- THE BALA VOLCANIC SEEIES CAERNARVONSHIRE ASSOCIATED ROCKS. Hon&on: c. j. CLAY AND SONS, CAMBRIDGE UNIVERSITY PRESS WAREHOUSE, AYE MARIA LANE. CAMBRIDGE: DB1GHTON, BELL, AND CO. LEIPZIG: F. A. BEOCKHAUS. THE BALA VOLCANIC SEEIES OF CAERNARVONSHIRE AND ASSOCIATED HOCKS; THE SEDGWICK PRIZE ESSAY FOR 1888, BY ALFRED HARKER, M.A. F.G.S., FKLLOW OF ST JOHN'S COLLEGE, AND DEMONSTRATOR IN GEOLOGY (PETROLOGY) IN THE UNIVERSITY OF CAMBRIDGE. Sane multum illi egerunt, qui ante, nos fuerunt, sed non peregerunt ; multum adhuc restat opens, multumque retabit, nee ulli nato post mille sccula praecludetur occasio adjiciendi. SENECA. CAMBRIDGE : AT THE UNIVERSITY PRESS. ambrt&gc : PRINTED BY C. J. CLAY, M.A. AND SONS, AT THE UNIVERSITY PRESS. LIBRARY Q E UNIVERSITY OF CALIFORNIA 4 _. . SANTA BARBARA &*? H3 PREFACE. HHHE following pages are, in substance, the Sedgwick Prize Essay -- for the year 1888, the subject proposed by the Examiners being " The Petrology of the Igneous Rocks associated with the Cambrian (Sedgwick) System of Caernarvonshire." Since the award, I have completely re-written the essay, and have chosen a title which more accurately expresses its actual scope. I am indebted to Professor T. M c Kenny Hughes for the use of specimens in the Woodwardian Museum and for kind assistance in other ways. I have also to thank the other Examiners, Professor T. G. Bonney and Mr J. J. Harris Teall, for many valuable suggestions, and Mr E. Hamilton Acton for kindly supplying me with chemical analyses of a number of Caernarvonshire rocks. All the specimens described are in the Woodwardian Museum collections, and to facilitate reference to the original sections I have given, in square brackets, the numbers of the slides in the Museum Cabinet. ALFRED BARKER. ST JOHN'S COLLEGE, CAMBKIDGE, October, 1889. TABLE OF CONTENTS. I. Introductory: Three divisions of the county, 1; specimens, 3; litera- ture, 3. II. Rkyolitic lavas: Groups of flows in eastern Caernarvonshire, 8; Lleyn rhy elites, 10; the Welsh felstones true lavas, 12; chemical composition, 13 ; characters in the field and in hand-specimens, 16 ; microscopical study, 18 ; ashes and agglomerates, 25. III. Nodular rhyolites: Wide distribution, 28; general characters, 29; discussion of theories, 30 ; description of specimens, 35. IV. The add intrusives: Distribution of the rocks, 41; chemical composi- tion, 46 ; characters in hand-specimens, 47 ; microscopical study, 48 ; riebeckite, 50; types of micro-structure in the rocks, 53; 'con- cretionary patches,' 57 ; granite of Sarn, 58. V. Intermediate rocks : Quartz-dolerite, etc., of Penmaenmawr, 62 ; micro- scopical study, 65 ; augite-andesites of Lleyn, 66 ; pyroxene-audesite of Cam Boduan, 69 ; microscopical study, 70 ; andesitic rocks near Y Foel Fras, 71 ; syenite of Llanfaglen, 72. VI. The diabase sills, etc. : Mode of occurrence and date of intrusion, 75 ; general characters, 77 ; contact-alteration, 78 ; microscopical study, 79 ; basalts of Lleyn, 86. VII. Other basic intrusions: Gabbro of Craig-y-fael and horublendic rock of Llangwnadl, 89 ; hornblende-diabases of Sarn district, 92 ; horn- blende-picrite of Penarfynydd, 97 ; hornblende-diabase with olivine of Clynog-fawr, 102 ; the post-Carboniferous dykes of Caernarvon- shire, 106. VIII. Review of vulcanicity in Caernarvonshire: Interdependence of all the igneous rocks, 112; nature of crust-movements at close of Bala period, 113; relation of volcanic phenomena to crust-movements and to the Llyn Padarn ridge, 119; distribution of groups of lavas, 120; lavas probably submarine, 121; separation of subterranean magma under gravity, 122; succession of lavas, 123; injection of basic rocks, 125; sketch of history of Bala vulcaiiicity, 126. MAPS. PAGE FIG. 1. Sketch-map to shew the three divisions of Caernarvonshire . 1 FIG. 2. Neighbourhood of Llanbedrog and Madryn .... 27 FIG. 3. Neighbourhood of Pwllheli 46 FIG. 4. Neighbourhood of Sarn 59 FIG. 5. Sketch-map to illustrate the distribution of the Cleavage . .116 FIG. 6. Sketch-map to shew the limits of the Groups of Lavas . .120 THE THREE DIVISIONS OF CAERNARVONSHIRE. i^ Carboniferous. iZHEJ Silurian . \ \ Or do vie ia,n . HH Cambrian. 3 Archaean . MERIONETH B. JBangor . Co. Conwy . C. Caernarvon . L. ,Llan,T)eris Cf. Clynog-fawr . Lr. Llanrwst; Cambridge Ur N Nevirt Pw. Pwllheli, . T. TreTTtadoc . Fig. I I. INTRODUCTION. THE county of Caernarvon, excluding the Great Orme's Head and another small outlying patch near Abergele, has a length of about 52 miles, with a greatest breadth of 20 miles. More than half of its border is washed by the sea, from Portmadoc to Conwy ; the valley of the Conwy river forms a natural boundary on the east; while from near the head-waters of that river the county- limit runs along an elevated and sinuous line of water-shed to Llyn Edno, thence westward to Llyn-y-ddinas, and down the river Glaslyn to Traeth Mawr. The general geology of the county is too well known to re- quire description here, but a few words with respect to its physical features will prepare the way for our subject. Caernarvonshire falls naturally into three divisions of widely different physical character, and in each the scenery is easily correlated with the geological structure and especially with the arrangement of the igneous rocks. (See map, fig. 1.) The largest division, which we shall name the eastern, is limited on the north-west side by the Llyn Padarn ridge, which extends from Llanllyfni to St Anne's Chapel, and which we may imagine prolonged to near Penmaenmawr. Excepting the alluvial flats bordering the Conwy and Traeth-mawr this part of the county presents everywhere a mountainous type of scenery. The mountains culminate in Snowdon and Y Glyder Fawr, and in the Carneddau Dafydd and Llewellyn, the four highest peaks in Wales. The rocks from which the loftiest elevations are carved out consist largely of massive acid lavas, with some tuffs, ribbed by great intrusive sheets of diabase Only in two or three H.E. 1 2 INTRODUCTION. localities, as at Penmaenraawr, Bera-mawr and Y Foel Fras, and Mynydd Mawr, do large intrusive masses of igneous rocks form striking features of the landscape. The western division of the county, consisting of the peninsula of Lleyn, is very different. Here the hills, though sometimes rising, as in Yr Eifl, to nearly 1900 feet, owe their imposing appearance chiefly to relief, and, as compared with the eastern division, the general surface is level. The influence on the scenery of masses of igneous rock, almost all intrusive, is much more apparent ; and in a bird's-eye view of the Lleyn from some commanding station it is easy to identify each bolder eminence with its corresponding red patch on the geological map. The north-western part of Caernarvonshire is without either the great mountains of the eastern division or the abrupt isolated hills of the western. With a single exception, it con- tains no igneous rocks of importance to our subject. All the igneous rocks alluded to above, whether extruded or intruded, are to be assigned on good evidence to the Bala period, and it will be our business in the present essay to describe their nature and discuss their mutual relations. To the geologist in Caernarvonshire the admirable maps of the Geological Survey, supplemented by Sir Andrew Ramsay's Memoir, are of course invaluable, although they give but meagre and not always accurate information with regard to the petrology of the igneous rocks. The researches of Mr F. Rutley, Professor Bonney, and Mr Grenville Cole have taught us much with respect to the rhyolitic lavas of Snowdonia and their secondary modifica- tions ; Mr Tawney's Woodwardian Museum Notes treat of many of the intrusive rocks of the Lleyn peninsula ; useful contributions dealing with various other rocks in the county are due to Mr J. A. Phillips, Professor Bonney, Dr Hicks, and other authors ; and Mr Teall's valuable British Petrography affords a large fund of information on the subject ; but the majority of the igneous rocks of Caernarvonshire are yet undescribed, and such an essay as the present cannot of course aim at exhausting so wide a subject. The cabinets of the Woodwardian Museum contain many rocks collected in Caernarvonshire by Hailstone, Henslow, and especially Sedgwick, besides those gathered by the late Mr Tawney, the present Woodwardian Professor, Mr Marr, and others. The SPECIMENS. 3 specimens of the earlier geologists are, however, of little value for microscopic examination, being selected rather to illustrate the appearance of the rock in the field than to reveal its essential characters. They are mostly taken from exposed places such as the tops of hills, and so are usually deeply weathered. Again, it is not always possible to learn the precise localities of specimens obtained by the hand of others, more particularly in the case of the older collectors, who had to work with very imperfect maps, and did not always select rocks in situ. For these reasons a personal visit has been made to the home of almost every rock mentioned in this paper, to procure representative specimens from definite spots. This was also necessary in many cases in order to form an opinion on the mode of occurrence of these igneous masses, and their geological relations, often a subject of dispute. For convenience of reference, a list is given below of the chief books and papers dealing directly with the igneous rocks of Caernarvonshire, so far as they are embraced in the present essay. The abbreviations employed here and throughout the following pages are : Q. J. G. S. Quarterly Journal of the Geological Society of London. G. M. Geological Magazine. Maps of the county on the scale of one inch to a mile are published with geological colouring. Most of the county is included in the quarter-sheets 75 N.E., N.W., S.W., 76 N. and S., 78 N.E., S.E., S.W. I shall assume these maps to be accessible to any one desirous of following the field-geology of the county. The chief horizontal sections are contained in sheets 28 and 31. Since this essay was first written, the whole of the Ordnance Survey of Caernarvonshire on the six-inch scale has been pub- lished, and field-geology in that district is now very much facilitated. Blake, Rev. J. F. On the Igneous Rocks of Llyn Padarn, Yr Eifl, and Boduan. Rep. Brit. Assoc. for 1886, p. 669 (1887). On the Cambrian and Associated Rocks in North-West Caernar- vonshire. Q. J. G. S., vol. XLIV., pp. 271290 (1888). 12 4 INTRODUCTION. On the Monian System of Rocks. Q. J. G. S., vol. XLIV., pp. 463 546 [530534] (1888). Bonney, Prof. T. G. Notes on the Microscopic Structure of some Rocks from Caernar- vonshire and Anglesey. Q. J. G. S., vol. xxxv., pp. 305 308 (1879). On the Serpentine and Associated Rocks of Anglesey ; with a Note on the so-called Serpentine of Porth-dinlleyn, Caernarvonshire. Q. J. G. S., vol. xxxvii., pp. 4050, [4850] (1881). On some Nodular Felsites in the Bala Group of North Wales. Q. J. G. S., vol. XXXVIIL, pp. 289 296, pi. x. (1882). On the so-called Diorite of Little Knott (Cumberland), with further Remarks on the Occurrence of Picrites in Wales. Q. J. G. /3 a. Sometimes its colour is greenish-brown or even green, but this is evidently the beginning of a secondary change. The hornblende, as well as the augite, is often converted into a feebly-polarising chloritoid. In other places, probably in conjunction with the fel- spar, it has given rise to the formation of bundles of epidote crystals [647]. Still another mode of alteration is that which has resulted in the formation of a dirty-brown mica. Quartz is present in all the slides. From the manner in which it replaces decaying felspars and hornblende, it cannot be doubted that much of this mineral is secondary [644] ; but there is always primary quartz in addition. In a few places hexagonal sections are seen [646], but the quartz is usually interstitial relatively to the other constituents. In one slide from Cefn-coed [645] there is a kind of ground-mass composed of a micro-pegmatite of quartz and orthoclase, but this is rather an unusual feature in the rocks. The general structure of the rock, the behaviour of the iron- ores, the ophitic occurrence of the hornblende, the varying re- lations of that mineral to the augite, with the presence of the latter constituent, all conflict with the usual characters of a quartz- syenite, and shew some curious points of resemblance to the hornblende-diabase laccolites of Caernarvonshire and Anglesey to be mentioned below. The date of intrusion of the Llanfaglen syenites can only be conjectured. We know only that they are post-Arenig 1 . I shall, however, refer again to the question of the age of these rocks in connection with the probable sequence of events during the great age of igneous activity in Caernarvonshire. 1 Marr, Q. J. G. S., vol. xxxn., p. 134; 1876: also Ramsay, p. 204. The shales were formerly described as Lingula Flags (Catalogue, p. 40), but no strata of that age have yet been recognised in this part of Caernarvonshire. VI. THE DIABASE SILLS, ETC. THE rocks which next demand notice are the diabases. On looking at a geological map of Caernarvonshire, or indeed of North Wales, one cannot fail to be struck by the remarkable coincidence between the distribution of the ' felstones ' and that of the ' green- stones.' In the eastern division of Caernarvonshire especially and for the moment the Lleyn district may be left out of con- sideration the basic rocks hardly ever occur except in close association with the rhyolitic lavas, and particularly with the later flows. A closer examination, however, reveals the fact pointed out by Sedgwick, and so admirably proved by the work of the Geological Survey, that while the acid rocks form large con- tinuous sheets which have certainly been poured out in the form of true superficial lava-flows, the diabases, on the other hand, have in- variably been intruded among preexisting lavas,, ashes, and sedi- mentary strata, without ever reaching the surface. No evidence of any true lava-flow of basic composition has been discovered in Caernarvonshire 1 . The diabases occur, in fact, in the form of sheets or 'sills' of varying thickness and extent, often following very closely the bedding of the rocks among which they have been injected, but in places breaking across to a new horizon. Occasion- ally the sheets swell into larger masses and assume rather irregular forms, becoming at the same time more clearly transgressive ; again, they may take a general lenticular shape while still occupy- ing a definite horizon between the strata; but as a rule the tabular form is well maintained. The contact-alteration of the beds above and below can often be clearly verified. 1 We reserve the possible exception of the basaltic rocks at Porth-dinlleyn. 76 THE DIABASE SILLS, ETC. It is easy to fix limits to the date of intrusion of these rocks, and the evidence is conclusive that all the diabases associated with the Bala strata of eastern Caernarvonshire were injected during the Bala age itself. It is very noticeable that although these scattered inclusions approach close to the border of the Silurian 1 formations, not one of them penetrates Silurian rocks. Besides this, it is manifest that the intrusions preceded in general the crust-movements which so profoundly affected this district about the close of the Bala age 2 . In many of the sedimentary rocks the bedding-lamination became at that epoch almost com- pletely obliterated by a newly-impressed cleavage-structure; but the diabase sills are always guided by the bedding, never by the cleavage. These sheets have everywhere partaken of the con- tortion of the strata among which they lie, and it is even possible in the case of the larger intrusions to trace the manner in which the folding has been modified by the presence of these stubborn masses of rock among less resisting materials. It may be asserted with some confidence that the latest of the basic sheets were injected almost at the time when the plication of the strata was in progress, so that their form and distribution are partly determined by the folds. Many of the sheets situated at the higher horizons in the eastern part of the county are, in fact, rather to be regarded as somewhat attenuated laccolites lying in the troughs of synclinal folds or sometimes occupying the summit of an anticlinal dome. Numerous instances are seen between Capel Curig and Llanrwst and in the country farther north, about Lakes Cawlyd and Eigiau : their relations are clearly indicated on the Survey maps. The injection of igneous rocks into the cavities which tend to form during the folding of strata is probably a not uncommon phenomenon : it has been described in the Corndon district, Shropshire 3 , and the Leadville district, Colorado 4 , and the laccolites of Wexford and Wicklow 5 are perhaps an instance of the same action. 1 This name is restricted to the strata from the Mayhill to the Ludlow inclusive. 3 Cf. Sbarpe, Q. J. G. S., vol. n., p. 311; 1846. 3 Watts, Eep. Brit. Assoc. for 1886, p. 670. 4 Bulkley, Trans. Amer. Inst. Min. Eng., vol. xin., p. 384; 1885: Neu Jahrb. f. Min., &c. 1888, vol. i., p. 411. 5 Kinahan, G. M., 1881, p. 134. FIELD-RELATIONS. 77 A significant circumstance in connection with the foregoing remarks is the almost complete absence of dykes from the district under consideration. Excluding, as before, the Lleyn peninsula, hardly any instances can be cited of dykes associated with the Bala lavas and the diabase sheets. Dykes of basic material do indeed occur in northern and central Caernarvonshire, but we shall see that these were injected at a much later epoch, and their course determined by joints formed during post-Carboniferous crust-movements. When the Bala strata were injected with the diabasic magma, no such joints existed in the sedimentary rocks, and the igneous material accordingly forced its way along the surfaces of weakness offered by bedding-planes. The few excep- tions strongly confirm this view, for they are almost exclusively confined to the diabases intruded into the highest strata in the Llanrwst district, which intrusions we suppose to have proceeded concurrently with the beginning of post-Bala folding in this area. In Lleyn, too, the diabase sometimes forms dykes cutting nearly perpendicularly through the strata ; as at Llangian and Llanengan near Abersoch, and about a quarter of a mile west of Sarn. A large number of dykes are marked on the map in the green schists at the extremity of the peninsula ; but it is probable that most of these belong to the post-Carboniferous group of intrusions, to be discussed later, and we shall not include them in this place. We also exclude from present consideration certain laccolitic intrusions of hornblende-diabase, which occur in the neighbour- hood of Penarfynydd and Rhiw and near Clynog-fawr. Although free from olivine, the Caernarvonshire diabases are decidedly basic in constitution. In a specimen from Pant Evan near Tremadoc Mr Acton finds 47'4 per cent, of silica, and the percentages of the alkalies (potash 1'5 and soda 3 '9) are quite in accord with the normal basic character of the rock. A little titanic acid is present, but no chromium or manganese. Phos- phoric acid (0'2) is present as apatite, while water (2'8) and carbonic acid (1*6) indicate some amount of secondary change. The specific gravity of this rock is 2'900, which may be taken as an average figure for the diabases of the county. Petrologically the rocks under discussion present few features of novelty. They have attracted little attention, and almost the only published notices of them are those contained in Mr Tawney's 78 THE DIABASE SILLS, ETC. Woodwardian Laboratory Notes, and Mr Teall's British Petro- graphy. A few of the Lleyn rocks have been described by Mr Elsden 1 and by the writer 2 . To the eye the rocks appear as ordinary ophitic diabases stained with the greenish hue which comes from chloritoid and other decomposition-products. Sometimes they are spotted with little black patches, seemingly delessite (well seen between Aberdaron and Pen-y-cil) : and in exposed places the destruction of these spots or of augite grains gives a curious pseud o- vesicular appearance to the weathered surfaces, as may be noticed on Mynydd-y-Rhiw near Sarn and in other places. Only occasionally, e.g. to the north of Pwllheli, do the diabases become visibly por- phyritic. In a few localities there is a local schistose structure due to shearing and internal slipping in the solid rock, as at Rhyd-ddu and some spots near Tremadoc. The jointing is rarely columnar (Clip-y-cilfinhir near Sarn 3 ) ; sometimes platy (Careg-y-rimbill at Pwllheli 4 ) ; rarely cylindrical (Craig-y-fael near Sarn) ; often roughly cuboidal or quite irregular. A very common feature, especially in the Lleyn diabases, is a pronounced spheroidal jointing, accompanied usually by shelly exfoliation. This may be studied in endless variety near Sarn, at Ty-rutten, in all parts of the Mynydd-y-Rhiw mass, and in the sheet at Tyn-y-rhedyn near Llanfaelrhys. The spheroids vary in size from two or three inches to as many feet in diameter. In some localities, as on the beach west of the Mynydd Penar- fynydd headland, they occupy the blocks marked out by three sets of plane joints ; but this is not always the case, and it is not uncommon to find one large spheroid enclosing a number of others like eggs in a nest (Coch-y-foel near Sarn). Wherever the diabases are seen in conjunction with sedi- mentary strata, the contact-alteration of the latter is observable ; but it varies very much in degree and in the distance to which it extends from the actual junction. In the field, indurated slates or shales often assume a rather arenaceous appearance, and have sometimes been erroneously mapped as sandstones, as is the case 1 G.M., 1888, pp. 303308. 2 Q. J. G. 8., voL XLIV., pp. 448450; 1888. 8 Cf. Blake, Q. J. G. S., vol. XLIV., p. 531 ; 1888. 4 Cf. Bonney, Q. J. G. S., vol. xxxn., p. 145; 1876. CONTACT-ALTERATION. 79 in the Tremadoc district. One effect of the alteration has been to prevent the impression of the cleavage-structure on the argillaceous strata. Rocks which are strongly cleaved at a short distance from the intrusion are found to have resisted this modification completely in the immediate neighbourhood of the diabase. This phenomenon is rather different from that seen in the Llanberis slate-zone, when the cleavage has been ob- literated by the effects of the intrusion of later (post-Carboni- ferous) dykes. It is well seen in Bwlch-y-ddeufaen and at other localities between Conwy and Snowdon. A banded appearance is sometimes set up in the baked shales by the development of some alternation of lithological characters not perceptible in the unaltered rock. Some of the best examples of contact-metamorphism in Caernarvonshire are seen in the Tremadoc district, especially under the diabase outlier on Moel-y- gest. Here indistinct darkish spots make their appearance in the gray slates, which, as we approach the diabase, become smaller but more distinct. Still nearer, these spots, presumably incipient crystals of andalusite, seem to be reabsorbed and disappear, the rock becoming porcellanous with a subconchoidal fracture. Mr Teall (Brit. Petr. pp. 217220) has compared altered rocks from this neighbourhood with the spilosite and desmoisite of con- tinental geologists. As a rule, each sheet or mass of diabase is fairly uniform to the eye ; but some shew segregation- or contemporaneous veins, which represent the latest phase in the process of consolidation. The general characters of these veins are, that they are more felspathic and so of lighter colour than the rest of the rock ; they are of coarser grain ; and they frequently have the augite partly idiomorphic, while in the matrix it is ' intersertal ' or usually ophitic. Such veins sometimes occur pretty numerously, and have a roughly parallel or subparallel arrangement. Good ex- amples are seen at Moel-y-gest and Pant-evan near Tremadoc, and in the boss known as the 'Gimlet Rock,' Careg-y-rimbill, at Pwllheli. Examined microscopically 1 , the diabases exhibit as a rule only the essential constituents, being decidedly poor in accessories. Any clear trace of olivine is almost unknown. Mr Teall (p. 215) 1 Sixty-five slides have been studied. 80 THE DIABASE SILLS, ETC. has remarked the absence of olivine as one point of distinction between these rocks and those associated with the Carboniferous and Tertiary strata of Britain ; but of the diabases and dolerites which may be regarded as post-Carboniferous in this part of the county only a small part contain olivine, and a separation on petrological grounds between the two sets of rocks becomes often impossible. The original minerals composing the Caernarvonshire diabases are the usual iron-ores, triclinic felspars, and augite, with apatite in small quantity, possibly olivine in rare examples, and in one or two cases hornblende, biotite, and perhaps quartz. A very characteristic feature in these rocks is the common association of magnetite and ilmenite. The original magnetite usually occurs in grains shewing little regularity of outline, in rods, or in tree-like growths produced by a series of minute rods set at right angles to a larger one which serves as a stem [580, 599, 733]. In several slides the magnetite rods, or perhaps sections of plates, shew a parallel arrangement over a considerable area [732], although the interspaces are filled by later minerals, and often each bar of the series is interrupted by a preexisting crystal of felspar. This peculiarity of orientation, though found in the magnetite of some other rocks 1 , is apparently not met with in other minerals. It is certainly not connected with fluxional movement, and may possibly be a magnetic rather than a crystalline effect. Occasionally magnetite rods appear to be built up from a series of imperfect octahedral crystals. More rarely this mineral, apparently titaniferous or intergrown with ilmenite, forms good skeleton octahedra like that figured by Zirkel from re-fused Mount-Sorrel granite [811]. The ilmenite commonly forms ragged plates, often arranged in parallel sets in a kind of frame-work. A section often shews two or three intersecting sets of parallel bars [731, etc.]. One or other of the two iron-ores is invariably present, and very frequently both may be recognised in the same slide. They do not entirely separate out at one stage of the consolidation, but in most cases the bulk of these minerals is later than the felspars 1 Eeusch, The Microscopic Texture of Basalts from Jan Mayen (from The Norwegian North- Atlantic Expedition, 187678), p. 5 and fig. 4; 1882. Judd, Q. J. G. S., vol. XLII., pi. vi. fig. 7; 1886. See also G. M., 1887, p. 413. MINERALOGICAL CONSTITUTION. 81 and earlier than the augite. We do not refer here to the secondary magnetite which often results from the alteration or destruction of the augite. Apatite occurs in slender needles in many of the diabases, but only locally and capriciously. It cannot be reckoned as an essential component. Sphene, excluding the amorphous leucoxene, has been found in one locality only, at Pant-Evan, Tremadoc. It forms clusters of light-brown grains with high refractive index. By comparison of different slides [576, 577, 578] this mineral appears to have been produced by reaction between the ilmenite and felspar. If this be so, it cannot properly be ranked among the original minerals, being probably connected with the mechanical forces which have deeply affected the rocks of that neighbourhood. The felspar of the Caernarvonshire diabases belongs always to triclinic species, and appears to range from andesine to anorthite, varieties of the andesine-labradorite series being the most usual. In the majority of cases the felspar clearly antedates the augite in the process of consolidation, and builds idiomorphic, though not always well-bounded, crystals from 0'05 to 0'5 inch in length. The crystals are commonly of rather elongated form, but sometimes there are a few of squarer shape. These last often shew pericline twinning in addition to that on the albite plan, but most of the felspars have only the common albite-lamellation. In certain rocks Carlsbad twinning modifies the appearance of the striation in polarised light. Not infrequently some of the lamella project beyond the others, producing a ragged appearance at the termi- nations. Only in a few cases is there any suggestion of the twinning being a secondary structure related to strains in the crystals. In many of the rocks the smaller felspars are only once twinned. A zonary banding is sometimes seen in the felspars, when examined in polarised light. Augite is invariably a plentiful constituent in the diabases. Only rarely does it occur in idiomorphic crystals, shewing in cross- section the prism and pinacoid forms about equally developed, as at Llanrwst [566] and at Pen-y-dre near Aberdaron [598], or the pinacoids only slightly truncated by the prism, as in some diabases north-west of Pwllheli [602], or again rather rounded outlines, as at the top of Carnedd Llewellyn [563]. Sometimes the augite is H. E. 6 82 THE DIABASE SILLS, ETC. partly idiomorphic, presenting crystal contours to the felspar on one side but moulding it on the other, as near Llyn Teyrn [569], Sarn [731], and in the Pen-y-cil sheet near Aberdaron [594, 599]. Most usually the augite is entirely posterior to the felspar, and forms either grains filling the remaining spaces or sub-ophitic or ophitic plates moulding and enclosing the felspar crystals. Twinning on the orthopinacoid is not infrequent, and sometimes twin lamellation is seen [568, 580, 602]. The colour of the augite in thin sections varies from light brown to colourless, and is usually very pale, especially in the Lleyn diabases. Occasionally a crystal or plate of augite is sepa- rated by definite lines into fields of slightly differing optical characters. Such lines may be quite irregular, or may correspond to crystal outlines, as at Deneio near Pwllheli [603]. Sometimes they give rise to a well-developed hour-glass structure. In certain cases this is seen only in polarised light, when a clinopinacoidal section may shew a difference of perhaps 5 between the extinction- angles for the two parts of the crystal, as at Castell Caeron and other places near Sarn [580, 811]. In other specimens there is a difference in natural light between the augite-substance within and without the ' hour-glass,' the former being colourless, while the latter is brown and more or less pleochroic. This is seen at Pen-y-dre, near Aberdaron [598], Deneio, near Pwllheli [603], and Porth-lleiddiad, near Llanfaelrhys [734]. It is to be noticed that the hour-glass structure is rarely seen except in idiomorphic crystals, and that the external part is always more coloured, pre- sumably richer in iron, than the internal. It may be suggested with some reason, that the structure is produced by a crystal being corroded by magmatic resorption, and subsequently built up by an accretion of new substance from the magma, which has become altered in composition by the resorption of some of the earlier iron-ores. The interesting synthetic studies of Famintzin 1 have shewn that an hour-glass form is readily produced in crystals cor- roded by their mother-liquor. The augite of the Caernarvonshire diabases always shews good prismatic cleavage, and rarely others parallel to the two pinacoids, 1 Studien fiber Krystalle und Krystallite, M<*m. Acad. St. Petersb., vol. xxxn., no. 10; 1884: see plate i., figs. 14, 15. MINERALOGICAL CONSTITUTION. 83 as near Pen-y-cil, and Deneio, and in a sheet near Sarn [594, 602, 731]. Very rarely is there any diallagic structure, and that only locally [Deneio, 603]. In one type of diabase the augite is tra- versed by numerous roughly parallel cracks and rows of thickly- set minute inclusions. This is seen in the rocks of Careg-y-rimbill, near Pwllheli, Llanengan, near Abersoch, Careg-dinas, north of Cam Fadryn, and Trefgraig, four miles west of Sarn, rocks which are also united by other characters [136, 582, 135, 592]. The ordinary decomposition-products of the augite will be noted below. The augite in the diabases very rarely shews any trace of bodily conversion into hornblende ; but a fringing growth of amphibole bordering the augite, traversing it in strings, and extending into the place of decomposed felspar and possibly olivine, is an exceedingly common feature in the diabases of the eastern half of Caernarvonshire. This secondary amphibole is always in crystalline relation with the augite on which it grows. It is colourless or rarely pale green 1 , and has commonly a fibrous structure [566, 568, 570, 571, 573, 687]. The Lleyn diabases never have this 'secondary enlargement' of augite by hornblende. At Pant-evan, near Tremadoc, a pale green amphibole with similar characters and relations forms not only fibrous fringes, but round compact patches with hornblende-cleavage, partly filling cavities in the augite which also contain serpentine [576]. A few scraps of brown hornblende and brown biotite associated with the augite may probably be reckoned as original. They are seen occasionally in the Careg-y-rimbill type of diabase already alluded to, with which perhaps we may include the occurrences near Tremadoc [136, 583, 135, 577]. Only one slide, from Llyn- yr-afon under YFoel Fras [855], shews a few small idiomorphic crystals of pale brown hornblende, of earlier formation than the augite. The general structure of the rocks shews certain varieties, but as a rule the whole of the felspar has been of earlier consolidation than the augite, and the various types of 'diabasic' structure result. There is every gradation, from large ophitic plates of augite including many felspar crystals [567, 136, 576, 580] to small ophitic patches just moulding the felspars [602, 731], and augite grains wedged into the interspaces of the earlier-formed 1 Very rarely brown [855]. 62 84 THE DIABASE SILLS, ETC. minerals [728, 600, 589]; but the ophitic structure is more common than the so-called ' granulitic.' Any approach to the gabbro-type is rare, but there are some diabases in which the felspar and augite seem to have crystallised concurrently, and both have more or less idiomorphic contours, as at Llyn Teyrn, and in some sheets near Sarn [769, 731]. Again, the augite may occur in shapeless grains, and yet penetrate, as well as being penetrated by the felspar prisms, as between Pen-y-cil and Aberdaron [599]. A rare variety is one near Llanrwst, in which the augite occurs in two distinct generations, viz. in long twinned prisms older than the felspar and in ophitic plates moulding that mineral [566]. One or two specimens come near to the basaltic type : a rock from the summit of Carnedd Llewellyn, for example, has rather rounded crystals of twinned augite, with felspars in smaller number, im- bedded in a fine-grained ground which is now much decomposed [563]. The numerous decomposition-products met with in the weathered diabases of Caernarvonshire will be only briefly noticed. The ilmenite and perhaps titaniferous magnetite give rise invariably to the usual leucoxene, clinging in cloudy semi-opaque masses to the remains of the iron ore. In some of the diabases from the eastern half of the county the leucoxene becomes light-brown and slightly doubly refracting, but distinctly granular sphene has not been found except in the Pant-evan rock, as remarked above. The pale secondary amphibole already described seems to come less from the augite than from the destruction of the other con- stituents. A fibrous pilitic-looking substance forms felted masses in some of the round pseudomorphs which may represent olivine, e.g. on Moel Hebog [565], A bastite-like material at Pant-evan, etc., may possibly come from a rhombic pyroxene, and, if so, the association of the minerals would indicate the same source for the compact variety of the pale secondary amphibole, which would thus seem to pseudomorph an enstatite mineral originally inter- grown in the augite plates [576]. The pale green decomposition-products so prominent in many of the slides are doubtless to be referred to more than one mineral of the chlorite and chloritoid (or saponite) families, besides a little serpentine. Professor Heddle 1 , who has devoted much attention 1 Trans. Eoy. Soc. Edinb., vol. xxix., pp. 55118; 1879. SECONDARY MINERALS. 85 to the 'chloritic minerals,' divides them into the 'chlorites' and the ' saponites/ and makes it appear that the ordinary decomposi- tion-products of augite, hornblende, etc., belong to the latter division. The four chief species of this division are celadonite (bright green), delessite (dark green), chlorophaeite (usually black), and saponite (translucent). The first two have a comparatively low content of water, with a high specific gravity ; the last two a high percentage of water and a low specific gravity. All are dissolved or destroyed by hydrochloric acid except celadonite, which in this respect resembles the chlorites proper. The sub- stances met with in the Welsh diabases seem, from the criteria here cited to be delessite, and perhaps sometimes chlorophaeite. Unfortunately the optical properties of the several species are not thoroughly known, and indeed small reliance can be placed on statements, unless chemical and optical examinations have been made on the same specimens. The common ' viridite ' of our diabases has usually a low refractive index and very slight double- refraction, with varying degrees of colour and pleochroism in thin sections. These, according to Rosenbusch 1 , are the characters of delessite. Sometimes, however, and especially in the radial- fibrous aggregates, the double refraction is much more marked. The vermicular structure is rare [Pant-evan, 578]. These green decomposition-products often form evident pseudo- morphs after the augite, especially after the ophitic plates ; but sometimes a sensibly isotropic substance, probably the same, impregnates the rock generally, and is even injected into the cleavage-cracks of the felspar, as in the Pen-y-cil sheet near Aberdaron [594]. Doubtless the production of this substance from the augite was attended by an increase of volume. Serpentine occurs in many of the slides, either mingled with the presumed delessite or alone. In the Deneio rock are numerous reticulating veins of fibrous chrysotile, the fibres set perpendicular to the walls of the fissure [603]. The rock tends to break under the hammer along these serpentinous veins, shewing a beautiful velvet-like surface. Magnetite dust is not infrequently produced by the alteration of the augite in all the diabases, and is seen either in the augite or among its decomposition-products. The 1 Mikr. Phys. d. petr. wichtig. Miner., p. 560 (2nd ed.), 1885. 86 THE DIABASE SILLS, ETC. separation of this dust seems to be quite an early stage in the decomposition of the augite. Epidote in brightly polarizing grains is a frequent secondary product in the diabases of Caernarvonshire 1 , although restricted to the eastern half of the county. It appears to be connected with the felspar rather than the augite, and sometimes builds more or less perfect pseudomorphs after the felspar crystals [568 571 and 856]. In other cases the distribution of the grains is less regular ; and when the epidote is present in large quantity, it forms large fan -like bundles of prisms. Calcite is common in dust or fine granules : only in the more weathered rocks does it build larger grains, and it is then sometimes accompanied by secondary quartz with fluid-cavities enclosing moving bubbles. Some of the diabases of Deneio and a few other localities have fan-shaped aggregates of a zeolitic mineral with high polarization- tints [602]. Among clearly introduced constituents may be men- tioned pyrites, sometimes in small cubic crystals, as at Tremadoc and at Pen-y-dre, near Aberdaron. In a few cases only has the diabase undergone so much alteration as to be completely disintegrated. This is the case north of Coch-y-foel, near Sarn, where the rock is converted into a soft brown ochreous substance. In the extreme west of the Lleyn peninsula occur certain basic rocks of volcanic habitus, which are too much decomposed to admit of profitable examination, but which we may fairly refer to the basaltic family. They are associated with the doubtful series of green schists, etc., which stretch in a broad band from Porth- dinlleyn, near Nevin, to Braichy-y-pwll and Bardsey. This re- markable series of rocks, the ' altered Cambrian ' of the Geological Surveyors and the ' Pebidian ' of Dr Hicks, is frequently schistose and often closely contorted ; usually green from chloritic sub- stances ; in some places quartzose, in others felspathic, occasionally micaceous, often including streaks and nodules of red jasper. It not infrequently includes calcareous beds, and even more or less 1 Cf. Cole and Jennings on the Cader Idris diabases, which, however, appear to be less basic than those under discussion. Q. J. G. S., vol. XLV., p. 432; 1889. THE BASALTS OF LLEYN. 87 crystalline limestones and ophicalcites, and is intersected by count- less dykes well exposed on the sea-shore. The stratigraphical position of these rocks cannot be discussed here. They exhibit, however, the closest resemblance to those exposed in various parts of Anglesey, particularly near Menai Bridge, Holyhead, and Am- Iwch. These Anglesey strata Professor Hughes 1 has proposed to refer, at least in part, to the Bala series ; and Professor Blake, while placing them beneath the Cambrian system, admits their continuity with the conglomerates at Porth-wen on the north coast, which have yielded Orthis Bailyana! Within the tract mainly occupied by these rocks in western Lleyn a number of small patches are coloured on the Survey maps as ' serpentine.' As remarked by Professor Bonney", Mr Elsden, and Professor Blake 3 , they have no real claim to this title. In part they are diabases, and have been included among the rocks described above; while the remainder are much decomposed basalts. The 'serpentines' of Methlan, Ty-hen, Hendrefor, and Trefgraig [592] are diabases ; those of Porth-dinlleyn and Careg- fawr are basalts, although a diabasic rock is associated with the occurrence in the former locality. In appearance these weathered basalts are compact, brittle rocks of green to purplish red colour, veined with calcite, quartz, epidote, and sometimes asbestos and chrysotile. In the Careg- fawr patch near Aberdaron there are abundant veins lined with epidote and containing asbestos : the former mineral grows per- pendicular, the latter parallel to the walls of the fissure. In thin sections from either locality [591, 612, 613] we see a decomposed mass, in which small felspar prisms and sometimes augite granules are still recognisable. There appears to have been an isotropic base, and small curving cracks, occasionally outlined in secondary magnetite dust, may represent a perlitic structure. Larger cracks, producing a brecciated appearance in the slides, are filled either with calcite or with a mixture of serpentine and a chloritoid ; and it is the fracture of hand-speci- mens along these cracks which gives a false serpentinous aspect to the rock in the field. The deception is heightened by the 1 Proc. Camb. Phil. Soc., vol. in., pp. 341348; 1880. 2 Q. J. G. S., vol. xxxvn., p. 48; 1881. 3 Q. J. O. S., vol. XLIV., p. 531 ; 1888. 88 THE DIABASE SILLS, ETC. mottled green and red appearance of the rock, due to some capricious oxidation of the contained iron-compounds. Professor Bonney described the basalt at Porth-dinlleyn as associated with volcanic agglomerate and breccia *; but he informs me that he did not regard this as evidence that the volcanic rocks were contemporaneous with the green schist group, the appearances being more suggestive of a ' neck,' like those of Fifeshire. The basalt at Careg-fawr, where the relations are more evident, is clearly intruded through the surrounding rocks. The age of these latter being so dubious, we cannot pretend to fix the date of the basalts, and the above notice of the existence of such rocks in a couple of small exposures will suffice. 1 See also Harrison, I. c. VII. OTHER BASIC INTRUSIONS. AMONG the smaller eruptive masses of Lleyn, two intrusions of basic rocks, breaking through the granite of the Sarn district, are deserving of notice 1 . The localities are Craig-y-fael, two miles south-west of Sarn, and Plas Llangwnadl at the same distance west-by-north of that village, the rocks being quarried in both places. (See map, fig. 4.) In the Craig-y-fael quarries the rock exposed shews a gabbro- like structure, and appears to consist of darkish grass-green bisilicates with a rather sub-metallic or diallagic lustre and light- green altered felspars. These two constituents are present in about equal proportions, and neither shews any idiomorphic contour. At the bridge above Plas Llangwnadl, on the main- road from Nevin to Aberdaron, the exposures are very similar, though sometimes with apparently less felspar in the rock. Tracing the mass down the stream, past the Plas, the character seems to vary, with some tendency to a gneissose structure. One variety with large white felspars has to the eye a suggestion of the aw/en-structure. Going on towards Llangwnadl itself, the rocks become distinctly schistose; and about a hundred yards before reaching the church they suddenly disappear, doubtless against the same fault which limits the granite in this direction, being succeeded by schistose, agglomeratic-looking rocks belonging appa- rently to the green schist series, and shewing evidence of fault- brecciation. The Craig-y-fael type may be described as a partially am- phibolised gabbro. The minerals seen in a thin section are augite and diallage, hornblende and actinolite, felspar and opaque 1 See Q. J. G. S., vol. XLIV., pp. 447, 448; 1888. 90 OTHER BASIC INTRUSIONS. iron-ore, with some pale-green decomposition-product which has the characters of a chloritoid [706], The felspar is a triclinic one with the usual albite-twinning, sometimes crossed by pericline lamellae, and the extinction angles seem to point to labradorite. The crystalline plates are much strained and bent, the lamellae being curved. It is possible, though not quite clear from the specimens, that the twinning may be in part induced by the strain. The augite is in long plates, which exhibit sometimes the ordinary prismatic cleavage, sometimes a distinct diallagic struc- ture. They have a slight greenish tint, and, apart from the ordinary decomposition, are seen to pass at the margin and in irregular veins and patches into a dull green hornblende. The latter mineral gives the absorption-formula a. very pale to almost colourless ; $. rather pale grass-green ; 7. a rather deeper and bluer green. Some of this hornblende is uralitic in appearance, but there are more compact portions shewing the prismatic cleavage ; and, in addition, granular patches and imperfect blades of actinolite are met with. There appears to be no augite or diallage remaining in the Llangwnadl rocks, but their essential identity with the partially amphibolised gabbro of Craig-y-fael can scarcely be doubted after an examination of the specimens and slides. The alteration of the original type has evidently proceeded farther near the margin of the district than in the central part, and the resulting rocks are such as would be termed epidiorites by some petrologists. The massive type of these completely amphibolised rocks is illustrated in the quarries near the bridge [707]. The microscope shews but little of the black iron-ore, its place being occupied by granules of brownish sphene. These granules, with cleavage- traces and more irregular fissures, are aggregated in small patches and strings between the grains of felspar and hornblende, but they seem to be clearly connected with the diminishing iron-ore, and may be taken to prove that the original mineral was a titaniferous one. Possibly some part of the iron-oxide has been absorbed by the hornblende, which here presents a greenish-brown to brownish- green colour, with pleochroism : AMPHIBOLISED ROCKS OF LLANGWNADL. 91 a. very pale to almost colourless : /S. rather deep brown with a greenish tinge ; 7. a slightly deeper tint of greenish brown : the absorption being, as before, 7 > /3 a. The felspar is partly in granular patches, but too much decomposed to shew any structure. Near Plas Llangwnadl the rock is still more altered, and has in places the quasi-porphyritic aspect already noted, which is due to granular felspar forming large patches. There is but little opaque iron-ore, and sphene is more abundant. The felspar, where it has preserved something of its original structure, appears to be labradorite [708]. Still further down the stream, approaching the boundary of the area, the schistose structure becomes more pronounced, and the rock resembles the ' amphibolites ' of some Continental geo- logists. A slide from this portion shews no new characters, except that the iron-ore has entirely disappeared [709]. Again, agneissic structure is sometimes to be seen, and this rock is illustrated by another slide, the material, however, not obtained in place [710]. The mode of occurrence of the rocks in both the small masses proves them to have been intruded through the granite, doubtless consolidating in the form of gabbro. The various mineralogical changes must be referred to mechanical influences, and indeed are well-known results of dynamic metamorphism. The manner in which these modifications increase towards the boundary of the northerly intrusion, and are increasingly attended by structural re-arrangement of the mass, clearly point to a relation between these changes and the forces which produced the Llangwnadl fault. The strike of the schistose structure in the amphibolite- like and gneissic rocks is parallel to the fault, as is also the local schistosity in the rocks on the other side of the dislocation. As regards the age of the intrusions, it can only be said that they are posterior to the granite and anterior to the fault, which leaves the upper limit of age rather indefinite. No post-Silurian or post-Carboniferous crust-movements appear to have been ac- companied in North Wales by structural metamorphism, and so 92 OTHER BASIC INTRUSIONS. the gabbros are most probably assigned to the Bala age, which we know to have been closed by extensive movements of the earth's crust in the Caernarvonshire area. Professor Blake 1 includes the rocks in question with the granite of Cefn-amwlch, etc., but has withdrawn his former opinion assigning the whole to his 'Monian' system. The gabbro and its derivative rocks are, however, as the above description indicates, thoroughly basic rocks, while the granite is always highly acid ; and there is no appearance of passage between the two. In the southern part of the Sarn district, an area remarkable for the variety of its petrological types, occurs a considerable development of coarsely crystalline hornblendic rocks, differing entirely from the neighbouring diabase. I have elsewhere de- scribed these rocks under the title hornblende-diabase, and given some account of their mode of occurrence 2 . They were first noticed by Mr Tawney 3 , and Mr Elsden * has also remarked upon them. They occur in three apparently distinct masses, each having a roughly oval outline on the map (fig. 4), with its long axis bearing N.N.E.-S.S.W., that is parallel to the local strike of the strata. The mode of occurrence of these masses and the contact-alteration of the adjacent shales attest their truly in- trusive nature. The first forms the hill Mynydd Penarfynydd, and has a length of about three-quarters of a mile. The second builds Careg-llefain and Mynydd-y-graig, being rather larger than the preceding. The third, less freely exposed, lies to the east of Rhiw, and extends from Tyn-y-borth to Tyddyn-y-corn, a distance of more than a mile. The Mynydd Penarfynydd mass is underlain by a considerable thickness of hornblende-picrite, exposed on the west and south- west flanks of the hill, which is some 360 feet in height, and the whole, with a small thickness of hornblende-diabase at the bottom, 1 Q. J. G. S., vol. XLIV., pp. 531, 532 ; 1888. 2 Q. J. G. S., vol. XLIV., pp. 450459; 1888. 3 G. M. 1880, pp. 207213. 4 G. M. 1888, p. 305. HORNBLENDE-DIABASES. 93 rests on shales, in which Mr Tawney 1 found Upper Arenig fossils at Penarfynydd farm. A comparison of the cliff-sections with the outcrops in the hill shews that the intrusion is to be regarded as a laccolite, probably more than 1000 feet thick, injected be- tween the strata. The second mass rises to more than 700 feet in the serrate ridge of Mynydd-y-graig, but unfortunately neither this intrusion nor the one near Rhiw, which occupies lower ground, shews any clear junction with the adjacent rocks. They may be laccolites of larger size than that of Penarfynydd and injected at a rather higher horizon. The manner in which all these rocks have modified the folding and cleavage of the district conclusively proves them to be of Bala age. In hand-specimens the rocks are always visibly crystalline, of medium to rather coarse grain, and shew a mixture of hornblende and felspar ; but the variation in colour and general appearance is so great that no single description will serve for the whole of the rocks. As in some other hornblende-diabases, and notably those of central Anglesey, the diversity of appearance is due to the varieties and varying relations of the hornblende rather than to any essential differences in constitution among the rocks them- selves. The structure is normally ' hypidiomorphic', though in one or two localities (north of Rhiw Church) the rock contains more or less well-formed crystals of hornblende some two inches in length. Mr Acton's analysis shews 44'9 per cent, of silica, thus placing the rock naturally between the ordinary diabases (4 7 '4 per cent.} and the hornblende-picrite (41 - 8). In many respects the horn- blende-diabase is allied to the latter, but differs from it mineralogi- cally in the absence of olivine and the presence of plenty of felspar. There is a considerable quantity of alkalies, the percentage of pot- ash being 21, and of soda 3'6. A loss of 1*5 per cent, on ignition at red heat represents the water in the serpentinous and chloritoid decomposition-prod acts. An examination of slides of the hornblende-diabase from various localities shews, among the original minerals, apatite, magnetite, picotite, ilmenite, felspar, olivine, augite, and horn- blende. Apatite occurs only rarely in rather large prisms of a pinkish 1 G. M. 1880, p. 211. 94 OTHER BASIC INTRUSIONS. tint with transverse jointing (east flank of Mynydd Penarfynydd [713]). Magnetite, in grains of primary origin, is frequent, and sometimes abundant. One specimen (west of YGraig-ddu [716]) has deep brown rounded grains of picotite ; another (north of Rhiw [719]) has ilmenite in skeletons of intersecting rods, with grey leucoxene ; but in the other slides magnetite is the only iron- ore present. The above minerals are the first products of con- solidation. Olivine is not actually found in any of the slides, but in one or two some serpentinous pseudomorphs included in the hornblende seem to represent that mineral. The specimens are from near Ffynnon-cefn, north of Rhiw [718], and from the top of Mynydd- y-graig [715]. Augite occurs sometimes in idiomorphic crystals, but usually in ophitic plates of varying extent or in shapeless crystalline grains. The forms seen in cross-sections are, as usual, the pinacoids truncated by the prism [718, 712]. The prismatic cleavage is constant, pinacoidal cleavages rare ; while the diallagic structure is never found. Twinning on the orthopinacoid is not infrequent. The colour of the mineral is very pale brown to colourless : in one specimen from the top of Mynydd-y-graig [715] this passes into violet-brown with slight pleochroism. The hour- glass structure is occasionally met with, as near Ffynnon-cefn [718]. The extinction-angle in a clinopinacoidal section is 39 or 40. Hornblende in these rocks is both an original and a se- condary mineral. It very rarely shews idiomorphic outlines, viz. the prism and clinopinacoid without terminal planes [719]. The mineral almost always occurs in ophitic plates or in patches or borders associated with the augite. Twinning is rarely seen : it is on the common law, the orthopinacoid being twin-plane and plane of composition. Whether primary or secondary, the horn- blende is normally compact and shews the prismatic cleavage well marked : a fibrous structure is only occasionally met with. Both original and derivative hornblende are usually of a fine brown colour, giving a, deep chestnut brown ; ft, a less deep tone ; 7, pale brown : RELATIONS OF THE HORNBLENDE. 95 absorption 7 > /3 a : fS coincides with b, and the extinction-angle 07 is a rather large one, measuring 18 or 20. The colour some- times passes into greenish-brown or green. As regards its mode of growth and relation to the augite, the ordinary brown hornblende shews some interesting differences. A careful examination of the sections enables us to discriminate four different cases 1 . (i) Original idiomorphic hornblende, which is only rarely met with. (ii) 'Complementary' hornblende, forming a border to original augite, the marginal hornblende of each plate or crystal being in crystallographic relation to the augite nucleus; i.e. the two minerals have the vertical axis and the plane of symmetry common. This must be regarded as an original ' intergrowth ' of augite and hornblende, in which, however, the hornblende is entirely posterior to the augite [716]. This relation is not un- common in hornblende-diabases (proterobases) from other localities, and is most beautifully exhibited in one from Delancy Hill, Guernsey [431]. In the rocks under discussion it is less obvious, but very common. In one case [717] both hornblende and augite are twinned, and the twin-planes of margin and nucleus, though of course parallel, are not coincident, as is the case when the horn- blende is pseudomorphic after augite. (iii) Ophitic hornblende, which forms plates moulding and including the felspars, magnetite-granules, and grains of augite. Sometimes even well-formed augite crystals are thus enveloped by hornblende, as at the trigonometrical station on Mynydd Penar- fynydd [712]. When augite grains are included, their boundaries are comparatively smooth and rounded, and those within the same plate of hornblende have different orientations, there being no crystallographic relation between the augite and the investing hornblende. (iv) Pseudomorphic (' paramorphic ') hornblende, formed at the expense of augite, portions of which often remain unaltered. Here, as in the complementary hornblende, the two minerals have the vertical axis (c) and the plane of symmetry in common ; but the line of demarcation is an exceedingly irregular and intricate one. The change in the augite begins sometimes at the margin, 1 Cf. also Min. Mag., vol. vin., pp. 3033; 1888. 96 OTHER BASIC INTRUSIONS. sometimes along cleavage-planes, sometimes along planes parallel to the pinacoids, or again in quite capricious patches. In some of the slides every stage of the process can be studied, from grains or idiomorphic crystals of augite with small patches of hornblende, mostly near the margin, to plates of hornblende enclosing mere shreds of augite. Here all the patches of augite within one horn- blende plate have, of course, one orientation, being relics of one original crystal. This pseudomorphic hornblende is doubtless for the most part truly secondary in the ordinary sense of the word. In one or two examples, on the other hand, there are appearances which suggest that the amphibolisation may have begun before the final solidifi- cation of the rock ; as, e.g., when a grain of augite is seen partly pseudomorphed by hornblende which is in crystalline continuity with hornblende moulding the grain [714]. The pseudomorphic hornblende in these rocks, though some- times of a rather paler brown, is usually indistinguishable in colour and structure from the original hornblende in the same slides. Another kind of secondary hornblende has different characters, and is probably of later formation. It is seen as a colourless or green extension in crystalline continuity with older hornblende, but occupying the place of other minerals which have perished. This is seen in a number of sections [712, 717, 718]; but the secondary fibrous amphibole fringing plates of augite, which is so constant a feature in the diabases of eastern Caernarvonshire is absent in these hornblende-diabases. The brown hornblende occasionally passes over into uralite, as on the south-east slopes of Mynydd-y-graig, near Y Graig-ddu [716], or into actinolite, as on the east flank of Mynydd Penar- fynydd [713]. Felspar is always an abundant constituent of the hornblende- diabases, though often much affected by decomposition. It belongs to basic varieties between labradorite and anorthite, or to the latter species, as shewn by the wide extinction-angles [717]. The crystals are sometimes simple, generally twinned once or twice, and in some cases finely lamellated : pericline twinning is, how- ever, very rarely seen. In almost every case the felspar is of earlier consolidation than the bisilicates, which frequently mould it in ophitic fashion. HORNBLENDE PICRITE OF PENARFYNYDD. 97 The decomposition-products of these rocks call for no special notice. The augite and hornblende often pass into a substance of the chloritoid family (delessite ?), but both minerals also give rise to serpentine. The augite is also converted into hornblende, as described above, while in some cases fibrous uralite, blade-like actinolite, or a dark micaceous mineral result from the alteration of the compact brown hornblende. Calcite dust usually accom- panies the decomposition of the felspars, and secondary quartz occurs in a few cases. One or two slides [719] have fan-like bundles of a zeolitic mineral polarizing in colours of a high order. The separation of magnetite dust often marks the first stage of alteration in the brown hornblende. It will be seen that these hornblendic rocks belong to a type widely different from the ordinary diabases of the county, which do not contain original hornblende. A rather different type of hornblende-diabase, with olivine, will be described below, but we shall first discuss the ultra-basic rocks, also hornblendic, which occur in close association with the Penarfynydd hornblende- diabase. The hornblende-picrite of Mynydd Penarfynydd forms the basal portion of the laccolite already mentioned. The rock occurs in a succession of parallel banks or quasi-strata to a total thickness of 200 or 250 feet. Owing to a steady dip of 40 to S. 30 E., it is exposed only on the west and south-west slopes of the hill, being succeeded above by the hornblende-diabase which builds the bulk of the laccolite, while there is a thin band of a rather different hornblende-diabase [711] below, in contact with the sub- jacent Arenig shales 1 . (See map, fig. 4.) The rock may be regarded as the type of hornblende-picrite. It was first described by Professor Bonney 2 , and studied in the field by Mr Tawney 8 , who was doubtfully of opinion that the rock is intrusive in the neighbouring hornblende-diabase. An examination of their junction below the trigonometrical station on 1 For further description, see Q. J. G. S., vol. XLIV., pp. 454459; 1888. 2 G. M., 1880, p. 208: Q. J. G. S., vol. XLI., p. 517; 1885. 3 G. M., 1880, pp. 208211. H.E. 7 98 OTHER BASIC INTRUSIONS. Mynydd Penarfynydd shews, however, that although there is no very gradual transition between the two rocks, they are evidently in close connection, and indeed have segregation-veins passing from one to the other. These veins are of coarser grain than the normal type of the rock, and more felspathic : unlike either of the normal rocks, they frequently contain well-formed crystals of hornblende with terminal faces : they have no olivine. Similar veins occur in the heart of the hornblende-picrite itself, and have usually a direction roughly parallel to the quasi-strata already alluded to. They are doubtless in a general sense contemporaneous veins, though representing the latest phase in the consolidation of the magma. They are most abundant in the upper part of the picrite. In the field and in hand-specimens the hornblende-picrite has a very striking appearance. The most prominent mineral is horn- blende, the lustrous cleavage-planes of which are seen to be studded with dull round spots representing grains of partially serpentinised olivine. One common variety contains a golden- brown mica, the flakes lying mostly along the cleavage-planes of the hornblende. White felspar crystals are abundant in some parts of the rock, but wanting or almost wanting in others. (It will be remembered that Tschermak's 1 picrite contained felspar up to about 25 per cent.} The several varieties form clearly marked banks like beds, all parallel to the plane base of the laccolite and to the stratification of the shales beneath. The bedded aspect is very remarkable in the field, being strongly brought out by differential weathering. There are two main types of structure: the first, most noticeable in the felspathic variety of the rock, exhibits a partial separation into patches of the component minerals, which produces a mottled or spotted appear- ance on smooth faces, as in the boulders on the beach, and a pitted or honey-combed effect on a weathered surface. The second type, seen when the constituents are more evenly distributed, shews a compact appearance, with usually a smoother weathered surface, though there is often a grooved or fluted aspect, due to the alternation of bands rich and poor in olivine, and even thin seams composed mainly of the decomposition-products of that 1 Sitz. k. k. Akad. Wien, vol. LIH., (1), p. 260; 1866. DESCRIPTION OF THE HORNBLENDE-PICRITE. 99 mineral. Somewhat similar alternations of more and less basic layers of rock are described by Keusch in the ' saussurite-gabbros ' of the Bergen district 1 . In Mynydd Penarfynydd the mottled or honeycombed type of structure is seen chiefly in the middle half of the mass, though it contains bands and layers of the compact and fluted varieties. Mr Acton has kindly analysed the hornblende-picrite (a fels- pathic variety [725]). Its ultra-basic nature is shewn by the silica-percentage, 41 '8, which corresponds very closely with Fuchs' analysis of the Schriesheim hornblende-picrite 2 , (41'44 per cent.}, a rock presenting close resemblance to the Caernarvonshire one. No titanic acid, chromium, or manganese was found, and only a trace of phosphoric acid. The percentages of the alkalies are, as might be expected, very low, viz. 0'2 of potash and 0'5 of soda: in the Schriesheim rock they are 0'93 and 0'24 respectively. Owing chiefly to the quantity of serpentine present, there is a loss of 3'6 per cent, on ignition at a red heat (Schriesheim 5 '60). The Penarfynydd hornblende-picrite is rather more basic than that of Professor Bonney from Ty-croes in Anglesey, in which Mr J. A. Phillips 3 found 42'94 and 4279 per cent, of silica. The microscope reveals magnetite, olivine, felspar, augite, horn- blende, and brown mica, besides serpentine, actinolite, asbestos, a chloritoid substance, calcite, dolomite, and other secondary products. Here, as in many other basic and ultrabasic plutonic rocks, original magnetite is not common ; but good crystals sometimes occur as the first-formed constituent [723]. As usual, granular secondary magnetite is abundant as an alteration-product of the olivine. Olivine is always present in force, rarely presenting an idio- morphic outline, but commonly in rounded grains imbedded in augite or hornblende. It sometimes has minute, flat, rectangular cavities or ' negative crystals ', containing dendritic growths of magnetite of the kind referred by Professor Judd 4 to secondary 1 Silurfossiler og pressede Konglomerater, Christiania, 1882: (trans. Fossilien- fiihrenden kryst. schiefer, Leipzig, 1883). 2 Neu. Jahrb. f. Hin. etc., 1864, p. 326. 3 Q. J. G. S., vol. xxxix., p. 256 ; 1883. 4 Q. J. G. S., vol. XLI., p. 385 and pi. xii., fig. 5; 1885. 72 100 OTHER BASIC INTRUSIONS. schillerisation [726]. The fissures which traverse the olivine- grains rarely follow any definite cleavage-direction. So far as I have noticed in this and other rocks, the most regular cleavage is found in those olivines which shew the most perfect crystal forms [721]. It is along the cracks that the process of serpentinisation, which may be observed in all its stages, first takes effect [720, 726, etc.] Felspar, when present in the rock, stands next in the order of crystallization. It occurs either in small slender crystals, or in irregular plates moulding the olivine. Its extinction-angles agree with those of anorthite. The augite is very pale brown or almost colourless, with pro- nounced prismatic cleavage. It shews in rare cases an octagonal cross-section, but is almost always in irregularly- shaped plates. The hornblende has the rich brown colour and the other optical characters already noticed in the hornblende-diabases. Occasion- ally it passes into a green variety giving for vibrations parallel to a, very pale brown ; /3, pale olive-green ; 7, rather pale grass-green. It also passes into a colourless variety. The usual prismatic cleavage is in a few cases supplemented by one parallel to the clinopinacoid. The hornblende never shews idiomorphic contours, and is usually in close connection with augite, which it includes or borders. In each plate the two minerals have the c-axis and the plane of symmetry in common, and a section parallel to this plane gives extinction-angles of 20 for the hornblende and 40 for the augite on the same side of the vertical axis. It seems probable that much of the hornblende is pseudomorphic after augite, the boundary between the two being commonly very intricate and ragged [720, 722, 725]; but it is also possible that there is some original intergrowth of the two minerals, as already remarked in the case of the hornblende-diabases. Again there is original hornblende which seems to be quite independent of augite [723]. The brown mica has the characters of biotite, including the usual absorption and dichroism, but it often becomes paler. It is in part clearly original, being then later than the felspar and COMPARISON WITH OTHER PICRITES. 101 earlier than the hornblende. Another portion, however, is pro- duced at the expense of the hornblende, being sometimes parallel to the orthopinacoid, but more commonly lying on the prismatic cleavage-planes of that mineral. The alteration of the hornblende-picrite presents no peculiar features. The olivine, quite fresh in many of the specimens, shews, when a number of slides are compared, the ordinary series of changes : first the separation of granular magnetite in strings following the cleavage or other less regular fissures; next the production of fibrous, unaxial serpentine on the walls of these fissures; then the gradual encroachment of the serpentine net- work upon the meshes of olivine, the fibrous structure being wanting in this later serpentine, which has a confused arrange- ment, and is often sensibly isotropic ; and finally in many cases a resorption of the iron-oxide, resulting in a green coloration of the serpentinous pseudomorph 1 . The other constituents of the rock give rise to the products already noted in the case of the hornblende-diabases. The comparison of the Mynydd Penarfynydd rock with other picrites has already been made by Professor Bonney. It resembles perhaps most closely the well-known type from Schriesheim near Heidelberg [169], although in the latter the hornblende is of a much paler colour and the mica quite bleached, while the olivine, in the specimens I have seen, is more decomposed. The rocks from Ty-croes in Anglesey and Little Knott in Cumberland [170] have less olivine than the one described above. In the St David's boulder [439], noticed and described by Professor Bonney, the olivine and augite are more decidedly idiomorphic than in our type. None of these seem to have such a variety of relation between the pyroxene and amphibole as the Penarfynydd speci- mens exhibit; but they shew many minor points in common, such as the secondary enlargement of the hornblende, the gradual passage of the brown variety into green or colourless, and the bleaching of the biotite. Of the Peekskill hornblende-picrite, 'hudsonite' of Cohen 2 , and 'hornblende-peridotite' or 'cortlandtite' of Professor G. H. Williams 3 , who has fully described its characters, 1 Cf. Wadsworth, Lithological Studies, Cambridge (Mass.), 1884. 2 Neu. Jahrb.f. Hin. etc., 1885, vol. n., p. 242. 8 Amer. Journ. Sci. (3), vol. xxx., p. 29; 1886. LIBRARY UNIVERSITY OF CALIFORNIA SANTA BARBARA 102 OTHER BASIC INTRUSIONS. I have seen no specimens. The same remark applies to the Gipps'-land picrite, the green hornblende of which is considered by Professor Bonney 1 to be largely secondary after pyroxene, and to the Cornish picrites. The Inchcolm rock [820] seems to differ 2 in many respects from the Caernarvonshire one, and the Shrop- shire picrite 3 represents a widely divergent type. A well-marked rock-type, though with some characteristic variations, is seen in the hornblendic rocks which occur in the neighbourhood of Clynog-fawr. They form two intrusive masses. One builds a low ridge running for about 300 yards, in the direction of strike of the adjacent rocks, from the farm of Pen-y-rhiwiau 4 nearly to the cliff. The other crosses the road eastward from Clynog-fawr, just east of the word 'Tanrallt' on the Survey map. Their intrusive nature is proved by the contact-alteration of the neighbouring slates, which, as is often the case, assume to the eye a more arenaceous aspect near the junction. These slates are probably referable to the Arenig stage, though doubtfully, in the absence of fossils. The association of these coarse-grained horn- blendic rocks with Arenig strata is seen not only here, but at Penarfynydd, numerous localities near Llanerchymedd and Am- Iwch in Anglesey, and probably on Little Knott in Cumberland. Many of these masses have the general appearance of laccolites. The intrusions near Clynog-fawr are doubtless the source of the numerous erratics which are scattered about in the neighbour- hood and to the south-west, as far at least as the base of Moel Penllechog. Professor Sedgwick collected specimens from these boulders and from the Pen-y-rhiwiau mass itself, as well as a rock labelled " one mile N.E. of Clynog ", which is probably from the dyke or mass near Tanrallt. Professor Bonney s had slices cut 1 Min. Mag., vol. vi., p. 54; 1886. 2 A. Geikie, Trans. Royal Soc. Edin., vol. xxix., p. 506; 1880. Teall, Brit. Petr., p. 94, pi. vn.; 1888. 3 Watts, Report Brit. Assoc. for 1887. 4 Pen-y-rhiwau on the Survey map ; Pen-y-rhiwan in Mr Tawney's paper. On the maps of the Geological Survey the mass is carried a little too far to the east, and the Tanrallt intrusion is not marked. B G. M., 1880, pp. 457, 458. ^. J. G. S., vol. XLI., p. 517; 1885. HORNBLENDE-DIABASES NEAR CLYNOG-FAWR. 103 from some of Sedgwick's specimens (01. 15 [132], 01. 17 [102], 01. 24 [133]), which he afterwards designated hornblende-picrite, and Mr Tawney 1 also examined the rock of Pen-y-rhiwiau. The general character of the mass hovers between hornblende-diabase and hornblende-picrite, olivine being either wanting or present in varying degree. If it be necessary to describe all the varieties under one name, that of olivine-hornblende-diabase conveys per- haps the best description of the average character. Different specimens shew some differences in structure as well as constitution. The commonest rock is a coarse-grained and black one in which the eye recognises only large hornblende crystals with discoloured felspars and an occasional flake of golden- brown mica. The boulders, selected by a process of survival, are not typical of the whole mass, being mostly derived from the coarser portions of the outcrop. At Pen-y-rhiwiau the darker and more peridotic varieties occur chiefly near the west end of the mass, where the rock sometimes verges on the type of the Penar- fynydd picrite. Many of the paler specimens from near the farm have little or no olivine. These bear the closest possible resem- blance to the hornblende-diabases near Llanerchymedd in the centre of Anglesey 2 , which I have described elsewhere. These latter too have furnished a plentiful supply of boulders scattered to the south-west, and shewing the same selection of the coarser and more basic variety. Examined with a low objective, the sections are seen to consist mainly of amphibole in brown ophitic plates and green actinolitic crystals. The other constituents, however, are numerous. Apatite occurs occasionally, but never abundantly, building slender needles or narrow hexagonal prisms, and being always the first product of consolidation [102]. A few grains of pale brown sphene are rarely seen [102]. Original iron-ores are sparingly present or entirely absent [102, 627]; but some slides shew mag- netite in grains or in little octahedral crystals [586]. Secondary magnetite is an abundant product of the decomposition of horn- blende and olivine. Copper pyrites [132] is occasionally present in imperfect crystals, and may be recognised in hand-specimens. 1 G. M., 1882, p. 548. 2 a. M., 1887, p. 546. See also Bonney, Q. J. G. S., vol. xxxvu., p. 137; 1881 : vol. xxxix., p. 254; 1883: vol. XLI., p. 515; 1885. Teall, Brit. Petr., pp. 81, 82, plates iv., vi. ; 1888. 104 OTHER BASIC INTRUSIONS. Olivine has probably been present in most of the rocks ex- amined, though it is often completely destroyed. As already noted, its quantity is very variable [585, 587]. The alteration is of the usual kind, giving rise to serpentine with the mesh-structure and strings of magnetite granules or dust [586, 587]. Some calcite and dolomite are sometimes associated with the serpentine [102, 586]. Felspar occurs in variable quantity, being never a dominant constituent, and sometimes scarce or absent. It is too much de- composed into calcite, etc., to afford any certain information as to its nature, but is presumably a variety rich in lime. Unequivocal augite is mostly uncommon : possibly original augite has been amphibolised, and indeed the manner in which the grains are sometimes imbedded in brown hornblende, with the usual crystallographic relation, distinctly suggests this [586]. In one boulder, however the only specimen containing much augite it occurs in good crystals moulded by a similar brown horn- blende. It is of a very pale yellowish brown tint or nearly colour- less, with good cleavage and the octagonal cross-section due to the prism-faces and piaacoids [132]. It is often twinned on the ortho- pinacoid, and has an imperfect hour-glass structure. A few colourless grains with good cleavage, imbedded in horn- blende, may possibly be enstatite, and occasionally some part of the serpentine has a structure rather suggestive of derivation from a rhombic pyroxene; but this is far from certain. A similar mineral occurs in some of the Anglesey intrusions. The ophitic plates of hornblende are, for the most part, deep brown, but in some places pass, either gradually or rather abruptly, into dull green or almost colourless [133, 627], sometimes with a zonary arrangement [132]. The absorption is expressible by the formula Almost colourless. and the pleochroism by the following scheme : Brown hornblende. a, pale brown ; (3, deep clove brown; y, deep clove brown ; Green hornblende. very pale rose-pink ; colourless ; pale olive green ; i very pale grey-brown ; rather pale grass- ' very pale grey-brown, green; CLYNOG-FAWR HORNBLENDE-DIABASES DESCRIBED. 105 The mineral usually shews perfect prismatic cleavage, and is some- times twinned on the orthopinacoid [102], occasionally with repe- tition [627]. Rarely the prismatic cleavage is locally replaced by one parallel to the orthopinacoid [585]. In some of the boulders the hornblende shews crystal outlines the prism and clinopinacoid [132, 133]. It sometimes [133] exhibits what appear to be 'solu- tion-planes ' parallel to the clinopinacoid, marked by discoloration and magnetite interpositions : a similar thing is seen in some of the Anglesey rocks [171, etc.]. Another common feature of these latter [535, 536, 539] is sometimes seen in the slides from Pen-y- rhiwiau. This is the ' secondary enlargement ' of the original or pseudomorphic hornblende plates by a marginal growth of pale secondary amphibole-substance, often fibrous but in crystalline continuity with the compact hornblende. The same thing is well seen in the Little Knott rock [170], and has indeed been figured by Professor Bonney 1 , though without any remark as to the secondary origin of the growth. It must be regarded as long pos- terior to the consolidation of the rock. All the slides contain a considerable quantity of actinolite in rather pale grass-green blades, frequently with a fan-like arrange- ment. The crystals are pleochroic, the greatest absorption being for the 7-axis, which makes an angle of about 18 with the c-axis of crystallography. Some crystals shew a lamellar twinning, the twin-plane being the orthopinacoid [586]. The cleavage is apparently orthopinacoidal [585, 586], and there is an occasional cross-jointing [585]. The actinolite never includes olivine, but sometimes minute octahedra of magnetite. There can be little doubt that these green blades of actinolite are developed at the expense of the hornblende [585], and the passage already noted in this mineral from a prismatic to a pinacoidal cleavage appears to be an early stage of the transformation. Actinolite is much less abundant in the Llanerchymedd rocks. Biotite occurs in irregular flakes, which sometimes give evi- dence of mechanical disturbance by shearing along 'gliding-planes' [133]. In ordinary light it is of a rather paler brown than usual, but there is the characteristically intense absorption for vibrations parallel to the cleavage-traces. This mica is sometimes an original constituent, being then newer than the augite, but older than the 1 Q. J. G. S., vol. XLI., pi. xvi.,fig. 2; 1885. 106 OTHER BASIC INTRUSIONS. hornblende [132]. Most of the mineral, however, in the slides examined, seems to be secondary after the hornblende [586]. In some cases a partially disintegrated crystal of the latter mineral shews scales of biotite forming on its cleavage-planes ; and, again, a mass of biotite is seen to enclose a kernel of decomposing horn- blende [627]. Taking note of all the hornblendic rocks of North Wales, we may remark that the Clynog-fawr rocks are on the whole nearer to the Llanerchymedd hornblende-diabases than to those of the Penarfynydd and Rhiw district, and they are of interest as indi- cating an intermediate variety between the former and hornblende- picrite of the Penarfynydd type. The Tanrallt intrusion I have not examined so carefully as that of Pen-y-rhiwiau ; but in hand-specimens and in a section [588] it exhibits very similar characters. In various parts of Caernarvonshire occur a large number of dykes of sub-basic or sometimes basic composition, which may safely be separated from the Bala igneous rocks described in the foregoing pages, and assigned to a later age. They appear to be unconnected with visible plutonic masses or with volcanic out- bursts properly so-called, and have a bearing usually between south-east and east-south-east, so that they cross the strike of the strata nearly at right angles. Many such dykes are noted on the Survey-map on the sea-coast and in the slate-quarries, and many others, in localities less freely exposed, may be found by searching. It may be observed that as we pass southward from the Menai Straits, the bearing of the dykes tends to be more easterly. Their width varies from a few inches to fifty or sixty feet. The age of these dykes is a matter for inference only, but if we assume that they all belong to one period, they must be set down as post- Carboniferous, since many of them on the shores of the Menai Straits intersect the Mountain Limestone. The date may be more precisely fixed by comparison with the dykes of Anglesey 1 . Several of the Caernarvonshire dykes can be traced on the other 1 G. M., 1887, pp. 409416; 1888, p. 267. POST-CARBONIFEROUS DYKES. 10? side of the Straits, and the dykes of this district thoroughly re- semble others having the same direction in the Anglesey coal-field, which latter are known with certainty to be post-Carboniferous but pre-Permian 1 . They are sometimes injected in lines of fault con- nected with post-Carboniferous crust-movements. Such evidence as we have, then, points to the probable supposition that all the Caernarvonshire dykes here referred to belong to the interval which separated the Carboniferous and New Red Sandstone periods in this area. The age of the dykes cutting the Llanberis slates seems at first sight a more difficult question than that of the Menai Straits dykes, with which they do not always agree in petrological cha- racter. But, in the first place, the strike of the dykes seems to indicate their relation to the later folding ; secondly, Sir A. Ramsay (pp. 102, 236) has pointed out that some of the dykes in Penrhyn Quarry contain fragments of cleaved slate, proving the intrusion to be later than the cleavage ; and, finally, the district shews no manifestation of igneous activity between the close of the Bala age and the close of the Carboniferous. The view that these dykes, like the others, are post-Carboniferous seems, therefore, to be well- founded. The bearing of the dykes is, as already indicated, almost always at right angles to the strike of the axes of post-Carboniferous earth-movements in this district, and the igneous magma has doubtless been injected into dip-joints. The persistent manner in which the dykes maintain their course, both vertically and hori- zontally, must be explained here, as elsewhere, by the fact that the pressure of the molten matter in a fissure must itself exert a power- ful influence tending to rend the adjacent rocks apart in the direc- tion perpendicular to the walls of the fissure, so that the dyke constantly tends to propagate itself in its own plane, as was long ago pointed out by Hopkins 2 . In the green schists between Porth- dinlleyn and Bardsey the dykes run, on the whole, with the dip of the strata, but it is quite in accordance with the difference between the country -rocks that the regularity should be less marked here than on the Menai Straits. 1 Ramsay, pp. 205, 204. Cf. Sedgwick, Q. J. G. S., Proc. G. S., vol. lv., p. 214; 1846. 2 "Researches in Physical Geology," Camb. Phil. Trans., vol. vi., p. 1, 108 OTHER BASIC INTRUSIONS. The only published analysis of any of these rocks is one of a dyke in the Penrhyn Quarry by Dr Voelcker, quoted by Mr Maw 1 , which gave the following figures : Si0 2 47-47 Ti0 2 2-51 A1 2 O 3 5-80 Fe 2 O 3 1-97 FeQ 3 10-22 FeS 2 0-23 CaS0 4 0-08 CaC0 3 14-85 MgCO, 14-59 K 2 O 0-43 N 2 a 2 0-70 H 2 O (combd.) 1-99 100-84 The rock is evidently deeply altered, and indeed the carbonates were present as visible crystals. Eliminating the carbonic acid, etc., the silica-percentage is raised to 56'24 ; but the other figures are still abnormal, and the analysis gives no clear idea of the ori- ginal nature of the rock. Mr Maw remarks how the blue slates are bleached in the neigh- bourhood of the dyke, and ascribes this circumstance to the reduc- tion of most of the iron-peroxide by the heat consequent upon the intrusion. The slates at the junction with this and similar dykes are indurated, and their cleavage often impaired 2 an additional proof of the post-Bala age of the dykes. Petrologically these rocks are not all of one type, but grada- tions can be made out, and there seems to be no objection to treating them as a whole. Some of those found on the Menai Straits I have described in the paper already referred to 3 . The rock of the smallest dykes is invariably an augite-andesite of very fine texture, though no unindividualised basis can with certainty be detected. A hand-specimen, when fresh, is black 1 0. M., 1868, p. 125. 2 Cf. Brit. Assoc. Beport for 1885, p. 834. 3 Numerous dykes in this district were recorded by Henslow, Trimmer, and Haugliton in the papers quoted in our Introdution. DESCRIPTION OF DYKES. 109 and compact, occasionally shewing clear felspars up to one-tenth of an inch in length with sometimes a fluxional disposition. Under the microscope these felspars shew albite and Carlsbad twinning, and appear to be labradorite or andesine. These felspars, with abundant crystals of magnetite, are imbedded in a ground-mass composed of felspar microlites, magnetite, and rounded augite granules. Olivine is not found. Dykes of this kind occur near Bangor, at Llanfair-is-gaer, on the shores of Llyn Padarn, and in other places. They are usually from four to eight inches in width. The same rock, however, often constitutes the marginal portion of larger dolerite dykes. Many of the larger dykes are of dolerite of a moderately coarse grain, shewing even to the eye a well-marked ophitic structure. The microscope shews felspars of two generations, magnetite, and augite, with sometimes a little apatite. Ilmenite is not found, and olivine too is usually absent. The magnetite frequently builds rods or branching aggregates, and is commonly later than the first generation of felspars. The felspars, tested by their extinc- tion-angles, seem to range from labradorite to andesine, the later ones being more acid, as a rule, than the earlier. The earlier crystals give long rectangular sections with fine albite-lamellation, often combined with a further twinning on either the Carlsbad or the pericline law. The later felspars are less elongated and in general quite allotriomorphic. They have less close twin-lamella- tion than the others, and are further distinguished by a strongly marked zonary banding in polarised light, indicating by the test of extinction-angles that the outer zones are more acid in composition than the inner. The augite shews in thin sections a pale brown tint, though not quite so pale as that of the common diabases of Caernarvonshire. In rare cases [736] there is a scarcely perceptible pleochroism, the colour varying from a rosy to a yellowish tinge. The augite forms ophitic plates moulding the earlier felspars and magnetite. It is sometimes slightly earlier, sometimes slightly later than the second generation of felspars : on the whole the two minerals are almost contemporaneous. Dykes of this doleritic type are found at Craig-y-fael [737], Plas Rhiw [736], and other places in the Sarn district 1 , as well as on 1 Cf. Q. J. G. S., vol. XLIV., pp. 460, 461; 1888. 110 OTHER BASIC INTRUSIONS. the Menai Straits. Some near Bangor are visibly porphyritic in hand-specimens, shewing scattered squarish felspars, white, pink, or liver-coloured, half-an-inch in length. One of these dykes, at Glan Adda [527], contains a few scattered flakes of biotite, with fine needles of apatite, but as a rule mica is rare. A few of the Menai Straits dykes are more distinctly basic than the foregoing, and contain rather abundant rounded grains of olivine. Such is the large dyke opposite Plas Newydd [529], which is a prolongation of the one on the Anglesey coast, and also a coarse-grained dyke exposed on the beach at Llanfair-is-gaer Church [559]. The latter contains a few flakes of biotite and some small patches of brown hornblende, included in the augite in the immediate vicinity of magnetite grains. The dykes of the Penrhyn, Llanberis, and Nantlle quarries exhibit a considerable variety of characters, with gradations from the doleritic to the diabasic type. As an example of the former, we may take a specimen from north-west of Cwm-y-glo [560]. The slide shews small magnetite crystals, large and small lath- shaped felspars with twin-lamellation, a few shapeless later fel- spars with the usual zonary banding, and ophitic augite, rather abundant and later than any of the felspar. A specimen from Llyn Padarn [561] is a diabase of rather ' granulitic ' habit. It contains abundant frame-works of ilmenite, one generation of lath-shaped felspars, and plenty of the usual pale augite \ The last mineral, though occasionally penetrated by the felspars, is in rather rounded grains, often polysomatic. The abundance of ilmenite in the diabases and its exclusion from the typical dolerites are highly characteristic. It should be noted, however, that I adopt here, as elsewhere, the structural distinction between diabase and dolerite, the former having but one genera- tion of each constituent mineral, while in the latter there is a re- currence of one or more constituents, usually the felspars, giving the ' porphyritic ' structure of Rosenbusch 2 . Some of the dykes contain a large quantity of epidote of secondary origin. A slice cut from a dyke nine feet wide in the Pen-y-bryn Quarry, Nantlle [562], affords a beautiful example of 1 The "hornblende" of Ramsey (Catalogue, p. 46) is evidently angite. 2 Neu. Jahrb. f. Min. etc., 1882, vol. n., p. 1. DYKES IN WESTERN LLEYN. Ill this mineral replacing felspar crystals by perfect pseudomorphs, but in the same rock are large nests replacing a considerable part of the whole mass. The dykes which intersect the green schists between Porth- dinlleyn and Bardsey seem to belong for the most part to the doleritic type, though some, especially the smaller ones, appear to be andesitic. The ophitic structure is sometimes very strikingly developed. In a dyke between Nevin and Porth-dinlleyn [126] the augite plates not only mould and enclose the felspars, but even protrude little tongues along fissures produced by corrosion along cleavage-planes of that mineral. The same specimen shews a curious effect of crushing, possibly connected with the proximity of the boundary-fault. The rock is traversed by very numerous fine cracks filled with a pale-green feebly-polarising chloritoid substance 1 . These veins run parallel to one another across the slice, except in the interior of the larger felspars, where they are deviated so as to follow the cleavage-planes of the crystals. These dykes in the green schists are assigned to a post-Carbo- niferous date necessarily with less confidence than the others enumerated above. It may be remarked, however, in support of this view that they have no apparent connexion with the intrusions in Lleyn which we have referred to the Bala age, and that un- doubted dykes of this age are decidedly rare in Caernarvonshire. i This is the slide described by Mr A. S. Reid, G. M., 1880, p. 457. VIII. REVIEW OF VuLCANicrrr IN CAERNARVONSHIRE. IT remains to be considered whether the facts brought for- ward in the preceding pages will serve as the material for any general conclusions with regard to the history of vulcanicity in North Wales during the Bala age. In this part of our subject we are necessarily on less firm ground than when merely recording petrological observations ; but, as summarising and adding point to the detailed study of the rocks, a few speculations, advanced under due reserve, though prompted directly by the observed pheno- mena, may be considered not out of place. Something at least will have been gained, if we can shew reasons for regarding the igneous rocks of this area as not capricious and disconnected effects of forces acting under no special law, but, rather, closely related manifestations of organised processes, which may some day be reduced to intelligible principles. Many geological text-books, and even treatises on special districts, treat igneous rocks in a way which leaves the student with a general impression that they are of no particular age, but constitute meaningless interpolations in the geological record, to be dismissed in a brief appendix after the sedimentary formations have been duly discussed. The legends accompanying geological maps often serve to confirm this impres- sion. The map illustrating a recent manual by a high authority indicates the volcanic rocks of Caernarvonshire and Antrim, of Iceland and the English lakes, by one common colour and legend. The clue to the igneous phenomena of our district seems to lie in their relation to stresses operating in the crust of the earth during their production. Before proceeding to the evidences of this important relation, it will be necessary to consider the character of the movements which brought the Bala age to a close in the area in question. That all the Bala and older rocks of North Wales were folded CHARACTER OF THE POST-BALA CRUST-MOVEMENTS. 113 and cleaved before the deposition of the succeeding Silurian strata is a fact which seems to admit no dispute. An examination, for instance, of the tract bordering the lower part of the Conwy valley reveals at once the unconformity between the two sets of rocks, and further a certain discordance in strike between the feeble and variable cleavage of the newer strata and the highly- developed and constant cleavage of the older. The facts have been sufficiently enforced by Sir A. Ramsay. A comparative study of the directions of the cleavage-planes over the Caernar- vonshire area enables us to realise in a simple manner the nature of the crust-movements which gave rise to the structure, and a glance at the chief undulations of the strata as marked on the Survey maps fully confirms the results so obtained. The succes- sion of events which have followed as effects of the gradually augmenting lateral thrust may be clearly made out : (i) folding, (ii) cleavage, (iii) foliation and metamorphism. This is the chronological order, as might be anticipated on remembering the different degrees of yielding indicated by these changes: (i) a rearrangement of the position of the rock-masses as a whole, (ii) a rearrangement of their constituent fragments, (iii) molecular and atomic rearrangements. An examination of each of these three phenomena leads to the same conclusions as to the direction and distribution of the forces to which they owe their common origin. A glance at the map is enough to shew that these forces acted in lines having an average N.W. S.E. direction. This, however, is not sufficient to define their action. All recent obser- vations and theories alike lead us to regard lateral crust-move- ments as of the nature of a comparatively superficial creep of the rocks affected over those underlying, and therefore as taking place not only parallel to a definite line, but in a definite direction in that line. In Shropshire the movement was roughly from east to west ; in the Lake District from south to north ; in the eastern half of Caernarvonshire, as will be clearly seen in the sequel, the movement had a general direction from south-east to north-west. Taking first the folding of the strata, and having regard more especially to the eastern division of the county, we note that the strike of the axes of disturbance is roughly N.E. S.W., but curves in such a manner as to bear more nearly N. S. towards the south-western limit of the district, and E. W. in the north- H. E. 8 114 REVIEW OF VULCANICITY IN CAERNARVONSHIRE. eastern corner. In brief, the strike, where most clearly marked, shews on the map as a rather flat curve with its concavity facing the south-east, a circumstance in itself sufficient to suggest that the thrust came from that direction and not from the north-west. The straightness of the strike throughout a great part of its extent at once connects itself with the Llyn Padarn ridge of crystalline and other hard materials, to which the folds of the strata are closely parallel. This ridge appears to have constituted a comparatively firm buttress, against which the whole of the eastern division of Caernarvonshire was pressed, and its influence is traceable not only in the folding, but still more in the direction and relative development of the cleavage and local micro-foliation of the slates. The Llyn Padarn ridge ranges in an approximately N.E. S.W. direction from St Anne's Chapel near Nant Ffrancon to Llanllyfni, and must be imagined as prolonged under-ground to some little distance in both directions. This ridge appears to have played a most important part in the geology of Caernarvon- shire, and it will be frequently referred to below. Its antiquity is proved by the occurrence of abundant pebbles of its characteristic quartz-porphyry in the succeeding conglomerate, and the idea of its intrusion into the overlying rocks must therefore be con- sidered as finally abandoned. Most geologists who have studied the district regard the quartz-porphyry and its associated rocks as older than the whole of the Cambrian formations. Their Archsean age has indeed been called in question by Mr Blake, who con- siders them to be interbedded igneous rocks forming part of the Cambrian itself. The fact that the conglomerate appears to rest on both flanks of the ridge 1 , and also the succession of the over- lying rocks, as exposed for instance in Nant Ffrancon, seem to me to support the Archaean theory. All this, however, is not material to our present purpose. The Llyn Padarn ridge existed as a buttress of relatively unyielding rocks during Bala and post-Bala times, and we shall see that it has exercised a ruling influence not only on the post-Bala folding and cleavage, but also on the manifestations of volcanic activity during the Bala age itself. The folding of the strata in the eastern division of Caernarvon- 1 The absence of the conglomerate near Cwm-y-glo and Pont-rhythallt is naturally explained by faulting, as shewn on the Survey map. INFLUENCE OF THE LLYN PADARN RIDGE. 115 shire increases in intensity on the whole as we pass from south- east to north-west. This is better seen in the field than on the maps, for abstraction must be made of the general south-easterly dip, and moreover the most marked folding is on a small scale. The maximum is reached in what we may name the Llanberis slate-zone, using the term in a topographical not a stratigraphical sense. This zone, lying immediately on the south-east side of the Llyn Padarn ridge, is where the thrust from the south-east encountered most resistance. Here the sections in the slate- quarries shew an intense plication of the strata, and here too are seen examples of the local thickening of diabase dykes by the compression of folds under extreme pressure *. In most disturbed districts the folding is a less accurate index of the distribution of stresses in the earth's crust than that obtained from a study of the slaty cleavage. Here, however, the two correspond in strike not only as regards their average direction, but also in their chief deviations from the general bear- ing, thus proving that the plication of the strata is strictly a result of the same forces that produced the cleavage-structure. Mere upheaval and depression have had no share in determining the strike of the beds in this district, nor have subsequent accidents occasioned any material change in this respect. Indeed the post-Silurian and post-Carboniferous crust-movements seem to have had the same general direction as those now under discussion, and the same is true of the movements which affected the Archaean rocks before the deposition of the oldest Cambrian strata. Over the greater part of the district the folding has not been accompanied to any considerable extent by faulting, and the great faults in the northern part of the county are clearly post- Carboniferous. Near the boundary between our eastern and western divisions, however, there are some interesting faults which must be assigned to the period of the folding, and these will be referred to below. The arrangement of the cleavage-planes may be easily realised by means of the rough sketch-map (fig. 5) drawn up from a large number of observations. The lines indicate the strike of the cleavage, while their closeness together is designed to give an 1 G. M., 1889, pp. 69, 70. 82 116 REVIEW OF VULCANICITY IN CAERNARVONSHIRE. idea of the relative development of the structure in different localities. The general bearing of the cleavage-strike is about N.E. and S.W., and this direction becomes more marked in pro- portion as we approach the Llyn Padarn ridge, which itself has the same strike. Again, as we pass from south-east to north- west across the eastern division of Caernarvonshire, the cleavage- structure becomes more highly developed, and culminates on reaching the Llanberis zone, on which are situated the famous quarries of Bethesda, Llanberis, and Nantlle. In other words, those rocks have been most crushed which lie nearest to the ridge on its south-east side. The rocks composing the Llyn Padarn ridge itself have been modified by the same forces, and the more so on the side facing the thrust. Witness, for instance, the small syncline extending north-easterly into the quartz-porphyry from the middle of the lake. In this neighbourhood, too, the less stubborn of the rocks constituting the ridge have received a cleavage-structure equal to that of the neighbouring Cambrian slates; while in some places, as near Pen-y-groes and Cilgwyn, even the hard quartz- porphyry on the south-eastern face of the ridge has been crushed into a schistose rock resembling the por- phyroldes of the Meuse valley. That the ridge yielded to some small extent as a whole is proved by the slates on the other side being cleaved, though far less perfectly than those in the quarries. The structure, however, dies out rapidly as we travel north-west, and the Arenig shales of Caernarvon and Bangor exhibit no trace of cleavage. They have been protected by the Llyn Padarn ridge from the powerful thrust which crushed, in varying degrees, all the strata of the south-eastern division. Over the greater part of the county the cleavage of the slates is of the type so lucidly explained by Dr Sorby, and is not attended by mineralogical changes. In the Llanberis slate-zone, however, the case is different. The glossy slates or phyllites of this tract have never been subjected to close study at the hands of petrologists ; but the microscopical examination of a few speci- mens is enough to confirm their general resemblance to the schistes de Fumay so minutely analysed by Professor Renard. They appear to consist in the main of quartz, white mica, chlorite, and sometimes haematite, all presumably authigenic, and these rocks must accordingly be ranked as highly metamorphic pro- SKETCH-MAP OF PART OF CAERNARVONSHIRE. / TO ILLUSTRATE THE DISTRIBUTION OF THE CLEAVAGE Scale: six miles io o-n. inch*. ^Archaean ridges. Chief igneous intrusions Fig.5. CLEAVAGE AND DYNAMO-METAMORPHISM. 117 ducts. It is significant enough that this micro-foliation is met with only in that tract which, as we have seen, was subjected to the most intense stress. Certain secondary changes in the diabase sheets of eastern Caernarvonshire are possibly connected with the mechanical stress. Such are the production of epidote pseudomorphing felspar crystals, and of veins and borders of colourless fibrous amphibole in crystal- line relation with the augite plates. These two features are seen in every slide cut from the diabases of the eastern division of the county, but never in those of the western, where, as we shall see, the lateral thrust was less powerful. Again, the ilmenite in the western diabases gives rise to the usual grey opaque leucoxene ; but in the diabases of the eastern division the decomposition- product has a brown tint with a certain translucency and some- times the double-refraction of sphene. This is a change which may very plausibly be referred to the dynamo-metamorphic effect of the more intense pressure in the latter case. The great igneous sheets in eastern Caernarvonshire have exercised but little influence upon the direction or the quality of the cleavage in the adjacent slates. The solid plugs and bosses of igneous rocks at Mynydd Mawr 1 and Y Foel Fras, on the other hand, have produced a very marked effect. The lines of cleavage- strike in each case may be seen to wind round the obstruction, as they would on a smaller scale round a hard imbedded nodule. The phenomena can be explained only on the supposition that the obstacle was there during the disturbance of the region, and are sufficient in themselves to establish the Bala age of the intrusions in question. Passing now into the Lleyn district, we find the cleavage- structure much less strongly developed than in the eastern division of the county, and also shewing much less uniformity of strike. This is due to two reasons. In the first place, the Llyn Padarn ridge, which in the eastern district offered the requisite resistance to the thrust, and regulated all its effects, dies away when traced towards the south-west. But what is more important is that the direct thrust itself was in great measure confined to the eastern division of Caernarvonshire. To the west it had a diminished intensity and took in general a more westerly direction. Indeed 1 G. M. t 1888, pp. 223, 224. 118 REVIEW OF VULCANICITY IN CAERNARVONSHIRE. if we follow the lines of cleavage-strike in a south-westerly direc- tion from the Llanberis slate-zone, we find them diverging like stream-lines on escaping from a confined channel. On the one hand they curve round westerly towards Clynog-fawr, owing to the sinking of the Llyn Padarn ridge ; and on the other they curve southerly towards Tremadoc, owing to the diminished intensity of the thrust and its more westerly direction in this part. By this curving round of the ' stream-lines ' the little tract to the north of Tremadoc is made to occupy a peculiar position, being indeed, to maintain the simile, a kind of eddy at the side of the stream. To understand this fully it is necessary to realise the magnitude of the horizontal displacement effected by the thrust in the eastern division. The deformation of the green spots in the Llanberis slates, first pointed out by Dr Sorby 1 and confirmed by many other circumstances, proves that the strata occupying some parts of that zone have been compressed in a nearly horizontal direction into about one-fourth of their original horizontal breadth. The compression in other parts of the district is of course less ; but there is no escape from the conclusion that the earth's crust in this region has suffered a bodily horizontal displacement measurable in some localities by miles. The western division of the county has been much less affected, and has, so to speak, lagged behind the eastern. In other words, the eastern district has been relatively thrust forward past the western. The bordering tract, to the north of Tremadoc, was thus subjected to powerful shearing stresses acting horizontally along vertical planes, and these stresses resulted in the formation of a series of north and south faults. The faults, as laid down on Mr Salter's map 2 , seem to betoken exactly the kind of relative dis- placement here imagined. The modification of the rocks in this tract is precisely in accord with the theory offered. Since the intense stress induced was mainly a shearing stress, we should 1 Edin. New Phil Jour., vol. LV., p. 137; 1853. Cf. Barker, Report of Brit. Assoc. for 1885, pp. 813852 ; 1886. Mr Fisher (G. M., 1884, p. 402) has suggested that these spots may be posterior to the cleavage. The regularity of their ellipsoidal form is against this view ; but what is more conclusive is that the spots are often cut and displaced by small faults, which are quite obsolete as structural planes and must themselves be older than the cleavage-structure. An instance is figured by Mr Maw (G. M., 1868, pi. vii.). * Catalogue of Coll. of Camb. and Sil. Foss., p. 9 ; 1873. BALA VOLCANOES CONNECTED WITH CRUST-MOVEMENTS. 119 expect the argillaceous strata to yield to it without appreciable mineral change, while the more rigid diabases, resisting the defor- mation, would cause a conversion of the mechanical work of shearing into chemical energy, and so induce mineralogical transformations 1 . Accordingly we find that the slates of this neighbourhood offer no special peculiarity, but the diabases present much more decided evidence of dynamo-metamorphism than any others in Caernarvon- shire. Secondary amphibole occurs not only in fibrous fringes and veins but in compact pellets with good hornblende-cleavage, and the titaniferous minerals have given rise to abundant clear grains of sphene, a feature not observed in any other Caernarvonshire diabases. We are now in a position to realise the character of the folding and cleavage of the Caernarvonshire strata, and to some extent also the nature and distribution of the forces which brought about those phenomena. A much more detailed examination might easily be made, but would have no special application to our sub- ject. The impression of the cleavage-structure, though doubtless in some degree a cumulative process, is as a whole clearly posterior to the plication of the strata. This is evident from the fact that the cleavage-dip is everywhere independent of the dip of the beds, and the fact is in agreement with what is seen in other areas of cleaved rocks. There can be no doubt, however, that the two phenomena are successive results of the same set of forces, and we are at liberty to regard these forces as growing gradually to a climax throughout a very long period of time. I hope to be able to shew some reason for considering the outburst of volcanic ma- terial during the Bala age as an earlier consequence of the same stresses in the earth's crust. It cannot be regarded as in itself improbable, that the growing forces should affect subterranean bodies of fluid, or potentially fluid, material, before they had gathered sufficient intensity to accomplish the folding of the solid rocks above. Keturning then to the igneous rocks, it must be noticed at the outset that their distribution presents striking relations with the position of the Llyn Padarn ridge, which seems to have played so important a part in the succeeding phenomena of folding and cleavage. Recalling our three-fold division of the county, it is of i G. M. t 1889, pp. 18, 19. 120 REVIEW OF VULCANICITY IN CAERNARVONSHIRE. sufficiently obvious significance that on one side of the ridge we find an ordered succession of lavas and intruded masses, but on the other side no igneous rocks of Bala age, with the possible excep- tion of the Llanfaglen syenite ; while in the Lleyn, where the hard axis is wanting, this simplicity of arrangement is entirely lost. The next point which strikes the eye on glancing over a geo- logical map of Caernarvonshire is that those intrusions of acid rocks, which have been indicated above as the centres of eruption, are situated very nearly on one line of strike drawn a little to the south-east of the Llyn Padarn ridge and its prolongation (see map, fig. 5). This is precisely the position they should occupy on the theory that the outbursts were due to a pressure from the south- east met by the resistance of the ridge. Such would be the line of weakness, through which at certain points the fluid lava would be forced, being lifted by the pressure into the sphere of action of volcanic agency, or perhaps actually squeezed out by this crust- pressure itself. Indeed the comparative scarcity of ashes seems to point to a steady, though intermittent, welling up of the fluid, with little intervention of explosive action ; and the wide extent and great aggregate thickness of the main Snowdonian flows, apparently produced by simultaneous emission of lava from three or four vents, confirms this view. The individual flows, however, are never of extreme thickness, and the appearances lend no sup- port to any idea of fissure-eruptions in this district. The several groups of flows into which we have divided the volcanic series shew a very evident distribution around certain centres, and these centres correspond to those intrusive masses which we have already been led by other considerations to regard as marking the probable vents. The sketch-map (fig. 6), designed to illustrate this distribu- tion of the lavas, must of course be considered as rather a diagram- matic one. It is impossible to assign limits to the lavas with much accuracy, the continuity of the flows being so frequently lost, where they have been removed by denudation or are still thickly covered by newer rocks. It appears quite clear, nevertheless, that the Arenig lavas cannot extend far north of Moel-wyn and certainly not west of Tremadoc; that, of the Bala lavas, the three lowest groups must have come from one or more volcanoes in the neigh- bourhood of Y Foel Eras; the uppermost group can be ascribed only to the Mynydd Mawr vent ; while the main Snowdon group SKETCH-MAP OF EASTERN CAERNARVONSHIRE TO SHOW THE LIMITS OF THE CROUPS OF LAVAS. T~ Cambridge University Pres Fig.6. THE LAVAS IN GENERAL SUBMARINE. 121 probably results from the commingling and overlapping of flows poured out from all the volcanoes which its range embraces. There is one point of importance in connexion with the Bala volcanic series of Caernarvonshire which must not be lost sight of. It was long ago maintained by Sedgwick 1 that the whole of the lavas are of subaqueous origin, and the same view seems to be adopted by the author of the Survey memoir. Unfortunately the physics of lava-flows under water is a rather obscure subject. Scrope 2 considered that the molten material, owing to the pressure of the superincumbent water, will retain its included vapour, and therefore preserve its liquidity, for a long time, and will travel, under an instantaneously solidified crust, to greater distances than those reached by subaerial flows. A similar result was arrived at from various considerations by Daubeny 8 and others. The wide extent of some of the Caernarvonshire lava-flows and their general characters seem to be consistent with a submarine origin. It is by no means impossible that during some portion of the Middle Bala age the actual vents rose above sea-level, but the lavas themselves were probably for the most part solidified under the pressure of some depth of water. The alternative would require us to postu- late a long succession of upheavals and depressions to account for the interbedded sedimentary strata. The volcanic material con- tained in the 'calcareous ashes' of Snowdon was probably ejected from a subaerial volcano, and fell into the sea to mingle with ordinary sediment and organic remains; but even this is by no means certain. If during the latter phases of their activity the volcanoes became elevated above sea-level, Caernarvonshire re- sembled in this respect the nearly coeval volcanic area of the English Lake District 4 . Turning now to the diabases, we see on any geological map that, as already pointed out, these rocks are for the most part intruded fairly accurately along the bedding-planes of the strata. It is true that most ' ideal sections ' pourtray the sheets breaking across the strata and becoming irregular in form as soon as they disappear beneath the surface; but very little reflection will 1 Q. J. G. S., vol. in., p. 135; 1847. 2 Volcanoes, pp. 244, etc. (2nd ed.), 1862. 3 A Description of active and extinct Volcanoes, pp. 678, etc. (2nd ed.), 1848. 4 J. Clifton Ward, Mem. GeoL Surv., 101 S.E., p. 70; 1876. 122 REVIEW OF VULCANICITY IN CAERNARVONSHIRE. make it clear, as already remarked in the case of laccolites, that when a number of intruded sheets crop out always with the same strike as the adjacent beds, it can only be because they are parallel to the planes of stratification. Excepting, then, such obviously trausgressive masses as those about Carnedd Dafydd, Cwm-dyll, and Moel Siabod, with several outcrops in the Lleyn district, we may say that the normal mode of occurrence of the Caernarvonshire diabases is as ' sills,' i.e. intrusive sheets injected mainly along the bedding-planes. We are naturally led to inquire the reason of this difference of behaviour between the acid and basic rocks of eastern Caernarvon- shire, the one group being always extrusive, the other intrusive. Scrope 1 laid special stress on the "greater liquidity" of fused basic rocks "enabling them to penetrate the seams of stratified rocks among which they were forced." But it is by no means clear that the acid magma, with its content of aqueous vapour, was neces- sarily less fluid than the basic ; nor that greater liquidity would cause a molten mass to force its way along bedding-planes rather than to the surface. Taking a different line, we may reasonably suppose, despite the rather extravagant speculations of Gilbert 2 , that, in the molten as in the solid condition, the basic rocks have a higher specific gravity than the acid. On this the usual assumption it is not difficult to understand why the lighter rhyolites rose to the surface, and were poured out as lava-flows, while the heavier diabases were injected laterally into the strata without ever reaching the surface. The specific gravities of the two groups of rocks in the solid state are about 2'6 and 2'9 respectively. Further, it appears highly probable that the diabasic magma came from a greater depth than the rhyolitic, and in fact under- lay it in deep-seated reservoirs from which both appear to have been derived. The constant association of the two groups of rocks has already been commented upon, and it has been proved that the diabases, like the rhyolites, are of Bala age. Although the sheets of basic rock are individually newer than the rhyolitic coulees with which they are immediately associated, the two groups of rocks are on the whole coeval, and it is difficult to resist 1 Volcanoes, 2nd ed., pp. 248, 249; 1862. 2 Geology of the Henry Mountains, p. 70; 1880. RELATION BETWEEN ACID AND BASIC ROCKS. 123 the conclusion that they are of common origin. The coincidence in their distribution, both in time and in space, seems too close to be explained by any hypothesis which makes the two magmas originally distinct and independent. We are therefore driven to the idea, which has been entertained by some geologists, of a process of separation under gravity taking place in the subter- ranean reservoirs which furnish the materials of both rocks. Such a theory the reverse of Durocher's pictures a molten rock- magma, originally of intermediate constitution, becoming separ- ated under the action of gravity into upper acid layers of less density and lower basic layers of greater density. If such a theory be admissible, the different modes of occurrence of the rhyolites and the diabases in Caernarvonshire are completely ac- counted for. In the English Lake District volcanic rocks of Bala age are largely developed, and the conditions of vulcanicity seem, despite some interesting differences, to have been very similar to those which obtained about the same time in Caernarvonshire. The Westmorland lavas, more intermixed with ashes and agglomerates than those in the Welsh area, divide generally into a lower andesitic and an upper rhyolitic series, and closely resemble the corresponding rocks described in the foregoing pages. The ande- sites may be supposed to represent approximately the original magma of intermediate composition, which was in part extra- vasated before the process of separation under gravity became effective 1 . In Caernarvonshire, on the other hand, it appears that the process of separation was in general well advanced prior to the outburst of any of the volcanoes. It is worthy of notice, how- ever, that the andesites and andesitic agglomerates of Penmaen near Pwllheli are older than the rhyolites of that district, and the same is true of the other small exposures in Lleyn, if they are to be regarded as bedded lavas. Cam Boduan, if intrusive, is later than the acid rocks, and would seem to indicate a fresh accession of the original intermediate magma. To determine the true re- lations of the andesites of the Y Foel Fras tract would require 1 The general succession referred to of intermediate followed by acid rocks is best seen in Westmorland. It should be noticed that basic lavas are not wanting in the Lake District: see analyses of specimens from Eycott Hill, Cumberland; J. C. Ward, Monthly Microscopical Journal, vol. xvin., p. 246 ; 1877. 124 REVIEW OF VULCANIC1TY IN CAERNARVONSHIRE. detailed mapping combined with the examination of a larger number of specimens than I have found opportunity to collect 1 . The agglomerates at Llyn-yr-afon can scarcely be intruded through the neighbouring granophyres, as the andesites of Bera-mawr might be supposed to be. From the manner in which the great boss of granophyre has encroached upon the lowest lavas, whose source must have been somewhere within the tract in question, we are justified in supposing that an earlier volcano in this vicinity has been destroyed by a later and more extensive invasion of the igneous magma. Other facts support this idea. The quartz- dolerites and andesites of Penmaenmawr and Tai-rhedyn, which may well be relics of one larger mass, also cut into the lowest lavas, and presumably represent a laccolitic intrusion of inter- mediate rock attendant upon the later outbreak of activity in the Y Foel Fras centre. From the position of the laccolite in the arch of an anticlinal fold, it even seems likely that when the intrusion was effected, a certain degree of folding was already in progress in this extreme northern part of the county. If we imagine a large reservoir of molten rock, already well separated into layers of different densities, to find exit above in obedience to increasing pressure within, it seems evident that, by a process of decanting, the upper and lighter layers will first be ejected, then the lower and heavier layers in order ; so that the lavas poured out will be each less acid than the preceding. On the other hand, while the differentiation of the original magma is still incomplete, the upper layers of our imaginary reservoir may be becoming progressively more acid, and must do so when the extravasation of material from the vent above is checked by some relief of the internal pressure. In this case the newer lavas will be more acid than the older. This latter set of conditions must be the more usual, and would probably be verified to some extent by a systematic series of analyses of lavas in the Caernarvonshire area. The lower groups are evidently less acid in constitution than the main Snowdon lavas, and, from such analyses as we can command, this group itself seems to grow more acid from the base upwards 2 ; while the small group of flows which we have named 1 Three attempts to traverse these mountains have been frustrated by the snow. 2 In the Cader Idris district Messrs Grenville Cole and Jennings find that the INJECTION OF THE BASIC MAGMA. 125 the upper Snowdon lavas, by a falling off in acidity, indicate the exhaustion of the lighter portions of the subterranean magma. The cessation of superficial volcanic action in any particular tract of the area must then be held to signify, not a relaxation of the crust-pressure to which all the phenomena are here referred, but a change in the mode of its manifestation. Instead of an ejection of the lighter magma there succeeded an injection of the heavier. The field-geology of the area sufficiently proves, as I have already insisted, that the diabasic intrusions followed in general the rhyolitic extrusions, and in the same tracts ; that the material came from greater depths ; and that when some of the diabases were being injected, certain parts of the area had already begun to experience a folding of their stratified rocks. The laccolitic intrusions of the most basic parts of the separated rock-magma afford evidence completely in accord with the ideas here put forward. These rocks, viz. the hornblende-diabases and picrites, are met with only in parts of the area where highly acid rhyolites attest a very perfect differentiation of the original magma, and they occupy a very low horizon, being found in- variably in strata which have been referred to the Arenig series, and far below the horizon of any extruded lavas in Caernarvon- shire. The high specific gravity of these rocks seems to have prevented their breaking upward through the strata even to the extent that the diabases have been able to do so. We have already had occasion to remark the want of igneous rocks of Bala age on the north-west side of the Llyn Padarn ridge. As regards the lava-flows, this may of course be explained by denudation, but the complete absence of acid intrusions, dia- base sheets, and injections of hornblendic rocks in this division of the county admits of no explanation but one similar to that advanced, taking account of the ridge itself as a barrier effectively dividing this district from eastern Caernarvonshire. The Llanfaglen rocks may conceivably be due to a stray intrusion of intermediate rock-magma connected with the igneous rocks of the other side of the ridge, but there is nothing to definitely fix their age. Being post-Arenig and having obviously no affinity with the post- tuffs and ashes become more acid from the base upwards. Q. J. G. S., vol. XLV., pp. 436, 437 ; 1889. 126 KEVIEW OF VULCANICITY IN CAERNARVONSHIRE. Carboniferous intrusions of the district, they may be placed pro- visionally among the Bala rocks. It is not necessary to suppose that the barrier was absolutely immoveable. We have seen that at a later time it yielded some- what to the augmented thrust from the south-east. Possibly during the volcanic age itself the ridge then covered with a great thickness of Cambrian (and Ordovician) strata may some- times even have risen above sea-level, the volcanoes fringing its unstable and oscillating shore-line. This would best explain some of the pyroclastic accumulations, and such an elevation would not leave any more decided traces of unconformity than we frequently see in the district in the local coalescence of distinct lava-flows separated elsewhere by sediments. A general elevation in the neighbourhood of the Llyn Padarn ridge during the period of the main Snowdon lavas, if sufficient to produce an appreciable slope of the sea-bottom, would help to explain the great distances to which some of the flows extended. In this case, the country to the north-west of the ridge may never have been covered by so great a development of lavas as is seen in the heart of the county. In order to connect these various speculations arising from the study of the volcanic series of Caernarvonshire, it may not be out of place to give a hypothetical sketch of the course of events in this area during the Bala period. With the close of the Arenig age, the volcanoes which had flooded the Merioneth area with their lavas, and sent a few flows over the Caernarvonshire border, became extinct, and they have probably had no revival. During the time immediately suc- ceeding, there was a cessation of igneous activity, and when the volcanic fires again broke out it was in a new area, that of Caernarvonshire. The transference of the seat of action in this direction seems to agree with the nature of the lateral thrusts which have been discussed above. As already pointed out, there is good reason to suppose that in central and eastern Caernarvon- shire the thrust was directed in general from south-east to north- west, but this was doubtless part of a divergent system of forces which took a more northerly direction in the north of the county and a more westerly direction in the district farther west. The rectilinear character of the folds subsequently formed in the SUMMARY. 127 central tract and the straight strike-line of the accompanying cleavage we have referred to the influence of the Llyn Padarn ridge; and it seems probable that this influence began to make itself felt at an early time in the Middle Bala age, being perhaps assisted by a certain amount of elevation of the Llyn Padarn axis as one of the first effects of the thrust. The first outbursts of lava in the eastern half of Caernarvon- shire were localised in the northern part. A volcano situated immediately north of Y Foel Fras was probably the source of the lowest flows those of Penmaenbach, Dwygyfylchi and the ridge of Y Drosgl. The magma erupted was of much less acid constitution than the typical rhyolites of Snowdon, etc. ; but the next lavas, those of Pen-yr-Oleu-wen and Carnedd Llewellyn, shew that the differentiation of the subterranean magma was steadily progressing. As in the former case, the vent of these lavas must have been situated within the area now occupied by the large composite igneous mass of Y Foel Fras, Bera Mawr, etc., and the flows were of limited extent. After a pause, a more extensive outpouring of lavas followed in the same northern dis- trict, the relics of which may be traced from Conwy Mountain to Capel Curig and Y Glyder Fach. The distribution of these lavas unmistakably points to the Y Foel Fras tract as the site of the volcano from which they were erupted ; while their extent over a distance of fourteen miles and the number and thickness of the flows seem to indicate that the volcano was one of con- siderable size. It is to this time, therefore, that we may most probably refer the breaking out of the large Y Foel Fras volcano, which destroyed the remains of the older vent, or perhaps the two older vents, from which the earlier lavas had issued. Mean- while, under the augmented thrust, an incipient anticline had been formed running eastward from Llanfairfechan, and into the arch of this had been injected as a rude laccolite a portion of the new intermediate magma, prior to any effective separation under gravity. In this way had been formed the intrusion of Penmaen- mawr and Tai-rhedyn, which, like the new volcano itself, is seen to encroach on the older lavas of Y Drosgl, etc. During this time it is probable that large reservoirs of molten rock-magma under the districts farther south-west had been undergoing a process of separation into lighter and heavier strata; 128 REVIEW OF VULCANICITY IN CAERNARVONSHIRE. so that the earliest products of the vents by which they found relief were very distinctly acid. It is evident indeed that con- siderable quantities of liquid rhyolitic material must have accu- mulated by the beginning of the next phase of vulcanicity, to which are due the main Snowdonian lavas. In the central district a portion of this acid material was injected among the sedimentary strata, forming the laccolites of Moel Perfedd and Bwlch-cywion, and here too a volcano was formed at what is now the head of Nant Ffrancon. From the slight interval which in this neighbour- hood divides the flows of the main Snowdon group from those preceding, it would appear that this vent was the first to become operative during the new phase of activity ; but there can be little doubt that the Y Foel Fras volcano subsequently resumed action, while a large new one broke out, the plug of which now forms Mynydd Mawr. The number of flows gives evidence of a pro- longed period of activity, while the thickness of some of them and the fact that they may be traced over a distance of more than thirty miles from north-east to south-west prove that the eruptions attained very large proportions, and were in part syn- chronous from the different vents. There can be no doubt that eruptions in the Lleyn peninsula were more or less contemporaneous with those that gave birth to the main Snowdon lavas in the east, but it would be difficult to fix the exact time at which volcanic outbursts first took place in this south-western division of the county. We know that the most important rhyolites were there preceded by some small andesitic flows, perhaps fairly representing the undivided inter- mediate magma. The situations of the vents must be sought among such intrusive masses as those near Clynog-fawr, Yr Eifl, and Cam Fadryn. When the volcanoes of the eastern division became extinct, the last to die was that of Mynydd Mawr, to which must be assigned the uppermost lavas. These were divided from the main Snowdon group by an interval in which the only volcanic products were of a fragmental character, and with these feeble flows the last phase of superficial vulcanicity came to a close. Most likely eruptions had ceased in Lleyn also, being succeeded there by sporadic intrusions of acid magma, which gave rise to rhyolite, quartz-porphyry, granophyre, or even granite, according to the SUMMARY. 129 conditions of consolidation. Possibly Carn Boduan, apparently an old vent plugged with andesitic lava, represents a late volcano arrested at an early stage of its progress ; but the denudation which removed the higher Bala beds from this part of the county, has left the history here incomplete. The outpouring of acid lavas being completed, the remaining heavier portion of the original magma was injected in the form of diabase sills and occasionally bosses. This may have been going on to some extent during the period of the volcanoes, but on the whole the diabases are obviously newer than the rhyolites. The strata began to yield more and more to the still increasing thrust from the south-east, and in the east of the county the injection of the diabases proceeded concurrently with the plication, the molten mass finding its most easy channel along the axes of the folds. The basic hornblendic rocks, intruded on low horizons, probably belong to a rather late phase of the igneous action. Meanwhile the Llyn Padarn ridge had risen into a more formidable barrier, and other features, such as the Cwm Tryfan and Capel Curig anticlines and the Snowdon, Cefn-y-capel, and Dolwyddelen synclines, had declared themselves. These were intensified by the growing thrust ; the cleavage of the more yield- ing beds accompanied and followed the severe plication of the strata ; and -mineral rearrangements were produced in the tracts where the most intense stresses were localised. Finally, the whole county having become elevated above sea-level, the rocks were subjected to ordinary denuding agencies, and the lateral thrusts, which had reached a climax during the cleavage and foliation of the slates, entirely died away. It is impossible to determine how much, if any, of the Upper Bala formation was ever deposited in Caernarvonshire; but the appearances seem to indicate that the break between the volcanic series and the succeeding Silurians of the eastern border is one of considerable magnitude, and it is possible that this county had already risen bodily above the water, while marine deposits were still forming in the country farther south. In conclusion, a few words on stratigraphical grouping may not be inappropriate. The question is one scarcely germane to our subject, although it is possibly touched by the title originally pro- posed for this essay. H. E. 9 130 REVIEW OF VULCANICITY IN CAERNARVONSHIRE. Any classification to be widely applied must be founded on the succession of palaeontological forms. The existence of three faunas in the Lower Palaeozoic rocks was clearly admitted in the discussion at the London Congress, and the general adoption of three distinct names for the three systems of rocks is probably only a question of time. The physical break at the summit of the Bala in this the Cambrian district of Sedgwick has been emphasised in the fore- going pages, and has been ascribed to the same general causes as the vulcanicity of the Bala age. It will be quite in accord with this theory if further investigation should reveal a break, though one of less magnitude, above the Arenig rocks on the southern borders of Caernarvonshire. The Ordovician period, including the Arenig and Bala (with Llandeilo), is distinguished physically in North Wales from the immediately preceding and succeeding periods by the prevalence of volcanic manifestations 1 . Since the Bala volcanoes became extinct, North Wales has known no im- portant outbreak of igneous activity. In the interval between the Carboniferous and Permian periods, when molten rock rose in numerous fissures within the Caernarvonshire area, we have no evidence of any superficial demonstration of the subterranean forces; while the great Tertiary outbursts which convulsed so large a portion of the British Isles did not extend into the principality. The history of vulcanicity in North Wales is the history of the pre-Cambrian and the Ordovician periods; and when this area comes to be studied in detail by abler geologists, we may hope to learn from it at least as much of the mechanism of igneous action and the internal economy of volcanoes as any other district, Palaeozoic or Neozoic, is able to teach us. 1 Messrs Grenville Cole and Jennings, however, place the earliest eruptions of the Cader Idris district as far back as the Tremadoc age or even earlier ; while Mr Blake regards the Llyn Padarn ridge as evidence of volcanic eruptions during the Harlech and Llanberis age. CAMBEIDGE: PRINTED BY c. j. CLAY, M.A. & SONS, AT THE UNIVERSITY PRESS. SOME PUBLICATIONS OF THE CAMBRIDGE UNIVERSITY PRESS. THE LIFE AND LETTERS OF THE REVEREND ADAM SEDGWICK, LL.D., F.E.S., Fellow of Trinity College, Cambridge, and Wood- wardian Professor of Geology from 1818 to 1873. (Dedicated, by special permission, to Her Majesty the Queen.) By JOHN WILLIS CLAEK, M.A., F.S.A., formerly Fellow of Trinity College, and THOMAS MKENNY HUGHES, M.A., Woodwardian Professor of Geology. 2 vols. Demy 8vo. [Nearly ready. A SYNOPSIS OF THE CLASSIFICATION OF BRITISH PALAEOZOIC BOCKS, by the Rev. ADAM SEDGWICK, LL.D., F.B.S., and FREDERICK M^CoY, F.G.S. One vol., Boyal 4to. Plates, 1. Is. A CATALOGUE OF THE COLLECTION OF CAMBRIAN AND SILURIAN FOSSILS contained in the Geological Museum of the University of Cambridge, by J. W. SALTER, F.G.S. With a Portrait of PROFESSOR SEDGWICK. Royal 4to. Is. 6d. THE FOSSILS AND PAL^EONTOLOGICAL AFFINITIES OF THE NEOCOMIAN DEPOSITS OF UPWARE AND BRICKHILL with Plates, being the Sedgwick Prize Essay for 1879. By the late W. KEEPING, M.A. Demy 8vo. 10s. 6d. A CATALOGUE OF AUSTRALIAN FOSSILS, Strati- graphically and Zoologically arranged, by R. ETHERIDGE, Jun., F.G.S. Demy 8vo. 10s. Qd. A CATALOGUE OF BOOKS AND PAPERS ON PROTOZOA, CCELENTERATES, WORMS, and certain smaller groups of animals, pub- lished during the years 18611883, by D'ARCY W. THOMPSON, M.A. Demy 8vo. 10s. 6d. A CATALOGUE OF THE COLLECTION OF BIRDS formed by the late H. E. STRICKLAND, now in the possession of the University of Cambridge. By 0. SALVIN, M.A. Demy 8vo. 1. Is. A CATALOGUE OF OSTEOLOGICAL SPECIMENS contained in the Anatomical Museum in the University of Cambridge. Demy 8vo. 2s. 6d ZLon&on: C. J. CLAY AND SONS, CAMBRIDGE UNIVERSITY PRESS WAREHOUSE, AVE MARIA LANE. 45 H3 THE LIBRARY UNIVERSITY OF CALIFORNIA Santa Barbara Goleta, California THIS BOOK IS DUE ON THE LAST DATE AVAILABLE FOB STAMPED BELOW. JLA 1 l\JS\ -H.r i c> DISPLAY FERIOD OCT 7 "SO 20m-8,'60(B2.594 8 4)476 000 990 958 1 45