opaque iron-oxide ; in felspar. h. Schiller-inclusions consisting of negative crystals partly occupied by a dendritic growth of iron-oxide ; in olivine. k. Zircon- crystal enclosed in quartz and itself enclosing an apatite-needle. 6 TABLE OF REFRACTIVE INDICES. illuminated than these, and this brightness is made more conspicuous by a dark boundary which is deeper in proportion to the difference in refractive index between the mineral in question and its surroundings. For these reasons a highly refringent crystal seems to stand out in relief against the rest of the slice (fig. Ik). Such considerations must be borne in mind in examining the minute inclusions in which many crystals abound. These inclusions may be of gas, of liquid (usually with a gaseous bubble), of glass, or a crystal of some other mineral, and these may be distinguished by observing that the depth of the dark border depends upon the difference in refractive index between the enclosing and the enclosed substance 1 (fig. 1). The most strongly marked border is seen when a gaseous is enclosed by a solid substance (a). A liquid-inclusion in a crystal has a less marked boundary, but a bubble of vapour in the liquid is strongly accentuated (b e). A glass-inclusion is still less strongly marked off from its enclosing crystal, while a gas- bubble contained in it shows a very deep black border (f). The refractive indices of the several rock-forming minerals may be found in the tables or books of reference, but the student will find it useful to carry in his mind such a list as that given below. Refractive 'indices of the common rock-forming minerals. Very low (1-43 1-51): tridymite, sodalite, most zeolites, (volcanic glasses), leucite. Low (1*52 1'63): felspars, nepheline, quartz, (Canada balsam), micas, calcite, dolomite, wollastonite, actinolite. Moderate (1*63 1'645): apatite, tourmaline, andalusite, horn- blende. Hiyh (1*68 1*8): olivine, sillimanite, pyroxenes, zoisite. idocrase, epidote, garnets. Very high (1*9 1'95) : sphene, zircon. Extremely high (2'0 2 -7) : chromite, rutile. Extinction between crossed nicols. When the polarizing and analysing Nicol's prisms are used together, with 1 For photographs of different kinds of inclusions, see Cohen, PI. IV. VIII. AXES OF EXTINCTION. 7 their planes of vibration at right angles to one another ('crossed nicols') 1 , if no object be interposed, there is total darkness ('extinction'), and the same is the case when a slice of any vitreous substance, such as obsidian, is placed on the stage. If, however, a slice of a crystal of any system other than the regular is interposed, there is in general more or less illumination transmitted, and often bright colours. On ro- tating the stage 2 carrying the object, it is found that extinction takes place for four positions during a complete rotation, these being at intervals of a right angle. In other words, there are two axes of extinction at right angles to one another and the slice remains dark only while these axes are parallel to the planes of vibration of the nicols, which are indicated by the cross-wires in the eye-piece. If we rotate the slice into a position of extinction and then remove the nicols, the cross- wires will mark the axes of extinction in the crystal-slice. Without attempting to deal fully with this branch of physical optics 3 , we may remark that all the optical properties of a crystal are related to three straight lines conceived as drawn within the crystal at right angles to one another (the axes of optic elasticity) and to a certain ellipsoid having these three straight lines for axes (the ellipsoid of optic elasticity}. The positions of the three axes may vary in different minerals, but they must always conform to the symmetry proper to the system, and the same is true of the relative lengths of the axes of the ellipsoid. The plane of section of any slice cuts the ellipsoid in an ellipse, the form and position of which depend upon the direction of the section (ellipse of optic elas- ticity), and the axes of extinction are the axes of this ellipse. In certain cases the ellipse of optic elasticity may be a 1 In using the two Nicol's prisms, it should always be ascertained that they are crossed. For this purpose the rotating prisms are usually provided with catches in the proper positions, but the true test is total darkness when no object is interposed. 2 In some microscopes, such as that devised by Mr A. Dick, the stape is fixed, and the two nicols rotate, retaining their relative position, an arrangement with several advantages. We shall assume for distinctness that the stage is made to rotate, as in the most usual models. 3 The student is referred for this to such a book as Kosenbusch (transl. Iddings), Microscopic Physiography of the Rock-making Minerals (1888), London. 8 STRAIGHT AND OBLIQUE EXTINCTION. circle. For this any direction is an axis, and accordingly we find that such a slice gives extinction throughout the complete rotation. In crystals of the triclinic, monoclinic, and rhombic systems there are two directions of section which give this result. They are perpendicular respectively to two straight lines in the crystal (the optic axes), which lie in the plane of two of the axes of optic elasticity, and are symmetrically disposed towards them. In crystals of the tetragonal and rhombohedral systems the two optic axes coincide with one another and with the unique crystallographic axis, and only slices perpendicular to this give total darkness. In the regular system, the ellipsoid being a sphere, the ellipse is always a circle, and all slices give total darkness between crossed nicols. Crystals of the regular system are spoken of as singly refracting or optically isotropic, and their optical properties 1 are similar to those of a glassy or colloid substance. Crystals of the other systems are doubly refracting or birefringent, and they are divided into uniaxial or biaxial according as they have one or two optic axes. It is evident that the chance of a slice cut at random from a birefringent crystal being perpendicular to an optic axis is very small. If more than one crystal of a given mineral in a rock-slice remain perfectly dark between crossed nicols through- out a rotation, it is a safe conclusion that the mineral is a singly refracting one. Straight and oblique extinction. By bearing in mind that the ellipsoid of optic elasticity, and consequently all the optical properties of a crystal, must conform to the laws of symmetry proper to the crystal -system of the mineral, we can foresee all the important points as regards the position of the axes of extinction in crystals of the different systems cut in various directions. For instance, a longitudinal section of a prism of apatite (a hexagonal mineral) will extinguish when its length is parallel to either of the cross-wires : this is straight extinction. A longitudinal section of a prism of albite (a triclinic mineral) will, on the other hand, have axes of extinction inclined at some angle to its length : this is 1 That is, such of them as we are here concerned with. MEASUREMENT OF EXTINCTION- ANGLES. 9 oblique extinction. It is to be noticed that these terms have no meaning unless it is stated or clearly understood from what direction in the crystal the obliquity is reckoned. In these examples we reckoned with reference to one of the crystallo- graphic axes defined by the traces of known crystal-faces. Another character often utilised is the cleavage. Thus in a nionoclinic mineral with prismatic cleavages, such as horn- blende, we select a crystal so cut that the two cleavages give only one set of parallel traces. These traces are then parallel to one of the crystallographic axes (the vertical axis), and we examine the position of extinction with reference to this. First we bring the cleavage-traces parallel to one of the cross-wires, removing if necessary for this purpose one or both of the nicols, and note the figure indicated on the graduated circle. Then, with crossed nicols, we rotate until the crystal becomes dark, and again note the figure. The angle through which we have turned is the extinction- angle. Observe that if a rotation through, say, 15 in one direction gives extinction, a rotation through 75 in the opposite direction would have given the same. For most purposes we do not need to distinguish between the two directions of rotation, but take merely the smaller of the two angles. To obtain a measurement of use in identifying a mineral we require more than the above. Slices of a crystal of hornblende cut in various directions along the vertical axis will give different extinction-angles, from zero (straight extinction) in a section parallel to the orthopinacoid to a maximum value in a section parallel to the clinopinacoid. This maximum extinction-angle is a character of specific value, being the angle between the vertical crystallographic axis and the nearest axis of optic elasticity. We may determine it wifjf sufficient accuracy for most purposes by noting the ex- tinction-angles in two or three vertical sections of the same mineral in a rock-slice and taking the largest value obtained 1 . By attention to the following points it is in most cases possible to refer to its crystal-system an unknown mineral of which several sections are presented in a rock-slice : 1 On the relation between this maximum extinction-angle and the extinction-angle measured in a cleavage-flake of hornblende or augite, see .V. M. (1893) x, 239, 240. 10 CRYSTALLOGRAPHIC SYSTEMS : TWINNING. Regular system : singly refracting ; all slices extinguish com- pletely between crossed nicols, as in glassy substances. Tetragonal and rhombohedral (including hexagonal) : bire- fringent and uniaxial ; straight extinction for longitudinal sections of crystals with prismatic habit and for any sections of crystals with tabular habit. The two systems cannot be distinguished from one another by optical tests, but in cross-sections of prisms the crystal outline or cleavages will usually suffice to discriminate. Rhombic (this and the remaining systems birefringent and biaxial) : straight extinction for longitudinal sections of crystals with prismatic habit ; sections perpendicular to the vertical axis have axes of extinction parallel to pinacoidal faces or cleavages and bisecting the angles between the traces of prism-faces or prismatic cleavages. A section nearly parallel to the vertical axis will give nearly straight extinction, except in minerals (e.g. olivine) which have a wide angle between the optic axes. Monoclinic : two important types may be noticed according as the intersection of the chief cleavages (and direction of elongation of the crystals) lies in or perpendicular to the plane of symmetry. In the former case longitudinal sections may give any extinction-angle from zero up to a maximum value characteristic of the species or variety : in the latter (e.g. epidote and wollastonite) longitudinal sections give straight extinction. The former case is the more frequent. Triclinic : no sections give systematically straight extinction. Twinning. The existence of twinning in a slice of a crystal is instantly revealed by an examination of the slice between crossed nicols, since the two individuals of the twin shew different interference-tints and extinguish in different positions 1 . When twin-plane and face of association coincide the most common case a slice perpendicular to the twin-plane will give in the two individuals of the twin extinction-angles which, reckoned from the line of junction, are equal but in 1 The only exceptions are in minerals, like the spinels, optically isotropic, and in cases in which the law of twinning is such that the directions of the axes of optical elasticity are not altered (e.g. quartz). EXTINCTION-ANGLES IN THE FELSPARS. 11 opposite directions. Conversely, a crystal which gives equal but opposite extinction-angles may be assumed to be cut very nearly perpendicularly to the twin-plane. If the plane of section cut the twin-plane of a crystal at a very small angle, the two individuals of the twin will overlap for a sensible width, and we shall see between the two a narrow band which does not behave optically with either. When repeated twinning occurs, as in felspars with albite- lamellation, the lamellae divide, as regards optical behaviour, into two sets arranged alternately. Extinction-angles in felspars. The discrimination of the several felspars by means of their extinction-angles measured on cleavage- flakes, as perfected by Schuster, is a method of great precision, but is not applicable to crystals in rock-slices. For these the method advocated by Michel Levy and others will often be found useful. There are two cases in which it is readily applied. (i) For crystals with albite-lamellation : Select sections cut approximately perpendicular to the lamellae. These are known by the extinction-angles in the two alternating sets of lamellae, reckoned from the twin-line, being in opposite direc- tions and nearly equal ; also by the illumination of the two sets of lamellae being not very different when the twin-line is parallel to a cross -wire. Measure the angles in question in three or four crystals so selected, and take the greatest value found. This will be very nearly the maximum angle for all such sections, which is a specific constant for each kind of felspar, as indicated below for certain types : Albite, pure, Ab Oligoclase of constitution Ab 4 An : Oligoclase ,, Ab 3 An! Oligoclase-andesine Ab 5 An 3 Labradorite, most acid type AbjAnj Labradorite, medium Ab 3 An 4 Anorthite, nearly pure An 1 5 16 27 38 53 The angles corresponding to other members of the albite- anorthite series can be interpolated by means of a curve such 12 APPLICATION OF EXTINCTION-ANGLES TO PLAGIOCLASES. as that on p. 13. The signs + and denote angles measured in opposite directions crystallographically. Unless other means of discrimination can be made use of, we have usually no way of distinguishing the two directions, and there is consequently an ambiguity between albite and the oligoclase-andesines. The other felspars have each a characteristic range of angles ; thus : to 5 oligoclase, the more acid types, 16 to 22 andesines, 27 to 45 labradorites, 45 to 50 bytownites, beyond 50 varieties near anorthite. It may be remarked further that, unless we discriminate the two directions (at right angles to one another) for which extinction occurs, there is an ambiguity between basic labra- dorite and anorthite ; for a crystal-lamella which extinguishes at, say, 40 to the right will also extinguish at 50 to the left. (ii) For microlites, assumed to have their length parallel to the intersection of the two principal cleavages: Here we measure extinction-angles from the long axis of the microlites, and select the highest angle obtained by measurements on several microlites. The following are the characteristic maxima for certain varieties of plagioclase : Albite, pure, Ab + 20 Oligoclase of constitution Ab 3 Anj Oligoclase-andesine Ab 5 An 3 7 Labradorite, most acid AbjAnj - 18 Labradorite, medium Ab 3 An 4 32 Anorthite, nearly pure, An - 64 ? The values corresponding to other members of the plagio- clase series can be roughly interpolated by means of the curve, p. 13. Here the ambiguity arising from positive and negative angles confuses albite with certain andesines and acid labradorites, while that arising from complementary angles confuses the more basic labradorites with the anorthites. The method, however, enables us to recognize at once by their low ex- tinction-angles (0 to 6) the oligoclases and oligoclase-andesines and by their high angles (beyond 20) the basic plagioclases. DIAGRAM OF EXTINCTION-ANGLES OF PLAGIOCLASES. 13 i i. 1 I J NJ <- < <) 3 . 4 " I 1 r 3- ! j 3 i 3 - ! i 1 t 3* q - S 1 J A 3 P 1 ' 5 J 1 d li 1 3 i-f - i i -- 7 -- | |j - H " r jr:r 1 I i : J *^ ? ^ 1 H | -1 "; ' -/'I M V fi 1 ^ % ' | '/ | / % \ N ^ |/ | % 4 1 M |- -| -- - -- i I - 1 | I i i i 1 ^ 1 ( 1 1 1 o y P~f -I \^ 9 t 1 1 1 ^--| - 1 1 1 1 '/^ y i ^ / '/ . ^ ^ / /^ 1 1 r-J ^ ^ ^ *Q/^ -- : r I--I H - M % % ^rX ^ / ^ ^ ^ % //. !\ Y \ , I i 1 ^ Y ^ i 1 IX i 1 1 ^ ?/ V $\ ^ a i? # $> ^ x -. ^ /y /^ $j ^ / 7 f v -'/ 1 i 1 1 7 ^ / >* : 1 1 v 'V ^\ ^s // ^ ^ 3 \ k. ^ 11 1 1 '- * \ 1 % ^ *f ^ 1 1 r 1 'r. '//. ^ ^ ^ ^ k & X^_ ^ ?f v/ 'x "^ /. > r ^ ^ "3 M i i - ^ ^ "|1 a z (D ^ ^\ _o 1 ^_ \o ^ laf ft! ^ K: ^ - I | Ml T t?J < tl\i 111 \.^ 1 r # "^ ' v" ^ 1 "" N ~ """^ i i | ; 1 ^ \ 1 1 1 1 > / ^ 1 1 \ / / ^ ^ y |" ~~1 I"-"! M 1 1 ^ \c ^ 1 7 ^ ~l x ? 1 \J 14 ZONARY BANDING IN FELSPARS. Zonary banding in felspars. In many rocks the felspars shew between crossed nicols concentric zones roughly parallel to the boundary of the crystal, the successive zones extin- guishing in different positions. (If there be albite-larnellation, we confine our attention to one of the two sets of lamella.) This difference in optical behaviour among the successive layers which build up the crystal may arise in two ways : firstly, from the successive zones being of different kinds of felspar- substance ; or, secondly, from ultra-microscopic twinning affecting in various degrees the different layers of a crystal chemically homogeneous. This has been pointed out by Michel Levy, and he gives a test which will resolve all except certain rare cases. It will be found, on rotating the slice between crossed nicols, that there are certain positions in which the zonary banding disappears. If simultaneously with this the albite-lamellation disappears also, so that the whole crystal 1 is uniformly illuminated, the appearances can be explained by ultra-microscopic twinning alone : if this is not the case, the zonary banding may be ascribed to the successive layers of felspar-substance in each crystal differing in chemical composition. When this occurs, the rule generally holds that the layers or zones become progressively more acid from the centre to the margin. Interference-tints. We have remarked that a thin slice of a doubly refracting crystal, examined between crossed nicols, is in general not dark except when placed in certain definite positions. In any other position it does not completely extinguish the light, but its effect, in conjunction with the nicols, is to partially suppress the several components of the white light in different degrees, so that in the emergent beam these components are no longer in the proportions to give white light. In this way arise polarization-tints or interference-tints. These belong to a definite scale, known as Newton's scale, on which the several tints (though graduating into one another) are distinguished by names and divided into several ' orders.' The student should learn the succession of these tints, in the first place from the coloured plates accompanying some mineral- 1 Or if there be Carlsbad twinning also, the whole of one individual of the Carlsbad twin. USE OF INTERFERENCE-TINTS. 15 ogical works 1 , but ultimately from the minerals themselves. The precise position in the scale of a given tint observed between crossed nicols can be fixed by means of a quartz- wedge or other contrivance for ' compensating ' or neutralising the birefringence of the slice ; but for ordinary purposes, at least with colourless or nearly colourless minerals, the inter- ference-tint can be judged by eye with sufficient accuracy. The most brilliant colours are those of the second order and at the top of the first ; the lowest colours of the first order are dull greys ; while in the third and fourth orders the tints become brighter but paler, ultimately approximating to white. The interference-tints given by a crystal-section depend (i) on the birefringence of the mineral, which is a specific character ; (ii) on the direction of the section relatively to the ellipsoid of optic elasticity, the tint being highest for a section parallel to the greatest and least axes of the ellipsoid ; (iii) on the thickness of the slice. These last two are disturbing factors, which must be eliminated before we can use the inter- ference-tints as an index of the birefringence of the crystal, and so as a useful criterion in identifying the mineral. The fact that the interference-tints depend in part on the direction of the section through the crystal will rarely be found to give rise to any difficulty in roughly estimating the birefringence of the mineral. If two or three crystals of the same mineral are contained in a rock-slice, it is sufficient to have regard to that one which gives the highest interference- tints. Even a single crystal will in the majority of cases give tints not so far below those proper to the mineral as to occasion error, but the possibility of the section having an unlucky direction must be borne in mind. Rock-slices prepared by a skilful operator are in most cases so nearly constant in thickness that variations in this respect may be left out of consideration. Any important difference is at once detected by well known minerals giving unusual inter- ference-tints. Thus if quartz or orthoclase give the yellow of the first order, the slice is rather a thick one ; if they give 1 Michel L6vy and Lacroix, Les Mineraux des Roches : Bosenbusch (transl. Iddings), Microscopical Physiography of the Rock-forming Minerals. 16 TABLE OF BIREFRINGENCE AND INTERFERENCE-TINTS. orange or red, the slice is considerably thicker than the average of good preparations. Knowing this, we can make allowance for it in estimating the birefringence of some doubtful mineral in the same slice. Such allowance can be roughly judged, or it can be made with considerable precision by means of the large coloured plate of Michel Levy and Lacroix. The actual birefringence (numerically expressed) of the several rock-forming minerals, and the interference-tints which they afford in slices of ordinary thickness, are given in numerous books and tables. For rough purposes the student will find it useful to remember about as much as is contained in the fol- lowing table. Birefringence and interference-tints of the commoner rock- forming minerals. (The colours given are for slices -001 inch in thickness.) Very weak (giving steel-grey tints) : leucite, apatite, nepheline. Weak (giving blue-grey to white of first order) : zoisite, micro- cliiie, orthoclase, albite, oligoclase, andesine, labradorite, quartz, bytownite, enstatite. Moderate (giving white, yellow, or orange of first order) : anda- lusite, chlorite, anorthite, hypersthene. Strong (giving red of first order to violet and blue of second): tourmaline, augite and diallage, common hornblende and actinolite. Very strong (giving green, yellow, or orange of second order) : olivine, epidote, talc, biotite, muscovite. Extremely strong (giving the pale colours of the third and fourth orders to almost pure white) : zircon, hornblende rich in iron, sphene, calcite and dolomite, rutile. Note that in minerals with strong absorption, such as the deep-coloured micas and hornblendes, the interference-colours are more or less masked by those due to absorption. Pleochroism. A character often useful in identifying minerals is pleochroism, the property of giving different ab- sorption-tints for different directions of vibration of the light within the crystal. To observe this property, we use the lower nicol only, and rotate either it or the stage. The direction of vibration is that of the shorter diagonal of the nicol. EXAMINATION OF PLEOCHROISM. 17 It is necessary not only to observe the changes of colour, if any, but also to note their relation to directions of vibration within the crystal. For example, elongated sections of biotite and hornblende, tourmaline and sphene, may be found to change from a deeper to a paler tint of brown on rotation ; but while in the first pair of minerals the direction of vibration most nearly coincident with the long axis of the section gives the deeper tone, in the second pair it gives the paler. To be more precise, we wish to know, for a specification of the pleochroism of a given mineral, the absorption-tints for vibrations in three definite directions within the crystal those of the three axes of optical elasticity. Taking a given mineral, say a hornblende, of which a number of crystals occur in our slice, we may proceed as follows. Select a crystal shewing only one set of cleavage-traces and giving the maximum extinction-angle : this section will be approximately parallel to the plane of symmetry, and will contain two of the required axes. These axes are the axes of extinction for the section, and their positions are thus easily found. The one nearest to the cleavage-traces is the y-axis, the other the a-axis. Bring the y-axis to coincide in direction with the shorter diagonal of the nicol, adjusting the position by obtaining extinction, and then removing the upper nicol. Observe the colour : then do the same for the a-axis. For the remaining /3-axis we must use another crystal. We may choose one shewing only a single set of cleavage-traces and giving straight extinction : the /3-axis is perpendicular to the cleavage-traces. Or we may choose a section shewing two sets of cleavage-traces intersecting at a good angle and extinguishing along the bisectors of the angles between the cleavage-traces : the /2-axis is the bisector of the acute angle. The results may be expressed thus in a ' scheme of pleochroism ' : a, pale straw, /3, deep brown with greenish tinge, y, deeper greenish-brown. Or we may use the ' absorption-scheme ' : y ^ ft a, signifying that the absorption parallel to y is slightly greater 18 METHOD OF STUDYING A ROCK-SLICE. than that parallel to (3, and this considerably greater than that parallel to a. Minerals of the rhombohedral and tetragonal systems can have only two distinct absorption-tints (dichroism), one for vibrations parallel to the longitudinal axis (extraordinary ray), the other for vibrations in any direction perpendicular to it (ordinary ray). Thus a particular variety of tourmaline may give E, colourless, 0, pale indigo ; or absorption-scheme 0>E. In minerals of the regular system there can be no pleochroism. In consequence of pleochroism the absorption-tints of a mineral vary in differently cut crystals seen in natural light, but the precise nature of the pleochroism can be investigated only with a polarized beam. If the pleochroism is feeble, it is best seen by rotating the nicol, not the stage. Examination of a rock-slice. In studying a rock- slice it is always well to proceed methodically. A low power should first be used : any object which it is desirable to examine under a higher magnification should be brought to the centre of the field before the objective is changed for a higher power. The slice should always be observed first in natural light : by their outline, relief, cleavages, inclusions, alteration-products, etc., all the ordinary rock -forming minerals can be identified in most cases without the use of polarized light. If the lower nicol is not readily movable it may be left in for many purposes, but it must be remembered that half the illumination is thus cut off, and for any but the lowest magnifying powers this is of importance. Opaque substances should always be viewed in reflected light. To examine the pleochroism of any coloured constituent, we put in the lower nicol, and rotate either it or the stage. For verifying feeble pleochroism the former plan is preferable, but the nicol must be rotated until its catch holds it before proceeding to the use of the two nicols, which will be the next act. DIFFICULTIES OF CLASSIFICATION. 19 For some purposes oblique illumination is advantageous. For instance, the extremely slender needles of apatite in certain lamprophyres and other rocks become visible only by this means. A ' spot-lens ' may be improvised by placing beneath the stage a convex lens of short focal length with its central part covered by a disc of black paper. In using a high power it will be noticed that the focus is very perceptibly different for the upper and lower surfaces of the slice. To make out the form of a body enclosed in the thickness of the slice the focus should be gradually moved, so as to bring different depths successively into view. It cannot be too strongly insisted that the identification of the component minerals of a rock is only a part of the examin- ation. The mutual relations of the minerals and their structural peculiarities must also be observed ; the order of crystallization, intergrowths, interpositions, decomposition- products, pseudomorphs, etc., as well as special rock-structures such as fluxion-phenomena, vesicles, effects of strain and fracture, etc. In short, the object of investigation should be not merely the composition of the rock, but its history. Classification and nomenclature of rocks. Petro- logy has not yet arrived at any philosophical classification of rocks 1 . Further, it is easy to see that no classification can be framed which shall possess the definiteness and precision found in some other branches of science. The mathematically exact laws of chemistry and physics which give individuality to mineral species do not help us in dealing with complex mineral aggregates, and any such fundamental principle as that of descent, which underlies classification in the organic world, has yet to be found in petrology. Rocks of different types are often connected by insensible gradations, so that any artificial classification with sharp divisional lines cannot truly correspond to nature. At present, therefore, the best arrange- ment is that which brings together as far as possible, for convenience of description, rocks which have characters in common, the characters to be first kept in view being those which depend most directly upon important genetic conditions. 1 See Science Progress (1896), iv, 469-490. 22 20 REMARKS ON NOMENCLATURE. The grouping adopted below must be regarded as one of con- venience rather than of principle. In a perfect system the nomenclature should correspond with the classification. This is of course impossible at present in petrology. Moreover great confusion has arisen in the nomenclature of rocks in consequence of the rapid growth of descriptive petrography. Many of the names still in use are older than the modern methods of investigation : they were given at a time when trivial distinctions were emphasized, while rocks essentially different were often classed together. Later writers, each in his own way, have arbitrarily extended, restricted, or changed the application of these older names, besides introducing new ones. The newer rock-names need cause no confusion, provided they are employed in a strict sense. Thus ' foyaite ' should be used for rocks like that of Foya, specimens of which are in every geological museum : to extend the name to all nepheline-bearing syenites is to intro- duce needless ambiguity. In practice perhaps the most con- venient usage is to speak of ' the Foya type,' ' the Ditro type,' etc., referring in each case to a described and well-known rock. There remain the names employed for families of rocks : some of these are old names, such as granite and syenite, which have come to have a tolerably well understood signification, not always that first attached to them ; others, such as peri- dotite, have been introduced to cover rocks not recognized as distinct families by the earlier geologists. A division of a family is often designated by prefixing the name of some characteristic mineral of that division; e.g. hornblende-granite, hypersthene-andesite, etc. These remarks apply more especially to igneous rocks, which we shall consider first. Such rocks, formed by the consolidation of molten ' magmas,' differ from one another in character, the differences depending partly on the composition of the magma in each case, partly on the conditions attending its consolidation. The composition is to some extent indicated by the essential minerals of the rock, which thus become an important, though not logically a prime, factor in any genetic classification. It is evident, however, that a mere enumeration of the minerals of a rock, without taking account of their MAIN DIVISIONS OF IGNEOUS ROCKS. 21 relative abundance, cannot give a very precise idea of the bulk-analysis ' ; while, on the other hand, it appears on examin- ation that magmas of very similar composition may, under different conditions of consolidation, give rise to widely different mineral-aggregates. Again, many rocks consist only in purt of definite minerals, the residue being of unindividualised matter or 'glass.' To diverse conditions of consolidation must be referred differences in coarseness or fineness of texture, the presence or absence of any glassy residue, the evidence of one or more than one distinct stage in the solidification, and, in general, the peculiarities in the mutual arrangement of the constituent minerals, which collectively are termed the ' structure ' of the rock. The massive igneous rocks will first be divided into three groups : abyssal or plutonic, hypabyssal, and superficial or volcanic. These names express the different geological rela- tions of the several groups as typically developed, but the divisions themselves are based upon the characteristic struct- ural features which different conditions of consolidation have impressed upon the rocks. Under each of these three heads the various rock-types will be grouped in families founded proximately on the mineralogical, ultimately on the chemical, composition, though this cannot be done without some few inconsistencies. The families will be arranged roughly in order from the more acid to the more basic, but it must be remembered that such an arrangement in linear series can represent only very imperfectly the manifold diversity met with among igneous rocks. 1 This difficulty is only partially evaded by ranking some of the con- stituent minerals as essential and others as accessory. A. PLUTONIC ROCKS. THE rock-types to be treated under the head of plutonic or abyssal are met with, in general, in large rock-masses which have evidently consolidated at considerable depths within the earth's crust. Transgressive as regards their actual upper boundary, their geological relations on a large scale are, as a rule, only imperfectly revealed by erosion; so that their actual form and extent are often matters of conjecture. Some of the masses seem to be of the nature of great laccolites; others have been supposed to mark reservoirs of molten magma, which once furnished the material of minor intrusions and surface volcanic ejectamenta. The immediate apophyses of the large masses have similar petrographical characters. The distinctive features of these rocks of deep-seated con- solidation are those which point to slow cooling (not necessarily slow consolidation) and great pressure. The rocks are without exception holocrystalline, i.e. they consist wholly of crystallized minerals with no glass. Even as microscopic inclusions in the crystals, glass is much less characteristic than water, which gives evidence of high pressure during the crystallization. The texture of plutonic rocks may be comparatively coarse, i.e. the individual crystals of the essential minerals may attain con- siderable dimensions. The typical structure is that known as hypidiomorphic, only a minor proportion of the crystals being ' idiomorphic ' (i.e. developing their external forms freely), while the majority, owing to mutual interference, are more or less ' allotriomorphic ' (taking their shape from their sur- roundings) 1 . 1 This is the terminology used by Eosenbusch. Zirkel has adopted Rohrbach's terms automorphic and xenomorphic in the same senses. SEQUENCE OF CRYSTALLIZATION. 23 Sequence of crystallization. The terms just intro- duced are used with a relative signification ; so that a given mineral in a rock may be allotriomorphic towards certain associated minerals and idiomorphic towards others. By observing such points we are able to make out the order in which the several minerals composing an igneous rock have crystallized out from the parent rock-magma. It is found that there exists in plutonic rocks a normal order of con- solidation for the several constituents, which holds good with a high degree of generality. It is in the main, as pointed out by Rosenbusch, a law of 'decreasing basicity.' The order is briefly as follows. I. Minor accessories (apatite, zircon, sphene, garnet, etc.) and iron-ores. II. Ferro-inagnesian minerals : olivine, rhombic pyroxenes, augite, eegirine, hornblende, biotite, muscovite. III. Felspathic minerals : plagioclase felspars (in order from anorthite to albite), orthoclase (and anorthoclase). IV. Quartz, and finally microcline. In most rocks such minerals as are present follow the above order. The most important exceptions are the inter- growth of orthoclase and quartz and the crystallization of quartz in advance of orthoclase in some acid rocks, and the rather variable relations between groups II. and III. in some more basic rocks. The order laid down applies in general to parallel intergrowths of allied minerals : thus when augite is intergrown with segirine or hornblende, the former mineral forms the kernel of the complex crystal and the latter the outer shell ; when a plagioclase crystal consists of successive layers of different compositions, the layers become progressively more acid from the centre to the margin. Certain constituents having variable relations are omitted from the foregoing list. Thus nepheline (elteolite) and sodalite belong to group III., but may crystallize out either before or after the felspars. Varieties of structure in plutonic rocks. The typical structure of rocks of plutonic habit is that implied in the foregoing remarks, and is known as the granitoid or 24 STRUCTURES OF PLUTONIC ROCKS. eugranitic structure. Among the more special modifications frequently met with are those depending upon the simul- taneous crystallization of two of the essential minerals, giving rise to the so-called 'graphic' intergrowths, usually on a micro- scopic scale. The resulting micrographic, micropegmatitic, or granophyric structure is most common in the quartz-bearing rocks, and arises there from an intimate interpenetration of part of the felspar by quartz (fig. 5 A ). Within a certain area of a slice the quartz of such an intergrowth behaves optically as if it were a single crystal, the whole becoming dark between crossed nicols in one position. On rotation the felspar can be made to extinguish in its turn. Intergrowths of other minerals (e.g. augite and felspar) are less common. In both granitoid and micrographic rocks there sometimes occur vacant interstitial spaces or little cavities of irregular shape, into which project the sharp angles of well-formed crystals. Such rocks are said to have a miarolitic or drusy structure, but this peculiarity is often obscured by secondary products occupying the druses. Opposed to the granitoid is the granulitic structure. In this a section of the rock appears as a mosaic of roughly equidimensional grains, usually of small size. There is only in some cases a tendency to crystallographic de- velopment (panidiomorphic structure) or again the earlier- formed minerals tend to take on rounded outlines. The structure probably results from movement during the process of consolidation, and we shall see that very similar appearances may be produced by the deformation and crushing of already solidified granitoid rock-masses. Both granitoid and granulitic rocks sometimes exhibit in greater or less degree a parallel disposition of elongated or tabular crystals of felspar, mica, etc., indicative of some flowing movement of the rock-magma subsequently to the separation of those crystals. With this there may be a certain banding of the rock due to alternations of slightly different types (mineralogically or structurally), which is known as a gneissic structure. These characters, however, may also have a quite different and secondary origin, and we shall defer notice of them to another place (Chap. XXII.). SPECIAL MODIFICATIONS OF PLUTONIC ROCKS. 25 Traversing plutonic rock-masses of normal structural types, or bordering them as an irregular fringe, may often be found strikingly coarse-textured or pegmatitic modifications with a strong tendency to graphic intergrowths 1 . While clearly related to the associated plutonic rock-masses, these pegmatitic rocks differ from them mineralogically in the sense of being somewhat more acid, and they are further chai-acterized by the frequent occurrence of special minerals, often including compounds of the rarer chemical elements. They are usually regarded as representing the final (pneumatolytic) phase of consolidation of the rock-magmas from which they were formed 2 . The lighter-coloured veins and streaks often seen traversing plutonic rocks are in many respects comparable with the pegmatites. They invariably shew a coarser texture and a more acid composition than the main mass in which they occur ; and, though they more or less clearly cut the latter, the relations are such as to prove that their origin is bound up with that of the main rock-mass. They are sometimes spoken of as (relatively) acid excretions from the crystallizing magma. Contrasted with these, there occur in many plutonic rocks darker and finer-textured ovoid or irregularly rounded patches which are usually considered as (relatively) basic secretions from the magma, belonging to an early stage in the history of consolidation. Composed in general of the same minerals as the enclosing rock, they are richer in the earlier-formed which are also the denser and more basic constituents. The lighter coloured veins, on the other hand, are relatively rich in the later-formed and more acid minerals. The typical plutonic rocks are non-porphyritic, i.e. there is evidence of but one continuous stage in the consolidation. In many intrusive and almost all volcanic rocks, some one, or more, constituent (usually a felspar) occurs in two distinct generations with different habits and characters, belonging 1 The original pegmatite of Hauy was such an intergrowth of quartz and felspar ('graphic granite'), but the modern usage of the name is more extended. 2 On this point see G. H. Williams, loth Ann. Rep. U. S. Geol. Sur. (1895), 675-684. 26 PORPHYRITIC STRUCTURE. to an earlier and a later stage of consolidation, in which quite different conditions prevailed. This is the ' porphyritic ' structure, and is typically wanting among plutonic rocks, which have what has been termed an ' even-grained ' character (' kb'rnig ' of Rosenbusch). In some of the plutonic rocks, however, and especially among the granites, occur relatively large crystals of felspar, which give a porphyritic character to the rock of which they form part, and perhaps point to different conditions from those under which the main mass of the rock consolidated; but even here there is no sharp division between an earlier and a later period of crystal- lization, such as is indicated in the volcanic rocks'. We shall consider the several families in an order which corresponds roughly with their chemical relationship, beginning with the acid rocks and ending with the ultrabasic. 1 Cf. Lawson on the Santa Lucia granite in California, B^dl. Dep. Geol. Univ. Gal. (1893), i. 915. CHAPTER II. GRANITES. THE granites are even-grained holocrystalline rocks com- posed of one or more alkali-felspars, quartz, and some ferro- magnesian mineral, besides accessory constituents. The rocks are generally of medium to rather coarse grain, and the tendency of the crystals as a whole to interfere with one another's free development gives what Rosenbusch styles the hypidiomorphic structure. According to their characteristic minerals, after felspars and quartz, the rocks are described as muscovite-, biotite-, hornblende-, and augite-granltes ; and this division corresponds roughly to different chemical compositions, from more to less acid types. Tourmaline-granite must be considered a special modification of the above, and, in particular, of the more acid kinds. With the granites we shall also include certain rocks (aplite, pegmatite, greisen) associated with granites but differ- ing from them in important structural and mineralogical characters, some of them never forming, like the true granites, large bodies of rock. Constituent minerals. Felspars make up the greater part of a granite, a potash- and a soda-bearing felspar commonly occurring together. The potash-felspar is often /orthoclase, either in simple crystals or in Carlsbad twins, the Baveno twin being uncommon 1 . When fresh, it shews its cleavages and sometimes a slight zonary banding, but these appearances are lost when the mineral is altered to any extent. The 1 Cohen, PL xxvm, fig. 2. 28 FELSPARS OF GRANITE. common decomposition- processes give rise either to finely divided kaolin or to minute flakes of mica. When the latter are large enough to be clearly distinguished, they are often seen to lie along the cleavage-planes of the felspar. Decom- position often begins in the interior of a crystal, which may be clouded or completely obscured while the margin remains clear. Instead of orthoclase we often find microcline, which is usually the last product of consolidation in the rock. When fresh, microcline shews its characteristic 'cross-hatched' struct- ure 1 and sometimes a vein-like intergrowth of albite 2 (fig. 2). a, FlG. 2. MlCBOCLINE FROM THE ' BAPAKIWI ' GRANITE OF FINLAND ; X 20. crossed nicols: shewing the characteristic 'cross-hatching.' It is traversed by veinlets of albite (a) intergrown with crystallographic relation to the microcline [1031]. Some petrologists hold that the peculiar microcline-structure, due to fine twin-lam ell ation in two directions, is not essential, and may be set up in some cases as a secondary effect of strain ; and that the quasi-monoclinic mineral orthoclase is merely microcline in which the twiii-lamellation is carried to an ultra-microscopic degree of fineness 3 . The alteration of 1 Cohen, PL xxxi, figs. 1, 2. 2 Rosenbusch-Iddings, PL xxv, fig. 1. 3 Cf. Teall, Ann. Rep. Geol. Sur. for 1895, p. 24; Keyes, 15th Ann. Rep. U. S. Geol. Sur. (1895), 711, 712. QUARTZ OF GRANITE. 29 microcline by weathering is similar to that of orthoclase. The soda-felspar of most granites ranges from nlbite to oligoclase. It has rather a tabular habit, giving rise to elongated rectangular sections. It is always twinned on the albite- and occasionally too on the pericline-law. The com- mon decomposition-products are kaolin, sometimes paragonite mica, and in the lime-bearing varieties some epidote 1 or calcite. Parallel intergrowths of orthoclase and plagioclase are sometimes found (microperthite). The felspars of granite are not rich in inclusions, but they may inclose sparingly microlites of the earlier constituents of the rock. The quartz of granites does not usually shew any crystal boundaries, except on the walls of drusy cavities (' miarolitic' structure), or less perfectly when the mineral is enclosed by microcline. Its most characteristic inclusions are fluid-cavities (fig. 1, b e); these are sometimes in the form of 'negative crystals,' either dehexahedral pyramids or elongated prisms ; more usually the shape is rounded or irregular. These fluid- pores often occur with a definite arrangement along certain planes, appearing in a section as lines 8 . The enclosed liquid does not fill the cavity, but leaves a bubble, which is mobile. In some cases the liquid is brine, and contains minute cubes of rock-salt (Dartmoor). In others liquid carbonic acid occurs instead of, or in addition to, water, and in some cases we see one bubble within another 3 . Glass- and stone-cavities are less abundant. Sometimes extremely fine needles are enclosed (Peterhead) : these seem to be rutile, and sometimes shew the characteristic knee-shaped twin. The dark micas pf granites are usually termed biotite. This may be considered to include varieties rich in ferrous oxide (the haughtonite of many Scottish and Irish granites), or in ferric oxide (lepidornelane). The mineral builds roughly hexagonal plates, which, cut across, give an elongated section shewing the strong basal cleavage. A lamellar twinning 1 On the epidotization of granitic rocks see Grimsley, Granites of Cecil Co., N. E. Md., Journ. Cincinnati Soc. Nat. Hist. 1894 ; Keyes, A. J. S. (1895), xv, 39-46; Reusch on the thulite-rock of Trondjhem, M. M. (1892), x. 40 (Abstr.) 2 Cohen, PI. vi, fig. 1. 3 Ibid. PI. vn, fig. 3. 30 MICAS OF GRANITE. parallel to the base is probably common, but, owing to the nearly straight extinction, this is not often conspicuous. The fresh biotite is deep brown with intense pleochroism. Its common inclusions are apatite, zircon, and magnetite, and the minute zircons are always surrounded by a 'halo' of extremely deep colour and intense pleochroism 1 (Skiddaw, Dartmoor, Dublin, etc.). Decomposition often produces a green colora- tion and ultimately a green chloritic pseudomorph with secondary magnetite- dust. This magnetite may be reabsorbed, restoring the brown colour but with less pleochroism and with loss of cleavage. The colourless, brilliantly-polarizing muscovite forms rather ragged flakes, posterior to the biotite or partly in parallel intergrowth with it (Dublin, etc., fig. 3, ). It is always clear, ^>^ B FIG. x20. GRANITE, NEAR DUBLIN ; A. Crystal of oligoclase shewing zonary structure and decomposition beginning in the interior [389]. B. Parallel intergrowth of biotite and muscovite [1774]. and is not susceptible to weathering. A lithia-mica, in large flakes, takes the place of muscovite in some greisens and peg- matites. 1 Cohen, PI. xxxvi, figs. 3, 4. MINERALS OF GRANITE. 31 The crystals of hornblende are irregularly bounded, or at least without terminal planes. They shew the prismatic cleavage, and occasionally lamellar twinning parallel to the orthopinacoid. The colour is green or brownish-green, with marked pleochroism, and the extinction-angle in longitudinal sections always low. Besides inclusions of earlier minerals, there may be an intergrowth with biotite. The common decomposition-products are a green chloritic substance or an epidote and quartz. When augite occurs, it is commonly the variety malacolite or salite, colourless in slices. It is not usually in perfect crystals, but an idiomorphic green augite is found in some coarsely granophyric types of rock (Mull). Augite may be either uralitized or decomposed into a green chloritoid product or serpentine arid calcite. The mineral is sometimes accom- panied by enstatite (Cheviot). Iron-ores are not plentiful in granites. Magnetite may occur or hcematite, either opaque or deep-red; pyrites is also found as an original mineral. Acute-angled crystals of light-brown pleochroic sphene are often seen, and in the less acid granites are abundant (fig. 5, B). Rounded grains may occur instead. The high refractive index and other optical properties enable the mineral to be readily identified. The little prisms of zircon are even more highly refractive (fig. 1, k), but when they occur, as they often do, enclosed in the biotite, the pleochroic halo is liable to obscure their nature. Apatite builds narrow colourless prisms, and often penetrates the biotite. Small reddish garnets occur in some muscovite-granites and aplites (Dublin) : other unusual minerals are cordierite, usually pseudomorphed by the micaceous substance termed pinite, and andalusite, coated with flakes of muscovite. In some granites from America and else- where allanite (orthite) is found 1 , while others contain epidote, often with an intergrown core of allanite 2 . Though epidote is a well-known weathering-product in granitic rocks, this relation to allanite and the occasional inclusion of good crystals 1 Iddings and Cross, A. J. S. (1885) xxx, 108-111. 2 Hobbs, ibid. (1889), xxxviii, 223-228; Joh. Hopk. Univ. Circ. No. 65 (1888) : Grimsley, Journ. Cincinnati Soc. Nat. Hist. 1894. 32 STRUCTURE OF GRANITE. of epidote in flakes of biotite seem to point to its primary origin in these Tourmaline characterizes a common modification of granite, especially near the margin of a mass. It may be in good crystals but has more frequently ragged outlines. The rude cross-fracture is often apparent. The colour is brown, some- times with patches of blue, and the dichroism is strong, the strongest absorption being for vibrations transverse to the long axis (the 'ordinary' ray). Structure. In the granites the normal order of crystal- lization of the constituent minerals rules with few exceptions. The minor accessory minerals crystallized out first, and are FIG. 4. BEVERSALS OF NORMAL ORDER OF CRYSTALLIZATION IN GRANITES ; X 20. A. Biotite moulded on muscovite, Rubislaw, Aberdeen [390 a]. B. Biotite moulded on quartz and felspars, Meillionydd, near Sarn j Caernarvonshire [814]. C. Orthoclase moulded on quartz, Shap [892]. thoroughly idiomorphic, i.e. have taken their shape without external interference. The ferro-magnesian minerals have in 1 Hobbs, Amer. Geol. (1893), xii, 218, 219; Keyes, Bull. Geol. Soc. Amer. (1893), vi, 305-312 ; Uth Ann. Rep. U.S. Geol. Sur. (1895), 704- 710, PI. xxxvm, figs. 1-4, xxxix, figs. 1-3, XL. MUSCOVITE-GRANITES. 33 general preceded the felspars, being often embraced or even enclosed by them, though the felspars may tend also to take on partial crystal-outlines. Rarely does, e.g., mica occur interstitially to felspar (fig. 4, B). Biotite moulded on muscovite 1 is not so rare (fig. 4, A). Apart from micro- graphic structures, the felspars, except microcline, have crystallized prior to the quartz, exceptions being infrequent (fig. 4, C). Where micrographic intergrowths occur, the felspar may be either orthoclase or a plagioclase (fig. 5, A). We need not further specify other structural peculiarities such as the miarolitic (Arran, Mull, Mourne Mts., etc.), the porphyritic (Dartmoor, Shap), the spheroidal (Mullaghderg in Donegal 2 ), the gneissic (Deeside), etc. Leading types. Almost all the true granites contain a brown mica. If a white mica be present in addition, we have muscovite-yranite ('two-mica granite' or 'granite proper ' of the Germans, 'granulite' of the French 3 , 'binary granite' of some American writers 4 . Such rocks are commonly somewhat more acid in composition than those with dark mica only. The Carboniferous granites of Cornwall and Devon afford good examples. They consist of orthoclase, a plagioclase, quartz, and two micas 5 , with the normal order of crystallization. The quartz has fluid-cavities, often enclosing minute cubes of rock-salt 6 (Dartmoor, fig. 1, c). Parallel intergrowths of biotite and muscovite are common. The minor constituents of the rock are magnetite, apatite, and zircon, the last when it is enclosed in the biotite being always encircled by the characteristic halo of intense pleochroism. More exceptional accessory minerals are andalusite, in pleo- 1 Keyes, 15th Ann. Rep. U. S. Geol. Sur. (1895) 703, PL xxxix, figs. 4-6 (Guilford, Md.). 2 Hatch, Q. J. G. S. (1888) xliv, 548-559, PI. xiv, with a summary of information on spheroidal granites in general. 3 The grauulite of German and English petrologists has a different signification. 4 This term, however, has also been applied to rocks consisting essen- tially of felspar and quartz, without mica. 5 Dr Haughton's analyses of the Trewavas Head rock proved the felspar to be albite, the dark mica lepidomelane, and the white mica lepidolite ; Q. J. G. S. (1869), xxv, 166, 167. 8 Hunt, G. M. 1894, 102-104, with figures. H. P. 3 34 MUSCOVITE- AND BIOTITE-GRANITES. chroic crystals coated by flakes of muscovite (Cheesewring), and 'pinite' pseudomorphs after cordierite (Land's End). Tourmaline is common, and the rocks graduate into tourma- line-granites, especially near the margin of an intrusion. The post-Ordovician granites which occupy so large a tract in Leinster 1 (e.g. Dalkey, near Dublin) are of a different type. They also have two micas, often in parallel intergrowth, and apatite and zircon are characteristic accessories ; but the potash-felspar is microcline, and is the latest product of crystal- lization. A plagioclase felspar is plentiful, and exceptionally albite is the only felspathic element present (Croghan Kinshela in Wexford). Little crystals of garnet occur in some instances (Three Rock Mountain near Dublin). This mineral is found also in the granite of Foxdale in the Isle of Man 2 , a closely similar rock, in which the dark mica is very subordinate to the white. Another well-known microcline-bearing rock is the ' grey Aberdeen granite ' of Rubislaw, etc. Similar rocks are found in Donegal. Rocks in which muscovite is only sparingly or occasionally present form a link with the next division. The Skiddaw granite is of this character 3 . Here the quartz is in great part of prior crystallization to the orthoclase, or there may be some micrographic intergrowth of the two minerals. Felspar- quartz-rocks free from mica are found among the pre- Cambrian intrusions of Ercal in the Wrekin district and of the Malverns. Here too the quartz has crystallized, or has finished crystal- lizing, before the dominant felspar, which is often microcline. These rocks seem to have affinities with the pegmatites. The commonest division of the granite family is perhaps biotite-granite (Fr. granite, Ger. Granitit), characterized by containing a brown mica to the exclusion of muscovite, horn- blende, or augite. . Such a rock may consist, e.g., of orthoclase, albite or oligoclase, quartz, biotite, and minor accessories, with the normal order of crystallization. The relative proportions of the several minerals vary 1 Sollas, Trans. Eoy. Ir. Acad. (1891) xxix, 427-512 ; Watts, Guide, 31-33. 2 Naturalist, 1894, 68 ; Q. J. G. S. (1895) li, 143. 3 Q. J. G. S. (1895) li, 140. BIOTITE-GRANITES. 35 considerably. In the granites (Ordovician and perhaps some older) of Wales 1 quartz is very abundant, and biotite (often chloritized) is only sparingly found. The dominant felspar B FIG. 5. A. Micrographic intergrowth of plagioclase felspar and quartz in granite, St David's ; x 20, crossed nicols [293]. B. Crystals of sphene (sjj) in dark basic secretion in Shap granite, Westmorland; x20 [1070]. is often a plagioclase (Caernarvon, St David's, etc.), and probably some of these rocks would be placed among the 'soda- granites ' of certain authors. The St David's rock shews a strong tendency to the micrographic structure (fig. 5, A). In the biotite-granite of Eskdale, Cumberland, the quartz is either intergrown in micrographic fashion with the ortho- clase, or has crystallized before it. The latter is the case too in the well-known porphyritic granite of Shap in Westmor- land 2 (fig. 4, (7), which is further noteworthy for its abundant sphene. Both micrographic and miarolitic structures charac- terize the probably Tertiary rocks of the Mourne Mountains 1 Q. J. G. S. (1888) xliv, 444, 445, and Bala Vole. Ser. Caern. 59, 61 (Sarn) ; Geikie, Q. J. G. S. (1883) xxxix, 314, PL x, fig. 11 (St David's) ; Jennings and Williams, ibid. (1891) xlvii, 380 (Ffestiniog). a Teall, PL xxxv, fig. 1 [395] ; Harker and Marr, Q. J. G. S. (1891) xlvii, 275-285, PL xi, fig. 1. 36 BIOTITE- AND HORNBLENDE-GRANITES. and the Carlingford district 1 , the quartz and felspars on the walls of the druses presenting very perfect crystal-boundaries. Biotite-granites are extensively developed in the Cairngorm and Monadhliath Mts. and other parts of the Scottish High- lands. In many British examples microcline partly or wholly takes the place of orthoclase (Malvern, Ross of Mull, Peter- head, etc.) Albite- veins intergrown in both orthoclase and microcline may sometimes be observed, e.g. in the Eskdale rock already alluded to. Among American examples may be cited the biotite-granites of Central Maryland 2 , in which the cross- hatching of the microcline seems to be a secondary strain- phenomenon. Biotite-granites rich in microcline are described from Alabama 3 . The rock of Pike's Peak, Colo. 4 , is rudely porphyritic through the development of large crystals of microcline. Here oligoclase and orthoclase occur subordinate to the dominant felspar, allotriomorphic quartz is abundant, and biotite only sparingly present in aggregates of small flakes. Less abundant than the types characterized by micas, and usually of less acid composition, is hornblende-granite (Ger. Amphibolgranit), in which the distinctive mineral is a green hornblende, usually with biotite in addition. Some of the newer Palaeozoic granites of Scotland are of. this kind, such as that of Lairg 5 and Ord Hill 6 in Sutherland and the Criffel rock at Dalbeattie 7 , in which, however, biotite is predominant. The rock quarried at Mount Sorrel in Charnwood Forest, Leicestershire 8 , is also in part a hornblende-granite, having that mineral associated with biotite. In Ireland a hornblende- granite has been described from Donegal 9 , and another is 1 Sollas, Trans. Eoy. Ir. Acad. (1894) xxx, 490. 2 Keyes, 15th Ann. Rep. U. S. Geol. Sur. (1895) 696-730. Some of these rocks (Ilchester, Ellicott County, etc.) contain original epidote and allanite: cf. Hobbs, A. J. S. (1889) xxxiii, 223-228. 3 Clements, Bull. No. 5 Geol. Sur. Ala. (1896) 139-142; Brooks, ibid. 185, 186. 4 E. B. Matthews, 16tli Ann Rep. U. S. Geol. Sur. (1895) Pt II. p. 22. 5 Cole's Stud. Micro. Sci. No. 42 (plate). 6 Ibid. No. 38 (plate). 7 Teall, Mem. Geol. Stir. Scot., Expl. of Sheet 5 (1896) 41-43. 8 Bonney. Q. J. G. S. (1878) xxxiv, 219. 9 Hatch, ibid. (1888) xliv, 548-551. HORNBLENDE- AND AUGITE-GRANITES. 37 associated with the Palaeozoic biotite-granites of Newry (at Goragh Wood). Hornblende-granites of Tertiary age are found in Skye, Mull, and Arran. In these the brownish green hornblende is associated with subordinate biotite. The rocks often shew a rude micrographic structure and graduate into typical granophyres, in which the biotite, and to some extent the hornblende, give place to a greenish augite. A miarolitic structure is common, the cavities being usually occupied by secondary calcite. Hornblende-granites, often rich in sphene are largely developed in Nevada and Utah 1 . Other American localities are the Shipton Range, Canada; Sauk Rapids, Minn.; Quincy, Mass. The Albany granite, in New Hampshire, carries porphyritic crystals of orthoclase with perthitic intergrowths of albite 2 . This rock has varying proportions of biotite and hornblende, and zircon is a conspicuous accessory mineral. The deep blue soda-amphibole riebeckite was first discovered in a granite from Socotra. It always forms exceedingly ragged shapeless crystals 3 . A riebeckite-granite has been described by Lacroix from St Peter's Dome, El Paso, Col. If we exclude the granophyric varieties, augite-granite is by no means an abundant rock-type. An example, of Old Red Sandstone age, occurs in the Cheviots 4 . This consists of orthoclase, plagioclase, quartz, augite, biotite, iron-ores, and apatite, the quartz and orthoclase sometimes shewing a micro- graphic intergrowth. Augite-granites with anorthoclase as the dominant felspar ('soda-granites') are described from Minnesota 5 , New Brunswick 6 , and other parts of North America. These rocks also tend strongly to micrographic structures, and graduate into typical granophyres. Granites in which a rhombic pyroxene is the dominant ferro-magnesian mineral seem to be very rare. An enstatite- Zirkel, Micro. Petrog. Fortieth Parallel (1876) 40-52. Hawes, A.J.S. (1881) xxi, 23. See figure by Bonney, Phil. Trans. (1883) clxxiv, PL vn, fig. 2. Teall, PL xxxix, fig. 2, and G. M. 1885, 112-116. Grant, 21st Ann. Rep. Geol. Sur. Minn. (1894) and Amer. Geol. (1893) xi, 383-388. 6 Matthew, Tr. N. Y. Acad. Sci. (1895) xiv, 204-208, PL xvi, xvn. 38 APLITES. granite has, however, been described from Soggendal in Norway, and Mr Kynaston finds a pale, faintly pleochroic enstatite as a frequent associate of augite in the Cheviot granite. Some granites, rich in soda, in the Christiania district carry segirine, usually in addition to hornblende or arfvedsonite. Closely related to the granites is the rock known as aplite (granite- aplite). It occurs as veins in granite, but cutting the latter and traversing adjacent rocks, and by some petrologists it would be placed in the hypabyssal division. It is a fine- textured rock with panidiomorphic to granulitic structure and is somewhat more acid than the associated granite. A charac- teristic type occurs in connection with the muscovite-granites near Dublin (Dalkey and Killiney). It consists of microclirie with some oligoclase, quartz, muscovite, and red garnet. An aplite at Meldon in Devonshire 1 is of similar Character, but instead of garnet contains topaz and some tourmaline. The Crosby dyke 2 in the Isle of Man may be referred here. It consists essentially of a granular mosaic of clear felspars, quartz, and white mica, the dominant felspar being an albite. Besides the abundant small flakes of white mica, some larger hexagonal plates occur, and sometimes scattered quartz- grains or larger felspars. There are also a few garnets of very irregular shapes, giving a sponge-like appearance in section. Aplites are described from various American localities. Turner has noted numerous dykes cutting the coarse horn- blende-granite of the Sierra Nevada. These are of fine grain, and contain as a rule very little mica or hornblende. A rock from the district south of Mariposa is a soda-aplite. Pirsson observed aplites on Coanicut Is., R.I. Many of the rocks termed granulites by German writers doubtless belong here. They will be noticed in a later chapter (Chap. XXII). The pegmatites belonging to this family of rocks (granite- pegmatite) consist essentially of microcline or orthoclase and quartz, often with white mica and sometimes red garnet. The 1 Teall, p. 316. 2 Hobson, Q. J. G. S. (1891) xlvii, 440. PEGMATITES ; TOURMALINE-GRANITES. 39 texture is often extremely coarse, and there is a frequent tendency to the graphic structure. Such rocks are extensively developed in connection with the Archaean gneiss of Suther- land. Others occur in Forf arshire ' : these are rich in muscovite, and locally carry garnet or tourmaline. It may be observed that these British pegmatites are not rich in rare or special minerals. In the United States, on the other hand, many of the most noted mineral-localities are furnished by pegmatites of this kind; e.g. Stoneham and Hebron in Maine, Chesterfield in Massachusetts, Haddam in Connecticut, Pike's Peak in Colorado, and Harney's Peak in the Black Hills of Dakota. Central Maryland is another district 2 . Pegmatitic and aplitic dykes, both carrying red garnet, occur in the Montara granite of San Francisco 3 , and such dykes, with only a small quantity of mica, are associated with the Santa Lucia granite near Monterey 4 . The tourmaline-granites appear as modifications of more normal granitic rocks 5 . The tourmaline seems to take the place of the mica. As a further modification, the felspars may be replaced partly or wholly by tourmaline and quartz, the former sometimes occurring in little needles with radiate group- ing imbedded in clear quartz. The extreme modification is a tourmaline-quartz-rock or schorl-rock, in which felspar is wholly wanting, while tourmaline may occur in two or more habits, as crystals or grains and as groups of needles. All these types are illustrated among the Cornish and Dartmoor granites. A curious variety known as luxullianite has been described by Prof. Bonney 6 . Here the conversion of felspars into clear quartz, crowded with radiate groups of tourmaline needles, can be traced in various stages, the little needles, about '03 inch in length, giving pale brown and light indigo colours for longi- tudinal and transverse vibrations respectively, while a brown tourmaline in distinct grains has been supposed to represent Barrow, G. M. 1892, 64; Q. J. G. S. (1893) xlix, 332-336. G. H. Williams, 15th Ann. Rep. U. S. Geol. Sur. (1895) 675-684. Lawson, ibid. 413. Lawson, Bull. I)ep. Geol. Univ. Gal. (1893) i, 16, 17. For coloured figure of a tourmaline-granite see Fouqu6 and Levy, PI. VIII, fig. 1. 6 M. M. (1877) i, 215-222. 40 GREISENS. the mica of the granite. A rock from Trowles worthy Tor shews a similar replacement of felspar (tig. 6, A), and has in addition irregular patches of isotropic fluor also enclosing needles of tourmaline 1 . The rock known as greisen (hyalomicte of French writers) consists essentially of quartz and white mica, which seems to be often a lithia-bearing variety. The Cornish greisens 2 are apparently a modification of the granite in the same sense as the tourmaline-rocks are, but with a different result. The place of the felspar is taken by mica and topaz, though tourmaline is also met with. It may be remarked that the topaz-rocks of Schneckenstein and Geyer 3 in Saxony are closely allied to greisen. Greisen is also found in connection with the granite of the Scilly Isles. In Grainsgill, Cumberland 4 , it has been formed at the expense of a pegmatitic modification of the B FIG. 6. MODIFICATIONS OF GRANITE ; x 20. A. Eeplacement of felspar by clear quartz full of tourmaline-needles, Trowlesworthy Tor, Cornwall : with remains of much-decomposed felspar [1361]. B. Greisen, Grainsgill, Cumberland: consisting of quartz and muscovite with only occasional relics of turbid felspar [1547]. " 1 Worth, Trans. Roy. Geol. Soe. Cornw. (1884) x, 177-188. a Teall, 315. 3 Salomon and His, M. M. viii, 282 (Abstr.). 4 Q. J. G. S. (1895) li, 141. BASIC SECRETIONS IN GRANITES. 41 Skiddaw granite, and the successive stages of the transforma- tion can be studied. The white mica builds sometimes rather large flakes (fig. 6, J3), sometimes aggregates of small scales, and in both cases is embraced or enclosed by a moderately coarse mosaic of clear quartz. Tourmaline is absent. A greisen very similar to this is associated with the Foxdale granite in the Isle of Man, and occurs there in the same manner as pegmatite, which is also developed. In conclusion we will note some examples of the dark, fine- grained, ovoid patches frequently enclosed in granitic rocks, and regarded as basic secretions separated out from the granite- magma at an early stage, not necessarily in situ. Mr J. A. Phillips 1 described such patches from the muscovite-granites of Gready in Cornwall and Foggen Tor on Dartmoor and the biotite-granites of Shap and Peterhead, and he distinguished them from foreign fragments caught up and metamorphosed by the magma. The characteristic of the true secretions is that they consist of the same minerals as the enveloping rock, but contain the earliest products of crystallization such as apatite, magnetite, and sphene in larger proportions, and are also richer in the ferro-magnesian relatively to the felspathic elements of the rock. Sometimes, as in the Criffel granite 2 , we may observe that hornblende is more plentiful as compared with biotite than in the normal rock, and similarly plagioclase felspar is more abundant relatively to orthoclase. The numerous dark patches in the Shap granite 3 , rich in sphene and biotite (fig. 5 7?), enclose, like the normal rock, large porphyritic crystals of orthoclase ; but these are partially rounded and corroded, the margin of each crystal being replaced by plagio- clase and quartz. Among American rocks a good illustration is afforded by the hornblende-granite of the Wahsatch Range (Little Cotton- wood Canon, Utah). Green hornblende, which in the normal rock is subordinate to biotite, becomes in the dark patches the dominant coloured mineral, while sphene is present in unusual abundance. 1 Q. J. G. S. (1880) xxxvi, 1-21 ; (1882) xxxviii, 216, 217. 2 Teall, Mem. Geol. Sur. Scot., Expl. of Sheet 5 (1896) 42. 3 Q. J. G. S. (1891) xlvii, 281, 282, PI. xi, fig. 2. CHAPTER III. SYENITES (including NEPHELINE-SYENITES). THE syenites are even-grained, holocrystalline rocks con- sisting essentially of alkali-felspars, and in one group fel- spathoid minerals, with ferro-magnesian constituents, typically in smaller proportion, and various minor accessories. The texture is often rather coarse to medium-grained, and the structure is that characteristic of plutonic rocks, the several minerals following the normal order of crystallization, and most of them having only imperfect crystal outlines (hypidiomorphic structure of Rosenbusch). In many syenites, however, the order of crystallization is modified by simultaneous irrtergrowths of different minerals. This family of rocks is less widely distributed and less abundant than the granites. Considered from a chemical point of view, it is characterized by an unusually high per- centage of alkalies. In the syenites which depart farthest in this respect from the commoner types of igneous rocks, the character shews itself in the presence of felspathoid con- stituents and soda-bearing ferro-magnesian minerals. The type characterized by hornblende and alkali-felspars is known as ' syenite proper ' 1 , or, for clearness, hornblende- syenite. When biotite more or less completely takes the place of hornblende, we have mica-syenite ; and when augite occurs prominently, often in company with one or both of the other 1 The original syenite of Werner was the hornblende-granite of Syene or Assouan on the Nile. The name, however, has come to be universally applied to the family under notice, rocks often hornblendic but typically free from quartz. MINERALS OF SYENITE. 43 coloured minerals, augite-syenite. The group characterized by the occurrence of nepheline or sodalite in addition to felspar is named nepheline-syenite, or often elseolite-syenite, without dis- tinction according to the dominant ferro-magnesian constituent, though several types, mostly of restricted occurrence, have re- ceived special names. A leucite-syenite is known only in the form of rocks with pseudomorphs of orthoclase, elseolite, musco- vite, etc., in the shape of leucite. The occurrence of subordinate quartz in some syenites gives rise to the varieties quartz-syenite, quartz-mica-syenite, and quartz-augite -syenite, but free silica never occurs in the nepheline-bearing group. On the other hand the coming in of a lime-soda-felspar as a prominent constituent in addition to the alkali-felspar gives rise to types intermediate between true syenites and diorites, and to these the name monzonite is sometimes given. Constituent Minerals! In mode of occurrence, in- clusions, alteration-products, etc., the felspars of syenites resemble those of granites. Besides orthoclase, microcline, and albite or oligoclase, there occur, especially in the augite- and nepheline-syenites, felspars rich in both potash and soda, known as soda-orthoclase, soda-microcline, anorthoclase, etc. These are regarded by some mineralogists as intergrowths on an ultra-microscopic scale of a potash- and a soda-felspar (cryptoperthite). An evident parallel intergrowth of albite and microcline or albite and orthoclase (microperthite) is also frequent in the same rocks. When nepheline occurs, it is of the variety known as elceolite, in larger and less perfect crystals than the nepheline of volcanic rocks. If idiomorphic, it forms hexagonal prisms with the basal plane bevelled by narrow pyramid-faces. In more shapeless crystals the straight extinction can be verified by reference to rows of inclusions which follow the direction of the vertical axis, and seem to determine the alteration of the mineral. The elseolite is colourless or often rather turbid. It gives rise by decomposition to various soda-zeolites or to moderately brightly polarizing prisms, fibres, and aggregates of cancrinite. A frequent associate of elseolite is sodalite, in dodecahedra or in allotriomorphic crystal-plates and wedges. 44 MINERALS OF SYENITE. It is colourless or faint blue in slices, and is easily recognized by its isotropic behaviour. It encloses fluid-pores, microlites of segirine, etc., and secondary products similar to those of elseolite. The common hornblende of syenites is partly idiomorphic but without terminal planes. It is of the green pleochroic variety, giving in vertical sections a maximum extinction- angle of 12 to 16. Its inclusions and alteration-products are the same as in granite. Some augite-syenites contain the socla-amphibole barkevicite with intense brown absorption and pleochroism and an extinction-angle of about 12. The auyite, when it occurs as an accessory, is colourless or very pale green with the same properties as in granite. In the augite-syenites it is sometimes pale green with faint pleochroism, sometimes pale brown to violet-brown with very distinct pleochroism. Various types of schiller- and diallage- structures are sometimes seen, and may affect only a portion usually the interior of a crystal (fig. 8). A green pleochroic ceyirine occurs in some augite-syenites and many nepheline- syenites, and intergrowths of this with augite are not uncommon. The biotite of the syenites is deep brown, becoming green only by secondary changes. In some augite- and nepheline- syenites vibrations parallel to the cleavage-traces are almost completely absorbed. The mineral is roughly idiomorphic, except when intergrown with hornblende or augite. When quartz occurs, it has the same characters as in granite, but is never very abundant. It does not occur in the nepheline-syenites and their allies. Most syenites contain plenty of sphene in good crystals shewing the cleavages and often the characteristic twinning 1 . Zircon is common in small prisms with pyramidal terminations, as in the granites. In some of the augite-syenites, however, it builds large crystals of simple pyramidal form. It is easily identified by its limpid appearance and extremely high refringence and bire- fringence. Apatite in colourless needles is widely distributed in syenites. The iron-ores are variable in quantity : they include magnetite, ilmenite, arid hwmatite, the last two often in thin flakes enclosed in the felspars. An occasional accessory is 1 See Rosenbusck-Iddings, PI. xi, fig. 3 ; xxrn, fig. 1. STRUCTURE OF SYENITE. 45 perqfskite in small octahedra, distinguished by their very high refractive index and feeble double refraction. Special types contain melanite garnet, brown in slices and always isotropic. Structure. The texture of the syenites and the mutual relations of their constituent minerals are normally similar to those observed in the granites, Rosenbusch's ' order of consolidation' being, as a rule, followed. In the typical hornblende-syenites there are few peculiarities. When quartz enters, it may be intergrown in micrographic fashion with part of the orthoclase, and this is specially the case in some augite-syenites. When plagioclase felspar is abundant, it is sometimes moulded by shapeless plates of orthoclase, and in the same rocks reversals of order between the bisilicates and the felspars may often be. noticed. Where the felspathoids occur, their place in the order of crystallization is a variable one. These minerals usually precede the felspars, but may continue to crystallize to a later stage. The nepheline-syenites not infrequently take on a porphyritic character. Some syenites contain basic secretions, acid veins, pegmatite fringes and other peculiarities noticed under the granites. Parallel and gneissic structures sometimes come in locally (e.g. Plauen'scher Grund). Leading types. Although typical hornblende-syenites probably occur in this country (e.g. Malvern), very little has been written about them, and for the type-rocks we must go to foreign occurrences. The name ' syenite ' as found in many of the earlier writings and maps in this country is to be under- stood in the old sense of hornblende-granite (including also granophyre, etc.) and the identification of hornblende is in very many cases erroneous. For example, the so-called 'syenites' of St David's, of Ennerdale, of Oarrock Fell, etc., have no claim to the title, whether the word be used in its original or its modern sense. The rock taken as the type of hornblende syenite is that of Plauen'scher Grund near Dresden ('plauenite' of Brb'gger, fig. 7.) It is composed essentially of orthoclase, with only subordinate oligoclase, and green hornblende. Apatite, 46 HORNBLENDE-SYENITES. magnetite, and sphene occur as accessories, and in places a little quartz. There is a variety in which biotite occurs in addition to the hornblende. The rock encloses dark basic secretions richer in plagioclase, hornblende, apatite, magnetite, and sphene. Further there are pegrnatoid acid veins of coarse texture, in which the more basic minerals occur only sparingly, while quartz is plentiful. Almost the same description applies to other Saxon syenites, such as that of Meissen, which, how- ever, has rather more brown mica, and further contains a little more quartz, either in grains or in micrographic intergrowth. ol FIG. 7. HORNBLENDE-SYENITE, PLAUEN'SCHER GRUND, DRESDEN ; x 20. shewing hornblende (h), orthoclase (or), subordinate oligoclase (ol), sphene (sp), and apatite (ap) [47]. There is rather more oligoclase than in the preceding, be- sides abundant sphene and apatite. A rock from Biella in Piedmont is closely similar to that of Plauen'scher Grund. sphene and apatite being rather plentiful. A syenite like that of Dresden, but sometimes rich in biotite, occurs near Salem, Mass 1 . An example from Ouster 1 Wadsworth, G. M. 1885, 207. MICA-SYENITES. 47 County, Colo. ', consists of orthoclase and oligoclase, the latter predominating, green and brown hornblendes, and dark green mica, with a little apatite. More felspathic varieties have been noted from Curtis Point, Beverley, Mass. 2 (with arfvedsonite-like hornblende) and Albany, N. H. (with blue riebeckite). Such rocks as that of Meissen may with propriety be termed quartz-syenite (quartz-hornblende-syenite), and form a connecting link with the hornblende-granites. Again, when a triclinic felspar becomes predominant we have transitions to quartz-diorite (e.g. Weinheim, in the Odenwald, near Heidelberg). Brogger's red quartz-syenite (Nordmark type) from the Christiania district, placed by Rosenbusch among the granites rich in alkali, also has oligoclase in addition to the dominant orthoclase, and sometimes a microperthitic intergrowth of albite and orthoclase. Biotite and hornblende are the chief ferro-magnesian constituents, but green augite and segirine also occur. The mica-syenite type, in which biotite predominates over hornblende, is of uncommon occurrence, except as a local variety of hornblende-syenite. More often there is some quartz present, and such rocks are found graduating into biotite-granite. In this latter connection may occur quite basic rocks, very rich in mica and allied to the lamprophyres. Such a rock has been described by Sauer 3 from the Black Forest (Durbach type). Rosenbusch mentions mica-syenites from Canada; one from Star Hill Mine, Portland West, P. Q., rich in apatite ; another from Blessington Mine, Inchin- brooke, P.O., with some augite as well as mica. These rocks are free from quartz or plagioclase. Among quartz-augite-syenites may be mentioned Brogger's Aker type from the Christiania district, which contains plenty of plagioclase as well as orthoclase, and has resemblances to the Monzoni rocks. Biotite occurs in addition to the pale green augite. Among British rocks we have that of Llan- faglen near Caernarvon 4 , which also carries two felspars. * Cross, Proc. Colo. Sci. Soc. (1887) 237-240. 2 Sears, Bull. Essex Inst. (1891) xxiii. 3 M. 31. x, 176 (Abttr.). * Bala Vole. Ser. Caern. (1889) 73, 74. 48 AUGITE- SYENITES. The augite, in imperfect crystals and grains, is colourless in section, but tends to pass into greenish uralitic hornblende. There is original hornblende, mostly of a brown variety, which forms sometimes good crystals, sometimes ophitic plates as in diabasic rocks : it is altered in places into brown mica. Other constituents are apatite, magnetite, ilmenite, pyrites, and quartz. Other quartz-syenites characterized by augite shew a strong tendency to micrographic intergrowth of quartz and felspar. This is seen in the larger pre-Carboniferous intrusions of Leicestershire (excepting the Mount Sorrel granite), which indeed may be classed as a less acid type of granophyre. The augite tends to pass into uralitic hornblende, and epidote is a characteristic secondary product in the rocks. Examples are seen at Groby, Bradgate Park, Markfield, and Garendon, all in the Charnwood Forest district 1 . A special type of augite-syenite is presented by the Triassic intrusions of Monzoni in the southern Tirol (monzonite of De Lapparent), which are associated with diabases and other basic- rocks. Orthoclase is sometimes the only felspar, but usually there is a plagioclase in addition, forming idiomorphic crystals enclosed with the other minerals by plates of orthoclase. The augite often passes over into green hornblende, but the latter mineral also occurs as an original constituent. Biotite is usually present, in flakes sometimes earlier, sometimes later, than the plagioclase. Sphene is frequent, and zircon is often enclosed by the mica. Other constituents are apatite, mag- netite, and pyrites, and in some examples a little interstitial quartz. In America Weed and Pirsson have described a rock closely resembling the typical monzonites from Yogo Peak, Montana 2 . This rock, with about equal amounts of felspar and augite, graduates on the one hand into a more felspathic augite-syenite and on the other into a thoroughly basic type very rich in augite. This last (Shonkin type) was first distinguished by the same writers at Square Butte in the 1 Bonney, Q. J. G. S. (1878) xxxiv, 214-218. 2 A. J. S. (1895) 1, 467-479. AUGITE-SYENITES. 49 Highwood Mts., Mont. 1 . It consists of predominant augite with orthoclase, albite, and anorthoclase, apatite, biotite, olivine, etc., and may be compared with the basic modifi- cations of the rocks of Monzoni (' pyroxenite ' of Brb'gger). A peculiar augite-syenite (Laurvig type), allied in some respects to the nepheline-syenites, occurs among the Devonian intrusions of the Christiania district. While augite is usually the dominant ferro-magnesian element, it is often accompanied by biotite, tegirine, hornblende, or arfvedsonite, and the rock thus passes into mica-syenite, etc. Alkali-felspars (orthoclase, microcline, albite, cryptoperthite, etc.) make up the bulk of the rock, and are often intergrown with one another. Not infrequently they have a schiller-structure. A little quartz is rarely present; on the other hand elseolite and sometimes olivine may occur as minor accessories. The augite is oc- casionally green, but commonly light brown with a violet tone and slight pleochroism : schiller-structure is common. The FIG. 8. An GITE- SYENITE (LAUBVIG TYPE) FROM A BOULDER ON THE YORKSHIRE COAST ; x 20. The minerals seen are cryptoperthite felspar (/) in large plates, augite (a) with schiller-structure in the interior of the crystal, deep brown biotite (b), magnetite (m), and apatite (op) [1841]. i Bull. Geol. Soc. Amer. (1895) vi, 408-415. 50 AUGITE- AND NEPHFLINE-SYENITES. hornblende is green or occasionally brown ; the biotite a very deep brown. The latter mineral is roughly idiomorphic, except when it is massed round magnetite or forms a marginal intergrowth with augite. The iron-ores are magnetite and sometimes haematite : apatite is universal, but sphene is typi- cally absent. Zircon is a constant accessory, and sometimes builds large crystals, giving the variety 'zircon-syenite' of von Buch and other early writers. These augite-syenites are common as boulders 1 on our East coast (fig. 8). A rock with a like richness in ferro-magnesian minerals and similar microperthitic intergrowths is described by Osann from the Sawtooth Mts. in Western Texas. Other augite-syenites rich in microperthitic felspars occur in New Hampshire (Jackson, Stark, and Columbia). A rock from Mosquez Canon, Tex., consists of orthoclase with micro- perthitic intergrowths of plagioclase, ajgirine-augite, apatite, iron-ore, and relatively abundant zircon, while a more typical trglrine-syenite is recorded from Fourche Mt., Ark. The nepheline-syenites are in part closely allied to certain types of augite-syenites, and in the Christiania area, for instance, the two rocks are closely associated. The nepheline- syenite of that district (Laurdal type) differs from the Laurvig rock chiefly in the presence of elaeolite and sometimes sodalite, the latter sometimes shewing a pale blue or violet tint. The same alkali-felspars as before are present, with frequent inter- growths. The abundant ferro-rnagnesian minerals embrace deep brown biotite, green hornblende, and light brown or purplish-brown pleochroic augite, either singly or in associa- tion. Apatite and magnetite are present, and occasionally a little olivine. A somewhat similar, though more acid, rock occurs in Arkansas 2 , where there is also one with a porphyr- itic tendency, with less nepheline, and with hornblende as the chief ferro-magnesian silicate (Pulaski type) 3 . A well-known nepheline-syenite is that of Sierra de Mon- 1 Proc. Yorks. Geol. Pol. Soc. (1889-90) xi, 303, 304, 410. 2 J. F. Williams, The Igneous Pocks of Arkansas, vol. u of Ann. Rep. Geol. Sur. Ark. for 1890, pp. 74-80, 130-135. 3 Ibid. 55-69. NEPHELINE-SYENITES. 51 chique in Portugal 1 (Foya type 2 ). Here the proportions of elaeolite and orthoclase vary ; sodalite is often present ; the coloured minerals are subordinate hornblende, augite edged with iegirine-augite, and biotite ; while apatite, magnetite, and abundant sphene are also present. Rocks generally comparable with this occur in Brazil 3 , near Montreal 4 , at Salem and Marblehead 5 (Mass.), at Red Hill (N. H.) 6 , in the Crazy Mts. (Mont.) 7 , in the Cripple Creek district (Colo.) 8 , at Mt. Ord and Paisano Pass 9 (Tex.), and at several localities in Arkansas. Some of the Arkansas rocks have porphyritic modifications. At Beemerville 10 (N. J.), again, occurs a variety with very large crystals of orthoclase, the interspaces tilled by little prisms of segirine and abundant elseolite, partly changed to cancrinite. Among other nepheline-syenites may be mentioned the Miask type from the Urals, in which a deep brown mica is the most prominent constituent, plagioclase is abundant, frequently intergrown with the orthoclase, and zircon is a characteristic accessory. The Ditro type from Transylvania, also carries mica, but much less plentifully : it is distinguished by its abundance of allotriomorphic sodalite and by the variety and intimate intergrowths of its felspars, which include microcline as well as orthoclase and oligoclase. Can- crinite, sphene, zircon, and perofskite also occur". In the Litchfield type 12 , from Maine, albite constitutes about half Sheibner, Q. J. G. S. (1879) xxxv, 42-47. The Pulaski type also occurs. This is Blum's foyaite, a name sometimes extended, however, to be syn nymous with nepheline-syenite. Graeff, M. M. vii, 231-234; Machado, ibid, viii, 168 (Abstracts). Lacroix, G. M. 1891, 216 and M. M. x, 42 (Abstracts). Wadsworth, Proc. Post. Soc. Nat. Hist. (1882) xxi, 406; G. M. 188 , 208, 209 ; Sears, Bull. Essex Inst. (1893) xxv. Bayley, Bull. Geol. Soc. Amer. (1892) iii, 245-253. Wolff and Tarr, Bull. Mm. Comp. Zool. Harv. (1893) xvi, 230, 231. Cross, Wth. Ann. Hep. U. S. Geol. Stir. (1895) Pt II, pp. 43, 44. Osann, 4tA Ann. Rep. Geol. Sur. Tex. J Emerson, A. J. S. (1882) xxiii, 302-308 ; Kemp, Trims. N. Y. Acad. Sci. (1892) xi, 63. 11 For coloured figures of these rocks see Fouque and Levy, PL XLV, fig. 1. 12 Bayley, Bull. Geol. Soc. Amer. (1892) iii, 235-241 ; M. M. x, 345 (Abstr.). 42 52 SODALITE- AND LEUCITE-SYENITES. of the rock, the other minerals being orthoclase, microcline, elseolite, sodalite, cancrinite, a deep green biotite (lepido- melane), and a little zircon. Plagioclastic nepheline-syenites are also described from Arkansas 1 ; these are segirine-bearing rocks. A variety from Dungannon 2 in Ontario resembles the Litchfield rock in the predominance of a soda-felspar, but is richer in nepheline. In one modification the felspar disappears, and the rock consists merely of nepheline with a little hornblende or mica. This corresponds with the ' ijolite ' of Ramsay and Berghell from Finland, a nepheline- pyroxene-rock free from felspar but sometimes rich in garnet. A sodalite-syenite, with little or no elaeolite, seems to be an uncommon type. It has been found in the Highwood Mts. Mont. 3 . Another peculiar rock -(Taimyr type), described by Chrustchoff 4 from northern Siberia may be styled nosean- syenite, consisting essentially of nosean and anorthoclase with some brown hornblende, biotite, zircon, etc. Altered leucite-syenites, containing pseudomorphs of orthoclase and elseolite in the form of leucite, have been described in Brazil 5 , Arkansas 6 , etc. They commonly shew porphyritic structure, and are perhaps more appropriately placed among the hyp- abyssal rocks. Although no genuine elseolite-bearing syenite is yet known from this country, an allied rock has been described by Mr Teall 7 from Loch Borolan in Sutherland. There the usual type consists essentially of orthoclase, a brown melanite garnet, and a green or green-brown biotite. A green monoclinic pyroxene is present in many examples : a brown pleochroic sphene, apatite, and magnetite occur as accessories. Nepheline is supposed to be represented by an alteration-product which often forms micrographic intergrowths with the orthoclase, in 1 J. F. Williams, The Igneous Eocks of Arkansas, vol. n of Ann. Rep. Geol. Sur. Ark. for 1890, 136-140. 2 Adams, A. J. S. (1894) xlviii, 10-16 ; M. M. xi, 46, 47 (Abstr.). Lindgren, Amer. Journ. Sci. (1893) xlv, 290-297; Weed and Pirsson, Bull. Geol. Sc.c. Amer. (1895) vi, 416, 417. * M. M. x, 259, 260 (Absti:). 5 Derby, Q. J. G. S. (1891) xlvii, 254-263. e J. F. Williams, I.e. 267-277. 7 Trans. Roy. Soc. Edin. (1892) xxxvii, 163-178, with plate. ROCKS ALLIED TO SYENITES. 53 patches giving a pseudo-porphyritic aspect to the rock. Another substance, in confused aggregates with a bluish tint in reflected light, is probably one of the sodalite minerals. It- is found especially in certain pegmatoid veins in the rock, consisting chiefly of orthoclase. The above rock, to which the name borolanite has been given, differs from certain garneti- ferous nepheline-syenites ' in having orthoclase dominant instead of elseolite. From Poohbah Lake in Ontario Lawson 2 has described some peculiar basic orthoclase-bearing rocks on which he proposes to found a new family with the name malignite. They are richer in lime than the Borolan rocks and richer in alkali than the Shonkin type. One variety consists about one-half of segirine-augite, with orthoclase, nepheline, abund- ant apatite, etc.. the felspar occurring interstitially or enclosing the other constituents. Another has large crystals of orthoclase (microperthite) in a granular aggregate of pegirine-augite, melanite, biotite, etc. In a third variety the pyroxene and garnet are wanting, being replaced by an arfvedsonite-like hornblende. Among special modifications of syenitic rocks may be mentioned the syenite- aplites and syenite-pegmatites described by Brogger as associated with the augite- and nepheline- syenites of the Christiania district. The pegmatites are remarkable not only for the frequent perthitic intergrowths of potash- and soda-felspars, but also for graphic intergrowths of the felspars with the ferro-magnesiau minerals and with elteolite and sodalite; and they are famous as the home of many rare minerals. Some of these features are reproduced in the pegmatites associated with the Arkansas nepheline- syenites 3 . 1 Cf. J. F. Williams, I.e., pp. 229-231. 2 Bull. Dep. Geol. Univ. Calif. (1896) i, 337-362. 3 J. F. Williams, I.e., 143-146, 238-258. CHAPTER IV. DIORITES. THE diorites are plutonic rocks of medium to coarse texture, consisting essentially of a soda-lime felspar and horn- blende, with less important constituents. The family so defined cannot be regarded as a natural one, its members ranging in chemical composition from sub-acid to thoroughly basic. The gabbros (characterized by pyroxenes in place of hornblende) also include intermediate as well as basic rocks, and the distinction between the hornblende- and augite-bearing types is rather an artificial one. It was established before the strong tendency of augite to pass over into hornblende was thoroughly appreciated : later research has shewn the certainty of some, and the possibility of many, of the 1'ocks that have been termed diorites being really amphibolized pyroxenic rocks. The more acid diorites contain free silica (quartz- dioritex), and, except for the smaller proportion of quartz and the nature of the felspars, do not differ much from the hornblende- granites 1 . They may have biotite in addition to hornblende (quartz-mica-diorites), or in some cases augite. In the diorites proper, without quartz, mica is not common, but the horn- blende may be accompanied by augite or sometimes enstatite. The hornblende is more abundant relatively to the felspar than in the preceding types, and some of the more basic diorites consist chiefly of hornblende. These are the ' amphi- bolites ' of some authors 2 . In some types olivine enters as a constituent (olivine-diorites). 1 See Berwerth, Lief 1, PI. n. 2 For a hornblende-rock (local modification of a diorite) see Fouque and Michel Levy, PI. xxm. FELSPARS OF DIORITE. 55 The occurrence of felspathoid minerals in clioritic rocks seems to be very exceptional. The theralites of Rosenbusch may be regarded as nepheline-diorites and nepheline-gabbros, but comparatively little is yet known of such rocks. n i FIG. 9. CRYSTALS OF PLAOIOCLASE FELSPAR IN QUARTZ-MICA-DIORITE, BEINN NEVIS ; x 20. Crossed nicols : the vibration-planes of the nicols are indicated by the lines (ni) [397]. A shews the association of twin-lamellation on the albite (a) and pericline (p) laws. B shews carlsbad twinning (c) com- bined with albite-twin-lamellation (a) and with zonary banding. Constituent Minerals. The felspar of the diorites is oligoclase, andesine, or labraclorite, or exceptionally a more basic variety. The twin-lamellation on the albite type is often accompanied by pericline- or carlsbad -twinning (fig. 9 A). In the quartz-diorites especially, the crystals frequently shew between crossed nicols a marked zonary banding, the central and marginal portions of a crystal often giving widely different extinction-angles, and the successive layers growing more acid from within outwards (fig. 9 B). In natural light the zones of growth may be indicated by the disposition of fluid-pores, minute scales of haematite, or other inclusions. The crystals are often clouded by a fine dust (probably kaolin), and may also furnish by their alteration scales of colourless mica 56 HORNBLENDE OF DIORITE. (paragonite ?), grains of epidote, calcite, etc. A little ortho- close may be present as an accessory, behaving in the quartz-diorites as in granites, while in typical diorites it occurs interstitially. The hornblende, when idiomorphic, shews the prism-faces and usually the clinopinacoid, and terminal planes are often present. Twinning is common, and the prismatic cleavage is always well pronounced. In the quartz-diorites the mineral, usually in imperfect crystals, is green, as in granites ; in more normal diorites it has brownish-green or greenish- brown colours ; and in the most basic types the original hornblende is usually of some greenish shade of brown, or even approaches the deep brown of 'basaltic hornblende.' Pale colours result from bleaching, or are found in secondary outgrowths of the brown crystals, and these are green rather than brown. Two kinds of outgrowth or enlargement of hornblende crystals are to be observed in some basic diorites, the new growth being in both cases in crystalline continuity with the old. In one case FIG. 10. BASIC DIOKITE, LLYS EINION, NEAB LLANERCHYMEDD, ANGLESEY ; x 20. The original idiomorphic brown hornblende has an extension of green hornblende on the clinopinacoid faces (h) and also a secondary fibrous outgrowth on the terminal planes (h'). The felspar (/) is much decomposed, and crystalline calcite (c) has been produced [539]. MICA AND PYROXENES OF DIORITE. 57 a growth of green hornblende takes place on the clinopinacoid faces so as to extend the crystal, with idiomorphic contour, in the direction of the orthodiagonal : in the other case pale green or colourless hornblende grows so as to extend a crystal in the direction of its length, and may present new crystal- faces, or abut on another crystal, or frequently terminate in a ragged fibrous fringe. The second type of outgrowth at least is of secondary origin, and is formed at the expense of other minerals (tig. 10). Besides more usual types of alteration 1 , the brown hornblende of diorites may shew bleaching, with separ- ation of magnetite, or it may be converted into a brown rnica or into green blades of actinolite. The deep brown biotite of the diorites occurs in idiomorphic flakes, or sometimes intergrown with hornblende. It is usually not rich in inclusions. It becomes green only by partial decoinposition. The rhombic pyroxene found in a few diorites is a variety poor in iron (enstatite) and is usually converted into pseudo- morphous pale bastite. When augite is present, it is of a variety sensibly colour- less in slices. If idiomorphic, it shews the octagonal cross- section due to equal development of the pinacoids and prism-faces, with good prismatic cleavage and not infrequently lamellar twinning parallel to the orthopinacoid. A not uncommon feature in diorites is a parallel growth of augite and hornblende, a crystal-grain of the former mineral con- stituting a kernel, round which a shell of brown hornblende has grown, and this seems to occur specially in the neighbour- hood of grains of iron-ore. This must be distinguished from another phenomenon frequent in the augite-bearing diorites, viz. the conversion of augite into brown hornblende as a secondary change. This process usually begins at the margin of a crystal or grain, but proceeds irregularly, shewing a very intricate boundary between the two minerals and often ragged scraps of one enclosed by the other. When the conversion is complete, the secondary hornblende can be distinguished from original only by inference, as, e.g. when it shews the external 1 Zirkel, Micro, Petr. Fortieth Parallel, PI. in, figs. 2, 3, 4. 58 STRUCTURES OF DIORITE. form of augite. In both phenomena the augite and hornblende have their plane of symmetry and longitudinal axis in common, and in longitudinal sections both extinguish on the same side of the axis. The quartz of quartz-diorites has the same general charac- ters as that of granites. The olivine which occurs in some basic diorites is often in rather rounded crystals moulded by the hornblende. It is easily recognized by its high refractive index and very strong double refraction. The mineral is readily altered into ser- pentine, carbonates, and especially pale fibrous amphibole, the last often grown in crystalline continuity with adjacent original hornblende. Among the iron-ores, magnetite is the most usual, but ilmenite is also found. Common accessories in some varieties are zircon and sphene in characteristic crystals. Apatite is general, and in some basic diorites abundant : in the coarse- grained rocks it sometimes builds rather large prisms. Structure. The structure of the dioritic rocks is variable. In the quartz-diorites 1 the mutual relations of the minerals are those noticed in granites, though sometimes a part of the felspar has crystallized before the ferro-magnesian minerals. A micrographic intergrowth of quartz and felspar is not infrequent. Many of the quartzless diorites also follow what may be called the normal order of crystallization. Rosenbusch points out that the most marked pauses in the process of consolidation have occurred before the separation of the ferro-magnesian minerals and after that of the plagioclase; so that while the apatite, sphene, etc., and the plagioclase may be markedly idiomorphic, the hornblende, biotite, and augite tend to occur in much more irregularly shaped crystals. When a miarolitic structure results from the tendency to idiomorphism in the latest crystallized elements, it is com- monly obscured by the cavities becoming filled by calcite and other secondary products. A different type of structure, though connected by transi- tions with the preceding, is found in many dioritic rocks. 1 Berwerth, Lief. 1, PL n. STRUCTURES OF DIORITE. 59 Here the plagioclase has crystallized earlier, or at least ceased to crystallize earlier, than the bisilicates ; so that the dominant felspar presents idiomorphic outlines to the hornblende and (if present) augite. These latter may wrap round, or even enclose, the felspar crystals, giving an ' ophitic ' structure identical with that described below as characteristic of the diabases, and the hornblendic rocks exhibiting this character have sometimes been termed hornblende-diabases. Such a structure is found more or less markedly in many of the more basic diorites, and is especially common in rocks in which the hornblende is in great part derivative after augite, but original hornblende moulded on felspar is also found 1 Pegmatoid and aplitic structures are less common in this family than in the granites and syenites. A porphyritic structure is not common in true diorites, but may come in as a marginal modification of a boss or stock, the porphyritic elements being crystals of hornblende or felspar. As a more special type of structure may be mentioned the orbicular (in the so-called corsite or napoleonite), where the bulk of the rock consists of spheroidal growths. These have a radial structure and consist of concentric shells composed essentially of hornblende and felspar in alteration. Leading Types. The quartz-mica-diorite of the Ada- mello Alps, on the border of Italy and the Tirol (Tonale type) comes very near in characters to some granites 2 , and has also points in common with the Monzoni syenites. The dominant felspar is a striated plagioclase, often shewing zonary banding and with a strong tendency to idiomorphic outlines; but there is frequently clear orthoclase in addition, in irregular crystal plates moulded on or enclosing the triclinic felspar. Biotite is the most constant coloured element, but hornblende is also abundant. The mutual relations of the two are variable, 1 Q. G,J. S. (1888) xliv, 450-453. 2 This is the 'tonalite' of vom Rath. Since it is an extreme type, and is classed by some petrologists with the granites, it is confusing to extend this name, as some writers have done, to all the quartz- diorites. Brogger restricts the term to the type free from any alkali-felspar; that with both an alkali- and a lime-soda-felspar he styles adame.llite. 60 QUARTZ-DIORITES. and both may enclose the plagioclase. Interstitial quartz is abundant ; patches of magnetite are often prominent ; and zircon in little well-built prisms is general. Schmidt describes the rock of the Yosemite Valley, California, as closely similar to that of the Adamello, and Matthew 1 has given an account of a porphyritic tonalite near St John, N. B. More typically intermediate quartz-diorites occur in Hun- gary (Banat type), where they are of Tertiary age, and are the plutonic equivalents of some of the andesitic lavas. Here quartz is less abundant, and orthoclase usually absent. The characteristic zonary banding of the felspars is strongly marked. Green or brown-green hornblende is the dominant coloured mineral, but brown biotite is also common, and the two are sometimes intergrown. Crystals of magnetite and other minor accessories are found. Some varieties of the rock tend to develope a porphyritic structure. Further examination will probably shew that some of the Scottish Carboniferous 'granites' are better classed as quartz- diorites (usually with mica). The zoned plagioclase crystals, the interstitial quartz, and other features are well exhibited (Beinn Nevis, etc.). There may be a rough micrographic struc- ture. Quartz-diorites and quartz-mica-diorites occur about Garabal Hill near the head of Loch Lomond 2 , and shew interesting gradations, on the one hand into granite, and on the other into quartzless diorites (mica-diorite, augite-diorite, etc.). Other quartz-diorites, usually with biotite as well as hornblende have been described from Arran, Glen Tilt, etc. In Wicklow, east of Rathdrum, occur quartz-diorites and quartz-mica-diorites, which seem to approach granites in their characters 3 . Subordinate orthoclase accompanies the dominant triclinic felspar. The other minerals are pale green hornblende, ragged flakes of biotite, abundant quartz, apatite, and some- times a little colourless augite (salite or malacolite). The augite-diorites, which are a common type in Wicklow, some- times have interstitial quartz in addition to the plagioclase, 1 Trans. N. Y. Acad. Sci. (1894) xiii, 188-191. 2 Dakyns and Teall, Q. J. G. S. (1892) xlviii, 104-120. 3 Hatch, G. M. 1889, 262, -263 ; see also Watts, Guide, 34. MICA-DIORITES. 61 hornblende and idiomorphic salite which are their essential constituents. In the United States, as in Britain, numerous rocks belonging to this type have been styled granite or 'granite- diorite.' As a typical quartz-diorite may be cited that described by Iddings 1 from Electric Peak in the Yellowstone Park. Here the dominant felspar ranges from oligoclase to labradorite, and there is sometimes orthoclase in addition ; the quartz is in allotriomorphic grains ; and the other con- stituents are biotite, hornblende, augite, hypersthene, and magnetite. Parallel intergrowths are frequent among the ferro-magnesian minerals, hypersthene being bordered by augite and the pyroxenes by biotite and hornblende 2 . A porphyritic quartz-mica-diorite was described by G. H. Wil- liams 3 among the varied dioritic rocks of the Cortlandt district. The large felspar crystals are strongly zoned, but only occasionally lamellated. A mica-diorite, without quartz, is not a common type. It is found as a local modification of biotite-granite between Carrick Mt. and Ark low, in Wicklow. Mr Teall 4 describes a good example from Pen Voose in the Lizard district, Cornwall. This consists essentially of felspar and a reddish brown mica with only quite subordinate green hornblende and accessory sphene. From Allt-a-Mhullin, south of Lochinver, Sutherland, the same author notes a mica-diorite with pcecilitic felspar. Among the Cortlandt rocks, on the Hudson River, a pure mica-diorite occurs, beside various mica-hornblende-diorites. It is a rather coarse-grained aggregate of plagioclase (oligo- clase-andesine) and very deeply coloured biotite, with accessory epidote, magnetite, abundant apatite, and sometimes a little quartz 5 . Mica-diorite has been noted near the Comstock Lode, Nevada. Of simple hornblende-diorite, without quartz, good examples, of Palaeozoic age, are found in Warwickshire and other parts 1 See 12th. Ann. Rep. U. S. Geol. Sur. (1892) 595-609. 2 Ibid. PI. L. 3 A . J. S. (1888) xxxv, 446. 4 PL xxxii., fig. 1 ; XLVII., fig. 3. s G. H. Williams, A. J. S. (1888) xxxv, 443-445 ; Kemp, ibid, xxxvi, 247-251. 62 HORNBLENDE-DIORITES OF THE MIDLANDS. of the Midlands. In the rock of Atherstone, Hartshill, the brown hornblende is in part idiomorphic towards the turbid felspar, but part of it, on the other hand, is derived from a colourless augite, and a kernel of the latter mineral sometimes remains unchanged. Grains of magnetite are present and abundant prisms of apatite (fig. 11). Mr Allport 1 also mentions pseudomorphs of calcite, etc., after olivine. The same writer describes a fine-textured diorite from Marston Jabet, in which idiomorphic brown hornblende is set in an aggregate of triclinic felspar. Rather coarse-grained diorites are met with in the m- ap- FIG. 11. DIORITE, ATHERSTONE, WARWICKSHIRE ; x 20. The figure shews idiomorphic hornblende (h), turbid felspar (/), magnetite (TO), and rather abundant prisms of apatite (ap). Cross- sections of the last shew the hexagonal shape, and longitudinal sections shew the cross- fracture [1608]. curious complex of igneous rocks in the Malvern district. A specimen taken near the New Reservoir consists essentially of idiomorphic greenish -brown hornblende and lahradorite felspar. The latter shews albite- and pericline-lamellatioii, and its de- composition has given rise to zeolites and paragonite mica. In the well-known diorite of Brazil Wood 2 in Cliarnwood 1 Q. J. G. S. (1879) xxxv, 637-641. 2 Hill and Bonuey, Q. J. G. S. (1878) xxxiv, 224. HORNBLENDE-DIORITES OF NORTH WALES. 63 Forest, Leicestershire, the hornblende tends to embrace the felspar, and this departure from the granitic type of struct- ure is observable in some other diorites from the Midland counties. Various diorites occur in the interior of Anglesey. One between Gwindu and Llanfaelog is a coarse-textured rock con- sisting of greenish brown hornblende and turbid felspar with magnetite and apatite. The minor intrusions near Llanerchy- medd 1 are of a rather different type. Brown hornblende occurs in well-formed crystals and also in shapeless plates which can sometimes be seen forming at the expense of a colourless augite. There is also hornblende of later growth than the crystals mentioned but not derived from augite. It occurs as a crystalline outgrowth of the original brown crystals. Part of it has grown upon the clinopinacoid faces and itself shews crystal boundaries ; this is green. Part has grown chiefly on the terminations of the original crystals and filled up interstices : this is pale or colourless (fig. 10). Some of these rocks contain a little olivine, or rather its alteration- products, and in certain specimens, not found in place, this mineral must have been abundant. With this richness in olivine goes a diminution in the amount of felspar, giving a transition from diorite to hornblende-picrite 2 . Other olivine- bearing hornblendic rocks occur near Clynog-fawr in Caernar- vonshire 3 . Here the hornblende occurs in ophitic plates, and the structure of the rocks closely resembles that of typical diabases. They have indeed been described under the pro- visional title of hornblende-diabases, and, although augite is not often seen in them, it is possible that much of the hornblende is derivative after that mineral. The same remark applies to certain rocks at Penarfynydd 4 in the Lleyn peninsula, where both ophitic and idiomorphic augite may be seen partly con- verted into brown hornblende. Olivine seems to have been rare in these rocks, but they are closely associated with a 1 G. M. 1887, 546-552. Other types of dioritic rocks from Central Anglesey are described by Mr Blake, Rep. Brit. Axsoc. for 1888, 403-406. 8 Bonney, Q. J. G. S. (1881) xxxvii, 137-139 ; (1883) xxxix, 254-256. 3 Bala Vole. Ser. Caern. 102-106. 4 Ibid. 92-97. 64 HORNBLENDE-DIORITES. hornblende-picrite rich in that mineral. Some thoroughly basic dioritic rocks, very like those of Anglesey, occur in the Lake District, e.g. at Little Knott 1 , White Hause, and Great Cockup 2 in the Skiddaw district. The rock at the first-named locality shews beautifully the pale fringes of hornblende which form a crystalline outgrowth of the original idiomorphic crystals. These fringes are clearly secondary, and occupy the place of destro} r ed felspar, etc. Some olivine has been present in some specimens. These Welsh and Cumbrian dioritic rocks occur usually in small laccolitic intrusions, probably of Ordovician age. In the Isle of Man several small masses of diorite are found on Langness. The hornblende, of a greenish brown tint, is perfectly idiomorphic, but often shews secondary out- growths. The felspars are much decomposed. Abundant zoisite, epidote, calcite, etc., have been produced, and the quartz which is always found is probably all secondary. Apatite is plentiful, but a little pyrites is usually the only iron-ore present. The diorites of the Scottish Highlands are not yet de- scribed in any detail. Those of the Garabal Hill district include mica-diorite and augite-diorite. The pale green augite is usually in allotriomorphic grains irregularly bordered by green hornblende. Diorites, with other hornblendic rocks, occur near Inchnadamft' in Sutherland 3 . Here the horn- blende is in unusually perfect crystals. In America the Cortlandt rocks include diorites consisting of brown hornblende, andesine, apatite, and magnetite, some- times with accessory hypersthene, and by failure of the felspar these rocks graduate into hornblende-rocks. There are also diorites with green hornblende 4 . From Alabama 5 are described both basic diorites and others of more acid nature, which con- tain a little quartz and orthoclase. The diorites described Bonney, Q. J. G. S. (1885) xli, 511-513, PI. xvi, fig. 2. Postlethwaite, Q. J. G. S. (1892) xlviii, 510. Teall, G. M. 1886, 346-353. G. H. Williams, A. J. S. (1888) xxxv, 441, 442. Clements, Bull. No. 5 Geol. Sur. Ala. (1896) 152-165 ; Brooks, ibid. 189, 190. AUGITE-DIORITES. 65 by Zirkel 1 from Nevada are chiefly of the more acid kind, sometimes carrying quartz or, again, passing into mica-diorite (Pah-Ute range). The diorites of the great laccolitic masses in Colorado, Utah, and Arizona, of which Cross 2 has given a full account, are also of relatively acid varieties, with quartz, and tend to take on a porphyritic structure, graduating into quartz-porphyrites. A number of dioritic rocks may be studied in the Channel Islands. A very fresh rock from the quarries of Delancy Hill, Guernsey, is an augite-diorite, with colourless augite as well as brown original hornblende. The latter mineral is moulded on the felspar-prisms, and often borders the augite with the usual crystallographic relation (fig. 12). A specimen from Rope- walk Quarry is also an augite-diorite with diabasic characters. The colourless augite is partly in rounded grains enclosed by FlG. 12. AUGITE-DIOBITE, DELANCY HlLL, GUERNSEY ; X 20. The augite shews either sharp octagonal cross-sections (a) or more rounded contours (a/). Hornblende (&), magnetite (m), and clear plagioclase felspar (/) are the other constituents. Much of the horn- blende occurs in marginal intergrowth with the augite, interposed between the latter mineral and the magnetite [431]. 1 Micro. Petrogr. Fortieth Parallel (1876), 85-93. 2 Laccolitic Mountain Groups, 14t/i Ann. Rep. U. S. Geol. Sur. (1895). 66 ESSEXITE. the felspar, partly in shapeless plates, and the brown horn- blende, apparently an original mineral, is clearly of posterior crystallization to the felspar. Magnetite is plentiful, and there are some large crystals of a rhombic pyroxene replaced by bastite. An augite-diorite from Fort Touraiile, in Alderney, gives evidence of the conversion of augite into hornblende. Some deep brown biotite is also present, and a little interstitial quartz is the last product of consolidation. A peculiar type of augite-diorite is described by Cross 1 from the Rosita Hills, Colorado. It consists mainly of labradorite and augite, with accessory olivine, magnetite, and apatite, but biotite is also present, usually surrounding augite or magnetite, and there is a little orthoclase, partly as a marginal intergrowth on the labradorite. The essexite of Sears 2 , occurring in association with the nepheline-syenite of Salem, Mass., may be regarded as a peculiar olivine-augite-diorite allied to the theralites. The pale green augite is bordered by brownish hornblende, and brown biotite is intimately associated with them. The rounded grains of olivine are often pseudomorphed by biotite-aggregates, green hornblende, and granular augite. The iron-ore is tita- niferous, and gives rise to secondary sphene. Apatite is abundant in irregular grains as well as in slender prisms. The felspar, in idiomorphic crystals, is labradorite, chiefly of an acid variety. Nepheline may be present, but is not certainly identified. A similar rock is found at Mount Royal, near Montreal. Here the augite is reddish violet, probably titaniferous. The rock passes into a theralite carrying both nepheline and sodalite. 1 Proc. Colo. ScL Soc. (1887) 246, 247. " Bull. Essex Instil. (1891) xxiii. CHAPTER V. GABBROS AND NORITES. THE gabbros and their allies are holocrystalline rocks, typically of plutonic habit, in which the essential constituents are a lime- soda-felspar and a pyroxene. Of intermediate to thoroughly basic character, they correspond partly with the diorites ; but the more acid, and especially the quartz-bear- ing types, are less represented in the pyroxenic than in the hornblendic series. According to the dominant pyroxene, we recognize aabbro proper (euphotide of Haiiy) with diallage or augite, and norjte (also called hypersthenite l or hyperite) with a rhombic pyroxene. A few of the more acid rocks contain free silica (quartz-gabbro and quartz-norite). Tn most of the more basic varieties olivine becomes a characteristic mineral (olivine-gabbro and olivine-norite). The majority of the rocks in this family contain more or less olivine, and the mineral may be present or absent in different specimens of the same mass. The gabbros and norites, indeed, shew considerable varia- tions in mineralogical constitution in parts of one mass, and most of the special types are probably to be regarded as merely local modifications. Thus, by the failure of one or other of the chief constituents of a gabbro, we may have an almost pure felspar-rock (labrador-rock, anorthosite) or pyroxene-rock 1 In many of the ' hypersthenites ' of the older writers the supposed hypersthene is only a highly schillerized diallage. 52 68 FELSPARS OF GABBRO. (diallage-rock, etc., pyroxenite 1 of Williams). By the dis- appearance of the pyroxene of an olivine gabbro, we have the so-called t.rnc.tnlit.p. (Ger. Forellenstein), composed essentially of felspar and olivine : with abundant olivine and diminishing felspar we have transitions to the succeeding family of peridotites. The name hornblende-gabbro has been used for rocks of this family which contain hornblende in addition to pyroxene, or in which original pyroxene is more or less completely replaced by hornblende 2 . When the conversion is complete we have no decisive criterion for verifying the derivative nature of the hornblende, and, as already remarked, the dis- tinction between diorite and gabbro is a somewhat artifk-i.-il oneT A historical account of the classification of the gabbros and allied rocks has been given by Bay ley 4 . Constituent minerals. The felspar of the gabbros and norites ranges in different examples usually from labra- dorite to anorthite. It builds large irregularly-shaped plates with, as a rule, rather broad lamellae 5 (albite twinning) often crossed by fine pericline-striation. The lamellae not infre- quently have something of a wedge-shape 6 . A crystal with broad albite lamellae, if cut nearly parallel to the brachy- pinacoid, may appear untwinned. It is not safe to assume that the most constant twin-lamellation necessarily corresponds with the albite law : the felspar of some rocks of this family has pericline-twinning alone or predominant (Skye, etc.). Zonary structure is typically not found. Besides fluid- pores and inclusions of earlier products of crystallization, the felspars often shew more or less marked schiller-structure 7 1 Amer. Geol, (1890) vi, 40-49. Williams regards his pyroxenites as a group coordinate with the peridotites. The name is ill-chosen, having been employed in two or three other quite different senses. 2 R. D. Irving, Copper-bearing Rocks of L. Suj>erior, 56-58, PI. vn. 3 Prof. Cole restricts the name gabbro to the olivine-bearing (corre- sponding roughly with the basic) division, and styles the intermediate felspar-augite-rocks ' augite-diorite.' 4 Journ. of Geol. (1893) i, 435-456. 5 Cohen, PI. xxix, fig. 3. Ibid. PI. xxx, fig. 2. 7 Ibid. PI. i, fig. 2. AUGITE OF GABBRO. 69 (fig. 1, g). The modes of alteration of the felspars are various : Rosenbusch notes the curious fact that calcite is seldom formed. The ' saussurite ' change seems to belong to dynamic metamorphism rather than weathering (see below, Chap. XXI.). Any plagioclase more acid than labradorite is exceptional, and so is the occurrence of orthodase (e.g. Carrock Fell, Lake Superior region 1 ). The augite of the gabbros builds irregular crystal -plates and wedges of very pale green or light brown colour. Besides the usual prismatic cleavage, an orthopinacoidal cleavage and diallage-structure are very common 2 . Instead of this, there is sometimes a very minute striation parallel to the basal plane (e.g. Skye, Carrock Fell). The common twin parallel to the orthopinacoid is often associated with this (fig. 13 A). FIG. 13. PYROXENES IN THE GABBRO OF CARROCK FELL, CUMBERLAND; x 20. The dominant variety is an augite with basal striation. A shews this structure combined with twinning on the orthopinacoid to give the ' herring-bone ' structure. The mineral is partly converted to green hornblende [1870]. B shews a parallel intergrowth of the augite with enstatite, the latter mineral forming the core and the former the outer shell, but with detached portions of augite enclosed in the enstatite in micrographic fashion [2279]. 1 E. D. Irving, Copper-bearing Rocks of L. Superior, 50-55, PI. v, vi. 3 Cohen, PL xvi, fig. 3. 70 HYPERSTHENE OF GABBRO AND NORITE. Decomposition gives a scaly or fibrous aggregate of chlorite and serpentine with other products. Another common altera- tion is the conversion to hornblende 1 , which may be light green and fibrous (uralite) or deep brown and compact. The rhombic pyroxenes, bronzite and hypersthene, occur as accessory minerals in rather rounded but allotriomorphic crystals, while in the norites they often shew but little crystal- outline. A schiller structure is common in many norites and gabbros (fig. 14). The most usual alteration is into distinct pseudomorphs of the serpentinous mineral bastite. This is pale green or yellowish with slight pleochroism and low polarization-tints. The pseudomorph is built of little fibres arranged longitudinally, and is traversed by irregular cracks op FIG. 14. NORITE (HYPEKSTHENITE), COAST OF LABRADOR ; x 20. Consisting of hypersthene (hy), felspar, (an), and apatite (ap). Schiller-inclusions are strongly developed in the hypersthene and to a less extent in the felspar [G 444]. which the fibres do not cross (see fig. 20). The individual fibres give straight extinction, but, as there is a slight departure from perfect parallelism in their arrangement, a > Cohen, PL xix, fig. 4. See G. H. Williams, A. J. S. (1884) xxviii, 261- 264 ; Bull. No. 28 U. S. Gtol. Sur. (1886) ; G. M. 1887, 87, 88 (Abstr.). OLIVINE OF GABBRO. 71 very characteristic appearance is offered. The rhombic pyro- xenes also shew uralitization. In the rocks here included original hornblende is found only as an occasional accessory : a deep brown variety occurs in some norites. Brown biotite may also occur as a minor accessory (e.g. Carrock Fell ; St David's Head), and it may be intergrown with augite (Stanner Rock) 1 . When olivine is present, it builds imperfect crystals or rounded grains, colourless in slices. Where it adjoins felspars, it is often bordered by a rim of hypersthene. The olivine sometimes has schiller-inclusions. The characteristic mode of alteration of olivine is ' serpen- tinization.' This process begins round the margin of the crystal-grain and along the, usually irregular, network of cracks which traverses it. Along these, as a first stage, strings of granular magnetite separate out. The immediate walls of the cracks are converted into pale greenish or yellowish fibrous serpentine, the fibres set perpendicularly to the crack, and giving straight extinction and low polarization-tints. At this stage the meshes of the network are occupied by unaltered remnants of olivine. These may be subsequently altered to serpentine, which is of a different character from that first formed, being often sensibly isotropic 2 . As a last stage, some of the magnetite may be reabsorbed, giving a deeper colour to the serpentine pseudomorph. The change from olivine to serpentine involves an increase of volume, which gives rise to numerous radiating cracks traversing adjacent minerals. These cracks are injected with serpentine, usually isotropic (fig. 15). Where original quartz occurs in gabbros, etc., it has the same properties as that in granites. Usually it forms part of a micrographic intergrowth. Original iron-ores occur only sparingly in some rocks of the gabbro family, but sometimes become abundant. They 1 Cole, G. M. 1886, p. 221, fig. 3. - This effect is possibly due to the overlapping of a crowd of minute fibres or scales without any definite orientation. For successive stages of serpentinization of olivine, see Geikie, p. 174, fig. 33 ; Cohen, PI. XLIV. 72 STRUCTURE OF GABBRO. are ilmenite (with leucoxene 1 as a decomposition-product) and magnetite. In some cases brown grains of picotite are found. FIG. 15. LABKADOKITE-OLIVINE-BOCK (TROCTOLITE), COVEBACK COVE, CORNWALL ; x 20. The olivine is almost wholly converted into serpentine (a few clear granules remaining), and the consequent expansion has caused radiating fissures through the surrounding felspar [1116]. The apatite builds the usual hexagonal prisms or sometimes short rounded grains (fig. 14). In other accessories the rocks are usually very poor, zircon and original sphene being absent and garnet uncommon. Structure. In texture the rocks of this family vary from medium to coarse grain. In some the individual crystals of felspar and pyroxene attain a large size, and they are then, as a rule, strongly affected by schiller-structures. Porphyritic structure is very rarely met with in the gabbrus and norites iSkvt- and Ardnamurchan). The order of crystallization is in general less decisively marked in basic than in acid rocks. This seems to be due to the periods of crystallization of the several minerals having in 1 For photographs of the leucoxenic alteration-product from various rocks, see Cohen, PI. XLV, figs. 3, 4. STRUCTURE OF GABBRO. 73 great measure overlapped. The relative idiomorphism of the crystals only indicates the order in which they ceased to form, not that in which the} 7 began. It is only with this under- standing that the rocks of the gabbro family can be said to follow the normal law. Apatite, iron-ores, and olivine, when present, are the earliest minerals and are clearly idiomorphic, while in the special types containing orthoclase and quartz these minerals have always crystallized last. But as regards the two main constituents^ auprite and plagioclasgj the mutual relations Me not always the same. In many gabbros the felspar is more or less distinctly embraced by the augite or diallage, but if this character becomes marked there are always other features which indicate a transition to the diabase type. The more typical gabbros are often thoroughly hypidiomorphic ; or the augitic constituent, especially if very abundant, may be embraced by the felspar. When a rhombic pyroxene enters, it is idiomorphic towards the monoclinic, and usually towards the felspar also. ho FIG. 16. OLIVINE-NORITE, SEILAND, NEAR HAMMERFEST ; x 15. A much-fissured crystal of olivine (ol) is surrounded by a continuous ring of hypersthene (hy) interposed between it and the anorthite felspar (/). There is a little brown hornblende (ho) and some brown biotite (b) clinging about the iron-ore grains [418]. 74 CORONA-STRUCTURE IN GABBRO. In many plutonic rocks there is an evident tendency for the earlier formed minerals to serve as nuclei round which the later ones have crystallized. This tendency is most marked in basic and ultrabasic rocks. Thus in gabbros and norites the pyroxenes often form a more or less continuous ring or 'corona' round olivine or iron-ores (fig. 16). Bayley 1 , while noting this feature, further describes fibrous intergrowths of felspar and augite surrounding olivine or magnetite. These seem to be original, but in other cases there is reason to believe that a mineral bordering another one is of secondary origin. Good examples are figured and described by G. H. Williams 2 in the hypersthene-gabbros of the Baltimore district. Here both hypersthene and diallage are surrounded by a double 'reaction-rim' of hornblende, interposed between the pyroxene and the felspar and due to a reaction between them. The inner zone of the rim is of fibrous, the outer of compact horn- blende. They are apparently the beginning of a process by which the pyroxenes are eventually wholly transformed into green hornblende, and the author named considers that they do not necessarily imply dynamic metamorphiam. In many cases there seems to be no decisive evidence as to the primary or secondary origin of the interposed minerals 3 . Gabbros with granulitic structure occur in many districts, sometimes in intimate association with those of more normal type 4 . Most of the rocks styled pyroxene-granulites probably 1 Amer. Journ. Sci. (1892) xliii, 515-518; Journ. of Geol. (1893) i, 702-710. 2 Bull. No. 28, U. S. Geol. Sur. (1886), with plates. 3 The following associations of minerals have been noticed in various districts. Augite; brown hornblende; felspar (Adirondacks). Olivine or magnetite; diallage or augite; felspar (Minnesota, etc.). Olivine; rhombic pyroxene ; actinolite ; felspar (Canadian anorthosites ; gabbros, New Brunswick). Olivine ; anthophyllite, actinolite, felspar (Lizard). Magnetite, biotite, felspar (New Hampshire, Minnesota, etc.). Pyroxene or magnetite ; garnet ; felspar (Canadian anorthosites). Magnetite or olivine ; brown hornblende ; garnet ; felspar : Magnetite ; biotite ; horn- blende; garnet; felspar: Magnetite; hornblende; quartz; garnet; felspar (Adirondacks). Hornblende; epidote; felspar (Baltimore). See Kemp, Bull. Geol.Soc. Amer. (1894) v, 218-221; Matthew, Trans. N.Y. Acad. Sci. (1894) xiii, 198-201 ; Hawes, Min. and Lith. of N. H. (1878) 205. 4 E.g. Druim.an-Eidhne in Skye, Geikie and Teall, Q. J. G. S. (1894) 1, 645-659 : N.E. Minnesota, Bayley, Journ. of Geol. (1895) iii, 1-16. QUARTZ-GABBROS. 75 fall under this head, but we defer noticing these until a later chapter (Chap. XXII.). Leading" types. We begin with the rather exceptional rocks in which free silica has been developed as an original constituent. A good example of a quartz-gabbro is that of Carrock Fell, in Cumberland 1 . It consists mainly of a some- what basic labradorite and an augite with basal striation. Imperfect prisms of enstatite also occur, and there is often a parallel intergrowth of the two pyroxenes (fig. 13 B}. The augite is often converted into a greenish fibrous hornblende and the enstatite into bastite. Biotite is found locally. Mag- netite and ilmenite occur, sometimes in evident intergrowths. Quartz is found partly in interstitial grains but chiefly in micrographic intergrowth with felspar, some of which is orthoclase. The rock varies much, the central part of the mass being rich in quartz, while the margin is highly basic, free from quartz and remarkably rich in iron-ores and apatite. The mutual relations of the felspar and augite vary, but on the whole the augite tends to envelope the felspar. Specimens of the gabbro of St David's Head, also intrusive in Lower Palaeozoic strata, are identical with the rock just described, except that the highly basic modification is not found. Biotite is rather more plentiful, and the quartz and micropegmatite occur rather more sparingly. The rhombic pyroxene is represented by pseudomorphs of pleochroic green bastite, always abundant. The mutual relations of the augite and labradorite vary, even in one slide : very fre- quently the former mineral is moulded on, or embraces, the latter. This tendency to the ' ophitic ' structure, together with the absence of diallagic structure in the augite, the t rather abundant occurrence of iron-ores, and other features, indicates an approach to the diabase type, which is also de- veloped in the district. It is noteworthy that the diabase of the Whin Sill, to be noticed below, in its coarse-grained central part, takes on characters almost indistinguishable from those of the Carrock Fell and St David's Head rocks. i Q. J. G. S. (1894) 1, 316-318, PL xvn ; (1895) li, 125. The rock has been termed hypersthenite, but the rhombic pyroxene is always subordi- nate to the mouoclinic and sometimes wanting. 76 GABBROS OF THE LIZARD. Hypersthene-bearing quartz-gabbros are extensively de- veloped near Wilmington, Delaware 1 . Some varieties have biotite, and by increase in the proportion of this mineral pass into the type which Chester styles 'gabbro-granite.' In other varieties brown hornblende becomes a conspicuous mineral, but this is probably formed at the expense of the diallage. The well-known rocks of the Lizard district 2 in Cornwall are, for the most part, simple gabbros without olivine, although that mineral occurs in some varieties. Judging from the cases in which precise determinations have been made, the felspar seems to be labradorite in the less basic rocks, anorth- ite in the most basic. It shews broad albite-lamellse, often crossed by others following the pericline law. The pyroxene varies from a pale green diopside, almost colourless in slices, to typical diallage, the diallagic structure being often seen to affect only part of a crystal. The enstatite-group is wanting or rare. When olivine occurs, it builds colourless grains shewing various stages of serpentinization. The Lizard gabbros exhibit, however, numerous modifica- tions which are ascribed to dynamic rnetamorphism, especially the conversion of the felspars to 'saussurite' and of the augite to amphibole. The minutely granular mineral-aggregate known as saussurite is opaque in any but the thinnest slices, and can be studied only under high magnifying powers. The change may be seen to begin in spots in the felspar crystals and spread to the whole. The pyroxene passes over into uralitic or actinolitic or compact hornblende in different cases 3 , the secondary amphibole being pale green or brown or colourless, or sometimes having a bright emerald-green colour (smarag- dite). According as one or both of these changes have affected the original felspar-pyroxene-rock, we have saussurite- diallage-gabbro, felspar-hornblende-gabbro, or saussurite-horn- blende-gabbro. At Karakclews occurs a rock consisting of a fine-grained aggregate of augit? (malacolite), labradorite, sphene, and an unknown substance, brown by transmitted and white 1 Chester, Bull. No. 59, U. S. Geol. Sur. (1890). 2 Teall, G. M. 1886, 483-485. For description of particular varieties see Bonney, Q. J. G. S. (1877) xxxiii, 884-915, and other papers. 3 Teall, PI. xvin, fig. 2. AMERICAN GABBROS. 77 by reflected light. Mr Teall 1 states that much of the so-called saussurite of the Lizard is similar to this rock in composition. Another mineral considered to be of secondary origin is the rhombic amphibole anthophyllite 2 . This sometimes occurs in colourless and rather fibrous crystals forming a zone round grains of altered olivine, and surrounded in turn by an outer zone of green actinolite. Gabbros without olivine are met with in Canada, New Hampshire, and other parts of America. Some from the north-western part of the Adirondacks, N.Y. 3 , consist essen- tially of felspar, in general labradorite, and augite,, often transformed to compact hornblende. Usually the ferro-mag- nesian mineral predominates, but there are rapid transitions to a highly felspathic type. Other gabbros in the same district have accessory hypersthene. An example from Iron Mountain, in the Laramie Hills, consists mainly of schillerized labradorite with some diallage and titaniferous iron-ore 4 . Gabbros without, as well as others with, olivine are largely developed in the Lake Superior region and the neighbouring parts of Minnesota, Wisconsin, etc. 5 An interesting type is the orthoclase-gabbro of Irving , in which the plagioclase felspar is oligoclase or an allied variety, and some orthoclase occurs in addition. The augite may be diallagic and is often uralitized ; apatite is abundant ; and the iron-ore is a highly titaniferous magnetite (Duluth and Lester River, Minn., etc.}. Among rocks which have been styled hornblende-gabbro, some examples from Guernsey (Bellegreve) exhibit very beautifully the conversion of colourless augite into brown or greenish brown compact hornblende, the process being seen in every stage. In some slides no augite remains, and, without comparison with other specimens, the rock might 1 M. M. (1888) viii, 118. 2 Teall, ibid. 119. Smyth, Butt. Geol. Soc. Amer. (1895) vi, 2.63-284. * Zirkel, Micro. Petrogr. Fortieth Parallel (1876) 107, 108. 6 Wadsworth, Prelim. Descr. of the Perid., Gabbros, etc., of Minn. (1887) ; R. D. Irving, Copper-bearing Rocks of L. Superior, Monog. No. 5, U. S. Geol. Sur. (1884). 6 R. D. Irving, I.e., 50-56, PI. v, vr. 78 SCOTTISH OLIVINE-GABBROS. be taken for a true diorite, but the hornblende is probably all derivative. The ferro-niagnesian silicates are often moulded on the felspar, which is of a basic variety. Magnetite and apatite are the only other constituents. Another good illustration is furnished by the rocks styled 'gabbro-diorite' by Williams, in the Baltimore district 1 . These have been originally hypersthene-bearing gabbro, and the transformation of both pyroxenes into green hornblende, fibrous or compact, can be traced step by step. The process is equally well displayed in some of the Cortlandt norites 2 . A good hornblende-gabbro, with compact brown hornblende, occurs near Bad River, Wisconsin 3 . The Tertiary igneous rocks of the Inner Hebrides (Skye, Rum, Mull, etc.) include numerous olivine-gabbros, and accord- ing to Prof. Judd 4 most of the gabbros there carry olivine, though that mineral may be obscured by secondary magnetite. The augite, as a rule, has a striation parallel to the basal plane with more or less marked schillerization ; but the author named has shewn how in the more deep-seated portions of the large rock-masses schiller-structures come in in more than one direc- tion, and affect the felspar and olivine as well as the pyroxene. A rhombic pyroxene is only locally present. The olivine is often of a variety rich in iron and gives rise to much magnetite-dust as an alteration-product. Original iron-ores and apatite may or may not be present. The felspar is usually a labradorite, and this, rather than the pyroxene, tends to assume crystal outlines, the structure of the rock being often ophitic, and the gabbro graduating into diabase. Very similar to the Scottish Tertiary gabbros are those of the Carlingford district in Ireland, probably of like age. Prof, von Lasaulx 5 described specimens consisting of anorthite, diallage, and olivine, and likened them to the gabbro of Store Bekkafjord in Norway. These were from Slieve Foy. From 1 Bull. No. 28, U. S. Geol. Sur. (1886) with plates : abstr. in G. M., 1887, 87, 88. See also Hobbs, Trans. Wise. Acad. Sci. (1890) viii, 157-159. 2 G. H. Williams, A. J. S. (1884) xxviii, 261-264, with figures. 3 E. D. Irving, I.e., 56-58, PI. vn. figs. 1-3. 4 Q. J. G. S. (1886) xlii, 49-89, PI. iv. 5 Sci. Proc. Roy. IhM. Hoc. (1878) ii, 31-33. OLIV1NE-GABBROS. 79 the neighbouring hill of Barnavarve Prof. Sollas 1 describes a gabbro free from olivine, consisting of a basic felspar (anorthite or bytownite) with rhombic and monoclinic pyroxenes which shew rather remarkable intergrowths. Here is also a variety of the rock containing interstitial micro-pegmatite, which the author named believes to be due to a later injection 2 . At Stanner Rock, near New Radnor, occur gabbros both with and without olivine. One of the latter shews parallel intergrowths of augite and biotite 3 . The olivine-gabbros described by Irving from the Lake Superior region tend to the ophitic type of structure. The felspar is usually anorthite or some other basic variety ; the augite sometimes, but not always, shews the diallage character; the iron-ore, often in large grains, is magnetite only slightly titaniferous ; and apatite is rare. A rock from Pigeon Point, Minn. 4 , consists of fresh labradorite, purplish pink titaniferous augite, olivine, titaniferous magnetite, and a little apatite. One modification contains large porphyritic crystals of the felspar. The large gabbro mass at the base of the Keweenaw formation in north-eastern Minnesota 5 consists essentially of a basic labradorite, augite, an olivine rich in iron (hyalo- siderite), and a non-titaniferous magnetite; but wide variations in the relative proportions of these constituents give rise to numerous special modifications. As already intimated, many of the rocks in this family contain both augite (or diallage) and hypersthene in varying proportions, and no hard line is to be drawn between gabbros and norites. In Sweden the rocks termed ' hyperite ' by Tornebohrn vary between olivirie-gabbro and norite, olivine and hypersthene appearing to replace one another, so that the total of the two remains about the same in the different varieties. The same thing is seen in the north of Norway and elsewhere. It is convenient to restrict the name norite to rocks in which the sole or dominant pyroxene is of a Trans. Roy. Ir. Acad. (1894) xxx, 482-486. L.c. 487, etc. Cole, G. M. 1886, 223-225, fig. 3. Bayley, Bull. No. 109, U. S. Geol. Sur. (1893) 32-38, PI. v. Bayley, Journ. of Geol. (1893) i, 696-714. 80 NORITES. rhombic variety, those in which the rhombic is only sub- ordinate to the monoclinic pyroxene being termed hypersthene- gabbro. Such rocks are represented among the gabbros of Skye and Mull. They occur also in the Baltimore district 1 , in the Adirondacks 2 , in Alabama 3 , etc. A good example of quartz-norite is described by Grant 4 from Mount Hope, near Baltimore. It consists of bytownite, quartz, and hypersthene, with accessory magnetite and apatite, and has a granitoid structure. A well-known example of norite comes from the island Hittero, off the west coast of Norway. The rhombic pyroxene is a hypersthene rich in iron, but, as is often the case, the ferriferous ingredient is concentrated in numerous deep brown schiller-inclusions, leaving the general mass of the crystal pale and scarcely pleochroic. Some specimens have a considerable amount of iron-ore (probably titaniferous) surrounded by green hornblende. The same strongly schillerized hypersthene is well exhibited by the norites of the Labrador coast 5 (fig. 14). Patches of brown hornblende and biotite are sometimes intergrown with it. In places the hypersthene becomes bleached, with a separation of granular magnetite. The other main constituent is felspar (usually typical labradorite but sometimes a more basic variety), moulded on the imperfect crystals of hyper- sthene. Stout prisms of apatite also occur, and sometimes patches of iron-ore bordered by brown mica. In Britain typical norites occur in Aberdeenshire (near Ellon) and Bantf'shire. Norites of several varieties are included among the Cortlandt rocks on the Hudson River 6 . The norite proper consists mainly of andesine and hypersthene, both shewing schiller-inclusions. There is accessory biotite, and a curious G. H. Williams, Bull. No.28, U. S. Geol. Sur. (1886). C. H. Smyth, jr., A. J. S. (1894) xlviii, 54-65 ; Bull. Geol. Soc. Amer. (1895) vi, 271. J. M. Clements, Bull. No. 5, Geol. Sur. Ala. (1896) 171, 172. Joh. Hopk. Univ. Giro. (1893) xii, 48. Cohen, PI. r, fig. 3. G. H. Williams, A.J.S. (1887) xxxiii, 135-144, 191-194. THERALITES. 81 feature is the occurrence of large crystals of orthoclase enclosing the other minerals in ' pcecilitic ' fashion. In other rock-types from this district the hypersthene is associated with green or brown hornblende (hornblende-norite), with biotite and magnetite (mica-norite), or with augite and biotite (augite^norite). Another rock, intermediate between norite and gabbro, is the hypersthene-gabbro described by the same author from Baltimore. This rock consists of bytownite, diallage, and hypersthene, with some magnetite and apatite. It shews the 'reaction-rims' already referred to, and passes over into a ' gabbro-diorite ' or hornblende-gabbro, in which the hornblende is derivative from the pyroxenic minerals. In this place may be included the rocks to which Rosenbusch has given the name theralite, and which may be considered as nepheline-gabbros. The original type is from the Crazy Mts in Montana 1 . Here olivine is only an occasional accessory. Hornblende is not present in the typical rock, but the idiomorphic augite, pale green to almost colourless in slices, is often surrounded by a narrow border of deep green pleochroic segirine. The felspar is partly unstriated plagio- clase, partly perhaps anorthoclase. It forms with nepheline a granular aggregate, in which either mineral' may be idiomorphic towards the other. The remaining constituents are biotite, apatite, and a little iron ore, with sometimes sodalite (Rock Creek) or haiiyne (Martinsdale). A purer theralite is described from Costa Rica 2 . This consists of augite, labradorite and a little orthoclase, nepheline and a mineral of the sodalite group, biotite, apatite, and magnetite, with secondary zeolites. A more remarkable rock is Pirsson's missourite 3 from the Highwood Mts, Montana, a leucite-gabbro, corresponding with the volcanic leucite-basalts. It is quite devoid of felspar, consisting of olivine, augite, biotite, and leucite, with some apatite and iron-ore. The structure is thoroughly allotrio- morphic. 1 Wolff, Notes on the Petrography of the Crazy Mts, etc. Northern Transcontinental Survey (1885). 2 Wolff, A. J. S. (1896) i, 271, 272. 3 A. J. S. (1896) ii, 317-323. H. P. 6 82 FELSPAR- AND PYROXENE-ROCKS. The felspar-rocks known in America as anorthosite must be regarded as peculiar members of the gabbro family. Such rocks, of pre-Cambrian age, occupy extensive tracts in Minne- sota 1 , etc., near Lake Superior. The felspar which makes up almost the whole of these coarse-textured aggregates varies from labradorite to anorthite in different localities. A little augite, of faint violet-brown tint in sections, is the only other original mineral, and this occurs both in grains and as minute parallel interpositions in the felspar. Similar rocks have been described by Adams 2 in the so-called Norian of several districts in Canada, by Kemp 3 in the Adirondacks, etc. In our country gabbros pass only locally into labradorite-rocks by the failure of the pyroxenic constituent (Lenkeilden Cove at the Lizard, Athenree in Tyrone 4 ). Contrasted with these are the pure pyroxene-rocks to which Williams in America has given the name ' pyroxenite.' The Webster type 6 is described from North Carolina and Maryland, and consists of a rhombic and a monoclinic pyroxene forming an even-grained crystalline aggregate. It is in fact a bronzite- diopside-rock. Another example, from Montana 6 , consists of light green diallage and colourless enstatite with some brown mica and only occasional felspar. From the same district comes a hypersthene-hornblende-rock, sometimes rich in green pleonaste, and from Alabama 7 an augite-hornblende-rock. Some British gabbros pass locally into augite- or diallage-rock (Lendalfoot in Ayrshire 8 ). By the dwindling and disappearance of the pyroxene, oliv- ine-gabbros pass into felspar-olivine-roc/c, known as troctolite 1 R. D. Irving, Copper-bearing Rocks of L. Superior, 59-61, PI. vn, fig. 4 ; Lawson, Bull. No. 8 Geol. and Nat. Hist. Sur. Minn. (1893) and abstr. in M. M. x, 263. The very coarse-textured felspar-rock of Labrador, with its beautiful schiller- structure, is in all mineralogical collections. Rep. Brit. Ass. for 1886, 666, 667. Bull. Geol. Soc. Amer. (1894) v, 215, 216 ; Geology of Moriah and We tport, Bull. N. Y. State Hus. (1895) iii, 337. Watts, Guide, 73. G. H. Williams, Amer. Geol. (1890) vi, 40-49, PI. n, fig. 2 (Abstr. in M. M. ix, 250, 251). 6 Merrill, Proc. U. S. Nat. Mm. (1894) xvii, 662, 657, 658. 7 Clements, Bull. No. 5 Geol. Sur. Ala. (1896) 163, 164. s Bonney, Q. J. G. S. (1878) xxxiv, 778-780. TROCTOLITES: IRON-ORE-ROCKS. 83 (Ger. Forellenstein). This consists essentially of labradorite or some more basic felspar, often anorthite, with a smaller proportion of olivine, which may be more or less serpentin- ized (fig. 15). Such rocks are known in the gabbro-district of the Lizard 1 and among the Tertiary intrusions of the Scottish islands. Prof. Judd 2 describes a fresh and rather tine-textured anorthite-olivine rock from Halival in Rum. Another example is described by Prof. Bonney 3 from Belhelvie in Aberdeenshire. This and the Cornish specimens have some small amount of diallage. In America troctolites have been noted in Minnesota 4 and other areas of gabbro-rocks. It has been noticed above that an ordinary gabbro may pass into a variety very rich in magnetite and ilmenite (e.g. Carrock Fell). Some gabbros and norites, in Scandinavia, in Minnesota 5 , in the Adirondacks B , etc. shew very basic modifica- tions which are almost pure iron-ore-rocks 7 . As a rule, they are highly titaniferous. An augite-magnetite-rock, consisting of crystal-grains of augite set in a framework of titaniferous magnetite, is one of the varieties of the curious banded gabbros of Skye 8 . 1 Teall, PI. vin, fig. 2. 2 Q. J. G. 8. (1885) xli, PI. xin, fig. 5. :i G. M. 1885, 441, 442. 4 Wadsworth, Prelim. Descr. of the Perid., Gabbros, etc., of Minn. (1887) 95, PI. v. 5 Ibid. 63, 64, PI. vi, fig. 1 ; Irving, Copper-bearing Rocks L. Superior, 51, 52; Wiiichell, 10th Ann. Rep. Minn. Geol. Sur. (1882), 80-83. 6 Kemp, Bull. Geol. Soc. Amer. (1894) v, 222. 7 Vogt, G.M. 1892, 82-86 (Abstract). For descriptions of iron-ore- rocks from Cumberland in Rhode Is. and Taberg in Sweden see Wads- worth, Bull. Mm. Camp. Zool. Harv. (1881) vii, 185-187; Lith. Stud. 75-81, PI. i, ii. 8 Geikie and Teall, Q. J. G. S. (1894) 1, PI. xxvm. 62 CHAPTER VI. PERIDOTITES (INCLUDING SERPENTINE-ROCKS). THE peridotites are holocrystalline rocks of ultrabasic composition, in which felspar is typically absent and olivine is the most prominent constituent. They were separated from the more normal basic rocks by Rosenbusch, but, though their marked characters make it desirable to discuss them apart, they do not constitute a family comparable, e.g., with that of the gabbros in importance. The peridotites do not usually occur in large bodies of uniform rock. In many localities they are seen to be only local modifications of olivine-gabbros, olivine-norites, or olivine-diorites, and they shew frequent transitions from one type to another. For so small a group a needless multiplicity of names has been created. The simple olivine-rock is the ' dunite ' of Hochstetter. "With the addition of enstatite we have the 'saxonite' of Wadsworth 1 , ' harzburgite ' of Rosenbusch; other types are styled 'Iherzolite,' 'eulysite,' etc., and the name ' picrite ' is used for those characterized by augite or horn- blende, usually with some felspar. For our purposes it will be sufficient to separate the picrites, rich in the bisilicate constituents and having usually subordinate plagioclase, from the more typical peridotites, very rich in olivine and non- felspathic. Different types may be specified by prefixes in the customary way (e.g. hornblende-picrite, enstatite-peridotite, etc.}. i Lithological Studies (1884, Camb. Mass.). This work contains many descriptions of peridotites and meteorites, with a number of useful coloured plates. MINERALS OF PERIDOTITE. 85 Many of the meteorites ('stony meteorites' as distin- guished from meteoric irons) have a mineral composition allied to that of the terrestrial peridotites, but often with special accessory minerals and peculiar structures. In consequence of the unstable nature of their principal constituent mineral, the peridotites are very readily decom- posed, and most of the serpentine-rocks have originated in this way. Constituent minerals. In the typical peridotites oiivine makes up from half to nearly the whole of the rock. If not so abundant that its crystals interfere with one another, it builds idiomorphic or rounded crystals. The mineral is colourless in thin slices, and shews either irregular cleavage-traces or a network of fissures. It often has schiller- inclusions of the nature of minute negative crystals enclosing dendritic growths of magnetite (fig. 1 h). Alteration along cracks gives rise to strings of magnetite granules, and complete destruction produces pseudomorphs of greenish or yellow ser- pentine, or sometimes colourless fibrous tremolite, etc. Of the other ferro-magnesian silicates the commonest in typical peridotites is a rhombic pyroxene ; either colourless or pale yellow (enstatite) or with faint green and rose pleochroism (bronzite) : varieties rich in iron do not often occur. The crystals often tend to be idiomorphic. Any marked schiller- structures are not very common. Decomposition results in pseudomorphs of bastite 1 . The augite is either light brown to colourless, with a high extinction-angle (about 40) as in many diabases, etc., or it may shew a faint green tint (chrome- diopside). A conversion to brown hornblende is common in the picrites, and so also are parallel growths of augite and brown hornblende, the former being the kernel. The hornblende may be a green or pale actinolitic variety, but in many of the picrites it is 'basaltic' hornblende with an extinction-angle of about 20 and colour varying from deep brown to colourless. The pale variety seems due to bleaching, often accompanied by a discharge of magnetite dust. The biotite of peridotites is also frequently of a pale tint. 1 Fouqu6 and Lvy, PI. LIU, LIV. 86 MINERALS OF PERIDOTITE. Some peridotites have little octahedra of magnetite, but some other spinellid mineral is more characteristic. It may be chromite (deep brown or opaque), picotite (coffee-brown), or pleonaste (green). These minerals usually build irregular rounded grains. In some of the rocks perofskite is a charac- teristic mineral, in minute crystals 1 . A basic felspar occurs in many of the picrites, but is wholly wanting in the more typical peridotites. Some types have accessory garnet, which is always the magnesiaii variety pyrope, red-brown in slices. Metallic nickeliferous iron occurs in some of the meteoric peridotites, besides special minerals, such as troilite. FlG. 17. PCECILITIC STRUCTURE IN HORNBLENDE-PICRITE, MYNYDD PENARFYNNYDD, CAERNARVONSHIRE ; x 20. The large plate enclosing olivine-grains and filling the field is a single crystal of hornblende. It is mostly colourless, but becomes deep brown in capriciously arranged patches round the edge [725]. * Structure. The constituents follow, as a rule, the normal order of crystallization, the olivine constantly preceding the bisilicates. In many picrites, and in other types not too rich in olivine, the more or less rounded crystals of olivine are 1 Cf. G. H. Williams on the serpentine of Syracuse, N. Y., A. J. S. (1887) xxxiv, 140-142. STRUCTURES OF PERIDOTITE. 87 enclosed by large plates of pyroxene or hornblende (pcecilitic structure 1 , fig. 17). When felspar occurs, it is later than the pyroxenes, but in the hornbleiide-picrites it is often moulded in ophitic fashion by part of the hornblende. In the most basic peridotites the largely predominant olivine builds a granular aggregate, in which may be im- bedded, with a pseudo-porphyritic appearance, relatively large FIG. 18. ENSTATITE-PEKIDOTITE WITH PSEUDO-PORPHYRITIC STRUCTURE, SKUTVIK, NEAR TROMSO, NORWAY ; x 20. Here olivine (o) is largely in excess, forming a granular aggregate in which are embedded large irregular crystals of a yellowish partly altered enstatite (e) [440]. crystals of enstatite, etc. (fig. 18). Any true porphyritic structure (i.e. some constituent occurring in two distinct generations) is rare in this family of rocks, the minerals usually forming an even-grained aggregate. The pyrope-bearing peridotites often shew a special type of structure, each garnet-crystal being surrounded by a broad border or shell known as celyphite* (Ger. Kelyphit). This 1 This is quite analogous to the ophitic structure of diabases, etc. See G. H. Williams, A. J. S. (1886) xxxi, 30, 31; Joum. of Geol. (1893) i, 176. 2 Kosenbusch-Iddings, PL xiv, fig. 4. 88 HORXBLEXDE-PICRITES. border is sharply divided from the garnet, and possesses a marked radial fibrous structure. The name is not applied to any particular mineral, and the so-called celyphite is not always of the same constitution. A pale or colourless augite is common, while brown hornblende and enstatite are some- times found, and brown picotite frequently accompanies the pyroxene. Again, brown biotite and magnetite have been observed 1 . A celyphite-border round garnet is also a charac- teristic feature in pyroxene-garnet-rocks (eclogites). Some petrologists have regarded it as a secondary ' reaction-rim, ; but there seems to be no decisive reason for rejecting the primary origin of the growth. Most of the meteoric peridotites have a peculiar structure termed chrondritic 2 . A fine-grained matrix of olivine, enstatite, chromite, etc., encloses numerous round grains (chondri) con- sisting of the same minerals. In these chondri the crystals very commonly have a tendency to diverge from a point on the circumference. Leading types. Numerous examples of rocks rich in olivine are known from the old gneiss area of Sutherland, from the western islands of Scotland, from North Wales. Cornwall, etc. There are frequent transitions from felspar- bearing picrites to thoroughly ultrabasic peridotites 3 . At Penarfynnydd 4 , on the south-west coast of Caernarvon- shire, is an Ordovician intrusion ranging from hornblende- picrite to a hornblende-peridotite very rich in olivine. The hornblende is either deep brown or colourless, in the same crystal, and it encloses the rounded grains of olivine with typical pcecilitic structure (fig. 17). A colourless augite and a deep brown biotite occur, with a little original mag- netite. Part of the hornblende is formed at the expense of augite. Anorthite is often present, usually embraced by the 1 Ciller, A. J. S. (1886) xxxii, 123 : Bull. No. 38 U. S. Geol. Sur. (1887) 15-17. 2 For figures see Wadsworth's ' Lithological Studies ' ; Lockyer, Nature (1890) xli, 306, 307. 3 For figures of several of these rocks, see Teall. 4 Q. G. J. S. (1888) xliv, 454-457. 'Bala Vole. Rocks of Caern.' 99- 101. HORNBLENDE- AND AUGITE-PICRITES. 89 hornblende. Similar rocks occur in central Anglesey, where secondary crystal-outgrowths from the hornblende are fre- quent 1 . Prof. Bonney 2 has described some of these rocks, which occur as boulders on the west coast of Anglesey. The same writer has described from Sark 3 a somewhat different type in which a pale altered mica is a prominent mineral, besides pale or greenish actinolite. This seems then to be a mica-hornblende-picrite, and Prof. Bonney compares it with the Scye type mentioned below. G. H. Williams 4 has given an interesting account of hornblende-picrites from the Cort- landt district on the Hudson River. They resemble very closely the British examples and a well-known rock from Schriesheim, near Heidelberg, the bleaching of the brown hornblende and subordinate brown biotite being a character- istic feature. "Williams used the name ' cortlandtite ' for these rocks, and they may conveniently be styled the Cortlandt type. Examples from Alabama 5 have either brown or very pale green hornblende, and contain abundant pleonaste. One from Montana 6 has accessory hypersthene. An augite-picrite of Carboniferous age is found at Inch- colm 7 , near Edinburgh, in which the dominant coloured mineral is a purplish-brown pleochroic augite, often with hour- glass structure 8 . Deep brown hornblende is also present, chiefly as a marginal intergrowth with the augite. Felspar and biotite are subordinate. Most of the olivine is con- verted into a yellow serpentine. Augite-picrites with typical pcecilitic structure occur in Shropshire 9 . Among examples from the Inner Hebrides Prof. Judd 10 notes one from the 1 Teall, PI. vi. 2 Q. G.J. S. xxxvii (1881) 137-140; xxxix (1883) 254-259. Also a similar rock from Alderney, ibid. (1889) xlv, 384. 3 G. M. 1889, 109-112. 4 A. J. S. (1886) xxxi, 31-37. 5 Clements, Bull. No. 5 Geol. Sur. Ala. (1896) 155-160. Merrill, Proc. U. S. Nat. Mm. (1894) xvii, 654. 7 Cole's Stud. Micro. Sci. (1882) No. 6 ; Teall, PL iv, fig. 2, vii. 8 The augite resembles that of some nepheline-dolerites, and the rock differs in other respects 'from true plutonic types. 9 Watts, Rep. Brit. Ass. for 1887, 700 ; Proc. Geol. Ass. (1894) xiii, 340, fig. 10 Q. J. G. S. (1885) xli, 393, PI. xm, fig. 4. 90 ENSTATITE-PERIDOTITES. Shiant Isles with fine pcecilitic structure. Busz 1 has described an augite-picrite with comparatively fresh olivine from High- week, near Newton Bushel, Devonshire : this has subordinate enstatite and biotite. Intrusions of enstatite-picrite occur in the old gneiss of the west of Sutherland. In one near Lochinver the slightly pleochroic enstatite or bronzite is moulded on the olivine, but shews good crystal-faces, being enclosed by large crystal-plates of felspar. There is a subordinate colourless augite and some brown hornblende, which is partly formed from the pyroxenes, partly original and later than the felspar. This rock is almost as much a norite as a picrite, but true enstatite-peridotites also occur in the district, consisting of about equal parts of olivine and a rhombic pyroxene, with grains of pleonaste (fig. 1 9).. Of mica-peridotite few examples are described. One from Elliot County, Kentucky 2 , consists of serpentinized olivine pi- pi FIG. 19. ENSTATITE-PERIDOTITE, ASSYNT LODGE, SUTHERLAND ; x 20. A granular aggregate of olivine (o), largely serpentinized, and a slightly pleochroic enstatite or bronzite (e). These two minerals are in about equal quantity : in addition there are little irregular grains of isotropic green pleonaste (pi) [1642]. 1 M. M. (1896) xi, 154, 155 (Abstr.) 2 Diller, A. J. S. (1892) xliv, 286-289. HORNBLENDE- AND AUGITE-PERIDOTITES. 91 and pale yellow-brown to colourless mica, with pcecilitic arrangement, besides crystals of perofskite, etc. Another example occurs in association with the gabbros of the Harz 1 . Prof. Judd 2 has described under the name ' scyelite ' a horn- blende-mica-peridotiie from the borders of Sutherland and Caithness (Loch Scye and Achavarasdale Moor). Here ser- pentinized grains of olivine are enclosed in pcecilitic fashion by a pale green to colourless hornblende, probably pseudo- morphous after cliallage, and a peculiar yellow mica. A similar rock is recorded at a point 2^ miles s. E. of Borgie Bridge. Prof. Sollas has remarked a hornblende-hypersthene-peridotite among the crystalline schists of Galway, at Derreennagusfoor. This consists of hypersthene, hornblende, olivine, magnetite, and a green spinel. Such rocks occur also in Ouster County, Colorado 3 , and in other localities 4 . Among hornblende-peridotites we may place the rock de- scribed as a hornblende-picrite from Grey stones in Wicklow 5 , which is non-felspathic. The dominant hornblende is green, and encloses in pcecilitic fashion the olivine pseudomorphs (of magnetite and a carbonate). It has cores and borders of colourless hornblende, and there is a third variety of this mineral with few cleavage-cracks and much magnetite-dust. Various augite-peridotites have been described. Specimens of these, as well as augite-picrites, are represented among the Tertiary eruptives of western Scotland 6 . One from the Isle of Hum shews fresh olivine set in a framework of green augite. Magnetite and chromite are accessories, and some- times hypersthene. Merrill has described an augite-peridotite from Little Deer Island, Maine 7 , and a diallage-mica-peridotite from Montana 8 . 1 Koch, M. M. ix, 41, 42 (Abstr.). 2 Q. G. J. S. (1885) xli, 401-407. Teall, PI. v. fig. 2. 3 Cross, Proc. Colo. Sci. Soc. (1887) 242-244. 4 Manbhum in India ; see Holland, Rec. Geol. Sur. Ind. (1894) xxvii, 142 : Kilimanjaro; see Hatch, G. M., 1888, 257-260. 5 Watts, Sep. Brit. Ass. for 1893, 767 ; Guide, 35. 6 Judd, Q. G. J. S. (1885) xli, 389-395. 7 Proc. U. S. Nat. Mus. (1888) xi, 192-195. s Ibid. (1894) xvii, 651, 652. 92 ENSTATITE-PERIDOTITES. A well-known enstatite- augite-peridotite occurs in the Pyre- nees and Ariege (Lherz type) 1 . About two-thirds of the rock consists of fresh olivine, the other minerals being a colourless enstatite, a faint green to colourless chrome-bearing diopside, and irregular grains of either brown picotite or green pleo- naste. As usual in types very rich in olivine, the structure is granular, not poecilitic. Such rocks, often serpentinized, are recorded from several other districts. A porphyritic bronzite- diallage-peridotite occurs in Maryland 2 , and a similar rock in Colusa County, California 8 . In some enstatite-peridotites the rhombic pyroxene is abundant, and forms a framework in which the somewhat rounded grains of olivine are set with poecilitic structure. A well-known representative comes from the Harz (Baste or Harzhurg type) 4 , where, however, both minerals are more or less completely serpentinized. A similar rock is described from Presque Isle, Michigan 5 . In another type olivine largely predominates, and the enstatite occurs in relatively large crystals, which, among the smaller grains of olivine, give a pseudo-porphyritic appearance to the rock. Good examples occur near Tromso, etc., in Norway (fig. 18), Inyo County, California 6 , etc 7 . In Maryland, Williams 8 has described similar rocks in which large crystals of bronzite or diallage, or both, are imbedded in a granular mass, mainly of olivine. In Montana 9 occurs a peridotite with abundant crystals of bronzite and olivine enclosed in a finely granular aggregate of enstatite, pale hornblende, some olivine, and green pleonaste. 1 The Iherzulite of Delametherie. SeeBonney, G. J/., 1877, 59-64, and for coloured figures Teall, PL i, fig. 1 ; Fouque and Levy, PI. LII, fig. 1. G. H. Williams, Amer. Geol. (1890) vi, 38, 39, PI. n, fig. 1. Wadsworth, Lith. Stud., PI. v, figs. 1-3. Wadsworth, Ibid. 133, 134, PL vin, figs. 1,2,5; Fouque and Levy, PL LIII, fig. 2. Wadsworth, Lith. Stud., 136-138, PL vn, figs. 3-5. Ibid. 132, PL vi, fig. 4. One from New Zealand carries grains of nickel-iron alloy (awaruite). Ulrich, Q. G. J. S. (1890) xlvi, 625-629, PL xxiv. 8 Bull. No. 28 U. S. Geol. Sur. (1886) 50-55; Amer. Geol. (1890) vi, 38, 39, PL ii, fig. 1. 9 Merrill, Proc. U. S. Nat. Mus. (1894) xvii, 655. VARIOUS PERIDOTITES. 93 From these rocks it is only a step to one composed wholly of olivine, with only a little accessory picotite or magnetite. Of this pure olivine-rock the type comes from New Zealand (Mount Dun), and is the ' dunite ' of Hochstetter. Other examples may be named, e.g. from Kraubath in Styria, from Franklin and Webster in North Carolina 1 , from Chalk Hills, near Salem, Madras 2 , from St Paul's Rocks in mid- Atlantic 3 , etc., with others more or less serpentinized. Of garnet-peridotites that from Elliott County, Kentucky 4 , is a good example. The pyrope crystals are surrounded by a ' celyphite ' border of brown mica with an outer ring of magnetite-dust, these minerals being supposed to be due to a reaction between the garnet and the olivine. The serpentine- rock of Zoblitz in Saxony is another example, in which, however, the olivine is wholly destroyed. Garnet occurs as an accessory in the diallage-peridotite of Tunaberg in Norway" (the ' eulysite ' of Erdrnann) and in other localities. Serpentine -rocks. Hitherto we have noticed only very briefly the secondary changes that affect the minerals of crystalline rocks. In the present family, however, the de- composition of a rock is often so complete that its original nature is detected only by careful study, arid the altered rock- masses are commonly denoted by a special name, serpentine- rocks or simply serpentines, expressing their dominant mineral composition. The mineral serpentine is the commonest de- composition-product of the non- aluminous magnesian silicates (olivine, the rhombic pyroxenes, and some of the augites and hornblendes), and the purest serpentine-rocks result from the alteration of peridotites 6 . Other decomposition-products occur i Wadsworth, Lith. Stud., 118, 119, PI. iv, figs. 2, 3. - Middlemiss, Eec. Geol. Sur. Ind. (1896) xxix, 33. 3 Renard, Voyage of 'Challenger,' Narrative, vol. ii, Append. B, with plate. ' * Diller, A. J. S. (1886) xxxii, 121-125 ; Bull. No. 38 U. S. Geol. Sur. (1887). 5 Wadsworth, Lith. Stud., 147. 6 For descriptions of coloured figures of numerous serpentine-rocks, see Wadsworth, Lithological Studies (1884). For a general sketch of observations and opinions on serpentine, see Teall, Chap. vi. On serpentine from diopside, see Merrill, Proc. U. S. Nat. Mus. (1888) xi, 105-109, PI. xxxii ; A. J. S. (1889) xxxvi, 189-191. 94 VARIETIES OF SERPENTINE. in the rocks, viz. iron-oxides (magnetite and limonite), steatite, carbonates (dolomite, etc.), chlorite, and tremolite, but the bulk is serpentine of various kinds, in which may be found undestroyed relics of the original minerals of the perfdotite (olivine, diopside, pyrope, chromite, etc.). Of the mineral serpentine some kinds are crystalline and doubly refracting with the interference-colours of quartz or felspar and faint pleochroism when the green tint is sufficiently pronounced. The habit is fibrous (chrysotile) or scaly (antigorite, etc.). Other kinds are amorphous and sensibly isotropic. Much of the serpentine occurs in definite pseudomorphs, and often retains something of the structure of the parent mineral to indicate its source. We may dis- tinguish four cases : (i) Serpentine derived from olivine, with the ' mesh- structure 1 ' (Tschermak's ' Maschenstructur ' ; see p. 63 and fig. 20). (ii) Serpentine derived from enstatite or bronzite, in distinct pseudomorphs with the bastite-structure (see p. 62 and fig. 20). (iii) Serpentine derived from a non-aluminous hornblende, with 'lattice-structure 2 ' ('Gitterstructur ' of Weigand). Here the cleavage of the hornblende is marked by veins of birefring- ent serpentine in two sets making the characteristic angle 55|. This serpentine is minutely fibrous, with the fibres set perpendicularly to the cleavage of the hornblende. The rest of the pseudomorph is of serpentine, giving no definite crystal- line reaction and consisting probably of a confusedly fibrous aggregate. (iv) Serpentine derived from a non-aluminous augite, with ' knitted structure*' (' gesterickte Structur' of Hussak). This consists chiefly of serpentine with scaly habit (antigorite). The scales give straight extinction and low polarization-tints. They occur in two closely interlacing sets parallel to the cleavage-planes of the augite, and so making an angle of about 87 with one another. 1 Eosenbusch-Iddings, PI. xxvi, fig. 4. 2 Ibid. fig. 5. 3 Ibid. fig. 6. SOURCES OF SERPENTINE. 95 The derivation of serpentine from pyroxene is very clearly exhibited in some American occurrences described by Merrill at Montville, N". J. l and in Essex County 2 and Warren County 3 , N. Y. The source of serpentine in rocks can often be made out by these various characters, and it is placed beyond doubt when any unaltered remnants of the parent mineral remain. In addition there may be serpentine encroaching upon con- tiguous minerals or traversing them in veins : this is, as a rule, sensibly isotropic. FIG. 20. SEBPENTINE-ROCK, COVERACK, CORNWALL ; x 20. A large bastite-pseudomorph after bronzite is seen on the right. The rest of the rock is of serpentine with mesh-structure, derived from olivine : it is stained in places with hydrated iron-oxide [1118]. It is commonly believed that the mineral serpentine is in all cases a decomposition-product of other magnesian silicates. Recently, however, Weinschenk has maintained that it may occur under certain conditions as an original constituent of a peridotite. His Stubach type, from the Venediger district in the East-Central Alps, consists essentially of olivine and 1 Proc. U. S. Nat. Mus. (1888) xi, 105-111, PI. xxxn. 2 Ibid. (1889) xii, 595-599. 3 A. J. S. (1889) xxxvii, 189-191. 96 SERPENTINE-ROCKS OF THE LIZARD. serpentine, intergrown in crystallographic relation, and he believes both minerals to be formed from igneous fusion 1 . The best-known serpentine-rocks in this country are those of the Lizard district in Cornwall 2 . The purer examples consist essentially of serpentine of various kinds, secondary iron-ore (often peroxidized), steatite, tremolite, etc., and often undestroyed relics of olivine or other original minerals of the peridotites. Professor Bonney has shewn that much of the serpentine has the character of that derived from olivine, and some of the original rocks were probably nearly pure olivine-rocks (Dun type). Others were enstatite- or bronzite- peridotites, and shew large bastite-pseudomorphs after a rhombic pyroxene (Cadgwith, Coverack, etc. ; fig. 20, cf. fig. 18) 3 . Others again are altered hornblende peridotites, some of the serpentine shewing the mesh- and some the lattice- structure, while relics of olivine, hornblende, and picotite may remain (Mullion Cove, Kynance Cove, etc.)*. Augite-picrites are also represented (Menheniot, etc.). Here felspar has been altered into a substance resembling serpentine, which Mr Teall thinks is probably that called pseud ophite-. Tremolite has been formed at the expense of olivine. The augite of the original rock is often preserved. Prof. Bonney and Gen. McMahon 5 , summarising the features of the Lizard serpentines, say that they " can be roughly separated into two groups : in the one a foliated mineral of the enstatite group is a conspicu- ous accessory ; in the other a colourless augite or hornblende, usually the latter. A few are non-porphyritic 6 , and in some cases exhibit no certain traces of any pyroxenic mineral, rhombic or monoclinic, though of course a spinellid or some iron oxide is always to be detected, and in one instance (at the Rill 7 , W. of Kynance Cove) the presence of a fair propor- tion of felspar has been asserted." 1 M. M. xi, 116 (Abstr.). 2 Bonney, Q. J. G. S. (1877) xxxiii, 915-923 ; and (1883) xxxix, 21-23 ; Teall, 115 et seqq. 3 See also Teall, PI. i, fig. 2; Cole's Micro. Stud. (1883) No. 50. 4 See Teall, PI. xv. 5 Q. J. G. S. (1891) xlvii, 466. 6 In the sense of containing no conspicuous crystals. 7 Teall, p. 119. "The original rock, therefore, was of the nature of a picrite." See also G. M. 1887, 137, 138. VARIOUS SERPENTINE-ROCKS. 97 Various serpentinous rocks are found near Holyhead and in neighbouring parts of Anglesey. That of Ty-ucha is regarded by Prof. Bonney 1 as an altered oli vine-rock : in other examples there is much calcite-veining, producing ' ophical- cite' (Cruglas) 2 . In rocks at Four-mile Bridge much of the serpentine has the character of that derived from augite, and the parent-rock seems to have been genetically connected with a gabbro mass. Mr Blake, however, finds indications of olivine- and enstatite-serpentine 3 . Of the numerous serpentine-rocks of Scotland 4 , one at Balhamie Hill in Ayrshire has been described by Prof. Bonney 5 as an altered olivine-bronzite-rock, closely resembling that of Cadgwith in Cornwall, the structure being of the pseudo- porphyritic type. Some near Belhelvie in Aberdeenshire 6 have also been enstatite-peridotites, but with the pcecilitic structure, and now shew pseudomorphs after olivine set in a framework of bastite, just as in the rock of Baste in the Harz, which has given its name to the latter mineral. Numerous serpentine-rocks are known in the United States. Wadsworth has described peridotite-serpentines from Min- nesota 7 , from Plumas County, California, from Westfield and Lynnfield, Mass., and other localities 8 . Near San Francisco occur some derived from peridotites (the Potrero 9 ), others from pyroxene -rocks (Angel Island 10 ). The rock at Syracuse, N. Y., was shewn by Williams 11 to be an altered peridotite, sharply defined pseudomorphs after olivine and enstatite being easily detected, while the remaining constituents, viz. brown mica, perofskite, and chromite, are still preserved. Q. J. G. S. (1881) xxx^i, 45. Blake, Eep. Brit. Assoc. for 1888, p. 409. Rep. Brit. Ass. for 1888, 408. For coloured plate of Portsoy serpentine see Cole's Micro. Stud. 52. Q. J. G. S. (1878) xxxiv, 770. Bonney, G. M. 1885, 439-448. Prelim. Descr. Perid. Gabb. etc. Minn. (1887) 29, PL i, fig. 1. Lithological Studies (1884) 158-160, PL vi, figs. 2, 5, vn, fig. 2. Palache, Bull. Geol. Dep. Univ. Cal. (1894) i, 165-169. > Ransome, Ibid. (1894) i, 220-222. 1 A. J. S. (1887) xxxiv, 140-142. H. P. 7 B. HYPABYSSAL ROCKS. SOME petrologists are content to divide the igneous rocks into two great groups, according as their structural characters indicate consolidation under deep-seated or under superficial conditions. Others, however, recognize another group inter- mediate between these two. Thus Rosenbusch inserts between his ' Tiefengesteine ' and ' Ergussgesteine ' a group ' Gang- gesteine ' or ' dyke-rocks.' The rocks to be treated under the present head correspond in a general way, though not precisely, with the last named, but Brogger's name ' hypabyssal ' is adopted as more accurately expressing the characters upon which the group is founded 1 . Although this threefold division seems to be necessitated by a comparative study of the great variety of rock-types met with in nature, it must be admitted that the hypabyssal group is a somewhat artificial one, the rocks included in it lacking any well defined set of common characteristics distinguishing them from the other two groups. Any definition would have to be framed chiefly in negative terms, and would bring together types presenting many points of difference from one another. Most of them are holocrystalline, but in some a glassy residue is found. In some families the porphyritic structure is characteristic 2 , as it is in the volcanic rocks; in others it is wanting or non-significant : but even the holo- crystalline non-porphyritic types have structural and mineral- ogical characters, to be noted below, yhich differentiate them from rocks of truly deep-seated 'origin. 1 The term 'intrusive' employed in the first edition of this book is abandoned, since it has been found to lead to misconception. - On the significance of this structure see Cross, 11th Ann. Eep. U. S. Geol. Sur. (1895) 232-235. CHAPTER VII. ACID INTRUSIVES. THE acid intrusive rocks embrace a considerable range of varieties, bridging over the difference between the even-grained, holocrystalline granites and the porphyritic, largely glassy rhyolites. The porphyritic character is almost universal, but the ground-mass which encloses the phenocrysts may be holo- crystalline, partly crystalline and partly glassy, or wholly glassy. On the nature and special structures of the ground- mass depend the several types usually recognized among these rocks. All agree in that the constituent minerals in so far as these are developed include in the first rank felspars rich in alkali and usually quartz, while ferro-magnesian minerals and free iron-ores occur only in relatively small quantity, and are sometimes wanting. From an examination of their mineral constitution and characteristic structures, the more crystalline types are readily referred to their proper positions ; but, in proportion as the bulk of the rock comes to consist of unindividualised glassy matter or an irresolvable cryptocrystalline ' base,' the criteria become fewer. In particular, the first stage of consolidation (that of the phenocrysts) may have been arrested before quartz (the last mineral) began to crystallize, and so, if the ground- mass consolidates as a glass, we may have a thoroughly acid rock without quartz. Thus the most glassy rocks (pitchstones) belonging to this family are not always to be distinguished by the microscope alone from less acid pitchstones. Again, they are scarcely divided from some glassy rhyolites (obsidians). 72 100 MINERALS OF THE ACID INTRUSIVES. The nomenclature of the acid intrusives is confused. The name 'felsite' or if containing evident phenocrysts of quartz ' quartz-felsite ' has been applied in this country not only to these rocks but also to many volcanic rocks (acid and intermediate) : and their usage lacks precision and significance. The name quartz-porphyry, borrowed from the German, covers most of the rocks, but not all, since porphyritic quartz may be wanting; this term is also used by Continental writers for the ' older ' acid lavas. For a type rich in soda, and having some miner- alogical peculiarities, the name quartz-ceratophyre (Ger. Quarz- keratophyr) has been used. It will be convenient to speak of the family, as a whole, as the acid intrusives. The names applied to particular types will be noticed in connection with the ground-mass. Constituent minerals. We notice here especially the minerals occurring as phenocrysts. Of these, the felspars include orthoclase (not microcline) and an acid plagioclase such as oligoclase. The two are commonly associated, and both build idiomorphic crystals with the usual types of twinning. A narrow zone of orthoclase surrounding each plagioclase crystal is seen in some rocks. The characteristic felspar of the quartz-ceratophyres is anorthoclase. The quartz has crystallized in the ordinary hexagonal pyramids, sometimes with narrow prism-faces, but the crystals are frequently rounded and eaten into, owing to corrosion by the ground-mass 1 , and may have lost all crystal outlines. In the rock-types most nearly approaching granites (granite- porphyries) the quartz contains fluid-pores 2 : in other types the inclusions are mostly of glass or portions of the ground-mass 3 . As already mentioned, quartz-phenocrysts are not always present. The brown biotite, which occurs in many of the rocks, has the same characters as in granites, and carries the same inclusions. It is usually in good hexagonal flakes. Less commonly, in the marginal part of an intrusion, it has a blade-like habit, due to extension along the a-axis. The usual 1 Cohen, PI. v, fig. 2. * Ibid. PL vn, fig. 4, vi, fig. 4. 3 Ibid. PI. v, fig. 4. STRUCTURES OF THE ACID INTRUSIVES. 101 mode of alteration is chloritization 1 . Hexagonal flakes of muscovite are found in a few of the granite-porphyries only. A green hornblende in well-built crystals is a rather excep- tional constituent. The deep blue soda-bearing amphibole riebeckite occurs in a few rocks, but always in very ragged allotriomorphic crystals. The augite of the acid intrusives is a pale greenish variety like that in some granites, but occurs here much more frequently. It builds good idiomorphic crystals in many granophyres and pitchstones. A few rocks rich in soda contain cegirine. A rhombic pyroxene (bronzite) is also known. As accessories, apatite and zircon are widely but sparingly distributed, while the iron-ores are usually represented only by a little magnetite. Such minerals as garnet, tourmaline, and pinite pseudomorphs after cordierite 2 occur in special localities. Some granite-porphyries carry tourmaline (Corn- wall ; Elba). Ground-mass and structures. The types which ap- proach most nearly to the plutonic habit are known as granite- porphyry. Here relatively large idiomorphic crystals of quartz and felspars, with mica or some other ferro-magnesian mineral, are enclosed in a fine-textured crystalline ground-mass of felspar and quartz. The structure of this ground may resemble that of a granite, or may be distinguished by a more marked idiomorphism of the lath-shaped felspars, usually untwinned. Mica may also occur in a second generation as part of the ground-mass. Very common are the types in which the phenocrysts, consisting of felspars, more or less corroded quartz, and biotite or some other constituent, are embedded in a very finely crystalline ground-mass of felspar and quartz. The elements of the ground-mass may have more or less idiomorphism. Quartz-porphyries having an evidently microcrystalline ground- mass of this kind are styled by Rosenbusch microgranites, the porphyritic character being understood 3 . 1 Cohen, PI. XLII. 2 Fouque and Levy, PI. xin, fig. 5. 3 For chromolithograph of a microgranitic quartz-porphyry, see Berwerth, Lief. i. 102 CRYPTOCRYSTALLINE AND VITROPHYRIC STRUCTURES. When the texture of the ground-mass sinks to such minute- ness as to be not clearly resolved under the microscope, it may be described as cryptocrystattine (' microfelsitic ' of some authors). For such rocks Rosenbusch uses the term felso- phyre 1 . Without entering into a discussion of an obscure subject, it may be said that this cryptocrystalline ground is * probably in some cases original, in other cases due to secondary change (devitrification) of a ground-mass originally glassy. The glassy (or ' vitrophyric ') type of ground-mass is seen in the rocks known as jntckstones. In some of these, pheno- crysts of felspar, etc., are only sparingly present, the great bulk of the rock consisting essentially of isotropic glass. This glassy ' base,' however, includes in many cases innumerable minute and imperfectly developed crystalline growths (crystal- lites) with regular grouping (fig. 23). These minute bodies will be more fully noticed in connection with the acid lavas. The pitchstones frequently shew perlitic cracks, and occasionally some of the flow-phenomena which are better exhibited in lavas. Typical pitchstones, excluding lava -flows, are of quite limited distribution. In the above types we have what may be regarded as a graduated transition from the granitic to the rhyolitic struct- ures, the only gap, that between cryptocrystalline matter and glass, being one which the instruments at our disposal do not enable us to bridge. There is, however, a second, more or less distinct, line of transition, parallel to the former but charac- terized by a different set of "Structures, viz. micrographic intergrowths of felspar and quartz arid regular crystalline aggregates of felspar fibres. To these structures Rosenbusch applies the somewhat inappropriate term ' granophyric,' in- cluding both micropegmatitic and microspherulitic : and the rocks having a ground-mass of this nature are very generally known as granophyres. We have already noticed in some granites a micrographic intergrowth of the kind named micropegmatite ; but when the whole mass of the rock, exclusive of crystals of certain minerals, takes on this character, we have a type characteristic 1 Cf. Teall, G. M. 1885, 108-111. MICROGRAPHIC STRUCTURES. 103 of hypabyssal rather than plutonic rocks as here understood. In such rocks the quartz and the greater part of the felspar form a micrographic ground-mass, which may enclose idio- morphic crystals of some ferro-magnesian mineral (augite or biotite) or of felspar (mostly plagioclase). Further, the micro- graphic intergrowth may come in to some extent in rocks which on the whole would be placed with the granite-porphyries or the microgranitic type. When the intergrowth is on a relatively coarse scale, it is often rude and irregular, but the finer-textured l micropegmatite' shews great regularity and B FIG. 21. GRANOPHYRES, SHEWING MICROGRAPHIC INTKRGROWTH OF FELSPAR AND QUARTZ ; X 20. Crossed nicols. A. Crug, near Caernarvon : shewing an intricate aggregate of rather delicate micropegmatite with a tendency to irregular 'centric' arrangement [17]. B. Carrock Fell, Cumberland, shewing part of a phenocryst of oligoclase with a fringe of micropegmatite. The felspar in this is in crystalline continuity with the phenocryst ; the quartz, shewn in the position of extinction, is continuous with a quartz- grain at the top of the figure [1545J. often a definite arrangement 1 (Fig. 21, A). In particular it frequently forms a regular frame surrounding phenocrysts of felspar 2 , and it can often be verified that the felspar of the 1 Cohen, PI. XL. 2 For good illustrations see Irving, Copper-bearing Rocks L. Superior, PI. xiv, figs, 1, 2. 104 SPHERULITIC STRUCTURES. intergrowth is in crystalline continuity with the felspar crystal which served as a nucleus (Fig. 21, E). The appear- ance is as if the original crystal had continued to grow throughout the final consolidation of the rock, enclosing the residual excess of silica as intergrown quartz. Sometimes a line of Carlsbad twinning can be traced from the crystal through the surrounding frame. There is no doubt that plagioclase felspar, as well as orthoclase, enters into such micrographic intergrowths. Less frequently the quartz of the intergrowth is seen to be in crystalline continuity with a quartz crystal or grain upon which it has grown. The finest micrographic intergrowth tends especially to a stellate or radiate ('centric') arrangement, with or without a nucleus of an earlier crystal. As the growth becomes very delicate in texture, the sectors within which the felspar extinguishes simultaneously become narrower, and are repre- sented between crossed nicols by dark rays when their direction makes a small angle with one of the cross-wires. When the structure is on too minute a scale to be resolved by tp FIG. 22. GRANOPHYKE (SPHERULITIC QUARTZ-PORPHYRY), ST DAVID'S ; x 20. Upper half in natural light, lower half between crossed nicols. A cryptographic intergrowth (pseudospherulitic of some authors) is grown round a corroded quartz-grain. The bundle of highly refracting crystals (ep) is secondary epidote [350]. BRITISH GRANITE-PORPHYRIES, ETC. 105 the microscope, it may be termed, by analogy, cryptographic. (Fig. 22). The optical characters of such an aggregate appear to be determined by the minute radially arranged fibres of felspar, which obscure the quartz. The structures known as microspherulitic and pseudo-spherulitic in acid rocks are probably of this nature. Between crossed nicols they shew characteristically a black cross, caused by extinction in those fibres which lie nearly parallel to one of the cross-wires. Such growths cluster round porphyritic crystals of quartz or felspar, or, as innumerable closely packed minute spherules, constitute almost the whole of the ground-mass 1 . Isolated spherulites or bands of spherulites may occur in a vitreous or de vitrified ground. Leading types. We proceed to select a few examples illustrating the several points indicated above. In view of the frequent association of the different types of ground-mass in one district or even in parts of one intrusion, we shall not find it convenient to follow any strict order. An intrusion near Dufton Pike 2 in Westmorland is a characteristic granite-porphyry with both white and dark micas, which occur both as phenocrysts and in the ground- mass. The other phenocrysts are idiomorphic quartz and felspar, chiefly plagioclase but with a few large sanidine- crystals. A marginal modification of the rock shews a blade- like habit of the biotite. The Carboniferous ' elvan ' dykes of Cornwall and Devon, as described by Mr J. A. Phillips 3 and by Mr Teall, have a microcrystalline to cryptocrystalline ground-mass enclosing large felspars, pyramidal or rounded quartz crystals, and often mica. The quartz contains fluid-pores, sometimes in the form of negative crystals 4 , which may enclose a salt-cube as well as a bubble 6 . Tourmaline is of frequent occurrence in crystals 1 For good figures of micrographic aud cryptographic structures, ranging from the micropegmatitic to the spherulitic, see Fouque and Levy, Plates x, fig. 2, xi, fig. 1, xn, xiv, xv, xvi. 2 Q. J. G. S. (1891) xlvii, 519. 3 Q. J. G. S. (1875) xxxi, 334-338, PI. xvi. 4 Cohen, PI. vi, fig. 4. 5 Ibid. PL vn, fig. 4. 106 ACID INTRUS1VES IN CAERNARVONSHIRE. or stellate groups of needles, and is sometimes seen to replace felspar. An occasional constituent is cordierite, represented by the so-called ' pinite ' pseudomorphs of yellowish green micaceous flakes (Sydney Cove 1 ). The varied group of Ordovician intrusive rocks in Caernar- vonshire 2 include some granite-porphyries of a well-marked type. Quartz is wanting among the phenocrysts, which are chiefly of oligoclase. One example at the head of Nant Ffrancon has a ground-mass of allotriomorphic quartz and felspar (chiefly orthoclase) : the ferro-magnesian constituent is biotite. Others, quarried at Yr Eifl and near Nevin, have a ground in which idiomorphic felspars are moulded by inter- stitial quartz : these contain augite, usually without biotite. Other rocks in the district, all augitic, shew more or less tendency to micrographic structures, and in many the whole ground-mass is of micropegmatite. Beautiful examples occur in the hills above Aber and at Moel Perfedd in Nant Ffrancon. The growth of the micropegmatite round felspar crystals is well exhibited, and in some cases a narrow zone of orthoclase is seen interposed between a plagioclase crystal and the surrounding growth. The structure is rarely so minute as to approximate to the spherulitic. Many of the smaller intrusions in the district, e.g. near Clynog-fawr, are of quartz- porphyry with a cryptocrystalline ground, which may possibly be due to devitrification. Porphyritic quartz, which is wanting in the more evidently crystalline types, appears here in corroded crystal-grains. A somewhat similar rock is that forming a low range in the neighbourhood of Llanberis. This exhibits flow-structure in places, and has been considered by Professor Bonney and others as a group of lavas. The complex group of acid rocks near Caernarvon and eastward, which some have supposed to be of pre- Cambrian age, affords examples of granite-porphyries, micrographic rocks (Fig. 21, A), microcrystalline and spherulitic quartz-porphyries, etc. The spherulitic growths often surround pyramids of quartz. The porphyritic felspars in all these rocks are mostly plagio- clase, and the ferro-magnesian mineral is biotite, often green 1 Teall, 334. 2 Bala Vole. Ser. Caern. 48-56. VARIOUS QUARTZ-PORPHYRIES, ETC. 107 from alteration. Various granophyres and, especially, beautiful spherulitic rocks, shewing the growth round pyramidal crystals of quartz, occur at St David's 1 . The structure is of the crypto- graphic type, not shewing a very perfect black cross (fig. 22). The Lake District contains examples of microgranites, such as the rock quarried at Threlkeld, while some minor intrusions shew a cryptocrystalline ground. Granophyres also occur, the large Butter-mere and Ennerdale intrusion being of a micropegmatitic rock with either biotite or augite, resembling some Caernarvonshire examples. The dykes of Armboth and Helvellyn have a spherulitic ground-mass enclosing idiomorphic crystals of quartz and felspar. The spherulitic growth, which does not always give a good black cross, is clustered especially about the quartz crystals. A few garnets occur. These rocks are probably all Ordovician. The Devonian dykes about Shap, in Ederiside, near Sedbergh, etc., have microcrystalline to cryptocrystalline grounds, and some of them contain biotite rather abundantly. Among American examples may be mentioned the horn- blende-granite-porphyries described by Zirkel 2 for the Fortieth Parallel Survey. These carry porphyritic quartz and felspars, plagioclase being prominent, hornblende, biotite, and often spheiie, with a microcrystalline ground-mass. The quartz has fluid-pores sometimes containing salt-cubes and other inclu- sions 3 . Typical examples occur at Franklin Buttes, Nevada, in the Oquirrh Mts, Utah, etc. Rocks with cryptocrystalline ground-mass ('felsite-porphyry ') also occur, though in less force 4 , and spherulitic varieties are found (Spruce Mt, Peoquop Range). A granite-porphyry similar to the above has been described in detail by Iddings 5 from the Eureka district, Nevada. The finest examples of pitchstoiies are those of Arran 6 , of 1 Geikie, Q. J. G. S. (1883) xxxix, 315, PL x, figs. 8, 9. 2 Micro. Petrogr. Fortieth Parallel (1876) 62-67. 3 Ibid. 63, 77, PI. i, fig. 5. 4 Ibid. 73-80. 5 Monog. xx U. S. Geol. Sur. (1893) 339-345. 6 Allport, G. M. 1872, 1-9 ; 1881, 438 ; Bonney, G.M. 1877, 499-511 : Judd, Q. J. G. S. (1893) xlix, 546-551, 559-561, PI. xix : Geikie, p. 116, fig. 14 : Teall, PI. xxxiv, figs. 3, 4. Cf. Sollas on Donegal pitchstones, Set. Proc. Roy. Dubl. Soc. viii, 87-91 (1893). 108 ARRAN PITCHSTONES. which some are of acid, others of subacid composition. They form dykes, probably of Tertiary age. The phenocrysts are of sanidine, quartz, plagioclase, and augite, varying in different examples and sometimes occurring very sparingly. The ground-mass is of glass crowded with crystallites', which often assume peculiar groupings. In one variety needle-shaped microlites (belonites) of hornblende occur, each forming the trunk of a delicate arborescent aggregate of more minute bodies (Corriegills, fig. 23, A and ). In another variety occur crosses, each of the four arms carrying a plume-like FIG. 23. PITCHSXONES, AKBAN. A. Arborescent grouping of crystallites, Corriegills ; x 20. B. The same, x 100 [57]. C. Plumed cross-like groupings, and growth of crystallites on a small felspar phenocryst, Tormore; x 20. D. The same, x 100 [G. 73]. growth (Tormore, fig. 23, C and Z>), Again, little rod-like bodies frequently occur as a fringe arranged perpendicularly on the faces of phenocrysts. The general mass of the glass is full of very minute crystallitic bodies, but around each grouping is a clear space, indicating that the tree-like or other growth has been built up at the expense of the surrounding part. Flow-structures are only occasionally met with, and perlitic cracks are not common. Dykes of pitchstone with 1 Cohen, PI. xi, fig. 1. BRITISH GRANOPHYRES. 109 various crystallitic growths occur also in Skye (Coirechatachan near Broadford). One of the most beautiful granophyres in this country is that of Carrock Fell, in Cumberland '. It contains a pale augite in good crystals, often uralitized or otherwise altered, and rarely a little biotite. There are also idiomorphic felspars, usually oligoclase, and some granules of iron-ore. The ground- mass shews in different specimens, or even in one slide, every gradation, from a coarse, irregular micropegmatite through exquisitely regular micrographic 2 and cryptographic structures to what would be described as spherulitic. These inter- growths usually make up the whole ground-mass, though sometimes part of the quartz forms irregular grains. The arrangement is sometimes centric, but more usually peripheral to the felspar phenocrysts, forming a regular border to them. It can often be seen that the felspar of the intergrowth is continuous with that of the crystal, and much of it must be plagioclase (Fig. 21, B). Other augite-granophyres are found among the Tertiary intrusions of Scotland and Ireland, e.g. in Mull 3 and in the Carlingford district. Some are quite coarse micropegmatites, or shew only a rude kind of intergrowth, and these rocks are frequently miarolitic. The more delicate micrographic and cryptographic growths are, however, represented. Some of the Skye granophyres have riebeckite instead of augite 4 . A remarkable rock from Corriegills in Arraii 5 appears as if divided into polygonal areas each enclosing a spherule with well marked boundary and radial structure. Dr Hyland 6 has described granophyre dykes in Co. Down. These contain apparently no augite, but a little green hornblende (Newcastle) or brown mica (Hilltown). The granophyre of Stanner Rock near New Radnor has perhaps been augitic, but the ferro-magnesian mineral is now 1 Q. J. - Teall, . J. G. S. (1895) vol. li, 126-129. PI. XLVII, fig. 5 (misplaced 4 in key-plate). 3 Teall, PI. xxxm, fig. 1. 4 Teall, Q. J. G. S. 1, 219. 5 Allport, G. M. 1872, 541 ; Bonney, G. M. 1877, 506-508. e Set. Proc. Roy. Dubl. Soc. (1890) vi, 420-430. 110 QUARTZ-CERATOPHYRES, ETC. a fibrous hornblende. From Prof. Cole's description 1 the rock seems to be of the cryptographic type. The biotite-bearing quartz-porphyries of the Cheviots 2 have sometimes granophyric structures, but are more commonly micro- or cryptocrystalline. Mr Kynaston has shewn me numerous examples in which the ground-mass encloses patches of micropegmatite like porphyr- itic crystals, sometimes shewing the outlines of idiomorphic felspar. Hornblende-granophyre occurs in the Grampians. A speci- men from Beinn Alder contains pheriocrysts of oligoclase, twinned green hornblende, biotite, and magnetite in a ground- mass of delicate micropegmatite. Acid intrusives rich in soda (quartz-ceratophyres) are not yet well known in this country. Probably some of the ' soda-felsites ' of Leinster 3 , of Ordovician age, are to be placed here. They are mostly crystalline rocks, with or without porphyritic structure, consisting essentially of predominating felspar and quartz. Plagioclase is much more abundant than orthoclase, and is sometimes albite, sometimes possibly anor- thoclase or cryptoperthite. The quartz-grains are often rounded and corroded. The microcrystalline ground does not apparently sink to the texture of cryptocrystalline. In America a number of anorthoclase-bearing rocks have been described which fall into this family. We may note especially some of the pre-Cambrian granophyres, passing into granites (soda-granite) in the Lake Superior region and else- where. Such rocks have been described and figured by Irving 4 and more particularly by Bayley 5 (Pigeon Point, Minn.). More remarkable are those rocks in which the ferro- inagnesian minerals are also of soda-bearing varieties. From the Apache Mts in Western Texas Osann has described a riebeckite-granite-porphyry (Paisani type), having scattered porphyritic felspars (microperthite) and quartz in a ground- G. M. 1886, 220-222, with figs. Teall, G. M. 1885, 111. Hatch, G. M. 1889, 70-73, 545-549. Copper-bearing Rocks of L. Superior, PI. xv. Bull. No. 109 % U. S. Geol. Sur. (1893). RIEBECKITE-QUARTZ-PORPHYRIES. Ill mass containing ragged crystal-grains of riebeckite with microperthitic felspar and quartz. Allied to this is the rock of Mynydd Mawr in Caernarvonshire 1 , which has ragged crystals of riebeckite with microperthitic felspars and some quartz-grains in a ground-mass of quartz and felspar with microlites of some unknown mineral. A somewhat similar rock occurs at Ailsa Craig 2 , and the occurrence of riebeckite in certain Skye granophyres has been mentioned above. An aggirine-granite-porphyry (Grorud type) has been de- scribed by Brogger 3 . 1 G. M. 1888, 225, 226; 1889, 455, 456; Bala Vole. Ser. Caern. 50-52. 2 Teall, M. M. (1891) ix, 219-221. 3 If. 31. (1895) xi, 115 (Abstr.). CHAPTER VIII. PORPHYRIES AND PORPHYRITES. THE rocks which are for convenience grouped together in this chapter belong to various hypabyssal types of intermediate chemical composition. They have not a very wide distribu- tion, and they graduate on the one hand into the acid intrusives already discussed, on the other into the more peculiar family of the lamprophyres. The porphyritic structure characterizes all the rocks in question, and in most of the types is marked by felspar phenocrysts of relatively large size. The ferro-magnesiau minerals are often confined to the elements of the earlier period of crystallization. Original quartz is found in the more acid types only, and is almost always restricted to the ground-mass. The rocks may be regarded as standing between the plu tonic syenites, diorites, etc., on the one hand, and the volcanic trachytes, dacites, and andesites on the other, just as the rocks treated in the preceding chapter stand between the granites and the rhyolites. According as the dominant con- stituent is an alkali-felspar or a soda-lime-felspar, they fall into two families, to be distinguished as porphyries and por- phyrites respectively. Under the former head we may recognise syenite-porphyry and orthoclase-porphyry (or simply porphyry), corresponding with granite-porphyry and quartz-porphyry among the acid rocks. From these orthoclase-bearing rocks have been separ- ated others characterized by a potash-soda-felspar, under the FELSPARS OF PORPHYRY, ETC. 113 name ceratopJiyre (Ger. Keratophyr). There are also nepheline- syenite-porphyry and nepheline-porphyry (tiriguaite, etc.), which are of very restricted occurrence. Of the rocks characterized by soda-lirne-felspars, the types most nearly approaching the plutonic have been styled diorite- porphyrite,, etc., the others being termed simply porphyrites. Since some ferro-magnesian mineral is usually a prominent constituent, we have the divisions mica-porphyrite, hornblende- porphyrite, and auyite-porphyrite. If a little porphyritic quartz be present, we have a quartz-porn fiy rite (quartz-rnica- porphynteyr It must be noted that writers who make no distinction in nomenclature between intrusive and volcanic rock-types use some of the above names in a more extended sense. Thus the , Continental petrologists include under the term porphyrite i . the ' older ' andesitic lavas, while some British authors apply ,(i/ythe same name to andesites modified by secondary changes ' (partial decomposition, etc.}. Some of the rocks styled propyl- ites belong to the division now to be considered. Constituent minerals. The orthodase phenocrysts of the porphyries are similar to those in the quartz-porphyries and other acid intrusives. In the porphyrites this mineral does not occur except in the ground-mass. A plagioclase felspar accompanies the porphyritic orthoclase in some of the por- phyries, and forms the most conspicuous phenocrysts in the porphyrites. Here it builds idiomorphic or rather rounded crystals, with twinning often on two or three different laws. It ranges in the porphyrites from oligoclase to labradorite, and frequently shews strong zoning between crossed nicols. A parallel iutergrowth of orthoclase and plagioclase is common in some porphyries. In certain types of that family also occurs a felspar which has been referred to anorthoclase, while it has also been explained as a minute parallel intergrowth of a potash- and a soda-liine- felspar. Viewed between crossed nicols, a crystal is often seen to be divided rather irregularly into portions with different optical behaviour, sometimes one part finely striated, another without visible stiiation. In certain special rocks (rhomb porphyries) the crystal has a 114 MINERALS OF PORPHYRY, ETC. peculiar habit, which gives a lozenge-shaped section ; in the ceratophyres it has the usual habit, giving rectangular sections. As phenocrysts quartz is found only sparingly in a few rocks, but it enters into the ground-mass of all the more acid of the porphyries and porphyrites, though less abundantly than in the true acid rocks. The most usual ferro-magnesian minerals are brown biotite and a pale or colourless idiomorphic augite. Some of the porphyrites have hornblende in sharply idiomorphic prisms, often twinned : it is more usually brown than green. In rocks rich in alkali the coloured constituent is often segirine- augite or true ceyirine. As accessories, apatite and iron-ores (often titariiferous) may occur in varying quantity, the latter not being abundant. Exceptionally olivine and other minerals are present. In the few rocks which contain nepheline or elyeolite, that mineral occurs in one or two generations. As phenocrysts it is idiomorphic, while the little crystals in the ground-mass may or may not have definite shape. The ' liebenerite ' pseudomorphs in certain porphyries have been supposed to represent nepheline. They consist essentially of a pale mica, and may with equal probability come from the destruction of cordierite. Some of these rocks rich in alkali carry melanite garnet. Ground-mass and Structures. In the great ma- jority of the rocks here considered the ground-mass is hoi ocrystal line, with a fine texture and with various types of structure. It consists essentially of felspar or, in the more acid members, of felspar and quartz. In the porphyries the felspar is usually in minute prisms, short in comparison with their length, and as a rule untwinned. Quartz, if present, occurs interstitially. The little prisms may have more or less of a parallel arrangement, clue to flow. Such short and relatively stout prisms are usually referred to orthoclase : if the crystals have the ' lath '-shape, they are probably of a plagioclastic variety. Any approach to an allotriomorphic character is uncommon, and the micrographic intergrowths so STRUCTURES OF PORPHYRY, ETC. 115 frequent among the acid intrusives are not found here. In the nepheline-bearing rocks a more allotriomorphic type of structure is often found ; while the bostonites and allied rocks shew an approach to the volcanic trachytes, often with marked flow-structure. The ground-mass of the porphyrites is also in general holocrystalline, consisting essentially of felspar, or, in the more acid varieties, of felspar and quartz. In this latter case the rocks may reproduce some of the characteristic structures noted in the preceding chapter, such as the cryptocrystalline and the micrographic. Other porphyrites have the ' ortho- phyric ' type of ground-mass (with short felspar-prisms), as in the porphyries, but there is every gradation from this to the allotriomorphic. In some of the more basic members the ground-mass consists of little lath-shaped plagioclase prisms with more or less noticeable flow-arrangement, an approach to the character of some andesites (' pilotaxitic ' structure). Glassy and vitrophyric rocks are not unknown in the families in question. Some of the Arran pitchstones, for example, have the composition of intermediate rather than acid rocks. Leading types. Only a few illustrative examples will be selected, chiefly from British and American rocks. As an example of syenite-porphyry may be noticed the rock quarried at Enderby in Leicestershire. It contains phenocrysts of a strongly zoned plagioclase felspar and of pale greenish brown hornblende with, more sparingly, flakes of biotite and round grains of quartz in a moderately tine-textured ground- mass of quartz and felspar, apparently orthoclase. A type richer in alkali is represented at several American localities. A rock from Coney Island, Salem, Mass., has. abundant phenocrysts of felspar (microperthite and crypto- perthite) in a ground-mass of similar felspars and needles of a greenish blue soda-am phi bole (catophorite), with fluxion- structure. Augite-syenite-porphyry has been noted at Lake Chatauqua, N.Y., Albany, N.H. (with accessory bronzite), and other places. 82 116 ORTHOCLASE- PORPHYRIES, ETC. Richer in alkali is the segirine-syenite-porphyry (Solvsberg type) described by Brogger 1 in the Christiania district. Here tKe structure of the ground-mass approaches that seen in the trachytic lavas and in the bostonites noticed below. A similar rock is that described (under the name ' acmite- trachyte ') by Wolff and Tarr 2 from the Crazy Mts in Montana. The pheno- crysts are of anorthoclase and augite (bordered by segirine) with occasional sodalite, and the ground-mass is of lath-shaped felspars (chiefly anorthoclase) and needles of segirine, with a variable amount of nepheline and secondary analcime. Another locality is the Apache Mts in western Texas. The most usual type of ortlwclase-porphyry (orthophyre of Rosenbusch) is exemplified by dykes and sills in the Carboni- ferous of Thuringia, in the Vosges, and in other districts. Besides the orthoclase phenocrysts there may be some of plagioclase. The ferro-magnesian minerals are only sparingly represented, and may be biotite, hornblende, or augite. The ground-mass is holocrystalline with the structure styled ortho- phyric, in which short prisms of untwinned felspar are associated with some interstitial quartz. Of a different type are the rocks forming North Berwick Law and the Bass Rock. They have been described 3 under the name trachyte, with the associated lavas which they closely resemble (see Chap, xn), but may be mentioned in this place. The felspar of which they are almost wholly com- posed is rich in soda as well as potash, and the non- porphyritic, trachytoid structure of the rocks allies them with the bostonites. The typical bostonites occur at Marblehead Neck near Boston, Mass., in the Adirondacks 4 , near Montreal, at Liver- more Falls and Shackford, N.H., in the Apache Mts, Tex., .etc., as dykes in connection with nepheline-syenite or other plutonic rocks, and especially in intimate association with 1 H. M. (1895) xi, 115 (Abstr.). 2 Bull. Mus. Comp. Zool. Han: (1893) xvi, 227-230. 3 Hatch, Trans. Roy. Soc. Edin. (1892) xxxvii, 123, 124, PI. i, figs. 3, 4. 4 Kemp and Marsters, Trans. N. Y. Acad. Sci. xi (1891) 14-16 ; Bull. Xo. 107 U. 8. Geol. Sur. (1893) 18-22. RHOMB-PORPHYRIES AND CERATOPHYRES. 117 dykes of lamprophyre (camptonite). The bostonites consist essentially of felspar, quartz being never abundant and the ferro-magnesian silicates typically absent. Phenocrysts may or may not be developed, the bulk of the rock being a ground- mass of little felspar rods, often with partial flow-disposition and recalling the structure of the trachytes. In many examples a high percentage of soda, with little or no plagioclase evident, points to a soda-orthoclase or anorthoclase, and indicates an affinity with the ceratophyres. Rocks approaching bostonite in character occur at some localities in Scotland and in the Limerick district 1 , but have not yet been studied in detail. Among the Devonian intrusions of the Christiania district occur the singular rocks known as rhomb-porpliyry (Ger. Rombenporphyr), and they may be -studied in numerous boulders in Holderness and the Eastern Counties. The pheno- crysts of potash-soda-felspar, with their unusual crystallo- graphic development, have been alluded to above. The crystals are often rounded and corroded, and they contain numerous inclusions of materials like the ground-mass. Some of the rocks contain pseudomorphs after olivine. The fine- textured hole-crystalline ground-mass consists of short prisms of felspar (probably orthoclase) with little granules of augite. Apatite is often plentiful, and grains of titaniferous iron-ore occur. Rhomb-porphyries have been discovered by Osann in the Apache Mts, Tex. The name ceratophyre was first used by von Giimbel for a rather varied group of rocks in the Fichtelgebirge. Some- what similar rocks have been described from Saxony, West- phalia, the Harz, and other areas. Porphyritic quartz does not occur in the ceratophyres proper, and felspar is the predominant mineral in both phenocrysts and ground-mass. The phenocrysts have the peculiarities attributed to anor- thoclase or to a cryptoperthite intergrowth. The commonest ferro-magnesian element is a pale augite (diopside)s The felspar prisms of the ground-mass may be short and unstriated or lath-shaped and striated, and the more acid members have a little interstitial quartz. 1 Watts, Guide, 93. 118 TINGUAITES. Rosenbusch has given the name tinguaite to certain 'dyke- rocks ' which have the composition of the (plutonic) nepheline- syenites and the (volcanic) phonolites, with structural cha- racters which place them between those two families. Such rocks are associated with nepheline syenites in the Serra do lingua and other places in Brazil, in southern Portugal, in the Christiania district 1 , in Arkansas 2 (Fourche Mt) and Texas (Apache Mts), etc. Phenocrysts of orthoclase, often with marked tabular habit and with the characters of sanidine, are embedded in a fine-textured holocrystalline ground-mass of orthoclase with elaeolite or nepheline, aegirine, etc. This ground is typically allotriomorphic : when the little felspars take on the lath-shape with fluxional arrangement, the rocks do not differ essentially from phonolites. There may be phen- ocrysts of nepheline, and in one type (leucite-tinguaite) large pseudomorphs of orthoclase and elaeolite occur in the form of leucite. This latter type has been described from Brazil 3 . Arkansas 4 (Magnet Cove), and Montana 5 . From the last- named state comes also a variety intermediate between the true tinguaite and the Solvsberg type mentioned above (Lan- dusky in the Little Rocky Mountains 6 ). A more basic nephel- ine-bearing type, on the other hand, occurs at Magnet Cove, Ark. 7 , and at Beemerville, N.J. 8 , having phenocrysts of nephe- line up to an inch in diameter in a tinguaitic ground-mass composed chiefly of nepheline, charged with segirine-needles, with some orthoclase, etc. This type is the ' sussexite ' of Brogger 9 , constituting the most basic member of a ' rock- series ' of which the other members are grorudite, solvsbergite, and tinguaite. Coming now to rocks of dioritic affinities, we may mention Brogger, M. M. (1895) xi, 116 (Abstr.). J. F. Williams, Igneous Rocks of Arkansas, vol. ii of Ann. Rep. Geol. Sur (Ab Ark. for 1890, 100-106. Derby, Q. J. G. S. (1891) xlvii, 251-265 ; Graeff, M. M. vii, 235 * J. F. Williams, I.e., 277-286. Pirsson, A. J. S. (1895) 1, 394-398. Weed and Pirssou, Journ. of Geol. (1896) iv, 419-421. J. F. Williams, I.e., 259-261. Kemp, Trans. N. Y. Acad. Sci. (1892) xi, 66, 67. M. M. (1895) xi, 114-116 (Abstr.). HORNBLENDE-PORPHYRITES. 119 a quartz-diorite-porphyrite from Sweet Grass Hills, Montana 1 , and a quartz-inica-diorite-porphyrite, approaching granite- porphyry, from Electric Peak in the Yellowstone Park 2 . This has abundant small pheriocrysts of felspars, quartz, and biotite, with a little hornblende, and a granular ground-mass of felspar and quartz. In the same district occur porphyrites, generally hornblende-porpkyrites*, carrying abundant phenocrysts of lime- soda-felspar and hornblende, with usually biotite and oc- casionally uralitized augite, in a fine-grained ground-mass. When the latter is rich in quartz, this mineral tends to form micropoecilitic patches enclosing the little felspar-prisms ; when quartz is scarce, the felspars, which are, at least in the main, plagioclase, tend to have a felted arrangement. The ground- mass also contains some hornblende and biotite. Resembling the Electric Peak rocks, and like them of somewhat acid character as a whole, are the hornblende-porphyrites and hornblende-mica-porphyrites described by Cross 4 from the laccolites and associated intrusions of the Henry and Abajo Mts in Utah, the West Elk and El Late Mts in Colorado, etc. Among the phenocrysts the dominant minerals after plagioclase felspar (oligoclase or andesine) are hornblende and to a less extent biotite, while augite and hypersthene occur only locally. Quartz is also developed porphyritically and in certain cases large crystals of orthoclase, which, however, seem to belong rather to the same stage of consolidation as the ground-mass (Mt Carbon and Gothic Mt, in the West Elk group, etc.}. The ground-mass is essentially an aggregate of orthoclase and quartz. Among British examples we may note some of the rocks which form sills of Lower Palaeozoic age in the Assynt district of Sutherland (Inchnadamff, etc. 5 ). 'Here the hornblende is green and in very perfect crystals, often twinned : they some- times shew zonary colouring, and are occasionally hollow. A colourless augite in imperfect crystals sometimes accompanies the hornblende. The plagioclase phenocrysts shew strong 1 Weed and Pirsson, A. J. S. (1895) 1, 311. 2 Iddings, 12th Ann. Rep. U. S. Geol. Sin: (1892), 617, 618. a Ibid. 588-594. 4 Cross, Uth Ann. Rep. U. S. Geol. Sur. (1891). 5 Teall, G. M. 1886, 346-350. 120 MICA-PORPHYR1TES. zonary banding between crossed nicols. Magnetite and apatite are present sparingly. The microcrystalline ground-mass is of felspar with subordinate quartz. These rocks are part of a variable set of intrusions. On the one hand is a non- porphyritic and coarser-textured type with allotriomorphic felspar (diorite), on the other a type with more abundant hornblende in two generations and with a panidiomorphic ground-mass (camptonite, see Chap, x and fig. 29). Diorite- porphyrites approaching camptonite have been described near St John, New Brunswick 1 and in other American localities. The dykes described by Mr Teall 2 in Kirkcudbrightshire are mostly mica-hornblende-porphyrites. The phenocrysts are of zoned plagioclase in large individuals, green hornblende and brown biotite, both in good crystals, and sometimes corroded grains of quartz, while the fine-textured ground-mass contains quartz and orthoclase in addition to the other minerals named. Numerous mica-porphyrite dykes, of Old Red Sandstone age, occur in the Cheviots 3 . The felspar phenocrysts (oligo- clase-andesine) are frequently rounded, and shew carlsbad and albite twinning. The biotite-flakes are often bent, and sometimes shew a resorption border. A colourless augite may also occur, and magnetite and apatite are minor constituents. The ground-mass is microcrystalline, fine-textured, and often obscured by decomposition. Quartz plays a variable part in it, and there are some transitions to granophyre and quartz- porphyry. Indeed the mica-porphyrites in general often carry a notable amount of quartz in their ground-mass. The rock which forms large intrusive sills in the Torridon Sandstone of Canisp, Sutherland, may also be placed here. It has large, frequently broken, phenocrysts of oligoclase, with carlsbad, albite-, and pericline-twinning. The dominant co- loured mineral is biotite, but Mr Teall also notes augite, either colourless or green or the former bordered by the latter. Calcite pseudomorphs in the form of augite are common. 1 Matthew, Trans. N. Y. Acad. Sci. (1895) xiv, 211, 212, PI. xiv. 2 Teall, Mem. Geol. Sur. Scot. , Expl. of Sheet 5 (1896) 44, 45. 3 Watts, Mem. Geol. Sur. Engl. and Wales, Expl. of Sheet 110 S. W., (N.S. 3) (1895) 62, 63. AUGITE-PORPHYRITES. 121 These minerals, with some magnetite, are set in a fine micro- crystalline ground-mass of felspar and quartz. A hornblende-porphyrite of basic composition is seen in the Mawddach valley, near Dolgelly. It contains large and rather irregularly bounded twin-crystals of brown hornblende in a much decomposed matrix. Mr Phillips 1 termed this hornblende uralite, but there is no clear evidence that it is other than an original mineral. The rocks to which the name augite-porphyrite has been applied by German petrologists seem to be for the most part old augitic lavas, though intrusive types are also included. Such rocks, probably of Triassic age, are represented in the Monzoni district in the southern Tirol. Augite, is, however, a frequent accessory mineral in the hornblende-porphyrites, and in particular occurrences may become the dominant coloured element of the rock. Thus in the Henry Mts Cross remarks augite-porphyrites at Mount Pennell and Mount Hillers, but they are mainly from sheets, while the great laccolites themselves are of the hornblendic type. 1 Q. J. G. S. (1877) xxxiii, 427-429, PI. xix. CHAPTER IX. DIABASES. THE larger intrusive bodies of hypabyssal pyroxenic rocks, whether intermediate or basic in composition, have petro- graphical features which characterize them as a group with considerable individuality. It is to these rocks that we shall apply the name dja ( b i a^e. Like their plutonic equivalents, the gabbros, they are holocrystalline and typically non-porphyritic, but they differ from the normal gabbros in their less coarse texture, in the absence of diallagic and other ' schiller ' struc- tures, and in the mutual relations of the felspar and augite which are their two chief constituents. In these respects there are, however, transitions between the two sets of rocks. The diabases occur as large dykes, sills, and laccolitic or other masses. Minor intrusions of rocks having a similar chemical composition commonly have more of the petro- graphical characters of volcanic rocks. For these we shall retain the names dolerite, andesite, basalt, etc., and they will be excluded from this place. The name diabase has been, and still is, employed in different senses. By the German school it is restricted to the older rocks, whether hypabyssal or volcanic, dolerite and basalt being terms reserved for rocks of Tertiary or later age. Mr Allport shewed very conclusively that such a distinction corresponds with no real difference between the older and the newer rocks, and he abandoned the name diabase in favour of dolerite for all. The rocks so designated by Allport include NOMENCLATURE OF DIABASES. 123 some of the hypabyssal and others of the volcanic type. English writers have followed him in admitting no criterion of geological age into their classification and nomenclature, but some of them have inconveniently employed the name diabase for a more or less decomposed dolerite. According to the absence or presence of the basic silicate olivine, the rocks of the present family are often divided into diabases proper and olivine-diabases. Olivine is in general found in the more basic members of the family, but this division does not correspond very exactly with the chemical division into intermediate (or sub-basic) and basic. By the presence of some other special mineral we may distinguish such types as quartz-diabase, bronzite-diabase, and hornblende- diabase ; or again quartz-bronzite-diabase and olivine-horn- blende- diabase. Various other names have been used for particular types of diabasic rocks. Among the hornblende-bearing diabases of the Fichtelgebirge von Giimbel distinguished two types ; proterobase, containing original hornblende in addition to augite, and epidiorite, in which the hornblende is all derived from augite. Some writers have extended these names to cover all diabasic rocks characterized by primary and secondary hornblende respectively. The old field-term 'greenstone,' referring to the staining of the rocks by chloritiCTnrtf^Jffie'r decomposition-products, included not only diabases but diorites, picrites, altered dolerites, etc., and so had no precise significa- tion. The picrites, included above among the plutonic rocks, have much in common with the diabases, and in some districts are closely associated with them. Constituent Minerals. The felspars of the diabases range from oligoclase to anorthite in different examples : varieties of labradorite are perhaps the most common. The crystals have a strong tendency to idiomorphism, with colum- nar or sometimes tabular habit. Twin-lamellation on the albite law is universal, and is often combined with carlsbad twinning, but the pericline law is not so common. Zonary growth is not often shewn, except when a later set of 124 PYROXENES OF DIABASE. felspars occurs, of shapeless outline and more acid composition ; these shew strong zoning between crossed nicols. Inclusions are not common, except glass-cavities and needles of apatite. Decomposition gives rise to calcite-dust, to finely divided material, which may be mica, to zeolites, or to granular epidote. The crystals also become charged with strings and patches of green chloritoid substance, probably derived in part from the pyroxene. The common pyroxenic constituent is an augite, usually without crystal outlines. It varies in thin slices from brown to nearly colourless, and rarely shews sensible pleochroism. Zonary and ' hour-glass ' structures are sometimes seen. The orthopinacoidal twin is common, and in some cases there is a fine basal lamination ' in addition (Whin Sill). The common- est decomposition-pi'oducts are pale green, fibrous or scaly aggregates of serpentinous and chloritic substances. The former may be recognized by their low refractive index and moderately high birefringence ; the latter are usually very feebly birefringent or sensibly isotropic, and shew distinct pleochroism. Delessite is probably a common product, besides chlorite proper, but the discrimination of this very ill-defined group of minerals is not easy. Another change to which augite is subject is that which results in a light-green 'uralitic' hornblende. This is usually, but not always, fibrous in structure. Some diabases contain bronzite in addition to augite. It is in more or less idiomorphic crystals, with faint pleochroism, and gives rise by alteration to pseudomorphs of light green fibrous bastite. Only occasionally does hornblende appear as an original constituent. It seems to be characteristically a brown variety. Brown biotite is also a rare accessory. A little quartz is found in some of the less basic diabases, occurring interstitially. Whether it is original or a decom- position-product is sometimes difficult to decide, but when the 1 Rosenbusch-Iddiiigs, PI. xix, fig. 6 ; Teall, Q. J. G. S. (1884) xl, PI. xxix, fig. 1. IRON-ORES OF DIABASE. 125 mineral forms part of a micrographic intergrowth with felspar its primary nature may safely be assumed. The olivine, which occurs in very many diabases, builds more or less rounded idiomorphic crystals or grains, sensibly colourless or very pale. It has the same mode of alteration as in the olivine-gabbros and peridotites. The iron-ores, which, in contrast with many gabbros, the diabases contain abundantly, include ilmenite and magnetite. The two are very commonly associated, and some so-called titaniferous magnetite has been supposed to be a minute inter- growth of the two. They are easily distinguished when they occur as crystals or skeleton-crystals. In most cases the ilmenite has given rise to more or less of its characteristic il FIG. 24. DECOMPOSING DIABASE, DENEIO, NEAR PWLLHELI, CAERNARVONSHIRE ; x 20. This shews decomposing felspar-crystals and ophitic augite, with ilmenite-skeletons (il), crusted with leucoxene, and patches of radiating fibres of a zeolitic mineral (z) [123]. decomposition-product, grey cloudy masses of semiopaque leucoxene 1 (fig. 24). 1 Rosenbusch-Iddings, PL xvi, fig. 2 ; Teall, PL xvn, fig. 2. 126 STRUCTURES OF DIABASE. Long columnar or needle-like crystals of apatite occur in most diabases, but in some are capriciously distributed. Structure. As regards structure, the diabases offer a contrast to normal plutonic rocks, owing mainly to the fact that the crystallization of the felspar has preceded that of the dominant ferro-magnesian constituent. As seen in a slice, the columnar crystals of felspar shew more or less elongated sections, with no law of arrangement, and around or between the.^e the augite is moulded. The last-named mineral in most cases distinctly wraps round the felspar crystals, and often forms plates of some extent, enclosing many of them. This is known as the ophitic structure 1 (fig. 25). In other cases the augite tends to form more or less rounded grains imbedded in a plexus of lath-shaped felspars, adjacent grains not being parts of one crystal but shewing different orientations. This is what Prof. Judd 2 has styled the granulitic structure: he considers it due to movement towards the end of the process of consolidation. In both types, if olivine is present it is always idiomorphic towards the augite, but may be penetrated by the felspar prisms. The rhombic pyroxene, too, is con- stantly of earlier crystallization than the augite, and may shew good outlines. The iron-ores are often idiomorphic, but magnetite may be in part later than the felspar. When, as is sometimes the case, a subordinate felspar, of later con- solidation than the dominant kind, is present, it lias crystallized with or after the augite, and is always shapeless. The typical diabases thus present a very uniform structural character, which in its best development is almost peculiar to them. In a few diabases, however, the augite, especially if not abundant, is partially idiomorphic, and the same is true of rocks which are on the border-line between diabase and gabbro. A porphyritic character, due to the development of relatively large crystals of felspar at an early stage, is not common : it is sometimes connected with an increasing fine- ness of texture of the rock on approaching the edge of an intrusive mass. Other occasional marginal peculiarities are flow-phenomena, vesicles or amygdules, and the development of 1 See chromolithograph of diabase in Berwerth, Lief. 1. 2 Q. J. G. S. (1886) xlii, pp. 68, 76, and figs. PI. v. QUARTZ-DIABASES. 127 a glassy base or sometimes of variolitic and allied structures. Rocks having these features and occurring as marginal modifi- cations of normal diabases do not differ in any essential from certain types of lavas, and will therefore not be noticed in this place. Leading types. A true quartz-diabase is not often met with. In any but quite fresh rocks, at least, it is not possible to be certain that quartz occurring interstitially is really an original constituent of igneous origin. Among the numerous dykes traversing the old gneiss of Sutherland are diabases of which some are quartz-bearing (Loch Glencoul, etc.). The chief constituent minerals are a basic plagioclase and a pale or colourless augite, the relations between the two being rather variable. A green or yellow-green hornblende occurs as a marginal alteration of the augite, especially around the grains of magnetite, and a little brown biotite is also associated with the latter. Apatite is the earliest and quartz the latest product of consolidation. The hornblende is connected with mechanical stress in the rock, and specimens may be collected to shew the complete amphibolization of the augite, as well as recrystallization of the felspar (see below, Chap. xxi). Another feature of these dykes is their fine-textured selvage, which seems in some cases to have been actually glassy. A well-known rock in the north of England is the Great Whin Sill 1 , which is intrusive in Lower Carboniferous strata, and ' extends from the Northumberland coast to the Eden valley. In its coarser central parts it sometimes approaches a gabbro in aspect, the augite becoming idiomorphic ; the fine- textured portions near the margin, on the other hand, take on an andesitic character, developing perhaps some glassy base; but the bulk of the intrusion is of diabase of a distinctive type. The normal structure is more or less ophitic, and the dominant constituents are a lath-shaped felspar, near andesine, and a pale brown augite, often with basal striation. The iron-ore is titaniferous, and may perhaps represent minute intergrowths of magnetite and ilmenite. Apatite occurs sparingly. An accessory mineral is bronzite, tending to be 1 Teall, Q. J. G. S. (1884) xl, 640-657, PI. xxix : also Brit. Petr. PI. xin, tig. 2. 128 QUARTZ-DIABASES. replaced in the usual fashion ; brown mica is occasionally seen, and a little brown hornblende is often present, bordering the augite with crystallographic relation. Quartz is detected in all the coarser varieties of the rock, and is at least in part original, since it frequently occurs in micrographic inter- growth with felspar. The rock is thus a quartz-diabase. Mr Teall 1 has described a similar rock from Ratho, near Edinburgh. With these rocks we may also compare that near Stirling 2 . The general mass of this is a simple diabase, the augite often shewing basal striation, but there are coarse- textured veins, which contain quartz in delicate micrographic intergrowth with part of the felspar. The Penmaenmawr 3 intrusion, probably of Ordovician age, is also characterized by quartz occurring interstitially in a micrographic intergrowth. In this rock bronzite becomes an essential constituent, being quite as abundant as the pale brown augite. The latter mineral often shews the delicate basal striation already noticed. Biotite is sometimes rather abundant, but the dominant type of rock is a quartz-bronzite- diabase. The structure of the rock is rather granulitic than ophitic, and it usually shews some approach to the characters of volcanic rocks in the occurrence of more than one generation of felspar. Some of the latest shapeless crystals are to be referred to orthoclase. The rock passes into a type which would be properly described as an andesite. The general body of the rock is traversed by comparatively coarse segrega- tion-veins of more acid composition 4 . Quartz-diabases are not unknown in America ; e.g., near St John, KB. 5 . The numerous sills of Ordovician age in Caernarvonshire 6 are of diabase without olivine, and have almost universally the ophitic structure. The felspar gives lath-shaped or rectangular sections from '05 to *5 inch long, with albite- but only occa- sionally pericline-lamellation : it often gives extinction-angles Teall, p. 190. 2 Monckton, Q. J. G. S. (1895) li, 480-491. Bala Vole. Ser. Caern. 65 ; Teall, PI. xxxv, fig. 2. Waller, Midland Naturalist (1885) viii, 1-7. Matthew, Trans. N. Y. Acad. Sci. (1895) xiv, 213, 214, PI. xv. fig. 2. Bala Vole. Ser. Caern. 75-86. VARIOUS BRITISH DIABASES. 129 indicating labradorite and neighbouring varieties. The augite is pale brown to almost colourless, and very rarely shews any approach to idiomorphisni. Besides the commoner decomposi- tion-products, there is often a fibrous colourless hornblende, fringing the augite but occupying the place of destroyed felspar, etc. The iron-ores include both magnetite and ilmenite, often together, and apatite is locally plentiful. Rhombic pyroxene is wanting, as well as olivine, while original hornblende and quartz are practically absent, and biotite very exceptional. These Caernarvonshire diabases are thus of very simple min- eralogical constitution. Despite the absence of olivine, they are of thoroughly basic composition. The diabases of similar age in Wicklow are also free from olivine, and are probably of more acid composition, some of them containing quartz. They are characterized by a partial or even total conversion of the ophitic augite into hornblende, with other changes ascribed to dynamic metamorphism 1 . In the Lake District diabases are not so largely deve- loped. The rock of Castle Head, Keswick, shews a divergence from the normal type in the presence of porphyritic idiomorphic crystals and crystal-groups of twinned augite. The general mass of the rock has had an ophitic structure, but is much decomposed, with the production of quartz, calcite, a feebly polarizing chloritoid substance, and little veins of fibrous serpentine (chrysotile). Numerous dykes of post-Carboniferous but pre-Permian age are found on the shores of the Menai Straits and in some other parts of Wales and England 2 . The smaller ones are augite-andesites, not calling for any special notice ; the larger may be classed as dolerites or as diabases shewing a tendency to a doleritic type. The dominant felspars give the usual rectangular section, and the light brown augite moulds round them in ophitic fashion, but the special feature of the rocks is the occurrence of a second and subordinate generation of felspar in allotriomorphic crystal-grains which have consolid- ated, on the whole, about simultaneously with the augite. They have less close twin-lamellation than the dominant 1 Hatch, G. M. 1889, 263-265. 2 Bala Vole. Ser. Caern. 109. 130 OLIVINE-DIABASES. felspars, are of more acid composition, and always shew a marked zonary banding between crossed nicols. These rocks contain magnetite, but not ilmenite. Numerous olivine-diabases are associated with the Car- boniferous strata of the Midlands. Good examples are seen in the Glee Hills, Shropshire 1 . The rock of Pouk Hill, near Walsall, is an ophitic olivine-diabase. In that of Rowley, near Birmingham, the augite occurs in little grains and tends to be idiomorphic 2 , or again there is a micrographic inter- growth of augite and felspar 3 . In this rock are relatively acid ol FIG. 25. OLIVINE-DIABASE, BONSALL, DEBBYSHIBE ; x 20. Shewing olivine- grains (ol), more or less completely serpentinized, magnetite (mg), and lath-shaped crystals of labradorite (Ib), set in a framework of crystalline augite (aw), which wraps round and encloses the felspar with typical ophitic structure [424]. segregation- veins, in which part of the felspar is orthoclase 4 . A few of the Derbyshire ' toad-stones ' have the structure of 1 This and many other British examples were noticed by Allport, Q. J. G. S. (1874) xxx, 529-567. 2 Teall, PI. xi. 3 Ibid. PL xxin, fig. 2. 4 Waller, Midland Naturalist (1885) viii, 261-266. OLIVINE-DIABASES. 131 ophitic diabases 1 (fig. 25), though Mr Arnold-Bemrose regards them as contemporaneous lavas. (See below, Chap. XIV.) Some of them contain peculiar pseudomorphs replacing olivine 2 and recalling the so-called iddingsite. Numerous intrusions of olivine-diabase, some of Carbon- iferous and others of Tertiary age occur in the southern half of Scotland 3 and in the western islands. As distinguished from the basalts and dolerites, they are typically ophitic rocks consisting of magnetite, olivine (often in fresh crystals), lath- shaped felspar, and crystal-plates of augite. Zeolites are frequent among the secondary products. Sills, dykes and rocks of ophitic olivine-diabase are abundant also in the Tertiary volcanic plateau of Antrim 4 . Without entering into an account of particular occur- rences in America, it may be stated that dykes of diabase, and especially of olivine-diabase, are widely distributed in the Archaean and other ancient formations of Canada and the northern United States 5 . Some of these rocks have special characters which connect them with the lamprophyres found in the same region. Emerson's mica-diabase from Franklin Furnace, N.J. 6 , is an example. Of hornblende-bearing-diabases a good example is found in a large dyke which runs on the east side of Holyhead Mountain 7 . The brown hornblende is very frequently in parallel intergrowth with augite, which it tends to envelope. The augite is a pale malacolite variety. Apatite and magnetite are abundant. The structure of this rock is very variable, sometimes the felspar, sometimes the augite, presenting idio- morphic boundaries to the other. 1 Teall, PL ix. 2 Arnold-Bemrose, Q. J. G. S. (1895) li, 613-616, PL xxiv. 3 The rock quarried at Corstorphine near Edinburgh is a good example of the earlier set : see coloured plate in Cole's Stud, in Micro. Sci. (1882) No. 32. For Tertiary examples see Teal], PL x. 4 Watts, Guide, 78. 5 A list of references to described examples is given by Kemp and Marsters, Bull. No. 107 U. S. Geol. Sur. (1893) 28, 29. 6 A. J. S. (1882) xxiii, 376-380. 7 G. M. 1888, 270, 271. 92 132 HORNBLENDE-DIABASES: TESCHENITES. Other examples occur in the neighbourhood of Penar- fynnydd 1 , near Sarn, in the south-west of Caernarvonshire, and apparently form laccolitic masses of Bala age. Here the brown hornblende is in part original, often enveloping the augite with parallel growth, but in part derived from the augite. By the coming in of abundant olivine and the dwindling of the felspar, the rock passes, though abruptly, into the hornblende-picrite already noticed. It has already been alluded to in connection with the diorites (p. 63). Of diabases containing derivative hornblende only (the epidiorites of some writers), we have numerous examples in this country. A good one is found at Guns Fell in the Cross Fell district 2 . Many of the 'greenstones' of Cornwall are much altered diabases shewing uralitization, chloritization, serpentinization, and other changes ; but the rocks so named in the field include also old basic lavas and other types 3 . We may briefly notice in this place the peculiar group of rocks, named teschenite by Hohenegger, occurring as intrusions in the Cretaceous of Silesia and Moravia (Teschen, Neutitschein, Sohla, etc.). They consist mainly of augite, brown hornblende, plagioclase, apatite, and analcime. The augite is often of a violet tint and strongly pleochroic, and it is frequently bor- dered by hornblende in parallel position. The apatite is very abundant and builds large prisms. The analcime is probably secondary, and has been supposed to be derived from nepheline, while some observers have recorded the presence of nepheline in the rocks. Teschenites occur in the Caucasus, in Portugal, etc., and a similar rock is found at Car Craig in the Firth of Forth*. It is rich in reddish brown, pleochroic augite, and contains altered felspar, analcime and other zeolites, iron ores, and brown mica (probably secondary). It presents points in common with the neighbouring picrite of Inchcolm. All these rocks are typically non-ophitic, but others more resembling normal diabases have also been included under the name 1 Q. J. G. S. (1888) xliv, 450-454 ; Bala Vole. Set: Caern. 92-97. 2 Q. J. G. S. (1891) xlvii, 524. 3 J. A. Phillips, Q. J. G. S. (1876) xxxii, 155-178; (1878) xxxiv, 471-496, PI. xx-xxn. 4 Teall, PI. xxn, fig. 1. AMERICAN TESCHENITES. 133 teschenite. In San Luis Obispo County, California, and at Point Sal in Santa Barbara County, Fairbanks 1 has described diabasic rocks with analcime, which he considers to be derived from nepheline. The analcime occurs interstitially or some- times in hexagonal pseudomorphs. The other constituents are zoned labradorite, pale augite (not segirine), magnetite, apatite, and in the former occurrences olivine. In general, we may use the term teschenite for a nepheline- bearing diabase or for a diabase which, from an abundance of secondary minerals rich in soda, may be supposed to have contained nepheline. Silurian intrusions of this type occur in association with the nepheline-syenite of Montreal. They contain both nepheline and sodalite, and some have olivine. 1 Bull. Dep. Geol. Univ. Calif. (1895) i, 278-300, PL xv, xvi ; (1896) ii, 19-38, PI. n. CHAPTER X. LAMPROPHYRES. THE lamprophyres are a peculiar group of rocks occurring typically as dykes or other small intrusions. Chemically they are characterized by containing, with a medium or low silica- percentage, a considerable relative quantity of alkalies (especi- ally potash), while the oxides of the diatomic elements are also abundantly represented. This shews itself in the commoner types of lamprophyres by an abundance of brown mica, and indeed the lamprophyres as a family are rich in ferro-magnesian silicates. They are fine-grained rocks, but almost always holocrystalline, and their structure is in some respects peculiar. Von Giimbel's name lamprophyre has been extended by Rosenbusch to cover the various members of this family. The best known varieties are mica-lam prophy res (' mica-traps ', trer. Glimmertrapp). Of these, two types have long been recog- nized, a chief point of distinction being the predominance of orthoclase in one and plagioclase in the other. To these types are given the names, respectively, minette (a word taken from the miners of the Vosges) and kersantite (from Kersanton, near Brest). To these Rosenbusch added two other types for rocks in which the place of biotite is taken by augite or hornblende. He separated those with dominant orthoclase (vogesite) from those with dominant plagioclase (camptonite). It should be noted that the criterion of the felspars does not lead in this family to a very natural division, especially when much of the potash in the rocks is present in mica. Further, BIOTITE OF LAMPROPHYRES. 135 the decomposition of the rocks often renders the identification of the felspars difficult. For most purposes it is perhaps sufficient to distinguish the rocks merely as mica-, hornblende-, and augite-lamprophyres. There are other types of very basic composition, which are devoid of felspar. The rocks of this family have a wide range of chemical composition. Their equivalents, from this point of view, among the volcanic types are chiefly basaltic rocks, and especially leucite- and nepheline-bearing basalts. From these the lamprophyres as a whole differ considerably in mineralo- gical composition, olivine being wanting or poorly represented in many of the types, and the felspathoid minerals occurring only very exceptionally ; while, on the other hand, brown mica, a mineral by no means characteristic of basaltic lavas, is a prominent constituent in many of the lamprophyres. Constituent minerals. The characteristic mineral of those lamprophyres most usually met with is biotite, which occurs in hexagonal flakes. The extinction-angle (3 or 4) is sufficient to shew frequently a lamellar twinning parallel to the basal cleavage. The flakes are very commonly bleached in the interior, retaining only at the margin the normal deep brown colour (fig. 26 A). With the bleaching there is a certain diminution in birefringence. More rarely we find a dark interior with a pale border, or a dark nucleus and border with a pale intermediate zone. Complete decomposition results in a pale, feebly polarizing substance as a pseudomorph. A greenish chloritic alteration is also found. Iron-oxide separates out, usually as limonite, and a carbonate (calcite or dolomite) is produced as little wedges or lenticles along the cleavages of the mica (tig. 26 A). The titanic acid of the mica separates out as rutile, in fine needles arranged in three sets at angles of 60 : this is well seen in basal sections (fig. 26, B). The original inclusions of the biotite are apatite, and sometimes magnetite and zircon. Short columnar crystals of augite occur in many lampro- phyres, shewing sharp outlines with an octagonal cross-section, and sometimes lamellar twinning. When fresh, the mineral is pale green or almost colourless in slices, but it is readily 136 MINERALS OF LAMPROPHYRES. replaced by serpentine, calcite, chlorite, etc., in good pseudo- raorphs. In other cases uralitization may be noticed. The FlG. 26. MlCA-LAMPBOPHYEE (MINETTE). A. Helm Gill, near Dent, Yorkshire ; x 20 : shewing the internal bleaching of the mica-flakes and the formation in them of lenticles of calcite (ca). The opaque iron-ore (py) is pyrites [444]. B. Decomposing biotite in mica-lamprophyre, Budlake, near Exeter ; x 100 : shewing the production of rutile-needles and patches of limonite [1346]. augite crystals are sometimes coated with flakes of biotite. The most usual occurrence of hornblende is in long well-shaped prisms, frequently twinned, but it has some variety of habit. The colour is brown or sometimes green. The mineral may be converted into a chloritic substance with separation of iron-oxides. A striking feature in the lamprophyres is that the felspars do not usually occur as phenocrysts. The nature of the felspar in the more altered rocks can be verified only after removing the carbonates from the slice with dilute acid. The small columnar or tabular crystals of plagioclase shew albite-lamella- tion and frequently zonary banding. They often have a kind of sheaf-like grouping. Decomposition, beginning in the interior, gives rise to abundant calcite. The ortkoclase, and MINERALS OF LAMPROPHYRES. 137 perhaps anorthoclase, build short rectangular crystals, simple or carlsbad twins, often clouded or with ferruginous staining. Some of the more acid lamprophyres have a certain amount of quartz, which is either the latest product of consolidation or is intergrown with a portion of the felspar with micro- graphic structure. A common accessory in some lamprophyres, and an essential in certain types, is olivine, which builds relatively large perfect crystals, or sometimes groups of rounded grains. It is occa- Ol* FlG. 27. MlCA-LAMPROPHYRE, BAWTHBY BRIDGE, NEAR SEDBERGH, YORKSHIRE ; x 20. There are abundant flakes of biotite, with internal bleaching, and octahedra of magnetite. The part shewn clear is an aggregate of felspar crystals obscured by secondary calcite dust. Olivine is represented by pseudomorphs (ol) of calcite coated with red iron-oxide. There is an enclosed grain of quartz (q) with a corrosion-border of augite, now decomposed [1598]. sionally found fresh, but very commonly represented by pseudomorphs of carbonates and serpentine (fig. 27). The iron-ores are not often very abundant, and may be quite wanting. The most usual is pyrites, but octahedra of magnetite are also found. 138 STRUCTURES OF LAMPROPHYRES. A constant and abundant accessory is apatite, but it is sometimes in such fine needles as to be invisible except by oblique illumination. Sphene and zircon are only exception- ally met with. Structures and peculiarities. Many of the lam- prophyres are non-porphyritic, with a rather exceptional structure due to a strong tendency to idiomorphism of all the constituent minerals. This is the panidiomorphic struct- ure of Rosenbusch 1 , The porphyritic members of the family, again, have a peculiarity, in that the porphyritic character is produced by a recurrence of the ferro-magnesian constituents, not of the felspars. Any recurrence of the latter, and especially of orthoclase, is rare, but two generations of biotite or of horn- blende are seen in many of the rocks. When olivine occurs, it is in conspicuous crystals, but only of one generation. Without shewing any real flow-structure, the felspars of the rock sometimes have a special grouping in sheaf-like or rudely radiating fashion. Exceptionally orthoclase is moulded on the other constituents : usually it is idiomorphic, save when it builds micrographic structures with quartz. There is little indication of any isotropic residue in the typical lamprophyres, though in some cases little ovoid vesicles, filled with secondary products, suggest the former presence of some glassy matter, now perhaps devitritied. In the most basic types of lamprophyres, however, there is what has been de- scribed as a glassy base. The rocks are remarkably prone to decomposition, and often have 20 or 30 per cent, of calcite and other secondary products. Grains of quartz and crystals of alkali-felspars are found, though very sparingly, in many lamprophyres. Their sporadic occurrence and, still more, some curious features which they invariably present compel us to regard them as something apart from the normal constitution of the rock and of quasi- foreign origin. The enclosed quartz grains (fig. 27) are of rounded form with evident signs of corrosion, and are seen to be surrounded by a narrow ring or shell due to a reaction 1 See chromolithograph of kersantite in Berwerth, Lief, i ; also Rosenbusch, Mass. Gest., PL in, fig. 1. QUARTZ AND FELSPARS ENCLOSED IN LAMPROPHYRES. 139 between the quartz and the surrounding magma. This shell is probably in the first place of augite, but it is often found to consist of minute flakes of greenish fibrous hornblende or of calcite and chloritoid products. The quartz having this mode of occurrence must be distinguished from genuine derived fragments torn from other rocks : these are of irregular form, often complex, and may contain inclusions unknown in the corroded quartz-grains. The two may sometimes be seen in the same slide. The enclosed felspar crystals are always of an acid species either orthoclase or a plagioclase rich in soda. The crystals are corroded so as to present a rounded outline, but not re- FIG. 28. OLIGOCLASE CRYSTAL ENCLOSED IN A LAMPBOPHYRE DYKE AT GILL FARM, NEAK SHAP WELLS ; x 20. Crossed nicols. The crystal is rounded by magmatic corrosion and bordered by a narrow margin of orthoclase (or). In addition to the albite-lamellation of the oligoclase (a), there is a carlsbad twinning (c) common to both felspars [1155]. duced to mere round grains. The plagioclase thus corroded is bordered by a narrow margin of orthoclase due to the action of the magma (tig. 28). 140 NORTH-COUNTRY MICA-LAMPROPHYRES. Illustrative examples. The best-known British examples occur as small dykes and sills in the north of England 1 , and are of an age between the Silurian and the Carboniferous. The dykes are numerous in the southern part of the Lake District, from Windermere to Shap and on to Sedbergh, and they are seen again in the Lower Palaeozoic inliers of Ingleton, Eden- side, and Teesdale. The rocks are mica-lamprophyres, but many of them contain subordinate augite, always in perfect crystals, but often decomposed. The relative proportions of orthoclase and plagioclase vary, so that some examples would be named minette and others kersantite, the latter being perhaps the commoner. Good pseudomorphs after olivine are seen in the dykes in the Sedbergh district (fig. 27). The dykes at Cronkley, in Teesdale, have abundant pseudomorphs with hexagonal and quadrangular outlines representing some mineral not yet certainly identified. Scattered quartz-grains with the characteristic corrosion- border occur in many of the dykes (fig. 27), and felspars, both orthoclase and oligoclase (fig. 28), are enclosed sporadically in the Edenside intrusions, and more abundantly in those to the south of the Shap granite. These rocks shew various transitions from typical lamprophyres to a micaceous quartz- porphyry of one of the less acid types, and indeed very different kinds of rocks occur imperfectly mingled in one and the same dyke. Quartz does not occur as a normal constituent in most of the north country lamprophyres, though it is found in the transitional rocks just mentioned. In an intrusion at Sale Fell, near Bassenthwaite, quartz occurs partly as interstitial grains, partly in micrographic intergrowth, and the rock shews considerable resemblance to the original kersantites of Brittany 2 . The last-named rocks are sometimes even-grained, sometimes porphyritic (' porphyrites micacees ' of Barrois). An augite-bearing minette seems to be one of the commonest types of lamprophyres. It is seen in Cornwall (Trelissick Creek, 1 G. J/. 1892, 199-206, with numerous references. 2 Barrois, M. M. vii, 166-168 (Abtsr.) Fouque and Le"vy, PI. ix : cf. Saxon kersantite, chromolitb. in Berwerth, Lief. i. VARIOUS BRITISH LAMPROPHYRES. 141 etc.}, in the Channel Islands (Doyle Monument, Guernsey), and at numerous foreign localities (e.g. Plauen'scher Grund, near Dresden). With more abundant augite (e.g. Weinheim in the Odenwald) it passes into the augite-vogesites. The typical vogesites of the Vosges, etc., have sometimes augite, sometimes hornblende, as the dominant coloured constituent. A good rnica-lamprophyre occurs at Peel Castle, in the Isle of Man. In Ireland numerous lamprophyres, both micaceous and hornblendic, are known from Galway, the Raphoe district, the coast of Down, etc. Mr Watts ' describes one from Lettery, near Clifden, as a camptonite, another from Clondermot, near Raphoe, as a vogesite, and most of those in Co. Down carry hornblende as well as biotite. The numerous lamprophyres of Scotland are for the most part not yet described. Some of the rocks occurring as sills FlG. 29. HORNBLENDE-LAMPKOPHYBE (APPBOACHING CAMPTONITE), FBOM INTRUSIVE SILL IN DURNESS LIMESTONE, LoCH AsSTTNT ; X 20. Shewing phenocrysts of green hornblende in a panidiomorphic ground- mass of plagioclase and hornblende, with a little magnetite and apatite [1687]. in the Assynt district of Sutherland 2 seem to be rather camptonites than diorites. They are characterized i)y green 1 Guide, 53, 73-75. - Teall, G. M. 1886, 346-353. 142 AMERICAN HORNBLENDE-LAMPROPHYRES. hornblende in rather slender twinned prisms (fig. 29). One sill, however, is a mica-lamprophyre with accessory augite and olivine (both destroyed). In America mica-lamprophyres of the minette type have been described from Coanicut Island, R. I. 1 , from the Sweet Grass Hills, Montana 2 (with augite), from Notre Dame Bay, Newfoundland 3 (with augite and hornblende), etc. In the north-eastern United States and in Canada horn- blende-lamprophyres of the camptonite type are widely dis- tributed. The name was first applied by Rosenbusch to rocks described by Hawes 4 from Campton Falls, N.H., while closely similar rocks are found near Montreal 5 , at Summit Station, Yt. 6 , at several points on the Hudson River highlands 7 and in the Lake Cham plain district 8 , and (with less abundant horn- blende) at the Forest of Dean iron-mine, N.Y. 9 In all these idiomorphic brown hornblende, usually in two generations, is the chief constituent, felspar is subordinate, and augite is at most an accessory. In other varieties augite becomes prominent in addition to the dominant hornblende, and occurs porphyritically, as at Lake. Memphremagog, Vt. 10 , and near Whitehall, N.Y. 11 One from Danbyborough, Vt. !2 , has no phenocrysts of hornblende, but two generations of augite and rather abundant biotite. When augite predominates the rock may be termed augite-camptonite, but such rocks shew an approach to diabase by the augite losing its sharply idio- morphic habit. Examples near Lewiston and Auburn, in Maine 13 , have abundant phenocrysts of olivine, more rarely Pirsson, A. J. S. (1893) xlvi, 374. Weed and Pirsson, A. J. S. (1895) 1, 313. Wadsworth, A. J. S. (1884) xxviii, 99, 100. A. J. S. (1879) xvii, 147-151. Harrington, Rep. Geol. Sur. Can. 1878. Nason, A. J. S. (1889) xxxviii, 229. Kemp, Amer. Naturalist, 1888, 694-696, PI. xn. Kemp and Marsters, Trans. N. Y. Acad. Sci. (1891) xi, 21, 22; Bull. No. 107 U. S. Geol. Sur. (1893) 29-32. 9 Kemp, A. J. S. (1888) xxxv, 331, 332. 10 Marsters, Amer. Geol. (1895) xvi, 25-39. 11 Kemp and Marsters, ibid. (1889) iv, 97-102. 12 Marsters, ibid. (1895) xv, 368-371. 13 Merrill, ibid. (1892) x, 49-55. VARIOUS AMERICAN LAMPROPHYRES. 143 of augite, and still more rarely of plagioclase in a ground-mass of icliomorphic brown hornblende, faintly purple augite, laths of felspar, some iron-ore, and apparently some glassy residue. Of the ultrabasic types of lamprophyre a brief notice will suffice. In these rocks felspar is typically absent, and olivine is often well represented in addition to the bisilicates, while there is also more or less colourless isotropic base, either a glass or perhaps analcime 1 . In the Monchique type 2 the characteristic minerals are olivine and augite, or in some varieties hornblende. Such rocks are met with as dykes in Portugal, Brazil, Montana, Arkansas 3 , and the Lake Champlain district 4 . In the Fourche type, from the two last-named regions 5 , olivine is wanting or poorly represented, and the bisilicates are almost the only minerals; while in the Ouachita type 6 , also free from olivine, biotite becomes the characteristic constituent. It will be observed that these various types, as well as the camptonites, are always found in association with nepheline- syenites and allied rocks, while the more ordinary mica-lam- prophyres occur in connection with granites, etc. 1 Pirsson, Journ. of Geol. (1896) iv, 679-690. 2 Hunter and Rosenbusch, M. M. x, 177, 178 (Abstr.). 3 J. F. Williams, Igneous Rocks of Arkansas, vol. n of Rep. Geol. Sur. Ark. for 1890, 151-157, 290-295, 353. 4 Kemp and Marsters, Trans. N. Y. Acad. fifci. (1891) xi, 22, 23 ; Bull. No. 107 U. 8. Geol. Sur. (1893) 32-35. 5 Williams, I.e., 107, 108, 290 ; Kemp and Marsters, I.e., 35, 36. 8 Kemp (Williams, I.e.), 394-398. C. VOLCANIC ROCKS. UNDER this head we shall treat only the solid rocks of volcanic origin (lavas), reserving the fragmental products of volcanic action for the sedimentary group. With the true extruded lava-flows will be included similar rocks occurring in the form of dykes, etc., in direct connection with volcanic centres, the common feature of all being that they have con- solidated from fusion under superficial conditions, i.e. by com- paratively rapid cooling under low pressure. This mode of origin has given the rocks as a whole characters which place them in contrast with the plutonic group, while the types treated above under the head of 'hypabyssal' have in some respects intermediate characters. Many volcanic outpourings have undoubtedly been submarine, and when these have taken place under a great depth of water the products may be ex- pected to approximate in some measure to the characters of rocks of deep-seated origin. In general, however, the contrast between volcanic and plutonic types of structure is well marked. The presence of a glassy (or devitrified) residue, though not ^eculiar^io"TTtfcanic rocks, ^gjhighly ^haracterisiacjoLtbam, and especially of the more acid members. Other features characteristic of lavas, though not confined to them, are the vesicular and amygdaloidal structures, and the various fluxion- phenomena, including flow-lines, parallel orientation of pheno- crysts, banding, drawing out of vesicles, etc. The great majority of the volcanic rocks have a porphyritic structure, i.e. their constituents belong to two distinct periods of consolidation, the earlier represented by the porphyritic CHARACTERISTICS OF VOLCANIC ROCKS. 145 crystals or ' phenocrysts ' 1 , and the later by the 'ground-mass,' which encloses them, and commonly makes up the bulk of the rock. This ground-mass may, and usually does, include some glassy residue or ' base ' : if the ground is wholly glassy, we have what is termed the ' vitrophyric ' structure. The same mineral say quartz or augJtB- !1 ^Hay"l>ccur both among the phenocrysts and as a constituent of the ground-mass. When such a recurrence is found, the crystals of the earlier genera- tion are distinguished from those of the later by their larger size, often by their more perfect idiomorphism, and in some cases by fracture, corrosion, or other evidence of vicissitudes in their history. The two periods of consolidation are styled by Rosenbusch the * intra telluric ' and the 'effusive.' the former being considered as tEe~result of crystallization prior to the pouring out of the lava, and so under more or less deep-seated conditions. When we speak of the consolidation of a lava at the earth's surface, we must be understood to refer to the ground-mass of the rock. In some few types of lavas the phenocrysts fail altogether, and the effusive period is the only one represented. The various types will be grouped under families, to be taken roughly in order, beginning with the most acid. It is customary to speak of the several families of lavas as answer- ing to the commonly recognized families of the plutonic rocks the rhyolites to the granites, the trachytes to the syenites, etc. but such a correspondence cannot be followed out with great exactness. It is certain that a given rock-magma may result in very different mineral-aggregates according as its consolidation is effected under deep-seated or under surface conditions ; and in the latter case, moreover, much of the rock produced may consist of unindividualised glass. It is more especially in the volcanic rocks that the Con- tinental petrologists have insisted upon a division into an 'older' and a 'younger' series (' palseovolcanic ' and ' neo- volcanic '), an arbitrary line being drawn between the pre- Tertiary lavas and the Tertiary and Recent. This distinction 1 This convenient term, due to Prof. Iddings, will be adopted here. Mr Blake has proposed the word 'inset,' as corresponding to the Ger. 'Einsprengling.' H. P. 10 146 NOMENCLATURE OF VOLCANIC ROCKS. is rejected by the British school, and will find no place in the following pages 1 . The simplified grouping of the volcanic rocks by their essential characters, without reference to their age or supposed age, involves some modification of the double nomenclature in use among the German and French writers. I The names employed by them for the younger lavas only will ' here be extended to all rocks of the same character, irrespective ' of their geological antiquity. The names applied by the Con- tinental writers to the pre-Tertiary lavas have also been used / habitually for hypabyssal rock-types, and may now be restricted ' to these latter. Some of them (quartz-porphyry, porphyrite, diabase, etc.) we have already used in this sense. 1 On this question see Sci. Progr. (1894), ii, 48-63. CHAPTER XI. RHYOLITES. IN the rhyolite family we include all tin; truly acid lavas ; rocks of porphyritic or vitrophyric structure, in which alkali- felspars and usually quartz figure as the chief constituent minerals. By the older writers most of these rocks were in- cluded, with others, under the large division ' trachyte.' The present family was separated by von Richthofen with the name 'rhyolite,' expressing the fact that flow-structures are commonly prominent in the rocks. Roth used the term ' liparite ' in nearly the same sense. The Continental petrographers, fol- lowing their regular principle, use these names for the Tertiary and Recent acid lavas only, the older (pre-Tertiary) being more or less arbitrarily separated and designated by other names (quartz-porphyry, porphyry, etc.); and some English geologists have tacitly adopted a like division, calling the older rhyolites, which have often suffered various secondary changes, quartz- felsites, felsites, etc. Some geologists distinguish between potash- and soda- rhyolites, according to the predominance of one or the other of the alkalies, but in fine-textured or glassy rocks this difference does not always express itself in the minerals evident. There is, however, a peculiar group of acid lavas very rich in alkalies, and especially in soda: these rocks, the ' pantellarites ' of Forstner, contain special characteristic minerals. We shall consider briefly the characters of the phenocrysts or enclosed crystals and of the ground-mass. \ In some rhyolites 102 148 MINERALS OF RHYOLITES. the phenocrysts occur only sparingly, or may even fail alto- gether. I Phenocrysts. Among the phenocrysts or porphyritically enclosed crystals of the rhyolites, the most constant are alkali felspars ; both orthoclase (including sanidine) in tabular or columnar crystals, simple or twinned, and an acid plagioclase, ranging from albite to oligoclase, in tabular crystals with the usual twin-lamellation. A parallel intergrowth of the mono- clinic and triclinic species is occasionally found. The felspars often contain glass- 1 and gas-cavities, but rarely fluid-pores : such minerals as apatite, magnetite, biotite, etc., may be sparingly enclosed. Certain rocks specially rich in soda (pan- tellarites, etc.] have anorthoclase. Quartz, when present, occurs in dihexahedral crystals, often corroded and with inlets of the ground-mass. Besides occasional inclusions of minerals of early consolidation, it contains glass- but rarely fluid-cavities. The more basic silicates are not present in great abundance. The most usual is biotite in deep-brown hexagonal flakes, with only occasional inclusions of apatite, zircon, or magnetite. A greenish augite with octagonal cross-section may be present, but brown hornblende is much less common. The pantellarites have the brown triclinic amphibole cossyrite, with intense ab- sorption and pleochroism. The most usual iron-ore is magnetite, but it is rarely abundant. Needles of apatite and minute crystals of highly refringent and birefringent zircon may also occur in small quantity. In rarer cases garnet is found instead of a ferro- magnesian bisilicate. Ground-mass and structures. The rhyolites exhibit in their ground-mass a great variety of texture and structure. The texture may be wholly or partly glassy; or cryptocrystalline, often with special structures ; or, again, evidently crystalline, though on a minute scale. Further, these several varieties of ground-mass may be associated in the same rock and in the same microscopical specimen. Fluxion is frequently marked by banding, successive bands being of different textures, so 1 Cohen, PI. iv, fig. 4. GLASSY GROUND-MASS OF OBSIDIAN. 149 that thin layers of glassy and stony or spherulitic nature altern- ate with one another. The vitreous type of ground-mass alone is found in the obsidians 1 . These rocks, colourless or very pale yellow in thin slices, afford good examples of structures common to all FIG. 30. GLASSY RHYOLITE (OBSIDIAN), TELKIBANYA, NEAR SCHEMNITZ, HUNGARY ; x 20. Shewing sinuous flow-lines traversed by a system of curving perlitic fissures [G. 329]. the natural glasses ; especially the perlitic cracks (fig. 30), produced by contraction of the homogeneous material 2 , and the vesicular structure due to the rock-magma having been distended by steam-bubbles. In extreme cases the cavities are so numerous as to make up the chief part of the volume of the rock, and we have the well-known pumice (Fr. ponce, Ger. Bimstein). The vesicles are commonly elongated in the direction of flow, and may even be drawn out into capillary 1 The less common glassy rocks of the trachyte and phonolite family and of the dacites are also termed obsidian. They are not easily distin- guished from the rhyolite-glasses. Some of the rocks styled pitchstones are lavas of the obsidian type, usually of acid composition. 2 Cohen, PI. xxxvn, figs. 1, 2. On artificial production of perlitic structure see Cole, G. M. 1880, 115-117 ; Chapman, ibid. 1890, 79, 80 ; fig. in Judd's Volcanoes, p. 109. 150 CRYSTALLITES IN OBSIDIAN. tubes. In the older lavas vesicles are usually filled by second- ary products, and become amygdules. In many cases a ground-mass consisting essentially of glass FIG. 31. CRYSTALLITES IN OBSIDIAN. A. Margarites, Obsidian Cliff, Yellowstone Park; x 400 [477]. B. Trichites, Telkibanya, Hungary ; x 100 [G. 327]. C. Longulites and swallow-tailed crystallites, Hlinik, Hungary ; x 200 [G. 70]. D. Flow-structure marked by arrangement of twisted trichites, Pra- bacti, Java ; x 200 [G. 64]. encloses minute bodies known as crystallites (fig. 31), which may be regarded as embryonic crystals 1 . They have definite forms, but no perfect crystal boundaries, and the more rudi- mentary types cannot be subjected to optical tests to determ- ine their nature. The simplest effort at individualisation from the vitreous mass results in globulites, minute spherical bodies without action on polarized light. They occur in profusion in many obsidians, either uniformly distributed or aggregated into cloudy patches (cumulites). From the partial coalescence of a series of globulites arranged in a line result margarites 2 , resembling strings of pearls. A high-power objective (say 1 inch) is often necessary to resolve this beaded 1 See Eutley, 31. 31. (1891) ix, 261-271, and Plate; Zirkel, Micro. Petr. Fortieth Parallel, PI. ix, figs. 1-4 ; Eosenbusch-Iddings, PI. n, in. 2 Cohen, PI. n, fig. 3. CRYPTOCRYSTALLINE STRUCTURE IN RHYOLITES. 151 structure. Long threads of this nature may extend in the direction of flow but with numerous little twists 1 . Similar threads with curved hair-like form, known as trichites 2 , often occur in groups originating in a common nuc'IeTTs'. These bodies, in which a beaded structure may or may not be observable, often seem to belong to a stage of development later than the cessation of flowing movement in the mass 3 . The small rod-like bodies known as lojjguiites 4 , sometimes slightly clubbed at the ends 5 , may be regarded as built up by the complete union of rows of globulites. They often occur in crowds, with a marked arrangement parallel to the direction of flow 6 . The transition from margarites to longulites is often seen, some of the little rods resolving into beaded strings, while others do not. The larger crystallitic bodies termed microlites are possibly to be conceived as built up from longu- lites. Various incomplete stages may be observed, the ends of the imperfect microlites having a brush-like form (scopulites of Rutley) or being forked in swallow-tail fashion. Fully developed microlites have an elongated form, and are indeed small crystals giving the optical reactions proper to the mineral (felspar, augite, hornblende, etc.) of which they consist. An original cryptocrystalline or ' microfelsitic ' ground-mass is found in some rhyolites, though it seems to be more charac- teristic of intrusive types (approaching what we have styled quartz-porphyries) than of true surface lavas. It consists in a granular mixture of felspar and quartz on so minute a scale that the individual grains cannot be resolved in a thin slice. There is no doubt, however, that in many old acid lavas a cryptocrystalline ground-mass has resulted from the devitrifi- cation (Ger. Entglasung) of a rock originally vitreous. The process has often begun along perlitic fissures, or flow-lines, and the earlier stages are beautifully displayed in such rocks as the Permian rhyolites ('pitchstones ') of Meissen in Saxony. No single criterion can be set up for distinguishing an original 1 Zirkel, Micro. Petrogr. Fortieth Parallel (1876), PI. ix, figs. 3, 4. 2 Cohen, PL n, fig. 2. 3 Zirkel, I.e., figs. 1, 2. 4 Cohen, PI. n, fig. 1. * Fouque and L6vy, PI. xvi, fig. 2. Cohen, PI. xn, figs. 1, 2. 152 MICROCRYSTALLINE STRUCTURE IN RHYOLITES. from a secondary cryptocrystalline structure. In a rock otherwise fresh, however, there will generally be no reason to suspect devitrification ; while, on the other hand, the presence of perlitic cracks is often taken to indicate that the rock in which they occur was originally glassy. A microcrystalline (as distinguished from cryptocrystalline) ground-mass is not very prevalent in true acid lavas, but may occur as bands alternating with glassy or microspherulitic bands, often on a small scale. When an evident microcrystal- line structure has been set up as a secondary alteration, it probably indicates, as a rule, -something more than the merely physical change of devitrification. It is often connected with an introduction of silica from an external source, and in the resulting microcrystalline mosaic quartz often plays a more important part than it would do in a normal igneous rock. In some of the partly silicified Ordovician rhyolites of West- morland a secondary quartz-mosaic still shews clear indication of former perlitic cracks, outlined by dust, as well as the characteristic banding. In these rocks, too, siliciftcation has sometimes affected not only the ground-mass but the felspar phenocrysts. Another change which has sometimes affected the ground- mass of rhyolites, as well as the felspar phenocrysts, is that which is characterized by an abundant production of epidote 1 . Spherulitic and allied structures. The spherulitic growths which are common in many acid lavas may be con- veniently divided into the larger and the smaller. Under the former head we have spherulites, often isolated, with diameters ranging from a fraction of an inch to several inches. They are best studied in certain obsidians, where they are usually of distinctly globular form and with well-defined boundary. The examples which have been most carefully examined, and may be taken as typical, consist mainly of extremely delicate fibres of felspar, arranged radially and on the whole straight, but 1 E.g. Rutley, Q. J. G. S. (1888) xliv, 740-744, PI. xvn. Most of the so-called ' epidosites ' (epidote-quartz -rocks) are, however, decomposed diabases, etc. The name was first used by Pilla for rocks of this nature associated with the gabbros of Tuscany. LARGE SPHERULITES IN RHYOLITES. 153 often forked or branching 1 . In the spherulites of perfectly fresh rocks the space between the fibres is found to be occupied in great part by aggregates of tridymite. In older spherulites, where tridymite is not recognized, quartz may perhaps be considered to represent it. In any case the structure is to be made out only in carefully prepared and very thin slices. It may often be observed that the flow-lines of the lava pass undisturbed through the spherulites, indicating that the latter crystallized after the cessation of movement. Spherulites are FIG. 32. OBSIDIAN, VULCANO, LIPAKI Is. ; x 20. The glassy matrix encloses isolated spherulites, with some tendency to coalesce in bands following the direction of flow. The flow-lines pass uninterruptedly through the spherulites [1785]. often developed along particular lines of flow, and may coalesce into bands (fig. 32). These larger spherulites shew many special peculiarities in different examples. Sometimes their outward extension has been effected in two or more stages, which are marked by a change in the character of the growth. Again, curious phenomena arise from the formation of shrinkage-cavities (lithophyses) in connection with spherulites. Some remarkable 1 See Cross, Bull. Phil. Soc. Washington (1891) xi, 411-444 ; Iddings, ibid. 445-464, with Plates. 154 SPHERULITES AND LITHOPHYSES. examples of lithophyses have been described from the Yellow- stone Park 1 and other districts in the United States 2 , from Hungary, and from Lipari 3 . A peculiar feature is the occurrence in the hollows of perfect crystals of the iron-olivine (fayalite), as well as aggregates of tridymite, and in some cases crystals of garnet, topaz, etc. The complex forms of these lithophyses can be realised only from specimens or figures. They must be distinguished from ordinary ovoid vesicles. The large spherulites are in some cases only skeleton- structures, the divergent rays being imbedded in glass. Such skeleton-spherulites, in a devitrified matrix, have been described by Prof. Cole 4 in the ' pyromerides ' of Wuenheim, in the Vosges. Examination of the older acid lavas shews that the large spherulites are specially susceptible to certain chemical changes. They are often found partly or totally replaced by flint or quartz, while their insoluble decomposition-products remain as roughly concentric shells of a chloritoid or pinitoid sub- stance. Again, a central hollow is often found, and it is not always clear whether this is due entirely to decomposition or partly represents an original lithophysal cavity 5 . The very minute spherulites commonly occur in large numbers, closely packed together, so as to constitute the chief bulk of particular bands, or even of the whole ground-mass of the rock. This is the microspherulitic structure 6 . The true nature of these very minute bodies, as composed of fine fibres of felspar with quartz, is a matter rather inferred than seen in any given case, but the radiate growth is detected by means of the ' black cross ' which each individual spherulite shews 1 Iddings, Obsidian Cliff, in 7th Ann. Hep. U. S. Geol. Surv. (1888) 265, 266, PI. xn-xiv ; A. J. S. (1887) xxxiii, 36-43 ; M. M. vii, 175-177 (Abstr.). 2 Nathrop, Colorado ; see Cross, Proc. Colo. Sci. Snc. 1886, 62-66. 3 Cole and Butler, Q. J. G. S. (1892) xlviii, 438-443, PI. xn ; John- ston-Lavis, G. M. 1892, 488-491. 4 G. M. 1887, 299-303. 5 See especially Cole, Q. J. G. S. (1886) xlii, 183-190; (1892) xlviii, 443-445. 6 See Teall, PI. xxxvn; Cohen, PI. xxxvin, figs. 1, 2. MICROSPHERULITIC STRUCTURE IN RHYOLITES. 155 between crossed nicols (figs. 33, 34 A). These minute spherul- ites seem to be much less readily destroyed than the larger FIG, 33. MICROSPHERULITIC RHYOLITE, GREAT YARLSIDE, WESTMORLAND ; x 20. Crossed nicols. Each little spherule shews a black cross [1813]. ones. The axiolites of Zirkel 1 seem to be of the nature of elongated spherulites, the fibres radiating not from a point but from an axis (fig. 34 A), or they may be conceived as representing the coalescence of a row of minute spherulites (c/. fig. 32). Any evident micrographic structure is not common in the ground -mass of rhyolites, though bands or streaks having this character are sometimes found. A curious feature, first de- scribed by Iddings in some obsidians from the Yellowstone Park 2 and rhyolites from the Eureka district of Nevada 3 , is the occurrence of porphyritic ' granophyre groups ' or micropegmatite phenocrysts in a glassy, cryptocrystalline, or microcrystalline ground-mass (see fig. 34 h). In these the 1 Micro. Petr. Fortieth Parallel, PI. vi, fig. 2. But compare Cole, M. M. (1891) ix, 271-274. 2 1th Ann. Rep. U. S. Geol. Sur. (1888) 274-276, PI. xv. 3 Monog. xx, U. S. Geol. Sur. (1893) 375, PI. v, fig. 2. 156 MICROPGECILITIC STRUCTURE IN RHYOLITES. quartz is subordinate to the felspar in quantity, and the micrographic groups often shew the crystal-boundaries of the ax. FIG. 34. SPECIAL, STRUCTURES IN RHYOLITES, x 20. Crossed nicols. A. Falls of Gibbon River, Yellowstone Park : different bands, following the flow-lines, shew microprecilitic (mpcj, axiolitic (ax), and microspherulitic (up) structures [1430]. B. Good- wick, near Fishguard, Pembrokeshire ; shewing micropegmatite pheno- crysts in a finely microcrystalline ground-mass [2289]. latter mineral. As a rule, however, there are several felspar crystals grouped together, the whole permeated by wedges of quartz, and the outline is complex or rather irregular. A structure met with in the ground of some rhyolites, and in certain bands of laminated rhyolites, differs essentially from the micrographic, in that it indicates the successive, instead of simultaneous, crystallization of the two constituent minerals. Minute felspar crystals with no orderly arrangement are en- closed in little ovoid or irregular areas of quartz, the whole of the quartz in such a little area being in crystalline continuity. This structure reproduces on a minute scale the ophitic and pcecilitic structures presented by different minerals in other rocks, and Prof. G. H. Williams adopted for it the term micro- pcecilitic 1 (fig. 34^1). i Journ. of Geol. (1893), i, 176-179. EXAMPLES OF OBSIDIANS. 157 A holocrystalline texture on other than a minute scale is rarely, if ever, met with in true rhyolites. The ' nevadite ' of Richthofen is exceptional in that the ground-mass is quite subordinate in quantity to the crowded phenocrysts, but this ground-mass is commonly glassy. In part, at least, these rocks belong to the dacites rather than the rhyolites. Leading types. The glassy type (obsidian) is exempli- fied by many of the rhyolites of Iceland and of Lipari 1 . (fig. 32) ; and in the latter locality pumice is extensively developed (Monte Chirica). The Hungarian rhyolites are not usually obsidians, but some good examples occur (Telkibanya) 2 with a rich variety of crystallites (fig. 31). Other well-known obsidians come from Ascension Is., Mexico, and the Yellow- stone Park. The rock of Obsidian Cliff' 3 in the last-named district frequently contains spherulites of some size, isolated or in bands, and remarkable chambered lithophyses, in which occur nests of tridymite and little crystals of the iron-olivine (fayalite). Very similar phenomena have been described from Lipari (Rocche Rosse) 4 , and some of the Hungarian lavas also contain small lithophyses, often of hemispherical form, cut off by the fluxion-banding of the lava. It was there that these curious structures were first observed by von Richthofen (Telkibanya, Goncz, etc.}. The more widely distributed types of rhyolites may be studied in rich variety from the Tertiary volcanic districts of Schemnitz in Hungary, of the Lipari group, of the Western States of America, etc. They differ in the nature of their phenocrysts and in the structure of their ground-mass. Many of them have a strongly marked banded structure, successive narrow bands, a fraction of an inch wide, being of different textures or structures (glassy, microspherulitic, axiolitic, microcrystalline, micropcecilitic). The most usual ferro-mag- nesian mineral is biotite, but it is never plentiful. When spherulitic structures are present they may be on a 1 Cohen, PI. xxxvm, fig. 3. 2 Cohen, PI. xxvn, fig. 2. 3 Iddings, 1th Ann. Rep. U. S. Geol. Sur. 249-295. 4 Cole and Butler, Q. J. G. S. (1892) xlviii, 438-445 ; Johnston-Lavis, G. M. 1892, 488-491. 158 VARIOUS TERTIARY RHYOLITES. more or less minute scale. Some flows in the Schemnitz district are built up almost wholly of very diminutive spherul- ites 1 , each giving a perfect black cross (Telkibanya, Sarospatak, Eisenbach, etc.). This microspherulitic type is also repre- sented among the rhyolites of the Yellowstone Park. In the typical 'perlites' of the Schemnitz district the individual spherulites are larger, with well-marked radial fibrous structure and globular form, sharply bounded, often by perlitic fissures (Hlinik, etc.). These contrast with a type in which the spherulites have an irregular outline, interlocking with one another or sending out processes into a glassy matrix. Zirkel 2 described from Nevada rhyolites (including ob- sidians) shewing a remarkable variety in the character of their ground-mass. Others, from the Eureka district, have been described by Iddings 3 . These carry a little biotite. In examples described by the same author 4 from New Mexico (Tewan Mts.) the ferro-magnesian mineral is augite. In these rocks plagioclase felspar is wanting : some contain spherulites and lithophyses. Rhyolites from Ouster County, Colorado, have no coloured constituent except a little red garnet 5 . The ground-mass is usually microcrystalline to cryptocrystalline, but sometimes spherulitic. The best British examples of fresh Tertiary rhyolites are found in Antrim. The Tardree rock is conspicuously por- phyritic, with a soda-bearing sanidine, corroded quartz-grains, and a little green augite in a ground-mass varying from microcrystalline to cryptocrystalline. Biotite and magnetite also occur sparingly, and von Lasaulx 6 detected scales of tridymite, chiefly in little nests occupying druses in the rock. Prof. Cole 7 regards the Tardree rock as marking the neck of a volcano, from which issued the obsidian and compact and banded rhyolites of the neighbouring locality Sandy Braes. He has also described 8 lithoidal varieties from Templepatrick, Fouque and L6vy, PI. xvn, fig. 1. Micro. Petrogr. Fortieth Parallel (1876) 163-205, PI. vi-ix. Monog. xx, U. S. Geol. Sur. (1893) 374-380, PL vin. Bull. No. 66 U. S. Geol. Sur. (1890) 10, 11. Cross, Proc. Colo. Set. Soc. 1887, 229-233. Journ. Geol. Soc. Irel. (1878) xiii, 25-31. G. M. 1895, 303-306. Sci. Tram. Roy. Dubl. Soc. (1896) vi, 77-118, PI. iv. OLD BRITISH RHYOLITES. 159 Clough water, etc. In the obsidian of Sandy Braes Mr Watts 1 has remarked perlitic cracks traversing both the brown glassy ground-mass and the quartz phenocrysts. Other rhyo- lites occur between Dromore and Moira, Co. Down. The most interesting British rhyolites, however, are those belonging to the Palaeozoic and older volcanic groups, and these have doubtless had their pristine characters modified in many instances by secondary physical and chemical changes. The Ordovician rhyolites of Caernarvonshire 2 are charac- terized by the general paucity of any phenocrysts, and especially of those of quartz 3 . Among the scattered felspar-crystals, a member of the albite-oligoclase series predominates over ortho- clase. Almost the only ferro-magnesian constituent is a little colourless augite, and even this is commonly wanting, though a pale green decomposition-product may perhaps represent it. In all these features the rocks closely resemble the Tertiary and Recent rhyolites of Iceland 4 , and probably the older rocks have once been largely glassy, as the younger are now. The usual texture of these old lavas is cryptocrystalline to micro- crystalline, sometimes shewing fluxion and banding, and oc- casionally good perlitic cracks. The vesicular structure is not very frequent. In some types the ground is partly micro- pcecilitic, minute felspar prisms being enclosed in quartz (Penmaenbach, etc.). Any approach to a microspherulitic structure of a perfect type is uncommon, but large isolated spherulites are abundant in many localities, and shew the various secondary alterations, concentric shell-structure, silici- ncation, etc., to which they are always prone 5 . The siliceous and other nodules which thus arise may reach several inches in diameter. Some of them have been supposed to represent lithophyses 6 . 1 Q. J. G. S. (1894) 1, 367-375, PI. xvm. 2 Bala Vole. Ser. of Caern. 18-23. 3 This is true more especially of central and eastern Caernarvonshire. The rhyolites of the Lleyn peninsula, many of which are intrusive, are richer in phenocrysts, including quartz. * Backstrom, M. M. (1894) x, 343, 344 (Abstr.). 5 Bala Vole. Ser. of Caern. 35-39. 6 Cole, Q. J. G. S. (1892) xlviii, 443-445, and references. 160 OLD BRITISH RHYOLITES. The Ordovician rhyolites of Westmorland 1 closely resemble the preceding, but in certain flows shew a very perfect micro- spherulitic structure. This is well seen in Long Sleddale' 2 and near Great Yarlside (tig. 33). The large altered spherul- ites or nodules also occur. Good examples of these, as well as of devitrified obsidian with perlitic structure, are found also at Bouley Bay, in Jersey 3 . As a secondary alteration some of these Westmorland lavas shew silicifi cation, in which much of the ground-mass and sometimes the porphyritic felspars are replaced by microcrystalline quartz. The same thing is seen in Caernarvonshire 4 , and even more markedly in some ancient rhyolites at Trefgarn and Roche Castle, in Pembrokeshire. Mr Allport was the first to give a clear account of some of the old altered volcanic glasses and to compare them with fresh Tertiary examples. He described what seems to be a devitrified and altered spherulitic rhyolite of pre-Cambrian age from Overley Hill or the Lea Rock near Wellington, Shropshire 5 . A few phenocrysts occur, but the bulk of the rock has been a glass enclosing numerous bands of spherulites. The glass is now devitrified, but perlitic cracks, marked by secondary products, are still evident. The spherulites too are for the most part much altered and stained red by iron- oxide. Rhyolites presenting many features of interest occur in the Ordovician of Fishguard, in Pembrokeshire 6 . Mr Reed, who has made a study of these rocks, has noticed in some of them beautiful examples of porphyritic micropegmatite, recalling the lavas described in America by Prof. Iddings (see fig. 34 B). The old rhyolites of the Malvern Hills shew some curious features. Specimens from the New Reservoir, Malvern, are es- sentially cryptocrystalline rocks (perhaps devitrified), sometimes 1 Q. J. G. S. (1891) xlvii, 303. 2 Rutley, Q. J. G. S. (1884) xl, PI. xvm, fig. 6, and Teall, PL xxxvin [1921]. 3 Davies, M. M. iii, 118, 119. * Miss Raisin, Q. J. G. S. (1889) xlv, 253, 254. 5 Q. J. G. S. (1877) xxxiii, 449-460; Teall, PI. xxxiv, figs. 1, 2. 6 Reed, Q. J. G. S. (1895) li, PI. vi, figs. 3-5. OLD AMERICAN RHYOLITES. 161 enclosing scattered phenocrysts of oligoclase. Narrow veins are occupied in some cases by infiltrations of calcite, in others by a clear mosaic of quartz, orthoclase, and plagioclase of secondary formation. Mr Rutley 1 has described examples from the Herefordshire Beacon, in which old perlitic cracks are marked out by secondary epidote. The chief alteration- products are epidote and quartz, and the author suggests that some of the so-called epidosites (quartz-epidote-rocks) may have originated in this way. Ancient acid lavas of Palaeozoic and pre-Palfeozoic ages occupy large tracts in the east of Canada and the United States. In spite of alteration, they have preserved many relics of original characteristic structures 2 . This is well illustrated by examples from South Mountain 3 (Penna.). Ancient de vitrified obsidians and rhyolites, some spherulitic, have been described from Vinal Haven 4 and North Haven 5 in Maine, from near St John, New Brunswick 6 , from the Michigamme district in Michigan 7 , etc. Although we have not made a distinct subfamily of soda- rhyolites, it may be remarked that there are among the acid lavas some characterized by anorthoclase felspar and even soda-bearing pyroxene or amphibole. Some of the ceratophyres and quartz-ceratophyres of certain authors belong here. One from Marblehead Neck, Mass., has phenocrysts of anortho- clase 8 . Another from Baraboo Bluffs, Wis., is also rich in soda 9 . An example from Berkeley, Cal., ranges from a porphyritic variety with microcrystalline ground-mass to a pure glass, but is usually microspherulitic 10 . 1 Q. J. G. S. (1888) xliv, 740-744, PI. xvn. 2 G. H. Williams, Journ. of Geol. (1894) ii, 1-31. 3 G. H. Williams, A. J. S. (1892) xliv, 482-496 ; F. Bascom, Journ. of Geol. (1893) i, 813-832. 4 G. H. Williams, Journ. of Geol. (1894) ii, 23 ; G. 0. Smith, Joh. Hopk. Univ. Circ. No. 121 (1895). 5 Bayley, Bull. Geol. Soc. Amer. (1894) vi, 474. 6 Matthew, Trans. N. Y. Acad. Sci. (1895) xiv, 197-200, PI. xn, xm. 7 Clements, Journ. of Geol. (1895) iii, 811-817. 8 Sears, Bull. Mus. Comp. Zool. Harv. (1890) xvi, 162-172. 9 Weidman, Bull. Univ. Wis. (1895) Sci. Ser. i. 10 Palache, Bull. Dep. Geol. Univ. Cal. (1893) i, 61-72. H. P. 11 162 PANTELLARITES. An segirine-bearing rhyolite is described by Bertolio from Sardinia (Oommende type). More remarkable are the rhyolites of Pantellaria, an island lying south-west of Sicily, a peculiar type rich in soda and iron (pantellarite). The pheno- crysts are of anorthoclase and soda-sanidine, a green pleochroic augite, and the deep-brown, intensely pleochroic cossyrite. The ground-mass varies from almost holocrystalline to almost wholly vitreous, a prevalent variety being a glass cro%vded with microlites of the above-mentioned minerals. CHAPTER XII. TRACHYTES AND PHONOLITES. THE trachytes are lavas which, with a lower percentage of silica than the rhyolites, have as much or more of the alkalies. Consequently the typical trachytes consist essentially of alkali-felspars with a relatively small amount of coloured minerals and without free quartz. The name trachyte (given by Haiiy to denote the rough aspect of the rocks in hand specimens) is used in the older literature to cover all the more acid half of the volcanic rocks. From it have been separated off, on the one .hand, the rhyolites of modern no- menclature and, on the other, some hornblende- and mica- andesites, etc. With the trachytes we shall treat some lavas of more peculiar constitution, in which a greater richness in alkalies has given rise to the formation of felspathoids as well as alkali-felspars : these are the phonolites and leucitophyres. The name phonolite (a translation of 'clinkstone,' from the supposed sonorous quality of the rock when struck) seems to have been in general use before the presence of microscopic nepheline in the rock was demonstrated, giving a character of precision to the definition. The original leucitophyres (of Coquand) were apparently any rocks with conspicuous crystals of leucite, but the name is now generally restricted to the types containing an alkali-felspar (sanidine) as an essential constituent. Trachytes and phonolites exhibiting clearly their character- istic features are hitherto known chiefly among Tertiary and 112 164 MINERALS OF TRACHYTES AND PHONOLITES. Recent volcanic products. It should be noticed, however, that some of these features are such as would readily be effaced in the older lavas. In any case, the Continental practice of restricting these families to Tertiary or later lavas rests on no philosophic ground, and indeed perfectly fresh trachytes and phonolites are known, e.g., in the Carboni- ferous of Scotland. The leucitophyres are a type of extremely restricted distribution, and the unstable nature of the cha- racteristic mineral must make such rocks difficult to detect among the older lavas. Constituent minerals. Felspars rich in potash or soda are by far the most abundant minerals in the rocks here considered. They occur both as phenocrysts and as the chief element in the ground-mass. The most prominent is usually orthoclase of the sanidine variety, often shewing a rough ortho- pinacoidal cleavage 1 . In phenocrysts it has either a tabular or a columnar habit, and both may occur in the same rock. Carlsbad twinning is frequent 2 , and in the larger crystals may shew the broken divisional line due to interpenetration. Some degree of zonary banding is sometimes found. The plagioclase felspar which occurs in many trachytes is usually oliyoclase, but in more basic rocks we may find varieties richer in lime instead. The phenocrysts often shew carlsbad- as well as albite-twinning; zonary banding is not uncommon; and parallel intergrowth with sanidine may be noted (fig. 35 B). In the true trachytes the most common ferro-magnesian element is perhaps brown biotite, in hexagonal flakes almost always affected by corrosion by the enclosing magma ('re- sorption'). This is shewn by a certain degree of rounding and the formation of a dark or opaque border, or even the total destruction of the flake, the resulting products being especially magnetite and sometimes greenish augite in minute granules. The frequent preservation of the original crystal- forms proves that the process is not one of fusion and re- crystallization, but rather pseudomorphism depending on changed physical conditions and chemical reactions with the fluid magma 3 . Brown hornblende is a less frequent constituent, 1 Cohen, PI. xx, fig. 3. 2 Ibid. PI. xxvm, fig. 1. 3 Washington, Journ. of Geol. (1896) iv, 257-282. MINERALS OF TRACHYTES AND PHONOLITES. 165 in idiomorphic crystals with similar resorption-phenomena. The augite, which is scarcely less common than biotite as a constituent of trachytes, never shews this feature. It is usually pale green in thin slices. In the phonolites and leucitophyres the crystal often shews a deeper tint at the margin, and is almost always sensibly pleochroic (segirine- augite), a character less common in the trachytes. Another pyroxene, cegirine, is characteristic of many phonolites and leucitophyres, but only occasionally present in the trachytes. It is green and pleochroic, with a much lower extinction-angle than the augites (5 or less in longitudinal sections). It some- times grows round a kernel of augite with parallel orientation. The rhombic pyroxene of certain trachytes is always of a deeply coloured and vividly pleochroic variety (liypersthene or ambly- stegite), giving red-brown, yellow-brown, and green colours for the several principal directions of absorption. The nepheline of the phonolites and leucitophyres occurs in minute crystals in the^ ground-mass, having the form of a short hexagonal prism with basal planes, and giving squarish or hexagonal sections. Owing to the small size of the crystals and the optical properties of the mineral, it is liable to be overlooked. Its decomposition gives rise to various soda- zeolites, which occur in nests and veins in many phonolites. The leucite of the leucitophyres is always idiomorphic, giving characteristic octagonal and rounded sections. Twin-lamella- tion is very frequent in the phenocrysts 1 , but the smaller crystals which may occur often behave almost as if isotropic. The leucite may enclose needles of augite and crystals of the earlier-formed minerals, but not of felspar. Minerals of the sodalite-group are found in certain trachytes and constantly in the phonolites and leucitophyres. They are almost always in idiomorphic dodecahedra. The sodalite is clear when fresh, but often turbid from alteration : zonary structure is frequent. The blue hauyne is less often met with, but nosean may be very plentiful, usually forming crystals of some size, and always shewing more or less plainly its characteristic structure and border 2 . The sodalite-minerals give rise by alteration to natrolite and other zeolites. 1 Cohen, PI. xxxi, fig. 4. 2 Teall, PI. XLI, fig. 1 ; XLVII, fig. 4. 166 GROUND-MASS OF TRACHYTES, ETC. Iron-ores (magnetite) occur but sparingly in these rocks. Yellowish sphene in good crystals is highly characteristic ; and apatite is common in colourless needles or sometimes in rather stouter prisms with violet dichroism. The trachytes often contain a little zircon in minute prisms. Among less common minerals may be mentioned the tri- dymite of certain trachytes, in aggregates of minute flakes : olivine, as a rare constituent except in certain basic trachytes ; and melanite garnet, which is found in some of the leucito- phyres and in certain phonolites as brown isotropic crystals belonging to an early stage of consolidation, sometimes shew- ing marked zonary banding 1 . As secondary products in trachytic, and also in andesitic, rocks, opal and other forms of soluble silica are not uncommon. Normally isotropic, these substances sometimes shew double refraction as a consequence of strain, usually about centres, so as to imitate a spherulitic structure. Opal sometimes en- closes little flakes or aggregates of* tridymite, or is coloured red by included scales of haematite. It occurs in the form of veins and irregular knots or patches. Ground-mass. In contrast with the rhyolites, the rocks under consideration have few glassy representatives, and the ground-mass is frequently holocrystalline or at least with no sensible amount of glassy residue. This is especially true of the typical trachytes, which, with a chemical composition not very different from that of a mixture of felspars, have a strong tendency to crystallize bodily. Fluxional phenomena are not conspicuous, and the characteristic banding of the rhyolites is here wanting. Vesicular structure is rare, and perlitic cracks are not formed ; but, in consequence of the crystalline nature of the ground, with a tendency to idiomorphism in its elements, a miarolitic or drusy structure may be met with. An} T structure comparable with the spherulitic is uncommon, though a rough radial grouping of felspar prisms is sometimes observable. Excluding the nepheline of the phonolites, non-felspathic constituents play in most cases a small part in the ground- 1 Rosenbusch-Iddings, PL v, fig. 4. EXAMPLES OF TRACHYTES. 167 mass of the rocks here considered. The ground consists, in the trachytes proper, essentially of minute felspars, which may, however, vary somewhat in habit. Most commonly they are ' lath-shaped ' microlites, with some degree of parallel dis- position in consequence of flow, and this type of ground is so characteristic of these rocks that it is often styled the trachytic 1 . On the other hand the minute felspars may have a shorter and stouter shape, recalling some of the rocks grouped above under the porphyries, and this structure is accordingly designated by Rosenbusch the orthophyric. Phonolites poor in nepheline do not differ essentially as regards structures from the trachytes, but when the character- istic mineral is plentiful, forming very numerous minute crystals in the ground-mass, the general aspect of the latter is somewhat altered 2 . The leucitophyres shew in their very variable structures further departures from the trachyte type ; but all the rocks included in the present chapter resemble one another in being normally holocrystalline. Leading types. Among the best known foreign trachytes are those of the Siebengebirge (Drachenfels type). Here a ground-mass of lath-shaped felspar microlites, with typical trachytic structure, encloses crystals of sanidine and oligoclase. The former are frequently of large size, and may shew carlsbad twinning. Biotite and magnetite occur sparingly. The trachyte of Kelberg in the High Eifel is very similar, but has a small amount of glassy base. In America a trachyte of the Drachen- fels type has been described by Cross 3 from the neighbourhood of Rosita in Colorado. The rock of Perlenhardt, in the Sieben- gebirge, exemplifies the orthophyric type of ground-mass of Rosenbusch. A little green augite accompanies the biotite, sphene is common, and sodalite occurs in crystals or crystal- line patches. Trachytes from Solfatara and Mte Olibano near Naples shew similar characters. Numerous augite-trachytes occur in the neighbourhood of Naples and the Phlegrsean Fields. Some very fresh augite-bearing trachytes occur as lava- flows and volcanic necks of Lower Carboniferous age in the 1 Berwerth, Lief, i ; Eosenbusch, Mass. Gest., PI. v, fig. 2. 2 Ibid. fig. 3. 3 Proc. Colo. Sci. Soc. 1887, 234. 168 OLD SCOTTISH TRACHYTES. Garlton Hills, Haddingtonshire 1 . These rocks consist of alkali- felspars with more or less of a bright to pale green, pleochroic augite, doubtless a soda-bearing variety. Specimens from Peppercraig (fig. 35) shew phenocrysts of sanidine, sometimes with intergrowths of oligoclase, in a holocrystalline ground- mass. The latter is chiefly of sanidine prisms, with a minor proportion of striated felspar. Augite builds imperfect crystals and grains and numerous smaller granules ; magnetite occurs sparingly in the same manner ; and occasional needles of apatite are seen. The rock which forms the volcanic necks FlG. 35. AUGITE-TKACHYTE, PEPPEKCKAIG, HADDINGTON J X 20. A in natural light, B between crossed nicols. Large phenocrysts of felspar are enclosed in a ground composed entirely of little felspar prisms and granules of augite [1980]. has been alluded to above (p. 116). A curious riebeckite-bearing trachyte, of Upper Old Red Sandstone age, occurs at Easter Eildon Hill, near Melrose 2 . Purely glassy varieties (trachyte-obsidian) are uncommon 1 Hatch, Trans. Roy. Soc. Edin. (1892) xxxvii, 115-126; see also Geikie, ibid. (1879) xxix, PI. xn, figs. 1, 2. The Carboniferous trachytes described by McMahon from Dartmoor seem to be much altered and their characters obscured ; Q. J. G. S. (1894) 1, 345, 346. 2 Ban-on, G. M. 1896, 376. ANDESITIC AND BASALTIC TRACHYTES. 169 in this family. In the localities where they are found, they are associated with trachytes wholly or mainly crystalline, or even narrow alternating bands occur of pure glass and of trachyte largely microcrystalline. Good examples of this occur in the Peak of Tenerife. It may be noted that a glassy variety of phonolite also is found in the Canaries, usually as a slaggy crust on the surface of a lava-flow. It is a brown or yellow glass with little development of crystallites. In those trachytes which in some respects approach the andesites, the coloured constituents, especially pyroxene, be- come relatively abundant, and plagioclase begins to pre- dominate over orthoclase among the phenocrysts. A type from Mte Amiata in Tuscany and Mt Dore in Auvergne con- tains a vividly pleochroic rhombic pyroxene (amblystegite) with subordinate biotite. Garnet and tridymite are accessories. The ground-mass of these rocks is of very variable character, even in the same flow, and is sometimes largely glassy. Similar trachytes occur at Mocsar in Hungary. Washington 1 has proposed the name vulsinite for a group of rocks intermediate between trachyte and andesite. They contain a considerable amount of a basic plagioclase in addition to the alkali-felspar, and the ferro-magnesian constituent is typically augite. In examples from Bolsena in Italy the phenocrysts are of alkali-felspar, anorthite, augite, and biotite, and the ground-mass is of soda-orthoclase, augite, etc., with trachytic structure. One from the Viterbo district has labradorite in place of anorthite 2 . The Arso-type of trachyte, the Ischia lava of A.D. 1302, approximates in some features to the basalts. The pheno- crysts include, in addition to sanidine and a plagioclase felspar, abundant augite and olivine. The ground-mass is of felspar microlites with interstitial glass, and is sometimes vesicular. Olivine-bearing trachytes occur also in the Azores. Other trachytes shew an approach to the characters of phonolites in the abundance of sodalite, the occurrence of regirine, etc. The trachytes of the Laacher See in the Eifel have crystals of sodalite and haiiyne, besides sanidine and 1 Journ. of Geol. (1896) iv, 547-554. 2 Ibidi 833< 170 SODA-TRACHYTES: TRACHYTIC PHONOLITES. oligoclase. Biotite, brown hornblende, segirine, sphene, mag- netite, etc., also occur, and the ground-mass is of the trachytic type. At the Laach volcano are found also ejected blocks of a rock named sanidine-trachyte or sanidinite. This consists essentially of sanidine with subordinate oligoclase, sodalite, occasional biotite, etc. Stellate groupings of crystals occur in both felspars, but on the whole the structure is that of a plutonic (syenitic) rather than a volcanic rock. While the dominant mineral of the trachytic lavas is commonly a potash-felspar, there are some types very rich in soda ; albite, anorthoclase, or some allied felspar occurring almost to the exclusion of sanidine or orthoclase. The 'quartz- less pantellarites ' of Pantellaria must be placed here, and the older equivalents of such types are to be sought among some of the rocks which have been styled quartzless cerato- phyres. A very interesting soda-felspar-rock has been described from Dinas Head on the north coast of Cornwall 1 . This is possibly to be regarded as an ancient lava 2 , and it consists almost wholly of albite. Besides a compact variety, there are others which are spherulitic and nodular. The centre of a spherule is cryptocrystalline while its outer portion consists of radiating blades of albite. Such rocks may be termed old soda-trachytes, corresponding with the soda-rhyolites which are also known in this country. Coming now to the phonolites, we notice first those in which nepheline is only sparingly present, and which thus stand in close relation with the trachytes. Such rocks, the ' trachytoid ' phonolites of Bosenbusch, are not the most characteristic type ; and the ' nephelinitoid ' group, in which the special mineral of the phonolites is more abundantly present, is commoner. Some of the Saxon phonolites are of the trachytoid type (Olbersdorf, near Zittau). A good example is found at Traprain Law in association with the trachytes of the Garlton Hills, Haddingtonshire 3 , and is of interest as being of Carboniferous age. It consists essentially of a mass of little sanidine prisms, with a fluxional arrangement, 1 Howard Fox, G. M. 1895, 13-20. 2 McMahon, hotfever, regards it as a metamorphosed sediment ; ibid. 257, 258. 3 Hatch, G. M. 1892, 149 ; Trans. Roy. Soc. Edin. (1892) xxxvii, 124. PHONOLITES. 171 in which lie ragged crystals of a bright green soda-augite. Small colourless patches are found on very close examination to consist of little crystals of nepheline with zeolitic decom- position-products. A lava from Middle Eildon Hill, near Melrose, is also a phonolite of trachytoid type, and is remark- able for having riebeckite instead of segirine. The mineral occurs, as usual, in irregularly shaped patches, moulded on the felspar. This rock is of Upper Old Red Sandstone age 1 . Of the commoner type of phonolite good examples occur in Bohemia (Briix, Teplitz, Marienberg, etc.), sanidine, nephel- ine, and segirine being the essential minerals. Some varieties have conspicuous phenocrysts of sanidine. At the Roche Sanadoire 2 in Auvergne the porphyritic sanidines have often a core of plagioclase in parallel intergrowth, and little lath- shaped crystals of plagioclase occur also in the ground-mass. Another British phonolite that of the Wolf Rock 3 off the coast of Cornwall is also a good example. It belongs FIG. 36. PHONOLITE, WOLF BOCK, COKNWALL ; A x 20, B x 100. The figure shews phenocrysts of sanidine (s) and nosean (no) in a ground-mass of sanidine, nepheline (ne) and aegirine (ce) [1771]. 1 Barren, G. M. 1896, 373-375. 2 Fouque and Levy, PI. XLVII, fig. 1 ; cf. fig. 2 and PI. XLVI. 3 Allport, G. M. 1871, 247-250 ; 1874, 462, 463 ; Teall. PI. XLI, fig. 1. 172 AMERICAN PHONOLITES. to the nosean-phonolites of some authors, that mineral being found plentifully in it, in addition to nepheline. The nosean occurs chiefly as phenocrysts with a dark interior and clear border 1 . Sanidine is also found as phenocrysts. The general mass of the rock consists of lath-shaped sanidine crystals, more or less idiomorphic crystals of nepheline, and little dirty green microlites of segirine. Iron-ores are scarcely represented, and there is little or no residual glass (fig. 36). Phonolites are only sparingly represented among the varied volcanic rocks of the United States. One from El Paso County, Colorado*, is essentially a finely granular aggregate of sanidine, nepheliue, and hornblende, with phenocrysts of the two former minerals. A similar rock, with the addition of a little nosean, is known from Black Butte in the Black Hills of Dakota 3 . The felspar phenocrysts are of soda-orthoclase or anorthoclase 4 . Phonolites occur as volcanic dykes and larger masses in the Cripple Creek mining-district, Colorado 5 . They are rich in alkali-felspars, and contain phenocrysts of soda- sanidine or anorthoclase. Nepheline occurs with variable habit, sometimes building small phenocrysts, while porphyritic nosean and minute crystals of sodalite are also found. JEgirine and fegirine-augite are the coloured minerals, or in certain cases a blue amphibole, and among the accessory minerals is analcime, believed to be of primary origin. Osann's rocks from western Texas (Apache type) are rich in hornblende, including a blue variety, and the felspars shew microperthitic intergrowths. The leucitophyres are a very small group of rocks known only from three or four districts 6 and best developed in the late Tertiary lavas of the Eifel. The leucite is often of two genera- tions, the larger crystals being frequently of irregular shape. It is always accompanied by nosean and sanidine (fig. 37). The ferro-magnesian mineral is a green pleochroic augite with 1 Teall, PI. XL vn, fig. 4 (misplaced 5 in key-plate). 2 Cross, Proc. Colo. Sci. Soc. 1887, 167, 168. 3 Caswell, Geol. of Black Hills, U. S. G. and G. Sur. Eocky Mts. (1880) 503. 4 Pirsson, A. J. S. (1894) xlvii, 341-346. 5 Cross, 16th Ann. Rep. U. S. Geol. Sur. (1895) Part II, 25-36. 6 On leucitpphyres from Bolsena, see Washington, Journ. of Geol. (1896) ii, 559-561; from the Viterbo district, ibid. 841-845. LEUCITOPHYRES. 173 zonary banding : the other constituents are sanidine, sphene, occasionally biotite, and often a little melanite. The struct- ure of the rocks is very variable 1 . In some there is a well- defined ground-mass of minute nepheline, sanidine, augite, and leucite, enclosing phenocrysts of leucite and nosean (Olbruck, -ce FIG. 37. LEUCITOPHYBE, RIEDEN, EIFEL ; x 20. The larger elements shewn are leucite (I), nosean (no), and segirine () [160]. etc.). In other varieties there is but little sanidine (Schoren- berg), while others again have sanidine in large shapeless plates enclosing the other constituents instead of a ground- mass (Perlerkopf). Pumiceous modifications of leucitophyre occur, especially as fragments in the associated tuffs. 1 A. Martin, M. M. (1891) ix, 251 (Abstr.). For figures of leucito- phyres see Fouque and Levy, PL XLVIII, fig. 1, and LI, fig. 1 ; Teall, PI. XLI, fig. 2, and XLVII, fig. 4 ; Bosenbusch-Iddings, PI. v, fig. 2 ; xv, fig. 1 ; xvi, fig. 6 ; xvn, fig. 1. CHAPTER XIII. AXDESITES. UNDER this family we include all the lavas of 'inter- mediate ' composition not embraced in the preceding chapter. The name andesite, first used by von Buch and derived from the prevalence of such rocks in the Andes, is roughly equi- valent to Abich's ' trachydolerite,' implying the intermediate position of these lavas between the acid ones (trachytes of older writers) and the basic (dolerites). The characteristic minerals are a soda-lime-felspar and one or more ferro- magnesian minerals. The alkali-felspars and quartz of the acid rocks are typically absent, as are also the lime-felspar and olivine of the basic rocks. The andesites are distinguished, according to the dominant ferro-magnesian constituent, as hombletide-, mica-, augite-, and hyperstkene-andesites. Further there is usually recognized a quartz-bearing and more acid division, known as dacltes or' quartz-andesites. Having re- gard to true lavas, these quartz-bearing andesites seem to be of somewhat limited distribution : many of the rocks described as ' dacites ' are of intrusive types, and belong to the less acid quartz-porphyries. Those petrologists who restrict the name andesite to rocks of late geological age, apply to their pre-Tertiary equivalents the name ' porphyrite 1 .' Under the same title they include various rocks of intrusive types, and it is to these latter that 1 Many of the rocks designated ' melaphyre ' are pyroxene-andesites, others being basalts. FELSPARS OF ANDESITES. 175 we have already confined the name. Again, certain English petrologists have used the name porphyrite for andesites which have undergone some degree of change by weathering, etc., a distinction which seems scarcely important enough to be recognized in classification or nomenclature. As regards the general affinities of the family, the dacites have features in common with the rhyolites, the hornblende- and mica-andesites with the trachytes, and the pyroxene- andesites with the basalts, marking thus the intermediate position held among the volcanic rocks by the lavas here considered. As regards the appropriateness of the name, it is remarkable that the lavas of the great volcanic belt of the Andes belong, in so far as they are known, almost exclusively to this family 1 . Phenocrysts. Soda-lime-felspars are the most abundant elements porphyritically developed in these rocks. They in- clude members varying from oligoclase to anorthite, but andesine and labradorite are the most common. As a rule the more acid plagioclase belongs to the hornblende- and mica-andesites and dacites, the more basic to the pyroxene- andesites 2 . The crystals, however, are often strongly zoned 3 , shewing a change from a more basic variety in the centre 'to a more acid at the margin. They are idiomorphic and of tabular habit. With albite-lamellation is frequently as- sociated twinning on the pericline or on the carlsbad law. The commonest inclusions are glass-cavities 4 , either as 'nega- tive crystals,' or rounded : sometimes large irregular cavities occupy much of the bulk of a crystal 5 . The decomposition- products of the felspars are calcite, finely divided kaolin or mica, epidote, quartz, etc. When an alkali-felspar occurs as an accessory, it has the same characters as in the rhyolites and trachytes. 1 Cf. Iddings, Journ. of Gcol. (1893) i, 164-175. 2 French petrologists recognize ' andesites ' and ' labradorites ' as dis- tinct groups, characterized by andesine and labrador felspar respectively, but this is with reference to the ground-mass. 3 Iddings, Monog. xx U. S. Geol. Sur. (1893) PI. v, figs. 1, 3, 4 ; vi, fig. 2. 4 See Zirkel, Micro. Petr. Fortieth Parall., PI. v, fig. 3 ; xi, fig. 2. 5 Cohen, PL in, fig. 1. 176 MINERALS OF ANDESITES. The hornblende of andesites is in idiomorphic prisms, often twinned 1 . It is usually a brown pleochroic variety with quite low extinction-angle, but green hornblende also occurs. The mica is a brown, strongly pleochroic biotite with extinction sometimes sufficiently oblique to shew lamellar twinning par- allel to the base. Both hornblende and biotite shew the same resorption-phenomena 2 as in the trachytes. It is possible that some part of the finely divided magnetite and granular augite in the ground-mass of certain andesites comes from the breaking up of hornblende altered in this way 3 . By de- composition of the ordinary kind the hornblende and mica of andesites give rise to chlorite, magnetite, carbonates, etc. The augite is in well-shaped crystals, light green and usually without sensible pleochroism. Twin-lamellation is common 4 . Alteration may give rise to chlorite, epidote, calcite, etc. The rhombic pyroxene in the andesites is usually hyper- sthene 5 , or at least a distinctly coloured and more or less pleochroic variety. It builds idiomorphic crystals, in which the pinacoid faces are more developed than the prism ; so that the cross-section is a square with truncated corners, as contrasted with the regular octagon of augite. In longitudinal sections the straight extinction is of course characteristic. The rhombic pyroxene is often converted in the older rocks to bastite 6 . The quartz of the dacites is either in good hexagonal pyramids or more or less rounded and corroded, with inlets of the ground-mass. Original iron-ores are usually not abundant : magnetite is the only one commonly found. Needles of apatite occur, and in the more acid andesites little zircons 7 . Some of the more 1 Cohen, PI. xv, fig. 1. 2 Ibid. PI. ix, fig. 4; Fouque and Levy, PI. xxvm, xxix; Zirkel, Micro. Petrogr. Fortieth Parallel, PI. v, fig. 2. 3 Washington, Journ. of Geol. (1896) iv, 273-278. 4 Kosenbusch-Iddings, PI. xix, fig. 5. 5 Cross, Bull. No. 1 U. S. Geol. Sur. (1883); A. J. S. (1883) xxv, 139; Teall, G. M. 1883, 145-148. 6 Rosenbusch-Iddings, PI. xvin, fig. 1. 7 Iddings, Monog. xx U. S. Geol. Sur. (1893), PI. m, figs. 15-20. GROUND- MASS OF ANDESITES. 177 basic rocks have sparingly phenocrysts of olivine 1 . As oc- casional accessories may be noted tridymite' 2 (in druses), garnet, and cordierite 3 . Structure of ground-mass. In many andesites the only mineral which occurs distinctly in two generations is the felspar. The felspar of the ground-mass builds little ' lath- shaped ' crystals, " of ten simple, sometimes twinned, but usually without repetition. It is probably, as a rule, of a more acid variety than the porphyritic felspar, andesine or oligoclase occurring in different cases. Augite also may be present as a constituent of the ground-mass, forming very small crystals of pale-green tint. Some of the hornblende- and mica-andesites have a trach- ytic type of ground-mass, composed essentially of very small felspar-laths with little or no glassy base, as in the Drachenfels trachyte. It is not always easy to ascertain whether any glass is present or not. From this type, as from the others, there are, however, transitions to rocks with a ground-mass mainly glassy. Less common is a 'microfelsitic' or cryptocrystalline struct- ure. This is seen in some of the dacites. In some cases spherulitic structures are found. In most typical andesites, and especially in the pyroxene- bearing kinds, the ground-mass has the very distinctive 'felted' character termed by Rosenbusch hyalopilitic 4 - This consists of innumerable small felspar-laths, simple or once twinned, often with evident flow-structure, and a residuum of glassy matter. 'Vesicles are common, and their infilling by secondary products gives rise to amygdules 5 . So character- istic is this type, that it is often spoken of as the ' andesitic ' ground-mass. When the little felspars are closely packed together, to the exclusion of any glassy base, we have the 1 Cohen, PI. xxvi, tig. 4. 2 Koto, Q. J. G. S. (1884) xl, 441, 444. 3 Osann, M. M. viii, 284, 285 (Abstr.). * See chromolithograph of augite-andesite (' augite-porphyrite '), Berwerth, Lief, i ; also Kosenbusch, Mass. Gest., PI. iv, tig. 1. 5 Very many of the amygdaloidal lavas (Ger. Mandelstein) belong here. H. P. 12 178 DACITES. jnlotaadtic structure of Rosenbusch. On the other hand, by increase in the proportion of isotropic base, these andesites graduate into more or less perfectly glassy forms. Wholly glassy types (andesite-obsidian, including andesite-pumice) are known in small development only, except in so far as they form part of tuffs, etc. Some andesitic rocks shew various kinds of variolitic structures 1 comparable with those seen in basalts (see fig. 41 A, on p. 192). Leading types. Among dacites the best known are those of Tertiary age in Transylvania 2 and Hungary and in some parts of the Andes. They include holocrystalline examples (Kis Sebes, Rodna) and others with cryptocrystal- line and microspherulitic ground-mass (Schemnitz district, etc.),, as well as those having the hyalopilitic structure so common among the andesites. Hornblende-dacite with micro- spherulitic structure occurs among the recent lavas of Santorin in the Grecian Archipelago 3 . Zeolites and isotropic opal are found as secondary products, or in other cases chalcedony 4 . Fresh andesite-glasses also occur at Santorin 5 , reproducing the perlitic fissures and other features of the acid obsidians. These seem to be in the main hornblende-bearing, but contain augite associated with that mineral. Vesicular and pumiceous modifications are found. The 'pitchstone' of the Sgurr of Eigg 6 and the neighbouring island Hysgeir 7 has the composition of a dacite. The pro- minent phenocrysts are of sanidine or perhaps anorthoelase. and are sometimes much corroded by the ground-mass. A little green augite and magnetite also occur. The ground is a brown glass rich in crystallitic growths or in minute felspar- microlites. The interesting rock described by Prof. Judd 8 from the 1 G. M. 1894, 551-553 (Carrock Fell dykes) ; Reed, Q. J. G. S. (1895) li, 183-187, PL vi, figs. 6, 7 [2292, 2293] (Fishguard). 2 The name was first used by Stacbe for quartz -bear ing andesites in Transylvania (Dacia). 3 Fouque and L6vy, PL xvm. 4 Ibid. PL xvn, fig. 2. 5 Ibid. PL xxx ; xxxi, fig. 1. 6 Judd, Q. J. G. S. (1890) xlvi, 380. ? Q. J. G. S. (1896) Hi, 372. 8 Q. J. G. S. (1886) xlii, 427-429, PL xiu, figs. 7, 8. BRITISH AND AMERICAN DACITES. 179 (probably) Old Red Sandstone breccia near Scroggieside Farm in N.E. Fife is on the border-line between rhyolite and dacite. It has a glassy modification, which the author styles mica- dacite-glass. Phenocrysts of oligoclase and deep brown biotite are embedded in a glassy ground-mass containing trichites, globulites, and imperfect microlites of felspar (perhaps ortho- clase). The glass shews beautiful perlitic fissures. Little is known of true dacites among the Lower Palaeo- zoic lavas of this country, though some of the rocks in- cluded above as rhyolites would probably be styled dacites by certain petrologists. The name has also, as remarked above, been applied loosely to some of the acid intrusives. A number of dacites were described from Nevada by Zirkel 1 , and some of Richthofen's ' glassy rhyolites ' from the same region seem to belong rather to this family 2 . Dacites are also well represented among the Tertiary and Recent lavas in California, Oregon, and Washington, and in San Salvador 3 . Biotite is prominent among the ferro-magnesian minerals, and sometimes hornblende. At Lassen's Peak in California 4 occurs a type rich in phenocrysts, which consist of plagioclase felspar, biotite, hornblende, and quartz, while the ground-mass is essentially of glass. This is one of the original ' nevadites,' and most rocks so styled are probably to be classed as dacites. The andesites characterized by biotite or hornblende have affinities, as already remarked, with the typical trachytes. A mica-andesite free from hornblende is exceptional, but the name may be applied to varieties in which biotite is the dominant, though not the sole, ferro-magnesian constituent. The rocks usually taken as the type of hornblende-andesite 5 are those of the Tertiary volcanic district of the Siebengebirge, near Bonn, already alluded to as the home of certain typical 1 Micro. Petrogr. Fortieth Parallel (1876) 134-142 : see also Iddings, Monog. xx U. S. Geol. Sur. (1893) 368-373 (Eureka district). 2 Hague and Iddings, A. J. S. (1884) xxvii, 460, 461. 8 Ibid. (1886) xxxii, 29, 30. 4 Ibid. (1883) xxvi, 231-233. 5 For coloured figures of several French examples see Fouqu6 and Levy, PL xxn, xxvui, xxix, xxxvm. 122 180 HORNBLENDE-ANDESITES. trachytes. In addition to abundant brown hornblende, these andesites contain more or less biotite and a few prisms or FIG. x20. HORNBLENDE-ANDESITE, STENZELBEKG, SIEBENGEBIRGE ; The hornblende (h) and subordinate biotite (b) shew the resorption- border ; the phenocrysts of felspar (p) shew zonary banding and glass- cavities ; the ground-mass is only diagrammatically represented [117 a]. grains of pale green augite. The two former minerals always shew the phenomenon of resorption. The Bolvershahn rock, with a considerable amount of deep brown biotite, may be called a hornblende-mica-andesite. The felspar phenocrysts shew very marked zonary banding in polarized light. The Wolkenburg rock is a characteristic hornblende-andesite. Its phenocrysts include the three ferro-magnesian minerals men- tioned, hornblende largely predominating, good crystals of andesine, and a little magnetite and apatite, while its ground- mass is of the trachytic type. A very similar rock is that of Stenzelberg (fig. 38), in which some of the hornblende crystals attain a conspicuous size. In America Iddings 1 has recorded mica-andesites, horn- blende-mica-andesites, and hornblende-pyroxene-andesites from the Tewan Mts in New Mexico. These rocks have a glassy 1 Bull. No. 66 U. S. Geol. Sur. (1890) 12-16. HORNBLENDE-ANDESITES. 181 base. Similar examples come from Lassen's Peak (Cal.), Mt. Hood (Ore.), and Mt. Rainier (Wash.) 1 . The phenocrysts often shew parallel intergrowths of hornblende, augite, and hypersthene. The 'trachytes' of Zirkel 2 and others in the Great Basin and elsewhere are in part hornblende-mica-andesites 3 . This type occurs with others at the Comstock Lode 4 , and an example with beautifully zoned felspar phenocrysts has been described by Iddings 5 from the Eureka district. Others occur in the Sierra Nevada of California 6 . In these districts horn- blende-andesites free from mica are also found. In our own country these rocks are very poorly represented. One good example occurs on the summit of Beinn Nevis 7 , and, though probably of Carboniferous age, it is fairly fresh. The phenocrysts are of light-brown idiomorphic hornblende and a plagioclase full of glass-inclusions, etc. The ground-mass is obscured by specks of iron-ore and alteration-products, but is seen to consist largely of densely packed, minute felspar- microlites. Little is known of hornblende-andesites among the Lower Palseozoic and pre-Palseozoic lavas of Britain or of America. An Ordovician hornblende-andesite of somewhat basic composition occurs near Kildare 8 . Good examples occur in Minesota 9 . Andesites having a pyroxene as their dominant non-fel- spathic constituent are perhaps more widely distributed than any other group of lavas, and are largely represented among the products of volcanoes now active. Since a rhombic and a monoclinic pyroxene are often associated, the rocks are spoken of as pyroxene-andesites, while the marked predominance of one or other of these minerals gives a hypersthene- or an augite-andesite. Iddings, 12th Ann. Eep. U. S. Geol. Sur. (1892) 610-612, PI. LI. Micro. Petrof/r. Fortieth Parallel (1876) 143-162. Hague and Iddings, A. J. S. (1883) xxvi, 460. Hague and Iddings, Bull. No. 17 U. S. Geol. Sur. (1885) 23. iMonog. xx U. S. Geol. Sur. (1893) 364-368, PI. v, figs. 1, 3, 4 ; vi, fi& 21 2. Turner, 14*7i Ann. Rep. U. S. Geol. Sur. (1894) 487, 488. Teall, PI. xxxvn, fig. 1. Reynolds and Gardiner, Q. J. G. S. (1896) lii, 602. Wadsworth, Bull. No. 2 Geol. Sur. Minn. (1887), PL x, xi ; Grant, Ann. Rep. Geol. Sur. Minn. (1894) 57, 58. 182 HYPERSTHENE-ANDESITES. Hypersthene-andesites, or hypersthene-augite-andesites in which the rhombic pyroxene predominates over the mono- clinic, are especially widely distributed among the lavas of different periods. Prof. Judd 1 has pointed out that the same general petrographical type is found in lavas ranging in chemical composition from basalt to dacite. Thus the basic dykes of Santorin, the lava of Buffalo Peaks in Colorado, the Cheviot rocks, the recent lavas of Santorin, and the rocks of Krakatau consist of the same minerals in a glassy base of the same general composition, but the relative proportions of the minerals (in the aggregate basic) to glass (decidedly acid) varies in the different cases from 9:1 to 1 : 9. This illustrates the impossibility of naturally classifying by mineral- ogical characters alone rocks which have a glassy base. The wide distribution of hypersthene-andesites in Europe and America was first insisted upon by Whitman Cross 2 , who shewed that in a very large number of andesitic lavas hyper- sthene had previously been mistaken for augite. The rock upon which his first observations were made was from Buffalo Peaks, Colorado. The 'augite-andesites' of ZirkeP from Nevada have both rhombic and monoclinic pyroxenes, but the former predominates 4 , and true augite-andesites seem to be unrepre- sented among the lavas of the Great Basin region. Hyper- sthene-andesites occur in great variety among the Recent lavas of Mt. Shasta (CaL), Mt. Rainier (Wash.), etc. These are crowded with phenocrysts of zoned plagioclase and pyroxenes, hypersthene predominating over augite, while the ground-mass varies from holocrystalline to vitreous 5 . Andesites carrying- hornblende in addition to hypersthene occur in the Eureka district 6 , the Sierra Nevada 7 , etc. Pyroxene-andesites are abundant among the older vol- canic rocks .of Britain. Some in the Lake District contain plenty of pseudomorphs after a rhombic pyroxene (Falcon G. M. 1888, 1-11. Bull. No. 1 U. S. Geol. Sur. (1883) ; A. J. S. (1883) xxv, 139. Micro. Petrogr. Fortieth Parallel (1876) 221-227, PI. xi, fig. 2. Cross, I.e. ; Hague and Iddings, A. J. S. (1884) xxvii, 457-460. Hague and Iddings, A. J. 8. (1883) xxvi, 222-235. Iddings, Monog. xx U. S. Geol. Sur. (1893) 348-364, PI. vn. fig. 1. Turner, lth Ann. Rep. U. S. Geol. Sur. (1894) 488. OLD BRITISH PYROXENE-ANDESITES. 183 Crag near Keswick, etc.), while many others are characterized by monoclinic pyroxene only (fig. 40). A few of these rocks have been described by Mr Clifton Ward, Prof. Bonney, and Mr Hutchings 1 . The ground-mass is usually typically hyalo- pilitic. In the Bala series of Caernarvonshire there are few an- desites. Some, with augite only, occur in the Lleyn district 2 , and one with dominant hypersthene forms an intrusive mass at Carn Boduan 3 in the same district (fig. 39). No detailed study has yet been made of the Arenig pyroxene-andesites of Merioneth, but lavas of the same approximate age in the FIG. 39. HYPEKSTHENE-ANDESITE, CAKN BODUAN, CAERNARVON- SHIRE ; x 20. Phenocrysts of felspar and bastite pseudomorphs after hypersthene are set in a ground-mass, in which felspar-microlites with partial fluxional arrangement are the most conspicuous element [643]. Stapeley Hills (Todleth, etc.) in Shropshire are of the same general type as the Cheviot rocks, containing both rhombic and monoclinic pyroxenes, and this is true also of the Bala lavas of the Breidden Hills (Moel-y-golfa, etc.) 4 . 1 G. 31. 1891, 539-544. 2 Bala Vole. Ser. Caern. 68. 3 Ibid. 69-71. 4 Watts, Q. J. G. S. (1885) xli, 539-543 ; Proc. Geol. Assoc. (1894) xiii, 337-339, with figures. 184 OLD SCOTTISH AUGITE-ANDESITES. Many of the old lavas loosely grouped under the field- term 'porphyrite' in the Old Red Sandstone and Carboni- ferous of Scotland are andesites, ranging in composition from a relatively acid type (dacite) to varieties verging on basalt. One of the former, from North-east Fife, has already been mentioned. In the same district are good examples of more basic types also 1 . One quarried at Northtield is an augite- andesite with phenocrysts of augite, and perhaps subordinate enstatite, in a ground-mass largely of glass filled with globul- ites and trichites and a profusion of felspar-microlites. An- other, from what seems to be an old volcanic neck, quarried at Causeway Head, is an enstatite-andesite of more crystalline FIG. 40. AMYGDALOIDAL AUGITE-ANDESITE, STOCKDALE, WESTMORLAND ; x 20. The rock is considerably weathered, the augite being wholly replaced by a green chloritoid mineral. The same substance, in bunches of little scales, lines the vesicle seen on the left of the figure, and a ring of opaque decomposition products borders the vesicle. On the right is a group of felspar phenocrysts [758]. type, and is an aggregate of little prisms of triclinic felspar (near andesine), prisms and granules of pale enstatite, and grains of magnetite, with very little glassy residue. There i Judd, Q. J. G. S. (1886) xlii, 425-427, PI. xin, figs. 1, 2. ENGLISH PYROXENE-ANDESITES. 185 are no porphyritic elements. Subordinate augite accompanies the rhombic pyroxene, and biotite is another accessory. The Old Red Sandstone lavas of the Cheviots' are mostly hypersthene-andesites, containing both rhombic and mono- clinic pyroxenes. The freshest type shews phenocrysts of labradorite, honeycombed with inclusions of ground-mass, crystals of hypersthene shewing distinct pleochroism, and crystals and grains of pale augite, in a ground-mass of pale brown glass and felspar microlites. The ground often has flow-structure, and shews varieties of the hyalopilitic type. The iron-ores are represented by magnetite and minute red scales of haematite. The rock is often veined by opal or chalcedony, stained red with ferric oxide. The more weathered lavas of the district (part of the ' porphyrites ' of some authors) have had similar characters, but the felspars and pyroxenes are more or less decomposed, and the ground ob- scured by ferruginous matter. There are sometimes vesicles, filled with chalcedony, etc. Fresh examples come from Kilham, Longknowe, Haddan, and Coldsmouth Hills. Certain dykes described by Mr Teall 2 in the North of England may be referred to here, being petrographically augite-andesites. Some of them (Cleveland, etc.) must be of Tertiary age, while others are probably late Palaeozoic. The Cleveland dyke is traced from near Whitby to Armathwaite near Carlisle, and perhaps farther. It contains porphyritic felspars, often broken, in a ground-mass composed of small felspar crystals, minute crystals and grains of augite, crystals of magnetite, and abundant interstitial matter. This last is sometimes glassy, but commonly charged with various products of devitrification, giving a decided reaction with polarized light. The Acklington dyke is similar, but usually without the porphyritic crystals and with less of the inter- stitial base. The Tynemouth dyke is less fine-textured. It contains porphyritic aggregates of anorthite crystals in a 1 Teall, PI. xxxvi, xxxvn, fig. 2 ; G. M. 1883, 102-106, 146-152, PI. iv, 252-254 ; Petersen, G. M. 1834, 226-234 (Abstr.) ; Watts, Mem. Geol. Sur. Eng. and Wales, Expl. of Quarter-sheet 110 S. W., N. S. sheet 3 (1895) 12, 13. 2 Q. J. G. S. (1884) xl, 209-247, PL xn, xm ; Brit. Petr. PL xir, xiv. 186 BRITISH TERTIARY ANDESITES. ground-mass of elongated lath-shaped felspars, grains of augite, magnetite, and a considerable amount of interstitial base with devitrification-products and microlites and skeletons of fel- spar 1 . The dykes of Hebburn, Brunton, Seaton, and Hartley are similar, though usually without the large felspars. The interstitial base of these rocks, with its enclosed cryst- allitic bodies and its devitrification phenomena, presents various points of interest. Mr Teall points out the resem- blance of the Cleveland dyke to the Eskdale dyke in the south of Scotland, which has been described by Sir A. Geikie 2 , and which in places consists very largely of glassy matter enclosing various crystallitic growths. Prof. Judd 3 , describing Tertiary dykes of this group in Arran, remarks as character- istic of them the tendency of the glassy residue to become separated from the crystalline portion of the rock, either as a selvage or as a central band, or in irregular patches and strings (Eskdale), or, again, wholly or partially filling vesicles in the rock, as already remarked by Mr Teall 4 in the Tyne- mouth dyke. The Tertiary andesitic rocks of the Western Isles of Scotland, as described by Prof. Judd 5 , are of somewhat peculiar character. There are more acid types with horn- blende or biotite, or both, and less acid with pyroxenes, augite predominating. Structures approaching the holocrystalline (doleritic) are common, though other kinds are found in great variety, and some have been largely glassy. The most striking feature, however, is the wide-spread chemical al- teration which has affected the rocks, obliterating to a great extent their original characters, and giving rise to abundant epidote, chlorite, and other secondary products, including sometimes pyrites. In their frequent dioritic or doleritic aspect and their peculiar mode of decomposition these lavas 1 The structure is the ' intersertal ' of Rosenbusch, who cites these dykes as examples of his ' tholeiite ' ; cf. Mass. Gest., PI. iv, fig. 2. 2 Proc. Eoy. Phtjs. Soc. Edin. (1880) v, 244-252, PI. v, vi ; Teall, p. 196, PI. xxiv, fig. 1. 3 Q. J. G. S. (1893) xlix, 541. 4 G. M. 1889, 481-483, PI. xiv. 5 Q. J. G. S. (1890) xlvi, 341-382. 1 PROPYLITES.' 187 resemble very closely the rocks to which the name l propylite ' has been given in Hungary and the Western States of America. It is now generally recognized that the rocks to which this name was applied by Richthofen, Zirkel, and others, do not constitute a distinct family, but are altered forms, partly of andesites, partly of various hypabyssal rocks. This appears, for instance, from Zirkel's own account of the dif- ferences between the ' propylites ' and andesites of the Western States 1 . The 'green hornblendes' supposed to characterize the former are, according to Becker, chloritic pseudomorphs 2 . 1 Micro. Petrogr. Fortieth Parallel, 132, 133. 2 On the question of propylite see Wadsworth, Proc. Bost. Soc. Nat. Hist. (1883) xxii, 416, 417. CHAPTER XIV. BASALTS. IN the basalt family we include all the basic lavas except those in which a relatively high content of alkalies has given rise to the formation of minerals of the felspathoid group. The rocks range in texture from vitreous to holocrystalline. Except in a few of the latter (dolerites), the distinction be- tween phenocrysts and ground-mass is commonly well marked, but the relative proportions of the two vary greatly in dif- ferent types. The characteristic minerals in this family of rocks are a felspar rich in lime, augite, and olivine. Following our principle, we shall make no distinction, as regards nomenclature and classification, between Tertiary and pre-Tertiary lavas. Foreign petrologists usually restrict the names basalt and dolerite to the newer examples, their older equivalents being denoted by such names as melaphyre, augite- porphyrite, diabase, etc., some of which are also applied to rocks of the hypabyssal division. Certain exceptional lavas (limburgites, etc.) which are of ultrabasic, rather than normally basic, composition will be briefly noticed. Some of them probably correspond rather with the nepheline-basalts, etc., treated in the succeeding chapter. Constituent minerals. The felspars of the basalts are of decidedly basic varieties. When distinctly porphyritic crystals occur, they seem to be usually bytownite or anorthite, while the felspars of the ground-mass are more commonly labradorite. The phenocrysts shew albite-lamellation, often MINERALS OF BASALT. 189 combined with pericline- and carlsbad-twinning. Zonary structure and zonary arrangement of glass- cavities are met with. The felspars of the ground-mass have the lath-shape, and are commonly too narrow to shew repeated twinning. Orthoclase is found only in certain abnormal types. The dominant pyroxenic constituent is an ordinary augite, and this too may occur in two generations. If so, the pheno- crysts often have good crystal-forms, with octagonal cross- section ; twinning is frequently seen ', and sometimes zoning and hour-glass structure. The colour is usually very pale, brownish or more rarely greenish, the latter especially in the interior of a crystal. The augite of the ground-mass is either in little idiomorphic prisms or in granules, and is often very abundant. Decomposition of the augite produces chloritoid substances, etc. 2 A rhombic pyroxene, hypersthene or bronz- ite, occurs only in certain basalts, where it seems to some extent to take the place of olivine. It is always in idio- morphic prisms, and in the older rocks is very generally serpentinized. Some basalts, again, contain corroded crystals of brown hornblende, and others a little brown mica. Octahedra and grains of magnetite are generally abundant, and this mineral frequently recurs in a second generation in little granules. Besides this, there are frequently little opaque or deep brown scales of ilmenite or deep red flakes of hannatite. Grains of native iron occur locally in a few basalts (Ovifak in Disco, Greenland) 3 . In the greater part of the basalts olivine 4 ' is an essential constituent, and in many it is abundant, though confined, as a rule, to phenocrysts. These are sometimes well shaped crystals, sometimes more or less rounded. The mineral is colourless or very pale green. It often shews serpentine- strings following cleavage- or other cracks 5 , and with further alteration passes into various secondary products, serpentine, 1 Cohen, Pl.'xxvm, fig. 1; xxix, fig..l. 2 Teall, PI. xxn, fig. 2. 3 Fouque' and Levy, PI. xxxvi, fig. 2 ; Steenstrup, M. M. i, 148, PI. vi. 4 Cohen, PL xxi, fig. 2. 5 Zirkel, Micro. Petrogr. Fortieth Parallel, PI. x, fig. 3 ; xi, fig. 3. 190 MINERALS OF BASALT. carbonates, etc. Another common change is the production of a red or brown margin to the olivine, due to iron-oxide, the olivine in basalts, and still more in limburgites, being often of a variety rich in iron. Another mode of alteration sometimes met with results in the formation of brown pleo- chroic pseudomorphs of a mineral with a perfect cleavage and the appearance of a mica. It seems to agree in general characters with the mineral described in California by Lawson 1 under the name iddingsite ; but the author named, regarding this as an original constituent, has made it the characteristic of a new group of lavas (carmeloites). Of other common minerals we need note only apatite, forming long needles, either colourless or of a faint violet or bluish tint. A peculiar feature in certain American basalts 2 is the occurrence of isolated grains of quartz. These are always corroded by the magma and generally surrounded by a ring of augite or its alteration-products, a character usually associated with foreign quartz-grains picked up by a basic magma. In this case, however, there is reason to believe that the mineral is an original, constituent formed under peculiar conditions. It is comparable with similar grains found in many lamprophyres (see above, p. 138). Structures. The rocks of the basalt family present a wide range of characters, from purely glassy examples at one extreme to wholly crystalline at the other. Rocks ex- hibiting such a range may occur, perhaps exceptionally, in one district, their petrological characters being correlated with their various modes of occurrence, as is well described by Prof. Judd 3 . On the whole, the tendency to crystallization is much stronger here than in the more acid families of lavas. Again, the order of crystallization of the several constituents is less strongly marked, the mutual relations between augite 1 Bull. Geol. Dep. Univ. Gal. (1893) i, 29-46, PL iv. 2 Ciller, A. J. S. (1887) xxxiii, 45-49 ; Bull. No. 79 U. S. Geol. Sur. (1891) 24-29; ladings, A. J. S. (1888) xxxvi, 209-213; Bull. No. 66 U. S. Geol. Sur. (1890) 16-31 ; Monog. xx U. S. Geol. Sur. (1893) 393, PL iv, fig. 1. Q. J. G. S. (1886) xlii, 66-82, PL v, vi. STRUCTURES OF TACHYLYTE AND VARIOLITE. 191 and felspar, in respect of priority, varying, while the iron- ores, though they commonly begin to crystallize at an early stage, may be in part rather late. These remarks are true of both the ' intratelluric ' and the ' effusive ' periods, when these are distinctly separable, but in some of the holocrystalline types the porphyritic character is not recognizable. Some of these rocks differ in no essential from those already described as diabases, the petrological distinction between the hypabyssal and the volcanic types not being marked by any hard and fast line. Except in the form of lapilli and fragments in tuffs, the purely vitreous type, tachylyte, is of very limited distribution, being found only as a very thin crust on some lava-flows or a narrow selvage to basalt-dykes. It consists of a brown or yellow glass densely charged with a separation of mag- netite. This is sometimes in globulites l disseminated through the glass so as to render it almost opaque, or collected in cloudy patches (cumulites); at other times it forms trichites or crystallites of minute size 2 . Perlitic structure is less common than in the obsidians. Interesting spherulitic struct- ures are met with in some examples 3 . When distinct pheno- crysts occur abundantly in the glassy ground-mass, we have what is sometimes called the ' vitrophyric ' structure. The basic glass is subject to secondary changes, probably involving, as a rule, hydration and other chemical changes, but the resulting substance, known as palagonite, is still an isotropic glass, yellow, brown, or sometimes green in sections. Radiate aggregates of felspar microlites or fibres, answering to the spherulites of acid rocks, occur in some basaltic glasses, which are known as variolitea. These aggregates vary in size arid in the regularity of their structure, which ranges from mere fan-like and sheaf- like groupings (cf. fig. 41 A) to spherules with a perfect radiate structure. They may occur isolated in a glassy matrix, or coalesce into bands, or form a densely packed mass with little or no interstitial matter. The variolites are very susceptible to alteration. 1 Cohen, PL n, fig. 4 ; xi, fig. 2. 2 Judd and Cole, Q. J. G. S. (1883) xxxix, PI. xiv. a Cole, ibid. (1888) xliv, 300-307, PL xi. 192 HYPO-CRYSTALLINE STRUCTURE IN BASALT. Leaving the glassy basalts, we note those in which the ground-mass enclosing the phenocrysts of olivine, augite, felspar, etc., is hypocry stall ine, consisting of lath-shaped felspar- microlites and granules or microlites of augite with more or less of a residual glassy base. Of this division there are various types, depending on the relative proportions of augite, felspar, and glass, and the mutual relations of the minerals. When the felspar-microlites preponderate, usually with a more or less fluxional arrangement, the ground-mass does not differ essentially from the ' hyalopilitic ' type so common in the pyroxene-andesites. Vesicles are frequent in such rocks. B GLU FIG. 41. A. Andesite vein approaching the structure of variolite, Carrock Fell, Cumberland ; x 20, crossed nicols. This is of the type which consists essentially of radiating felspar tibres grouped in sheaf-like bundles. There are also skeleton -prisms of a pyroxenic mineral, better been in natural light [1552]. B. Limburgite, Wbitelaw Hill, Had- dington ; x 20, natural light. Phenocrysts of olivine (ol), zoned augite (au), and magnetite are enclosed in a ground-mass of glass containing abundant prisms and granules of augite but no felspar. The glass, which constitutes the bulk of the ground, varies from brown to nearly colourless [1982]. More often, however, augite is abundantly represented in the basaltic ground-mass. Again, unindividualised glass may form HOLOCRYSTALLINE STRUCTURE IN BASALT. 193 the bulk of the ground, and this is especially the case in the limburgites (tig. 41 B). By the failure of the glassy residue we pass to those types of basalt in which the phenocrysts are enclosed in a holocrystalline ground-mass. Here again there are numerous varieties. Sometimes little eye-like or lenticular patches re- latively rich in augite are contrasted with adjacent patches rich in felspar. When felspar-microlites make up a large part of the ground-mass, we have a structure analogous to the ' pilotaxitic ' of some andesites and trachytes, the flow being more or less marked. On the other hand, the ground may consist mainly of small rounded granules of augite, between which the little felspars seem to be squeezed (fig. 42).. There remain the types distinguished as dolerites (usually olivine-dolerites), which, in the most typical examples, are holocrystalline rocks not conspicuously porphyritic, sometimes of coarse texture as compared with the generality of lavas. The chief structures are the granulitic and the ophitic, the f CLW FIG. 42. BASALT, ETNA LAVA OF 1669 EBUPTION, CATANIA ; x 20 : Shewing phenocrysts of augite (au), felspar (/), olivine (ol), and magnetite (m) in a holocrystalline ground-mass of little lath-shaped felspars and granules of augite and magnetite [131].. H. P. 13 194 GLOMEROPORPHYRITIC STRUCTURE IN BASALT. distinction between which has been noticed (p. 126) under the diabases. Typical ophitic structure is rare in true lava- flows. The ' intersertal ' structure of Rosenbusch corresponds in part with the granulitic, but it includes the type in which some residual glass, as well as augite and other minerals, occurs in the interstices between the lath-shaped felspar crystals. Only exceptionally in doleritic lavas do we find an idiomorphic development of the augite and an approach in structural characters to some plutonic types (e.g. the Lb'wen- burg olivine-dolerite in the Siebengebirge). Some dolerites enclose large scattered porphyritic crystals of felspar. In other cases there are porphyritic aggregates of crystals (felspar, olivine, augite, etc.} having the mutual relations characteristic of plutonic rocks : this is the glomero- porphyritic structure of Prof. Judd 1 . It is not confined to the holocrystalline dolerites. The crystals forming such a hypidiomorphic aggregate may still present idiomorphic out- lines towards the surrounding rock 2 . Many of the Tertiary basalts in Germany, etc., enclose so- called ' olivine-nodules,' which are hypidiomorphic aggregates of olivine with enstatite, diopside, etc. 3 By some they have been regarded as very early intratelluric formations from the magma, by others as actual enclosed pieces of peridotites. Leading' types. Some basalts, belonging in general to the less basic varieties, are free, or nearly free, from olivine. These rocks usually carry a rhombic as well as a monoclinic pyroxene, and here, as in some other families, hypersthene may be considered as, to some extent, taking the place of the more basic silicate olivine. Such rocks, which may be styled hypersthene-basalts, occur among the Tertiary lavas of the western United States. Examples have been noted by Iddings 4 from the Eureka mining district in Nevada. The Ordovician lavas of the English Lake District are chiefly of 1 Q. J. G. S. (1886) xlii, 71, PL vn, fig. 3. 2 Teall, ibid. (1884) xl, 235, PI. xm, fig. 1. 3 For coloured figures see A. Becker, Zeits. deutsch. geol. Ges. (1881) xxxtii, PI. m-v ; Fouque and Levy, PI. XL, fig. 1. 4 Monog. xx U. S. Geol. Sur. (1893) 386-394, PI. vn, fig. 2. HYPERSTHENE-BASALTS. 195 this type, though, as already noticed, rhyolites and pyroxene- andesites are likewise found. Here the hypersthene is always converted into a light green, pleochroic, serpentinous substance comparable with bastite. The most striking variety, repre- sented at Eycott Hill 1 and numerous other localities in the district and at Melmerby 2 near Cross Fell, has large rounded phenocrysts of labradorite with carlsbad and albite-twinning. These contain rather large opaque inclusions in the form of negative crystals and smaller enclosures with zonary dispos- ition. In other varieties of the lavas these large crystals are not present. The ground-mass consists of slender striated Ib FIG. 43. HYPERSTHENE-BASALT, EYCOTT HILL GROUP, MELMERBY, CUMBERLAND ; x 20. To the right is one of the large crystals of labradorite (Ib) with its peculiar inclusions. The hypersthene is represented by bastite pseudo- morphs (ba) : augite occurs in less abundance. These, with the little felspar-prisms, the granules of magnetite, and some residual glassy base, make up the bulk of the rock [1251]. prisms of plagioclase, crystals of hypersthene converted to pleochroic bastite, granules of augite, abundant magnetite, and 1 Ward, Monthly Micro. Journ. (1877) xvii, 240-245 ; Bonney, G. M. 1885, 76-80 ; Teall, 225-227. 2 Q. J. G. S. (1891) xlvii, 517. 132 196 AMERICAN OLI VINE-BASALTS. an isotropic base (fig. 43). In the basic lavas of the Lake District generally olivine is entirely wanting. Hypersthene, pseudomorphed by bastite, is frequently present, but rarely to the exclusion of augite. We come next to the more widely distributed olivine-basalts. Such rocks are extensively developed among the lavas of late geological age in America ; for instance, in the Great Basin region, lying between the Rocky Mts and the Sierra Nevada. Here they are mostly porphyritic, with relatively large pheno- crysts of olivine, plagioclase, and occasionally augite in a glassy, microlitic, or microcrystalline ground-mass. A smaller number are non-porphyritic, consisting of a uniform aggregate of plagioclase, augite, olivine, and magnetite, often with a con- siderable amount of glassy base 1 . Other examples have been described from the Sierra Nevada 2 , the Tewan Mts (KM.) 3 , and San Salvador 4 . In the latter region it has been remarked that the varieties poor in olivine carry hypersthene in addition to augite. Recent olivine-basalts occur at many localities in Colorado, New Mexico, Arizona, and about Mt. Shasta and Lassen's Peak in California. In this last district Diller 5 has described a quartz-bearing basalt in which the dominant pyroxene is hypersthene. The Tertiary basaltic rocks of the Inner Hebrides and various parts of the west and south of Scotland and the north- east of Ireland are olivine-basalts (including olivine-dolerites). They have been well described and figured by Prof. Judd 6 , who has pointed out how the varied series of structures which they present constitute intermediate types between the holocrystal- line plutonic rocks at the one extreme and the glassy basalts (tachylytes) at the other. He distinguishes two parallel lines of transition. One, characteristic of the true extruded lava- flows, includes the ' granulitic ' dolerites and the basalts in 1 Hague and ladings, A. J. S. (1884) xxvii, 456, 457; cf. Zirkel, Micro. Petrogr. Fortieth Parallel (1876) 229-254 ; PI. x, figs. 1, 3, 4 ; xi, fig. 3. Turner, 14t7i Ann. Rep. U. S. Geol. Sur. (1894) 490-492. Iddings, Bull. No. 66 U. S. Geol. Sur. (1890) 16. Hague and Iddings, A. J. S. (1886) xxxii, 27, 28. Bull. No. 79 U. S. Geol. Sur. (1891). Q. J. G. S. (1886) xlii, 49-95, PL iv-vn: see also Teall, PI. x. SCOTTISH TERTIARY BASALTS. 197 which the augite tends to form granules between the felspar prisms ('microgranulitic' structure). The other series of varieties includes the ophitic dolerites and the micro-ophitic basalts, in which the augite tends to enwrap and enclose the felspars : this seems to be the case especially in intrusive members of the group. The distinction is traceable even in those basalts which consist largely of a glassy base, the crystal- litic growths enclosed in the glass being in the one case in the form of granules and short microlites, often rounded, in the other case in the form of skeleton-crystals and more spreading growths. Many of the Scottish dolerites and most of the basalts are porphyritic, the felspar occurring in two generations, of which the earlier is a thoroughly basic variety, near anorthite, while the latter is less basic, usually labradorite. Por- phyritic augite, however, is not found, and this feature dis- tinguishes the group of rocks in question from the Tertiary basalts of various European areas and also from many Car- boniferous basalts of Scotland and Ireland. Professor Judd notes the absence of olivine-nodules as another distinctive feature of the British Tertiary basalts. Of the Tertiary olivine-dolerites of intrusive occurrence in the Western Isles and others, probably of like age, in the southern parts of Scotland, many have ophitic structures, and approach true diabases in their characters. Others, how- ever, are of the 'granulitic' type, and these, in addition to the dominant lath-shaped felspars, shew a later generation of more acid composition, in shapeless grains with marked zonary banding between crossed nicols (e.g. Craig Craggen in Mull, Muckraw in Linlithgowshire). Of the basalts of the Antrim plateau 1 , some have porphyr- itic felspars, but most are of quite compact character. Olivine grains are enclosed in a mass of elongated felspar-crystals and granules of augite, with occasionally a second generation of smaller olivines. The basic lavas of Carboniferous age in this country are also characteristically olivine-bearing rocks. Those of 1 Watts, Guide, 79. 198 BRITISH CARBONIFEROUS BASALTS. Derbyshire l are chiefly of doleritic type. The minerals present are idiomorphic olivine (sometimes replaced by a remarkable mica-like mineral 2 ), augite, exceptionally a rhombic pyroxene (Sandy Dale), labradorite or a more basic felspar, magnetite and ilmenite. Most of the rocks are olivine-dolerites of granul- itic structure, the augite occurring in grains (Castleton, Tides- well Dale 3 , Miller's Dale, etc.). A few are ophitic (Peak Forest and Bonsall, see fig. 25, p. 130). Rarely there are porphyritic olivine-basalts with olivine and large augite phenocrysts in a ground-mass of small felspar laths, augite grains and prisms, and iron-ores, with little interstitial matter (Blackwell Lane, Great Low). The Kelso lavas, in the Lower Carboniferous of the Cheviot district, are olivine-basalts with phenocrysts of anorthite. One from Stichill in Roxburghshire was described by Mr Teall 4 . In other examples, from Northumberland, Mr Watts 5 notes brown pleochroic pseudomorphs after olivine, which he identifies with iddingsite. The Carboniferous olivine-basalts of the southern half of Scotland present a considerable variety of characters 6 . The commonest type has rather abundant small olivines and grains of augite in a mesh of slender felspars with microlitic augite and minute granules of magnetite (Dalmeny, Bathgate Hills, etc.). In another type the olivine phenocrysts are large, and the felspar microlites are found only in small amount (lowest lavas of Bathgate Hills, Linlithgowshire). A well-known rock from the Lion's Haunch on Arthur's Seat, Edinburgh 7 , has numerous large, well-built crystals of augite, olivine, and felspar, with little crystals of magnetite, in a ground-mass of little crystals and microlites of felspar, granules of augite and 1 Arnold-Bemrose, Q. J. G. S. (1894) 1, 611-625. 2 Ibid. PI. xxiv, figs. 1-4. This pseudomorph, apparently closely allied to the iddingsite of Lawson, is not infrequent also in the basalts of Skye and Mull. 3 Teall, PI. ix. 4 G. M. (1883) 258-260, PI. vi. 5 Mem. Geol. Sur. Engl. and Wales, Expl. of Quarter-sheet 110 S. W., N. S. sheet 3 (1895) 14. 6 Geikie, Q. J. G. S. (1892) xlviii, Proc. 105, 106. 7 Teall, PI. xxin, fig. 1. BASALTS OF THE ISLE OF MAN. 199 magnetite, and some residual glass. In the lava of Craig- lockhart Hill the ground-mass is more glassy, while the phen- ocrysts are augite and olivine without felspar. On the other hand, there is a holocrystalline type, which is an olivine- dolerite with granulitic to sub-ophitic structure (Gallaston, N.W. of Kirkcaldy). A curious variety, very rich in felspar, comes from Markle quarry in the Garlton Hills, Haddington- shire 1 . Here olivine occurs only in small sporadic grains, while phenocrysts of labradorite are numerous, and the ground- mass consists of laths, microlites, and granules of felspar with dispersed magnetite and probably only a little augite. A rock very like that of Lion's Haunch occurs as a dyke near the Stack of Scarlet in the south of the Isle of Man 2 . The phenocrysts are large idiomorphic crystals of fresh plagio- clase and violet- brown augite, with pseudomorphs of calcite and serpentine after olivine. The ground-mass is of lath- shaped felspars, augite, and iron-ores. This is probably con- nected with the Carboniferous volcanic series of the Stack, which consists of tuffs with dykes and probably flows of a more compact basalt 3 . The latter is considerably decomposed, the augite being converted into chloritic and other products. Porphyritic felspars occur, and the little lath-shaped felspars of the ground-mass shew a fluxional arrangement. The much fresher basalt, which forms numerous small dykes in the south of the Isle of Man 4 , is probably of Tertiary age. The olivine here is abundant and fresh, with inclusions of picotite. This, and sometimes plagioclase, are the only phenocrysts. The ground-mass is in general holocrystalline with fine texture, consisting of felspar microlites, ophitic violet-pink augite, and magnetite. Analcime, sensibly isotropic, occurs as a decomposition-product. In the neighbourhood of Limerick is a considerable de- velopment of basaltic lavas of Carboniferous age. These differ from the Irish Tertiary basalts in various points, and especially in the frequent presence of augite among the phenocrysts. 1 Hatch, Trans. Roy. Soc. Edin. (1892) xxxvii, 119, PL i, fig. 2. 2 Hobson, Q. J. G. S. (1891) xlvii, 443, 444. 3 Ibid. 441. * Ibid. 445-447. 200 VARIOUS OLD BASALTS. Olivine-basalts do not, as a rule, figure largely in the great volcanic groups which characterize the Lower Palaeozoic in various parts of Britain. Sir A. Geikie 1 has noted oli vine- basalts of early Cambrian (or late pre-Cambrian) age near St David's (Rhosson, Clegyr Foig, etc.). The idiomorphic crystals of olivine in these rocks are replaced largely by haematite. The ground-mass consists of augite-granules, abund- ant octahedra of magnetite, and a base crowded with globul- ites and trichites, felspar being only occasionally recognized. These characters suggest a resemblance to the limburgite type, noticed below. Various basaltic lavas are intercalated in the Palaeozoic strata of Cornwall and Devon. Some have been largely vitreous, the glass being now represented by a greenish yellow to brownish yellow serpentinous-looking substance which seems to be identical with the so-called palagonite (Cant Hill, near St Minver) 2 . These rocks are often amygdaloidal. In America ancient olivine-basalts have been described from Notre Dame Bay in Newfoundland 3 , North Haven in Maine 4 , South Mountain in Pennsylvania 5 , the Penokee (Huronian) group 6 , Keweenaw Point, etc. (Mich.) and other localities in the Lake Superior region 7 , the Grand Canon of the Colorado 8 , and other districts of pre-Cambrian and Lower Palaeozoic rocks. The name tachylyte is commonly employed to cover the glassy representatives of both the basalts and the pyroxene- andesites. Examples occur at numerous places in the Tertiary volcanic districts of Skye, Mull, etc. 9 They usually enclose porphyritic crystals of olivine and magnetite, less commonly Q. J. G. S. (1883) xxxix, 304, PI. ix, fig. 4. Eutley, Q. J. G. S. (1886) xlii, PI. xn. Wadsworth, A. J. S. (1884) xxviii, 95. G. O. Smith, Joh. Hopk. Univ. Circ. No. 121 (1895). G. H. Williams, A. J. S. (1892) xliv, 490-492. Van Rise, Monog. xix U. S. Geol. Sur. (1892) 410. Pumpelly (Irving), Copper-bearing Rocks, etc., Monog. v U. S. Geol. Sur (1884) 69-77, PI. ix. ladings, Uth Ami. Rep. U. S. Geol. Sur. (1894) 520-524. Judd and Cole, Q. J. G. S. (1883) xxxix, 444-462, PI. xm, xiv. For localities of numerous other examples in Mull, see Kendall, G. M. 1888, 555-560. BRITISH TACHYLYTES AND VARIOLITES. 201 of augite and felspar. The glass is crowded with incipient growths of magnetite and occasionally of other minerals. These take the form of globulites, sometimes collected into cumulites (the Beal in Skye), of margarites (Lamlash near Arran), or of numerous minute opaque rods (Some in Mull, etc.), sometimes accompanied by transparent crystallites and belonites (Gribun in Mull). Spherulites occur in some in- stances. In the tachylyte of Ardtun in Mull 1 they are some- times isolated, sometimes in bands, sometimes packed together, with polygonal boundaries to the exclusion of any glassy matrix. When imperfect, they seem to consist of brown globulitic matter, which is more condensed towards the centres. When better developed, they shew radiating fibres arranged in sectors, some brown and others grey, with pleo- chroism in both cases. But little is known of tachylytes among the older volcanic rocks 2 . Closely allied to the spherulitic tachylytes are the rocks known as variolite, of which examples have been described from Anglesey, the Lleyn district of Caernarvonshire, and various parts of Ireland 3 . The spherules shew considerable variety of structure, ranging from mere fan-like groupings of felspar rnicrolites (cf. fig. 41 A, p. 192) or sheaf -like aggreg- ates with a lath-shaped crystal as nucleus (see Sollas) to very regular, radiate, spherulitic growths. They may be closely packed to make up the entire mass of a portion of the rock, or arranged in bands, or isolated in a matrix of brown or greenish glass with cumulites, globulites, etc. (see Cole). The individual spherules are commonly from one-tenth to one- half of an inch in diameter, but sometimes less or more. Secondary changes may cause devitrification of any glassy 1 Cole, Q. J. G. S. (1888) xliv, 300-307, PI. xi. 2 See Groom, Q. J. G. S. (1889) xlv, 298-304, PI. xn (Carrock Fell). :J Miss Raisin (Lleyn), Q. J. G. S. (1893) xlix, 145-159, PL i ; Cole (Careg Gwladys, Anglesey), Sci. Proc. Roy. Dubl. Soc. (1891) vii, 112-120, PL x; (Annalong, Co. Down) ibid. (1892) 511-519, PL xxi; (Dunmore Head, Co. Down) ibid. (1894) viii, 220-222; Sollas (Roundwood, Co. Wicklow) ibid. (1893) 99-106, figures. On foreign variolites see note 1, p. 202 ; also Lowinson-Lessing (' sordawalite ') M. M. viii, 164 (Abstr.) ; Brauns (Hesse) M. M. ix, 255, 256 (Abstr.). For coloured figure of the ' variolite of the Durance ' see Fouque and L