1863: Upper Oolite: Solenhofen. (The Archaeopteryx macrura, Owen. See above Chapter 19.)

REPTILIA:

1710: Permian (or Zechstein): Thuringia. (The fossil monitor of Thuringia (Protosaurus Speneri, V. Meyer) was figured by Spener of Berlin in 1810. (Miscel. Berlin.))

1844: Carboniferous: Saarbruck, near Treves. (See Chapter 23.)

PISCES:

1709: Permian (or Kupferschiefer): Thuringia. (Memorabilia Saxoniae Subterr. Leipsic 1709.)

1793: Carboniferous (Mountain Limestone): Glasgow. (History of Rutherglen by David Ure, 1793.)

1828: Devonian: Caithness. (Sedgwick and Murchison Geological Transactions second series volume 3 page 141 1828.)

1840: Upper Ludlow: Ludlow. (Sir R. Murchison. See Chapter 26.)

1859: Lower Ludlow: Leintwardine. (See Chapter 26.)

Obs.— The evidence derived from foot-prints, though often to be relied on, is omitted in the above table, as being less exact than that founded on bones and teeth.

In Table 26.2 a few dates are set before the reader of the discovery of different classes of animals in ancient rocks, to enable him to perceive at a glance how gradual has been our progress in tracing back the signs of vertebrata to formations of high antiquity. Such facts may be useful in warning us not to assume too hastily that the point which our retrospect may have reached at the present moment can be regarded as fixing the date of the first introduction of any one class of beings upon the earth.

2. WENLOCK FORMATION.

We next come to the Wenlock formation, which has been divided into Wenlock limestone, Wenlock shale, and Woolhope limestone and Denbighshire grits.

a. WENLOCK LIMESTONE.

This limestone, otherwise well known to collectors by the name of the Dudley Limestone, forms a continuous ridge in Shropshire, ranging for about 20 miles from S.W. to N.E., about a mile distant from the nearly parallel escarpment of the Aymestry limestone. This ridgy prominence is due to the solidity of the rock, and to the softness of the shales above and below it. Near Wenlock it consists of thick masses of grey subcrystalline limestone, replete with corals, encrinites, and trilobites. It is essentially of a concretionary nature; and the concretions, termed "ball-stones" in Shropshire, are often enormous, even 80 feet in diameter. They are of pure carbonate of lime, the surrounding rock being more or less argillaceous (Murchison's Siluria chapter 6.) Sometimes in the Malvern Hills this limestone, according to Professor Phillips, is oolitic.

(FIGURE 536. Halysites catenularius, Linn. sp. Upper and Lower Silurian.)

(FIGURE 537. Favosites Gothlandica, Lam. Dudley. a. Portion of a large mass; less than the natural size. b. Magnified portion, to show the pores and the partitions in the tubes.)

(FIGURE 538. Omphyma turbinatum, Linn. Sp. (Cyathophyllum, Goldfuss) Wenlock Limestone, Shropshire.)

Among the corals, in which this formation is so rich, 53 species being known, the "chain-coral," Halysites catenularius (Figure 536), may be pointed out as one very easily recognised, and widely spread in Europe, ranging through all parts of the Silurian group, from the Aymestry limestone to near the bottom of the Llandeilo rocks. Another coral, the Favosites Gothlandica (Figure 537), is also met with in profusion in large hemispherical masses, which break up into columnar and prismatic fragments, like that here figured (Figure 537, b). Another common form in the Wenlock limestone is the Omphyma turbinatum (Figure 538), which, like many of its modern companions, reminds us of some cup-corals; but all the Silurian genera belong to the palaeozoic type before mentioned (Chapter 24), exhibiting the quadripartite arrangement of the septalamellae within the cup.

(FIGURE 539. Pseudocrintes bifasciatus, Pearce. Wenlock Limestone, Dudley.)

Among the numerous Crinoids, several peculiar species of Cyathocrinus (for genus see Figures 478, 479) contribute their calcareous stems, arms, and cups towards the composition of the Wenlock limestone. Of Cystideans there are a few very remarkable forms, most of them peculiar to the Upper Silurian formation, as, for example, the Pseudocrinites, which was furnished with pinnated fixed arms, as represented in Figure 539. (E. Forbes Mem. Geological Survey volume 2 page 496.)

(FIGURE 540. Strophomena (Leptaena) depressa, Sowerby. Wenlock and Ludlow Rocks.)

The Brachiopoda are, many of them, of the same species as those of the Aymestry limestone; as, for example, Atrypa reticularis (Figure 532), and Strophomena depressa (Figure 540); but the latter species ranges also from the Ludlow rocks, through the Wenlock shale, to the Caradoc Sandstone.

(FIGURE 541. Calymene Blumenbachii, Brong. Ludlow, Wenlock, and Bala beds.)

(FIGURE 542. Phacops (Asaphus) caudatus, Brong. Wenlock and Ludlow Rocks.)

(FIGURE 543. Sphaerexochus mirus, Beyrich; coiled up. Wenlock Limestone, Dudley; also found in Ohio, North America.)

(FIGURE 544. Homalonotus delphinocephalus, Konig. Wenlock Limestone, Dudley Castle.)

The crustaceans are represented almost exclusively by Trilobites, which are very conspicuous, 22 being peculiar. The Calymene Blumenbachii (Figure 541), called the "Dudley Trilobite," was known to collectors long before its true place in the animal kingdom was ascertained. It is often found coiled up like the common Oniscus or wood-louse, and this is so usual a circumstance among certain genera of trilobites as to lead us to conclude that they must have habitually resorted to this mode of protecting themselves when alarmed. The other common species is the Phacops caudatus (Asaphus caudatus), Brong. (see Figure 542), and this is conspicuous for its large size and flattened form. Sphaerexochus mirus (Figure 543) is almost a globe when rolled up, the forehead or glabellum of this species being extremely inflated. The Homalonotus, a form of Trilobite in which the tripartite division of the dorsal crust is almost lost (see Figure 544), is very characteristic of this division of the Silurian series.

WENLOCK SHALE.

(FIGURE 545. Graptolithus priodon, Bronn. Ludlow and Wenlock shales.)

The Wenlock Shale, observes Sir R. Murchison, is infinitely the largest and most persistent member of the Wenlock formation, for the limestone often thins out and disappears. The shale, like the Lower Ludlow, often contains elliptical concretions of impure earthy limestone. In the Malvern district it is a mass of finely levigated argillaceous matter, attaining, according to Professor Phillips, a thickness of 640 feet, but it is sometimes more than 1000 feet thick in Wales, and is worked for flag-stones and slates. The prevailing fossils, besides corals and trilobites, and some crinoids, are several small species of Orthis, Cardiola, and numerous thin-shelled species of Orthoceratites.

About six species of Graptolite, a peculiar group of sertularian fossils before alluded to as being confined to Silurian rocks, occur in this shale. Of fossils of this genus, which is very characteristic of the Lower Silurian, I shall again speak in the sequel.

b. WOOLHOPE BEDS.

Though not always recognised as a separate subdivision of the Wenlock, the Woolhope beds, which underlie the Wenlock shale, are of great importance. Usually they occur as massive or nodular limestones, underlaid by a fine shale or flag-stone; and in other cases, as in the noted Denbighshire sandstones, as a coarse grit of very great thickness. This grit forms mountain ranges through North and South Wales, and is generally marked by the great sterility of the soil where it occurs. It contains the usual Wenlock fossils, but with the addition of some common in the uppermost Ludlow rock, such as Chonetes lata and Bellerophon trilobatus. The chief fossils of the Woolhope limestone are Illaenus Barriensis, Homalonotus delphinocephalus (Figure 544), Strophomena imbrex, and Rhynchonella Wilsoni (Figure 531). The latter attains in the Woolhope beds an unusual size for the species, the specimens being sometimes twice as large as those found in the Wenlock limestone.

In some places below the Wenlock formation there are shales of a pale or purple colour, which near Tarannon attain a thickness of about 1000 feet; they can be traced through Radnor and Montgomery to North Wales, according to Messrs. Jukes and Aveline. By the latter geologist they have been identified with certain shales above the May-Hill Sandstone, near Llandovery, but, owing to the extreme scarcity of fossils, their exact position remains doubtful.

3. LLANDOVERY GROUP— BEDS OF PASSAGE.

We now come to beds respecting the classification of which there has been much difference of opinion, and which in fact must be considered as beds of passage between Upper and Lower Silurian. I formerly adopted the plan of those who class them as Middle Silurian, but they are scarcely entitled to this distinction, since after about 1400 Silurian species have been compared the number peculiar to the group in question only gives them an importance equal to such minor subdivisions as the Ludlow or Bala groups. I therefore prefer to regard them as the base of the Upper Silurian, to which group they are linked by more than twice as many species as to the Lower Silurian. By this arrangement the line of demarkation between the two great divisions, though confessedly arbitrary, is less so than by any other. They are called Llandovery Rocks, from a town in South Wales, in the neighbourhood of which they are well developed, and where, especially at a hill called Noeth Grug, in spite of several faults, their relations to one another can be clearly seen.

a. UPPER LLANDOVERY OR MAY-HILL SANDSTONE.

(FIGURE 546. Pentamerus oblongus, Sowerby. Upper and Lower Llandovery beds. a, b. Views of the shell itself, from figures in Murchison's Sil. Syst. c. Cast with portion of shell remaining, and with the hollow of the central septum filled with spar. d. Internal cast of a valve, the space once occupied by the septum being represented by a hollow in which is seen a cast of the chamber within the septum.)

(FIGURE 547. Stricklandinia (Pentamerus) lirata, Sowerby.)

The May-Hill group, which has also been named "Upper Llandovery," by Sir R. Murchison, ranges from the west of the Longmynd to Builth, Llandovery, and Llandeilo, and to the sea in Marlow's Bay, where it is seen in the cliffs. It consists of brownish and yellow sandstones with calcareous nodules, having sometimes a conglomerate at the base derived from the waste of the Lower Silurian rocks. These May-Hill beds were formerly supposed to be part of the Caradoc formation, but their true position was determined by Professor Sedgwick to be at the base of the Upper Silurian proper. (Quarterly Geological Journal volume 4 page 215 1853.) The more calcareous portions of the rock have been called the Pentamerus limestone, because Pentamerus oblongus (Figure 546) is very abundant in them. It is usually accompanied by P. (Stricklandinia) lirata (Figure 547); both forms have a wide geographical range, being also met with in the same part of the Silurian series in Russia and the United States.

About 228 species of fossils are known in the May-Hill division, more than half of which are Wenlock species. They consist of trilobites of the genera Illaenus and Calymene; Brachiopods of the genera Orthis, Atrypa, Leptaena, Pentamerus, Strophomena, and others; Gasteropods of the genera Turbo, Murchisonia (for genus, see Figure 567), and Bellerophon; and Pteropods of the genus Conularia. The Brachiopods, of which there are 66 species, are almost all Upper Silurian.

(FIGURE 548. Tentaculites annulatus, Schlot. Interior casts in sandstone. Upper Llandovery, Eastnor Park, near Malvern. Natural size and magnified.)

Among the fossils of the May-Hill shelly sandstone at Malvern, Tentaculites annulatus (Figure 548), an annelid, probably allied to Serpula, is found.

LOWER LLANDOVERY ROCKS.

Below the May-Hill Group are the Lower Llandovery Rocks, which consist chiefly of hard slaty rocks, and beds of conglomerate from 600 to 1000 feet in thickness. The fossils, which are somewhat rare in the lower beds, consist of 128 known species, only eleven of which are peculiar, 83 being common to the May-Hill group above, and 93 common to the rocks below. Stricklandinia (Pentamerus) levis, which is common in the Lower Llandovery, becomes rare in the Upper, while Pentamerus oblongus (Figure 546), which is the characteristic shell of the Upper Llandovery, occurs but seldom in the Lower.

LOWER SILURIAN ROCKS.

The Lower Silurian has been divided into, first, the Bala Group; secondly, the Llandeilo flags; and, thirdly, the Arenig or Lower Llandeilo formation.

BALA AND CARADOC BEDS.

(FIGURE 549. Orthis tricenaria, Conrad. New York; Canada. 1/2 natural size.)

(FIGURE 550. Orthis vespertilio, Sowerby. Shropshire, N. and S. Wales. One-half natural size.)

(FIGURE 551. Orthis (Strophomena) grandis, Sowerby. Two-thirds natural size. Caradoc Beds, Horderley, Shropshire, and Coniston, Lancashire.)

The Caradoc sandstone was originally so named by Sir R.I. Murchison from the mountain called Caer Caradoc, in Shropshire; it consists of shelly sandstones of great thickness, and sometimes containing much calcareous matter. The rock is frequently laden with the beautiful trilobite called by Murchison Trinucleus Caractaci (see Figure 553), which ranges from the base to the summit of the formation, usually accompanied by Strophomena grandis (see Figure 551), and Orthis vespertilio (Figure 550), with many other fossils.

BRACHIOPODA.

Nothing is more remarkable in these beds, and in the Silurian strata generally of all countries, than the preponderance of brachiopoda over other forms of mollusca. Their proportional numbers can by no means be explained by supposing them to have inhabited seas of great depth, for the contrast between the palaeozoic and the present state of things has not been essentially altered by the late discoveries made in our deep-sea dredgings. We find the living brachiopoda so rare as to form about one forty-fourth of the whole bivalve fauna, whereas in the Lower Silurian rocks of which we are now about to treat, and where the brachiopoda reach their maximum, they are represented by more than twice as many species as the Lamellibranchiate bivalves.

There may, indeed, be said to be a continued decrease of the proportional number of this lower tribe of mollusca as we proceed from older to newer rocks. In the British Devonian, for example, the Brachiopoda number 99, the Lamellibranchiata 58; while in the Carboniferous their proportions are more than reversed, the Lamellibranchiata numbering 334 species, and the Brachiopoda only 157. In the Secondary or Cainozoic formations the preponderance of the higher grade of bivalves becomes more and more marked, till in the tertiary strata it approaches that observed in the living creation.

While on this subject, it may be useful to the student to know that a Brachiopod differs from ordinary bivalves, mussels, cockles, etc., in being always equal- sided and never quite equi-valved; the form of each valve being symmetrical, it may be divided into two equal parts by a line drawn from the apex to the centre of the margin.

TRILOBITES.

In the Bala and Caradoc beds the trilobites reach their maximum, being represented by 111 species referred to 23 genera.

(FIGURE 552. Young individuals of Trinucleus concentricus (T. ornatus, Barr.). a. Youngest state. Natural size and magnified; the body rings not at all developed. b. A little older. One thorax joint. c. Still more advanced. Three thorax joints. The fourth, fifth, and sixth segments are successively produced, probably each time the animal moulted its crust.)

(FIGURE 553. Trinucleus concentricus, Eaton. Syn. T. Caractaci, Murch. Ireland; Wales; Shropshire; North America; Bohemia.)

Burmeister, in his work on the organisation of trilobites, supposes that they swam at the surface of the water in the open sea and near coasts, feeding on smaller marine animals, and to have had the power of rolling themselves into a ball as a defence against injury. He was also of opinion that they underwent various transformations analogous to those of living crustaceans. M. Barrande, author of an admirable work on the Silurian rocks of Bohemia, confirms the doctrine of their metamorphosis, having traced more than twenty species through different stages of growth from the young state just after its escape from the egg to the adult form. He has followed some of them from a point in which they show no eyes, no joints, or body rings, and no distinct tail, up to the complete form with the full number of segments. This change is brought about before the animal has attained a tenth part of its full dimensions, and hence such minute and delicate specimens are rarely met with. Some of his figures of the metamorphoses of the common Trinucleus are copied in Figures 552 and 553. It was not till 1870 that Mr. Billings was enabled, by means of a specimen found in Canada, to prove that the trilobite was provided with eight legs.

(FIGURE 554. Palaeaster asperimus, Salt. Caradoc, Welshpool.)

(FIGURE 555. Echinosphaeronites balticus, Eichwald. (Of the family Cystideae.) a. Mouth. b. Point of attachment of stem. Lower Silurian S. and N. Wales.)

It has been ascertained that a great thickness of slaty and crystalline rocks of South Wales, as well as those of Snowdon and Bala, in North Wales, which were first supposed to be of older date than the Silurian sandstones and mudstones of Shropshire, are in fact identical in age, and contain the same organic remains. At Bala, in Merionethshire, a limestone rich in fossils occurs, in which two genera of star-fish, Protaster and Palaeaster, are found; the fossil specimen of the latter (Figure 554) being almost as uncompressed as if found just washed up on the sea-beach. Besides the star-fish there occur abundance of those peculiar bodies called Cystideae. They are the Sphaeronites of old authors, and were considered by Professor E. Forbes as intermediate between the crinoids and echinoderms. The Echinosphaeronite here represented (Figure 555) is characteristic of the Caradoc beds in Wales, and of their equivalents in Sweden and Russia.

With it have been found several other genera of the same family, such as Sphaeronites, Hemicosmites, etc. Among the mollusca are Pteropods of the genus Conularia of large size (for genus, see Figure 518). About eleven species of Graptolite are reckoned as belonging to this formation; they are chiefly found in peculiar localities where black mud abounded. The formation, when traced into South Wales and Ireland, assumes a greatly altered mineral aspect, but still retains its characteristic fossils. The known fauna of the Bala group comprises 565 species, 352 of which are peculiar, and 93, as before stated, are common to the overlying Llandovery rocks. It is worthy of remark that, when it occurs under the form of trappean tuff (volcanic ashes of De la Beche), as in the crest of Snowdon, the peculiar species which distinguish it from the Llandeilo beds are still observable. The formation generally appears to be of shallow-water origin, and in that respect is contrasted with the group next to be described. Professor Ramsay estimates the thickness of the Bala Beds, including the contemporaneous volcanic rocks, stratified and unstratified, as being from 10,000 to 12,000 feet.

LLANDEILO FLAGS.

(FIGURE 556. Didymograpsus (Graptolites) Murchisonii, Beck. Llandeilo flags, Wales.)

The Lower Silurian strata were originally divided by Sir R. Murchison into the upper group already described, under the name of Caradoc Sandstone, and a lower one, called, from a town in Carmarthenshire, the Llandeilo flags. The last mentioned strata consist of dark-coloured micaceous flags, frequently calcareous, with a great thickness of shales, generally black, below them. The same beds are also seen at Builth, in Radnorshire, where they are interstratified with volcanic matter.

(FIGURE 557. Diplograpsus pristis, Hisinger. Llandeilo beds, Waterford.)

(FIGURE 558. Rastrites peregrinus, Barrande. Scotland; Bohemia; Saxony. Llandeilo flags.)

(FIGURE 559. Diplograpsus folium, Hisinger. Dumfriesshire; Sweden. Llandeilo flags.)

A still lower part of the Llandeilo rocks consists of a black carbonaceous slate of great thickness, frequently containing sulphate of alumina, and sometimes, as in Dumfriesshire, beds of anthracite. It has been conjectured that this carbonaceous matter may be due in great measure to large quantities of imbedded animal remains, for the number of Graptolites included in these slates was certainly very great. In Great Britain eleven genera and about 40 species of Graptolites occur in the Llandeilo flags and underlying Arenig beds. The double Graptolites, or those with two rows of cells, such as Diplograpsus (Figure 557), are conspicuous.

The brachiopoda of the Llandeilo flags, which number 47 species, are in the main the same as those of the Caradoc Sandstone, but the other mollusca are in great part of different species.

(FIGURE 560. Orthoceras duplex, Wahlenberg. Russia and Sweden. (From Murchison's Siluria.))

(FIGURE 561. Asaphus tyrannus, Murchison. Llandeilo; Bishop's Castle; etc.)

(FIGURE 562. Ogygia Buchii, Burm. Syn. Asaphus Buchii, Brongn. Builth, Radnorshire; Llandeilo, Carmarthenshire.)

In Europe generally, as, for example, in Sweden and Russia, no shells are so characteristic of this formation as Orthoceratites, usually of great size, and with a wide siphuncle placed on one side instead of being central (see Figure 560). Among other Cephalopods in the Llandeilo flags is Cyrtoceras; in the same beds also are found Bellerophon (see Figure 488) and some Pteropod shells (Conularia, Theca, etc.), also in spots where sand abounded, lamellibranchiate bivalves of large size. The Crustaceans were plentifully represented by the Trilobites, which appear to have swarmed in the Silurian seas just as crabs and shrimps do in our own; no less than 263 species have been found in the British Silurian fauna. The genera Asaphus (Figure 561), Ogygia (Figure 562), and Trinucleus (Figures 552 and 553) form a marked feature of the rich and varied Trilobitic fauna of this age.

Beneath the black slates above described of the Llandeilo formation, Graptolites are still found in great variety and abundance, and the characteristic genera of shells and trilobites of the Lower Silurian rocks are still traceable downward, in Shropshire, Cumberland, and North and South Wales, through a vast depth of shaly beds, in some districts interstratified with trappean formations of contemporaneous origin; these consist of tuffs and lavas, the tuffs being formed of such materials as are ejected from craters and deposited immediately on the bed of the ocean, or washed into it from the land. According to Professor Ramsay, their thickness is about 3300 feet in North Wales, including those of the Lower Llandeilo. The lavas are feldspathic, and of porphyritic structure, and, according to the same authority, of an aggregate thickness of 2500 feet.

ARENIG OR STIPER-STONES GROUP (LOWER LLANDEILO OF MURCHISON).

(FIGURE 563. Arenicolites linearis, Hall. Arenig beds, Stiper-Stones. a. Parting between the beds, or planes of bedding.)

(FIGURE 564. Didymograpsus geminus, Hisinger, sp. Sweden.)

Next in the descending order are the shales and sandstones in which the quartzose rocks called Stiper-Stones in Shropshire occur. Originally these Stiper-Stones were only known as arenaceous quartzose strata in which no organic remains were conspicuous, except the tubular burrows of annelids (see Figure 563, Arenicolites linearis), which are remarkably common in the Lowest Silurian in Shropshire, and in the State of New York, in America. They have already been alluded to as occurring by thousands in the Silurian strata unconformably overlying the Cambrian, in the mountain of Queenaig, in Sutherlandshire (Figure 82). I have seen similar burrows now made on the retiring of the tides in the sands of the Bristol Channel, near Minehead, by lob-worms which are dug out by fishermen and used as bait. When the term Silurian was given by Sir R. Murchison, in 1835, to the whole series, he considered the Stiper-Stones as the base of the Silurian system, but no fossil fauna had then been obtained, such as could alone enable the geologist to draw a line between this member of the series and the Llandeilo flags above, or a vast thickness of rock below, which was seen to form the Longmynd hills, and was called "unfossiliferous graywacke." Professor Sedgwick had described, in 1843, strata now ascertained to be of the same age as largely developed in the Arenig mountain, in Merionethshire; and the Skiddaw slates in the Lake-District of Cumberland, studied by the same author, were of corresponding date, though the number of fossils was, in both cases, too few for the determination of their true chronological relations. The subsequent researches of Messrs. Sedgwick and Harkness, in Cumberland, and of Sir R.I. Murchison and the Government surveyors in Shropshire, have increased the species to more than sixty. These were examined by Mr. Salter, and shown in the third edition of "Siluria" (page 52, 1859) to be quite distinct from the fossils of the overlying Llandeilo flags. Among these the Obolella plumbea, Aeglina binodosa, Ogygia Selwynii, and Didymograpsus geminus (Figure 564), and D. Hirundo, are characteristic.

But, although the species are distinct, the genera are the same as those which characterise the Silurian rocks above, and none of the characteristic primordial or Cambrian forms, presently to be mentioned, are intermixed. The same may be said of a set of beds underlying the Arenig rocks at Ramsay Island and other places in the neighbourhood of St. David's. These beds, which have only lately become known to us through the labours of Dr. Hicks (Transactions of the British Association 1866. Proceedings of the Liverpool Geological Society 1869.), present already twenty new species, the greater part of them allied generically to the Arenig rocks. This Arenig group may therefore be conveniently regarded as the base of the great Silurian system, a system which, by the thickness of its strata and the changes in animal life of which it contains the record, is more than equal in value to the Devonian, or Carboniferous, or other principal divisions, whether of primary or secondary date.

It would be unsafe to rely on the mere thickness of the strata, considered apart from the great fluctuations in organic life which took place between the era of the Llandeilo and that of the Ludlow formation, especially as the enormous pile of Silurian rocks observed in Great Britain (in Wales more particularly) is derived in great part from igneous action, and is not confined to the ordinary deposition of sediment from rivers or the waste of cliffs.

In volcanic archipelagoes, such as the Canaries, we see the most active of all known causes, aqueous and igneous, simultaneously at work to produce great results in a comparatively moderate lapse of time. The outpouring of repeated streams of lava— the showering down upon land and sea of volcanic ashes— the sweeping seaward of loose sand and cinders, or of rocks ground down to pebbles and sand, by rivers and torrents descending steeply inclined channels— the undermining and eating away of long lines of sea-cliff exposed to the swell of a deep and open ocean— these operations combine to produce a considerable volume of superimposed matter, without there being time for any extensive change of species. Nevertheless, there would seem to be a limit to the thickness of stony masses formed even under such favourable circumstances, for the analogy of tertiary volcanic regions lends no countenance to the notion that sedimentary and igneous rocks 25,000, much less 45,000 feet thick, like those of Wales, could originate while one and the same fauna should continue to people the earth. If, then, we allow that about 25,000 feet of matter may be ascribed to one system, such as the Silurian, as above described, we may be prepared to discover in the next series of subjacent rocks a distinct assemblage of species, or even in great part of genera, of organic remains. Such appears to be the fact, and I shall therefore conclude with the Arenig beds my enumeration of the Silurian formations in Great Britain, and proceed to say something of their foreign equivalents, before treating of rocks older than the Silurian.

SILURIAN STRATA OF THE CONTINENT OF EUROPE.

When we turn to the continent of Europe, we discover the same ancient series occupying a wide area, but in no region as yet has it been observed to attain great thickness. Thus, in Norway and Sweden, the total thickness of strata of Silurian age is considerably less than 1000 feet, although the representatives both of the Upper and Lower Silurian of England are not wanting there. In Russia the Silurian strata, so far as they are yet known, seem to be even of smaller vertical dimensions than in Scandinavia, and they appear to consist chiefly of the Llandovery group, or of a limestone containing Pentamerus oblongus, below which are strata with fossils corresponding to those of the Llandeilo beds of England. The lowest rock with organic remains yet discovered is "the Ungulite or Obolus grit" of St. Petersburg, probably coeval with the Llandeilo flags of Wales.

(Figures 565 and 566. Shells of the lowest known Fossiliferous Beds in Russia.

(FIGURE 565. Siphonotreta unguiculata, Eichwald. From the Lowest Silurian Sandstone, "Obolus grits," of St. Petersburg. a. Outside of perforated valve. b. Interior of same, showing the termination of the foramen within. (Davidson.))

(FIGURE 566. Obolus Apollinis, Eichwald. From the same locality. a. Interior of the larger or ventral valve. b. Exterior of the upper (dorsal) valve. (Davidson, "Palaeontographic Monograph.")))

The shales and grits near St. Petersburg, above alluded to, contain green grains in their sandy layers, and are in a singularly unaltered state, taking into account their high antiquity. The prevailing Brachiopods consist of the Obolus or Ungulite of Pander, and a Siphonotreta (Figures 565, 566). Notwithstanding the antiquity of this Russian formation, it should be stated that both of these genera of brachiopods have been also found in the Upper Silurian of England, i.e. In the Wenlock limestone.

Among the green grains of the sandy strata above-mentioned, Professor Ehrenberg announced in 1854 his discovery of remains of foraminifera. These are casts of the cells; and among five or six forms three are considered by him as referable to existing genera (e.g., Textularia, Rotalia, and Guttulina).

SILURIAN STRATA OF THE UNITED STATES.

Table 26.3. SUBDIVISIONS OF THE SILURIAN STRATA OF NEW YORK. (Strata below the Oriskany sandstone or base of the Devonian.)

COLUMN 1: NEW YORK NAMES.

COLUMN 2: BRITISH EQUIVALENTS.

1. Upper Pentamerus Limestone: Upper Silurian (or Ludlow and Wenlock formations).

2. Encrinal Limestone: Upper Silurian (or Ludlow and Wenlock formations).

3. Delthyris Shaly Limestone: Upper Silurian (or Ludlow and Wenlock formations).

4. Pentamerus and Tentaculite Limestones: Upper Silurian (or Ludlow and Wenlock formations).

5. Water Lime Group: Upper Silurian (or Ludlow and Wenlock formations).

6. Onondaga Salt Group: Upper Silurian (or Ludlow and Wenlock formations).

7. Niagara Group: Upper Silurian (or Ludlow and Wenlock formations).

8. Clinton Group: Beds of Passage, Llandovery Group.

9. Medina Sandstone: Beds of Passage, Llandovery Group.

10. Oneida Conglomerate: Beds of Passage, Llandovery Group.

11. Gray Sandstone: Beds of Passage, Llandovery Group.

12. Hudson River Group: Lower Silurian (or Caradoc and Bala, Llandeilo and Arenig Formations).

13. Trenton Limestone: Lower Silurian (or Caradoc and Bala, Llandeilo and Arenig Formations).

14. Black-River Limestone: Lower Silurian (or Caradoc and Bala, Llandeilo and Arenig Formations).

15. Bird's-eye Limestone: Lower Silurian (or Caradoc and Bala, Llandeilo and Arenig Formations).

16. Chazy Limestone: Lower Silurian (or Caradoc and Bala, Llandeilo and Arenig Formations).

17. Calciferous Sandstone: Lower Silurian (or Caradoc and Bala, Llandeilo and Arenig Formations).

The Silurian formations can be advantageously studied in the States of New York, Ohio, and other regions north and south of the great Canadian lakes. Here they are often found, as in Russia, nearly in horizontal position, and are more rich in well-preserved fossils than in almost any spot in Europe. In the State of New York, where the succession of the beds and their fossils have been most carefully worked out by the Government surveyors, the subdivisions given in the first column of Table 26.3 have been adopted.

In the second column of the same table I have added the supposed British equivalents. All Palaeontologists, European and American, such as MM. De Verneuil, D. Sharpe, Professor Hall, E. Billings, and others, who have entered upon this comparison, admit that there is a marked general correspondence in the succession of fossil forms, and even species, as we trace the organic remains downward from the highest to the lowest beds; but it is impossible to parallel each minor subdivision.

That the Niagara Limestone, over which the river of that name is precipitated at the great cataract, together with its underlying shales, corresponds to the Wenlock limestone and shale of England there can be no doubt. Among the species common to this formation in America and Europe are Calymene Blumenbachii, Homalonotus delphinocephalus (Figure 544), with several other trilobites; Rhynchonella Wilsoni, Figure 531, and Retzia cuneata; Orthis elegantula, Pentamerus galeatus, with many more brachiopods; Orthoceras annulatum, among the cephalopodous shells; and Favosites gothlandica, with other large corals.

The Clinton Group, containing Pentamerus oblongus and Stricklandinia, and related more nearly by its fossil species with the beds above than with those below, is the equivalent of the Llandovery Group or beds of passage.

(FIGURE 567. Murchisonia gracilis, Hall. A fossil characteristic of the Trenton Limestone. The genus is common in Lower Silurian rocks.)

The Hudson River Group, and the Trenton Limestone, agree palaeontologically with the Caradoc or Bala group, containing in common with them several species of trilobites, such as Asaphus (Isotelus) gigas, Trinucleus concentricus (Figure 553); and various shells, such as Orthis striatula, Orthis biforata (or O. lynx), O. porcata (O. occidentalis of Hall), and Bellerophon bilobatus. In the Trenton limestone occurs Murchisonia gracilis, Figure 567, a fossil also common to the Llandeilo beds in England.

Mr. D. Sharpe, in his report on the mollusca collected by me from these strata in North America (Quarterly Geological Journal volume 4.), has concluded that the number of species common to the Silurian rocks on both sides of the Atlantic is between 30 and 40 per cent; a result which, although no doubt liable to future modification, when a larger comparison shall have been made, proves, nevertheless, that many of the species had a wide geographical range. It seems that comparatively few of the gasteropods and lamellibranchiate bivalves of North America can be identified specifically with European fossils, while no less than two-fifths of the brachiopoda, of which my collection chiefly consisted, are the same. In explanation of these facts, it is suggested that most of the recent brachiopoda (especially the orthidiform ones) are inhabitants of deep water, and that they may have had a wider geographical range than shells living near shore. The predominance of bivalve mollusca of this peculiar class has caused the Silurian period to be sometimes styled "the age of brachiopods."

In Canada, as in the State of New York, the Potsdam Sandstone underlies the above-mentioned calcareous rocks, but contains a different suite of fossils, as will be hereafter explained. In parts of the globe still more remote from Europe the Silurian strata have also been recognised, as in South America, Australia, and India. In all these regions the facies of the fauna, or the types of organic life, enable us to recognise the contemporaneous origin of the rocks; but the fossil species are distinct, showing that the old notion of a universal diffusion throughout the "primaeval seas" of one uniform specific fauna was quite unfounded, geographical provinces having evidently existed in the oldest as in the most modern times.

CHAPTER XXVII.

CAMBRIAN AND LAURENTIAN GROUPS.

Classification of the Cambrian Group, and its Equivalent in Bohemia. Upper Cambrian Rocks. Tremadoc Slates and their Fossils. Lingula Flags. Lower Cambrian Rocks. Menevian Beds. Longmynd Group. Harlech Grits with large Trilobites. Llanberis Slates. Cambrian Rocks of Bohemia. Primordial Zone of Barrande. Metamorphosis of Trilobites. Cambrian Rocks of Sweden and Norway. Cambrian Rocks of the United States and Canada. Potsdam Sandstone. Huronian Series. Laurentian Group, upper and lower. Eozoon Canadense, oldest known Fossil. Fundamental Gneiss of Scotland.

CAMBRIAN GROUP.

The characters of the Upper and Lower Silurian rocks were established so fully, both on stratigraphical and palaeontological data, by Sir Roderick Murchison after five years' labour, in 1839, when his "Silurian System" was published, that these formations could from that period be recognised and identified in all other parts of Europe and in North America, even in countries where most of the fossils differed specifically from those of the classical region in Britain, where they were first studied.

TABLE 27.1. SHOWING THE SUCCESSION OF THE STRATA IN ENGLAND AND WALES WHICH BELONG TO THE CAMBRIAN GROUP OR THE FOSSILIFEROUS ROCKS OLDER THAN THE ARENIG OR LOWER LLANDEILO ROCKS:

UPPER CAMBRIAN.

TREMADOC SLATES. (Primordial of Barrande in part.)

LINGULA FLAGS. (Primordial of Barrande.)

LOWER CAMBRIAN.

MENEVIAN BEDS. (Primordial of Barrande.)

LONGMYND GROUP. a. Harlech Grits. b. Llanberis slates.

While Sir R.I. Murchison was exploring in 1833, in Shropshire and the borders of Wales, the strata which in 1835 he first called Silurian, Professor Sedgwick was surveying the rocks of North Wales, which both these geologists considered at that period as of older date, and for which in 1836 Sedgwick proposed the name of Cambrian. It was afterwards found that a large portion of the slaty rocks of North Wales, which had been considered as more ancient than the Llandeilo beds and Stiper-Stones before alluded to, were, in reality, not inferior in position to those Lower Silurian beds of Murchison, but merely extensive undulations of the same, bearing fossils identical in species, though these were generally rarer and less perfectly preserved, owing to the changes which the rocks had undergone from metamorphic action. To such rocks the term "Cambrian" was no longer applicable, although it continued to be appropriate to strata inferior to the Stiper-Stones, and which were older than those of the Lower Silurian group as originally defined. It was not till 1846 that fossils were found in Wales in the Lingula flags, the place of which will be seen in Table 27.1. By this time Barrande had already published an account of a rich collection of fossils which he had discovered in Bohemia, portions of which he recognised as of corresponding age with Murchison's Upper and Lower Silurian, while others were more ancient, to which he gave the name of "Primordial," for the fossils were sufficiently distinct to entitle the rocks to be referred to a new period. They consisted chiefly of trilobites of genera distinct from those occurring in the overlying Silurian formations. These peculiar genera were afterwards found in rocks holding a corresponding position in Wales, and I shall retain for them the term Cambrian, as recent discoveries in our own country seem to carry the first fauna of Barrande, or his primordial type, even into older strata than any which he found to be fossiliferous in Bohemia.

The term primordial was intended to express M. Barrande's own belief that the fossils of the rocks so-called afforded evidence of the first appearance of vital phenomena on this planet, and that consequently no fossiliferous strata of older date would or could ever be discovered. The acceptance of such a nomenclature would seem to imply that we despaired of extending our discoveries of new and more ancient fossil groups at some future day when vast portions of the globe, hitherto unexplored, should have been thoroughly surveyed. Already the discovery of the Laurentian Eozoon in Canada, presently to be mentioned, discountenances such views.

UPPER CAMBRIAN.

TREMADOC SLATES.

(FIGURE 568. Theca (Cleidotheca) operculata. Lower Tremadoc beds. Tremadoc.)

The Tremadoc slates of Sedgwick are more than 1000 feet in thickness, and consist of dark earthy slates occurring near the little town of Tremadoc, situated on the north side of Cardigan Bay, in Carnarvonshire. These slates were first examined by Sedgwick in 1831, and were re-examined by him and described in 1846 (Quarterly Geological Journal volume 3 page 156.), after some fossils had been found in the underlying Lingula flags by Mr. Davis. The inferiority in position of these Lingula flags to the Tremadoc beds was at the same time established. The overlying Tremadoc beds were traced by their pisolitic ore from Tremadoc to Dolgelly. No fossils proper to the Tremadoc slates were then observed, but subsequently, thirty-six species of all classes have been found in them, thanks to the researches of Messrs. Salter, Homfray, and Ash. We have already seen that in the Arenig or Stiper-Stones group, where the species are distinct, the genera agree with Silurian types; but in these Tremadoc slates, where the species are also peculiar, there is about an equal admixture of Silurian types with those which Barrande has termed "primordial." Here, therefore, it may truly be said that we are entering upon a new domain of life in our retrospective survey of the past. The trilobites of new species, but of Lower Silurian genera, belong to Ogygia, Asaphus, and Cheirurus; whereas those belonging to primordial types, or Barrande's first fauna as well as to the Lingula flags of Wales, comprise Dikelocephalus, Conocoryphe (for genera see Figures 577 and 581 (This genus has been substituted for Barrande's Conocephalus, as the latter term had been preoccupied by the entomologists.)), Olenus, and Angelina. In the Tremadoc slates are found Bellerophon, Orthoceras, and Cyrtoceras, all specifically distinct from Lower Silurian fossils of the same genera: the Pteropods Theca (Figure 568) and Conularia range throughout these slates; there are no Graptolites. The Lingula (Lingulella) Davisii ranges from the top to the bottom of the formation, and links it with the zone next to be described. The Tremadoc slates are very local, and seem to be confined to a small part of North Wales; and Professor Ramsay supposes them to lie unconformably on the Lingula flags, and that a long interval of time elapsed between these formations. Cephalopoda have not yet been found lower than this group, but it will be observed that they occur here associated with genera of Trilobites considered by Barrande as characteristically Primordial, some of which belong to all the divisions of the British Cambrian about to be mentioned. This renders the absence of cephalopoda of less importance as bearing on the theory of development.

LINGULA FLAGS.

(FIGURES 569 to 571. "Lingula flags" of Dolgelly, and Ffestiniog; N. Wales.

(FIGURE 569. Hymenocaris vermicauda, Salter. A phyllopod crustacean. One-half natural size.)

(FIGURE 570. Lingulella Davisii, M'Coy. a. One-half natural size. b. Distorted by cleavage.)

(FIGURE 571. Olenus micrurus, Salter. One-half natural size.))

Next below the Tremadoc slates in North Wales lie micaceous flagstones and slates, in which, in 1846, Mr. E. Davis discovered the Lingula (Lingulella), Figure 570, named after him, and from which was derived the name of Lingula flags. These beds, which are palaeontologically the equivalents of Barrande's primordial zone, are represented by more than 5000 feet of strata, and have been studied chiefly in the neighbourhood of Dolgelly, Ffestiniog, and Portmadoc in North Wales, and at St. David's in South Wales. They have yielded about forty species of fossils, of which six only are common to the overlying Tremadoc rocks, but the two formations are closely allied by having several characteristic "primordial" genera in common. Dikelocephalus, Olenus (Figure 571), and Conocoryphe are prominent forms, as is also Hymenocaris (Figure 569), a genus of phyllopod crustacean entirely confined to the Lingula Flags. According to Mr. Belt, who has devoted much attention to these beds, there are already palaeontological data for subdividing the Lingula Flags into three sections. (Geological Magazine volume 4.)

In Merionethshire, according to Professor Ramsay, the Lingula Flags attain their greatest development; in Carnarvonshire they thin out so as to have lost two- thirds of their thickness in eleven miles, while in Anglesea and on the Menai Straits both they and the Tremadoc beds are entirely absent, and the Lower Silurian rests directly on Lower Cambrian strata.

LOWER CAMBRIAN.

MENEVIAN BEDS.

(FIGURE 572. Paradoxides Davidis, Salter. One-tenth natural size. Menevian beds. St. David's and Dolgelly.)

Immediately beneath the Lingula Flags there occurs a series of dark grey and black flags and slates alternating at the upper part with some beds of sandstone, the whole reaching a thickness of from 500 to 600 feet. These beds were formerly classed, on purely lithological grounds, as the base of the Lingula Flags, but Messrs. Hicks and Salter, to whose exertions we owe almost all our knowledge of the fossils, have pointed out that the most characteristic genera found in them are quite unknown in the Lingula Flags, while they possess many of the strictly Lower Cambrian genera, such as Microdiscus and Paradoxides. (British Association Report 1865, 1866, 1868 and Quarterly Geological Journal volumes 21, 25.) They therefore proposed to place them, and it seems to me with good reason, at the top of the Lower Cambrian under the term "Menevian," Menevia being the classical name of St. David's. The beds are well exhibited in the neighbourhood of St. David's in South Wales, and near Dolgelly and Maentwrog in North Wales. They are the equivalents of the lowest part of Barrande's Primordial Zone (Etage C). More than forty species have been found in them, and the group is altogether very rich in fossils for so early a period. The trilobites are of large size; Paradoxides Davidis (see Figure 572), the largest trilobite known in England, 22 inches or nearly two feet long, is peculiar to the Menevian Beds. By referring to the Bohemian trilobite of the same genus (Figure 576), the reader will at once see how these fossils (though of such different dimensions) resemble each other in Bohemia and Wales, and other closely allied species from the two regions might be added, besides some which are common to both countries. The Swedish fauna, presently to be mentioned, will be found to be still more nearly connected with the Welsh Menevian. In all these countries there is an equally marked difference between the Cambrian fossils and those of the Upper and Lower Silurian rocks. The trilobite with the largest number of rings, Erinnys venulosa, occurs here in conjunction with Agnostus and Microdiscus, the genera with the smallest number. Blind trilobites are also found as well as those which have the largest eyes, such as Microdiscus on the one hand, and Anoplenus on the other.

LONGMYND GROUP.

Older than the Menevian Beds are a thick series of olive green, purple, red and grey grits and conglomerates found in North and South Wales, Shropshire, and parts of Ireland and Scotland. They have been called by Professor Sedgwick the Longmynd or Bangor Group, comprising, first, the Harlech and Barmouth sandstones; and secondly, the Llanberis slates.

HARLECH GRITS.

(FIGURE 573. Histioderma Hibernica, Kinahan. Oldhamia beds. Bray Head, Ireland. 1. Showing opening of burrow, and tube with wrinklings or crossing ridges, probably produced by a tentacled sea worm or annelid. 2. Lower and curved extremity of tube with five transverse lines.)

The sandstones of this period attain in the Longmynd hills a thickness of no less than 6000 feet without any interposition of volcanic matter; in some places in Merionethshire they are still thicker. Until recently these rocks possessed but a very scanty fauna.

With the exception of five species of annelids (see Figure 460) brought to light by Mr. Salter in Shropshire, and Dr. Kinahan in Wicklow, and an obscure crustacean form, Palaeopyge Ramsayi, they were supposed to be barren of organic remains. Now, however, through the labours of Mr. Hicks, they have yielded at St. David's a rich fauna of trilobites, brachiopods, phyllopods, and pteropods, showing, together with other fossils, a by no means low state of organisation at this early period. (British Association Report 1868.) Already the fauna amounts to 20 species referred to 17 genera.

A new genus of trilobite called Plutonia Sedgwickii, not yet figured and described, has been met with in the Harlech grits. It is comparable in size to the large Paradoxides Davidis before mentioned, has well-developed eyes, and is covered all over with tubercles. In the same strata occur other genera of trilobites, namely, Conocoryphe, Paradoxides, Microdiscus, and the Pteropod Theca (Figure 568), all represented by species peculiar to the Harlech grits. The sands of this formation are often rippled, and were evidently left dry at low tides, so that the surface was dried by the sun and made to shrink and present sun-cracks. There are also distinct impressions of rain-drops on many surfaces, like those in Figures 444 and 445.

LANBERIS SLATES.

(FIGURE 574. Oldhamia radiata, Forbes. Wicklow, Ireland.)

(FIGURE 575. Oldhamia antiqua, Forbes. Wicklow, Ireland.)

The slates of Llanberis and Penrhyn in Carnarvonshire, with their associated sandy strata, attain a great thickness, sometimes about 3000 feet. They are perhaps not more ancient than the Harlech and Barmouth beds last mentioned, for they may represent the deposits of fine mud thrown down in the same sea, on the borders of which the sands above-mentioned were accumulating. In some of these slaty rocks in Ireland, immediately opposite Anglesea and Carnarvon, two species of fossils have been found, to which the late Professor E. Forbes gave the name of Oldhamia. The nature of these organisms is still a matter of discussion among naturalists.

CAMBRIAN ROCKS OF BOHEMIA (PRIMORDIAL ZONE OF BARRANDE).

In the year 1846, as before stated, M. Joachim Barrande, after ten years' exploration of Bohemia, and after collecting more than a thousand species of fossils, had ascertained the existence in that country of three distinct faunas below the Devonian. To his first fauna, which was older than any then known in this country, he gave the name of Etage C; his two first stages A and B consisting of crystalline and metamorphic rocks and unfossiliferous schists. This Etage C or primordial zone proved afterwards to be the equivalent of those subdivisions of the Cambrian groups which have been above described under the names of Menevian and Lingula Flags. The second fauna tallies with Murchison's Lower Silurian, as originally defined by him when no fossils had been discovered below the Stiper-Stones. The third fauna agrees with the Upper Silurian of the same author. Barrande, without government assistance, had undertaken single- handed the geological survey of Bohemia, the fossils previously obtained from that country having scarcely exceeded 20 in number, whereas he had already acquired, in 1850, no less than 1100 species, namely, 250 crustaceans (chiefly Trilobites), 250 Cephalopods, 160 gasteropods and pteropods, 130 acephalous mollusks, 210 brachiopods, and 110 corals and other fossils. These numbers have since been almost doubled by subsequent investigations in the same country.

(Figures 576 to 580. Fossils of the lowest Fossiliferous Beds in Bohemia, or "Primordial Zone" of Barrande.

(FIGURE 576. Paradoxides Bohemicus, Barr. About one-half natural size.)

(FIGURE 577. Conocoryphe striata. Syn. Conocephalus striatus, Emmrich. One-half natural size. Ginetz and Skrey.)

(FIGURE 578. Agnostus integer, Beyrich. Natural size and magnified.)

(FIGURE 579. Agnostus Rex, Barr. Natural size, Skrey.)

(FIGURE 580. Sao hirsuta, Barrande, in its various stages of growth. The small lines beneath indicate the true size. In the youngest state, a, no segments are visible; as the metamorphosis progresses, b, c, the body segments begin to be developed: in the stage d the eyes are introduced, but the facial sutures are not completed; at e the full-grown animal, half its true size, is shown.))

In the primordial zone C, he discovered trilobites of the genera Paradoxides, Conocoryphe, Ellipsocephalus, Sao, Arionellus, Hydrocephalus, and Agnostus. M. Barrande pointed out that these primordial trilobites have a peculiar facies of their own dependent on the multiplication of their thoracic segments and the diminution of their caudal shield or pygidium.

One of the "primordial" or Upper Cambrian Trilobites of the genus Sao, a form not found as yet elsewhere in the world, afforded M. Barrande a fine illustration of the metamorphosis of these creatures, for he traced them through no less than twenty stages of their development. A few of these changes have been selected for representation in Figure 580, that the reader may learn the gradual manner in which different segments of the body and the eyes make their appearance.

In Bohemia the primordial fauna of Barrande derived its importance exclusively from its numerous and peculiar trilobites. Besides these, however, the same ancient schists have yielded two genera of brachiopods, Orthis and Orbicula, a Pteropod of the genus Theca, and four echinoderms of the cystidean family.

CAMBRIAN OF SWEDEN AND NORWAY.

The Cambrian beds of Wales are represented in Sweden by strata the fossils of which have been described by a most able naturalist, M. Angelin, in his "Palaeontologica Suecica" (1852-4). The "alum-schists," as they are called in Sweden, are horizontal argillaceous rocks which underlie conformably certain Lower Silurian strata in the mountain called Kinnekulle, south of the great Wener Lake in Sweden. These schists contain trilobites belonging to the genera Paradoxides, Olenus, Agnostus, and others, some of which present rudimentary forms, like the genus last mentioned, without eyes, and with the body segments scarcely developed, and others, again, have the number of segments excessively multiplied, as in Paradoxides. Such peculiarities agree with the characters of the crustaceans met with in the Cambrian strata of Wales; and Dr. Torell has recently found in Sweden the Paradoxides Hicksii, a well-known Lower Cambrian fossil.

At the base of the Cambrian strata in Sweden, which in the neighbourhood of Lake Wener are perfectly horizontal, lie ripple-marked quartzose sandstones with worm-tracks and annelid borings, like some of those found in the Harlech grits of the Longmynd. Among these are some which have been referred doubtfully to plants. These sandstones have been called in Sweden "fucoid sandstones." The whole thickness of the Cambrian rocks of Sweden does not exceed 300 feet from the equivalents of the Tremadoc beds to these sandstones, which last seem to correspond with the Longmynd, and are regarded by Torell as older than any fossiliferous primordial rocks in Bohemia.

CAMBRIAN OF THE UNITED STATES AND CANADA (POTSDAM SANDSTONE).

(FIGURE 581. Dikelocephalus Minnesotensis. Dale Owen. One-third diameter. A large crustacean of the Olenoid group. Potsdam sandstone. Falls of St. Croix, on the Upper Mississippi.)

This formation, as we learn from Sir W. Logan, is 700 feet thick in Canada; the upper part consists of sandstone containing fucoids, and perforated by small vertical holes, which are very characteristic of the rock, and appear to have been made by annelids (Scolithus linearis). The lower portion is a conglomerate with quartz pebbles. I have seen the Potsdam sandstone on the banks of the St. Lawrence, and on the borders of Lake Champlain, where, as at Keesville, it is a white quartzose fine-grained grit, almost passing into quartzite. It is divided into horizontal ripple-marked beds, very like those of the Lingula Flags of Britain, and replete with a small round-shaped Obolella, in such numbers as to divide the rock into parallel planes, in the same manner as do the scales of mica in some micaceous sandstones. Among the shells of this formation in Wisconsin are species of Lingula and Orthis, and several trilobites of the primordial genus Dikelocephalus (Figure 581). On the banks of the St. Lawrence, near Beauharnois and elsewhere, many fossil footprints have been observed on the surface of the rippled layers. They are supposed by Professor Owen to be the trails of more than one species of articulate animal, probably allied to the King Crab, or Limulus.

Recent investigations by the naturalists of the Canadian survey have rendered it certain that below the level of the Potsdam Sandstone there are slates and schists extending from New York to Newfoundland, occupied by a series of trilobitic forms similar in genera, though not in species, to those found in the European Upper Cambrian strata.

HURONIAN SERIES.

Next below the Upper Cambrian occur strata called the Huronian by Sir W. Logan, which are of vast thickness, consisting chiefly of quartzite, with great masses of greenish chloritic slate, which sometimes include pebbles of crystalline rocks derived from the Laurentian formation, next to be described. Limestones are rare in this series, but one band of 300 feet in thickness has been traced for considerable distances to the north of Lake Huron. Beds of greenstone are intercalated conformably with the quartzose and argillaceous members of this series. No organic remains have yet been found in any of the beds, which are about 18,000 feet thick, and rest unconformably on the Laurentian rocks.

LAURENTIAN GROUP.

In the course of the geological survey carried on under the direction of Sir W.E. Logan, it has been shown that, northward of the river St. Lawrence, there is a vast series of crystalline rocks of gneiss, mica-schist, quartzite, and limestone, more than 30,000 feet in thickness, which have been called Laurentian, and which are already known to occupy an area of about 200,000 square miles. They are not only more ancient than the fossiliferous Cambrian formations above described, but are older than the Huronian last mentioned, and had undergone great disturbing movements before the Potsdam sandstone and the other "primordial" or Cambrian rocks were formed. The older half of this Laurentian series is unconformable to the newer portion of the same.

UPPER LAURENTIAN OR LABRADOR SERIES.

The Upper Group, more than 10,000 feet thick, consists of stratified crystalline rocks in which no organic remains have yet been found. They consist in great part of feldspars, which vary in composition from anorthite to andesine, or from those kinds in which there is less than one per cent of potash and soda to those in which there is more than seven per cent of these alkalies, the soda preponderating greatly. These feldsparites sometimes form mountain masses almost without any admixture of other minerals; but at other times they include augite, which passes into hypersthene. They are often granitoid in structure. One of the varieties is the same as the apolescent labradorite rock of Labrador. The Adirondack Mountains in the State of New York are referred to the same series, and it is conjectured that the hypersthene rocks of Skye, which resemble this formation in mineral character, may be of the same geological age.

LOWER LAURENTIAN.

This series, about 20,000 feet in thickness, is, as before stated, unconformable to that last mentioned; it consists in great part of gneiss of a reddish tint with orthoclase feldspar. Beds of nearly pure quartz, from 400 to 600 feet thick, occur in some places. Hornblendic and micaceous schists are often interstratified, and beds of limestone, usually crystalline. Beds of plumbago also occur. That this pure carbon may have been of organic origin before metamorphism has naturally been conjectured.

(FIGURES 582 and 583. Eozoon Canadense, Daw. (after Carpenter). Oldest known organic body.

(FIGURE 582. Eozoon Canadense, Daw. (after Carpenter). Oldest known organic body. a. Chambers of lower tier communicating at +, and separated from adjoining chambers at o by an intervening septum, traversed by passages. b. Chambers of an upper tier. c. Walls of the chambers traversed by fine tubules. (These tubules pass with uniform parallelism from the inner to the outer surface, opening at regular distances from each other.) d. Intermediate skeleton, composed of homogeneous shell substance, traversed by f. Stoloniferous passages connecting the chambers of the two tiers. e. Canal system in intermediate skeleton, showing the arborescent saceodic prolongations. (Figure 583 shows these bodies in a decalcified state.))

(FIGURE 583. Eozoon Canadense, Daw. (after Carpenter). Oldest known organic body. Decalcified portion of natural rock, showing CANAL SYSTEM and the several layers; the acuteness of the planes prevents more than one or two parallel tiers being observed. Natural size.))

There are several of these limestones which have been traced to great distances, and one of them is from 700 to 1500 feet thick. In the most massive of them Sir W. Logan observed, in 1859, what he considered to be an organic body much resembling the Silurian fossil called Stromatopora rugosa. It had been obtained the year before by Mr. J. MacMullen at the Grand Calumet, on the river Ottawa. This fossil was examined in 1864 by Dr. Dawson of Montreal, who detected in it, by aid of the microscope, the distinct structure of a Rhizopod or Foraminifer. Dr. Carpenter and Professor T. Rupert Jones have since confirmed this opinion, comparing the structure to that of the well-known nummulite. It appears to have grown one layer over another, and to have formed reefs of limestone as do the living coral-building polyp animals. Parts of the original skeleton, consisting of carbonate of lime, are still preserved; while certain inter-spaces in the calcareous fossil have been filled up with serpentine and white augite. On this oldest of known organic remains Dr. Dawson has conferred the name of Eozoon Canadense (see Figures 582, 583); its antiquity is such that the distance of time which separated it from the Upper Cambrian period, or that of the Potsdam sandstone, may, says Sir W. Logan, be equal to the time which elapsed between the Potsdam sandstone and the nummulitic limestones of the Tertiary period. The Laurentian and Huronian rocks united are about 50,000 feet in thickness, and the Lower Laurentian was disturbed before the newer series was deposited. We may naturally expect the other proofs of unconformability will hereafter be detected at more than one point in so vast a succession of strata.

The mineral character of the Upper Laurentian differs, as we have seen, from that of the Lower, and the pebbles of gneiss in the Huronian conglomerates are thought to prove that the Laurentian strata were already in a metamorphic state before they were broken up to supply materials for the Huronian. Even if we had not discovered the Eozoon, we might fairly have inferred from analogy that as the quartzites were once beds of sand, and the gneiss and mica-schist derived from shales and argillaceous sandstones, so the calcareous masses, from 400 to 1000 feet and more in thickness, were originally of organic origin. This is now generally believed to have been the case with the Silurian, Devonian, Carboniferous, Oolitic, and Cretaceous limestones and those nummulitic rocks of tertiary date which bear the closest affinity to the Eozoon reefs of the Lower Laurentian. The oldest stratified rock in Scotland is that called by Sir R. Murchison "the fundamental gneiss," which is found in the north-west of Ross- shire, and in Sutherlandshire (see Figure 82), and forms the whole of the adjoining island of Lewis, in the Hebrides. It has a strike from north-west to south-east, nearly at right angles to the metamorphic strata of the Grampians. On this Laurentian gneiss, in parts of the western Highlands, the Lower Cambrian and various metamorphic rocks rest unconformably. It seems highly probable that this ancient gneiss of Scotland may correspond in date with part of the great Laurentian group of North America.

CHAPTER XXVIII.

VOLCANIC ROCKS.

External Form, Structure, and Origin of Volcanic Mountains. Cones and Craters. Hypothesis of "Elevation Craters" considered. Trap Rocks. Name whence derived. Minerals most abundant in Volcanic Rocks. Table of the Analysis of Minerals in the Volcanic and Hypogene Rocks. Similar Minerals in Meteorites. Theory of Isomorphism. Basaltic Rocks. Trachytic Rocks. Special Forms of Structure. The columnar and globular Forms. Trap Dikes and Veins. Alteration of Rocks by volcanic Dikes. Conversion of Chalk into Marble. Intrusion of Trap between Strata. Relation of trappean Rocks to the Products of active Volcanoes.

(FIGURE 584. Section through formations from a, low, to c, high. a. Hypogene formations, stratified and unstratified. b. Aqueous formations. c. Volcanic rocks.)

The aqueous or fossiliferous rocks having now been described, we have next to examine those which may be called volcanic, in the most extended sense of that term. In the diagram (Figure 584) suppose a, a to represent the crystalline formations, such as the granitic and metamorphic; b, b the fossiliferous strata; and c, c the volcanic rocks. These last are sometimes found, as was explained in the first chapter, breaking through a and b, sometimes overlying both, and occasionally alternating with the strata b, b.

EXTERNAL FORM, STRUCTURE, AND ORIGIN OF VOLCANIC MOUNTAINS.

The origin of volcanic cones with crater-shaped summits has been explained in the "Principles of Geology" (Chapters 23 to 27), where Vesuvius, Etna, Santorin, and Barren Island are described. The more ancient portions of those mountains or islands, formed long before the times of history, exhibit the same external features and internal structure which belong to most of the extinct volcanoes of still higher antiquity; and these last have evidently been due to a complicated series of operations, varied in kind according to circumstances; as, for example, whether the accumulation took place above or below the level of the sea, whether the lava issued from one or several contiguous vents, and, lastly, whether the rocks reduced to fusion in the subterranean regions happened to have contained more or less silica, potash, soda, lime, iron, and other ingredients. We are best acquainted with the effects of eruptions above water, or those called subaerial or supramarine; yet the products even of these are arranged in so many ways that their interpretation has given rise to a variety of contradictory opinions, some of which will have to be considered in this chapter.

CONES AND CRATERS.

(FIGURE 585. Part of the chain of extinct volcanoes called the Monts Dome, Auvergne. (Scrope.))

In regions where the eruption of volcanic matter has taken place in the open air, and where the surface has never since been subjected to great aqueous denudation, cones and craters constitute the most striking peculiarity of this class of formations. Many hundreds of these cones are seen in central France, in the ancient provinces of Auvergne, Velay, and Vivarais, where they observe, for the most part, a linear arrangement, and form chains of hills. Although none of the eruptions have happened within the historical era, the streams of lava may still be traced distinctly descending from many of the craters, and following the lowest levels of the existing valleys. The origin of the cone and crater- shaped hill is well understood, the growth of many having been watched during volcanic eruptions. A chasm or fissure first opens in the earth, from which great volumes of steam are evolved. The explosions are so violent as to hurl up into the air fragments of broken stone, parts of which are shivered into minute atoms. At the same time melted stone or LAVA usually ascends through the chimney or vent by which the gases make their escape. Although extremely heavy, this lava is forced up by the expansive power of entangled gaseous fluids, chiefly steam or aqueous vapour, exactly in the same manner as water is made to boil over the edge of a vessel when steam has been generated at the bottom by heat. Large quantities of the lava are also shot up into the air, where it separates into fragments, and acquires a spongy texture by the sudden enlargement of the included gases, and thus forms SCORIAE, other portions being reduced to an impalpable powder or dust. The showering down of the various ejected materials round the orifice of eruption gives rise to a conical mound, in which the successive envelopes of sand and scoriae form layers, dipping on all sides from a central axis. In the mean time a hollow, called a CRATER, has been kept open in the middle of the mound by the continued passage upward of steam and other gaseous fluids. The lava sometimes flows over the edge of the crater, and thus thickens and strengthens the sides of the cone; but sometimes it breaks down the cone on one side (see Figure 585), and often it flows out from a fissure at the base of the hill, or at some distance from its base.

Some geologists had erroneously supposed, from observations made on recent cones of eruption, that lava which consolidates on steep slopes is always of a scoriaceous or vesicular structure, and never of that compact texture which we find in those rocks which are usually termed "trappean." Misled by this theory, they have gone so far as to believe that if melted matter has originally descended a slope at an angle exceeding four or five degrees, it never, on cooling, acquires a stony compact texture. Consequently, whenever they found in a volcanic mountain sheets of stony materials inclined at angles of from 5 degrees to 20 degrees or even more than 30 degrees, they thought themselves warranted in assuming that such rocks had been originally horizontal, or very slightly inclined, and had acquired their high inclination by subsequent upheaval. To such dome-shaped mountains with a cavity in the middle, and with the inclined beds having what was called a quaquaversal dip or a slope outward on all sides, they gave the name of "Elevation craters."

As the late Leopold Von Buch, the author of this theory, had selected the Isle of Palma, one of the Canaries, as a typical illustration of this form of volcanic mountain, I visited that island in 1854, in company with my friend Mr. Hartung, and I satisfied myself that it owes its origin to a series of eruptions of the same nature as those which formed the minor cones, already alluded to. In some of the more ancient or Miocene volcanic mountains, such as Mont Dor and Cantal in central France, the mode of origin by upheaval as above described is attributed to those dome-shaped masses, whether they possess or not a great central cavity, as in Palma. Where this cavity is present, it has probably been due to one or more great explosions similar to that which destroyed a great part of ancient Vesuvius in the time of Pliny. Similar paroxysmal catastrophes have caused in historical times the truncation on a grand scale of some large cones in Java and elsewhere. (Principles volume 2 pages 56 and 145.)

Among the objections which may be considered as fatal to Von Buch's doctrine of upheaval in these cases, I may state that a series of volcanic formations extending over an area six or seven miles in its shortest diameter, as in Palma, could not be accumulated in the form of lavas, tuffs, and volcanic breccias or agglomerates without producing a mountain as lofty as that which they now constitute. But assuming that they were first horizontal, and then lifted up by a force acting most powerfully in the centre and tilting the beds on all sides, a central crater having been formed by explosion or by a chasm opening in the middle, where the continuity of the rocks was interrupted, we should have a right to expect that the chief ravines or valleys would open towards the central cavity, instead of which the rim of the great crater in Palma and other similar ancient volcanoes is entire for more than three parts of the whole circumference.

If dikes are seen in the precipices surrounding such craters or central cavities, they certainly imply rents which were filled up with liquid matter. But none of the dislocations producing such rents can have belonged to the supposed period of terminal and paroxysmal upheaval, for had a great central crater been already formed before they originated, or at the time when they took place, the melted matter, instead of filling the narrow vents, would have flowed down into the bottom of the cavity, and would have obliterated it to a certain extent. Making due allowance for the quantity of matter removed by subaerial denudation in volcanic mountains of high antiquity, and for the grand explosions which are known to have caused truncation in active volcanoes, there is no reason for calling in the violent hypothesis of elevation craters to explain the structure of such mountains as Teneriffe, the Grand Canary, Palma, or those of central France, Etna, or Vesuvius, all of which I have examined. With regard to Etna, I have shown, from observations made by me in 1857, that modern lavas, several of them of known date, have formed continuous beds of compact stone even on slopes of 15, 36, and 38 degrees, and, in the case of the lava of 1852, more than 40 degrees. The thickness of these tabular layers varies from 1 1/2 foot to 26 feet. And their planes of stratification are parallel to those of the overlying and underlying scoriae which form part of the same currents. (Memoir on Mount Etna Philosophical Transactions 1858.)

NOMENCLATURE OF TRAPPEAN ROCKS.

When geologists first began to examine attentively the structure of the northern and western parts of Europe, they were almost entirely ignorant of the phenomena of existing volcanoes. They found certain rocks, for the most part without stratification, and of a peculiar mineral composition, to which they gave different names, such as basalt, greenstone, porphyry, trap tuff, and amygdaloid. All these, which were recognised as belonging to one family, were called "trap" by Bergmann, from trappa, Swedish for a flight of steps— a name since adopted very generally into the nomenclature of the science; for it was observed that many rocks of this class occurred in great tabular masses of unequal extent, so as to form a succession of terraces or steps. It was also felt that some general term was indispensable, because these rocks, although very diversified in form and composition, evidently belonged to one group, distinguishable from the Plutonic as well as from the non-volcanic fossiliferous rocks.

By degrees familiarity with the products of active volcanoes convinced geologists more and more that they were identical with the trappean rocks. In every stream of modern lava there is some variation in character and composition, and even where no important difference can be recognised in the proportions of silica, alumina, lime, potash, iron, and other elementary materials, the resulting materials are often not the same, for reasons which we are as yet unable to explain. The difference also of the lavas poured out from the same mountain at two distinct periods, especially in the quantity of silica which they contain, is often so great as to give rise to rocks which are regarded as forming distinct families, although there may be every intermediate gradation between the two extremes, and although some rocks, forming a transition from the one class to the other, may often be so abundant as to demand special names. These species might be multiplied indefinitely, and I can only afford space to name a few of the principal ones, about the composition and aspect of which there is the least discordance of opinion.

MINERALS MOST ABUNDANT IN VOLCANIC ROCKS.

TABLE 28.1. ANALYSIS OF MINERALS MOST ABUNDANT IN THE VOLCANIC AND HYPOGENE ROCKS.

COLUMN 1: SILICA.

COLUMN 2: ALUMINA.

COLUMN 3: SESQUIOXIDE OF IRON.

COLUMN 4: PROTOXIDES OF IRON AND MANGANESE.

COLUMN 5: LIME.

COLUMN 6: MAGNESIA.

COLUMN 7: POTASH.

COLUMN 8: SODA.

COLUMN 9: OTHER CONSTITUENTS. In this column the following signs are used: F. Fluorine; Li. Lithia; W. Loss on igniting the mineral, in most instances only Water.

COLUMN 10: SPECIFIC GRAVITY.

— — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — -

THE QUARTZ GROUP:

1 2 3 4 5 6 7 8 9 10.

Quartz: 100.0 2.6.

Tridymite: 100.0 2.3.

— — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — -

THE FELDSPAR GROUP:

1 2 3 4 5 6 7 8 9 10.

Orthoclase. Carlsbad, in granite (Bulk):

65.23 18.26 0.27 .... trace .... 14.66 1.45 .... 2.55.

Orthoclase. Sanadine, Drachenfels in trachyte (Rammelsberg).

65.87 18.53 .... .... 0.95 0.30 10.32 3.42 W 0.44 2.55.

Albite. Arendal, in granite (G. Rose).

68.46 19.30 .... 0.28 0.68 .... .... 11.27 .... 2.61.

Oligoclase. Ytterby, in granite (Berzelius).

61.55 23.80 .... .... 3.18 0.80 0.38 9.67 .... 2.65.

Oligoclase. Teneriffe, in trachyte (Deville).

61.55 22.03 .... .... 2.81 0.47 3.44 7.74 .... 2.59.

Labradorite. Hitteroe, in Labrador-Rock (Waage).

51.39 29.42 2.90 .... 9.44 0.37 1.10 5.03 W 0.71 2.72.

Labradorite. Iceland, in volcanic (Damour).

52.17 29.22 1.90 .... 13.11 .... .... 3.40 .... 2.71.

Anorthite. Harzburg, in diorite (Streng).

45.37 34.81 0.59 .... 16.52 0.83 0.40 1.45 W 0.87 2.74.

Anorthite. Hecla, in volcanic (Waltershausen).

45.14 32.10 2.03 0.78 18.32 .... 0.22 1.06 .... 2.74.

— — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — -

Leucite. Vesuvius, 1811, in lava (Rammelsberg).

56.10 23.22 .... .... .... .... 20.59 0.57 .... 2.48.

Nepheline. Miask, in Miascite (Scheerer).

44.30 33.25 0.82 .... 0.32 0.07 5.82 16.02 .... 2.59.

Nepheline. Vesuvius, in volcanic (Arfvedson).

44.11 33.73 .... .... .... .... .... 20.46 W 0.62 2.60. — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — -

THE MICA GROUP:

1 2 3 4 5 6 7 8 9 10.

Muscovite. Finland, in granite (Rose).

46.36 36.80 4.53 .... .... .... 9.22 .... F 0.67 2.90. W 1.84.

Lepidolite. Cornwall, in granite (Regnault).

52.40 26.80 .... 1.50 .... .... 9.14 .... F 4.18 2.90. Li 4.85.

Biotite. Bodennais (V. Kobell).

40.86 15.13 13.00 .... .... 22.00 8.83 .... W 0.44 2.70.

Biotite. Vesuvius, in volcanic (Chodnef).

40.91 17.71 11.02 .... 0.30 19.04 9.96 .... .... 2.75.

Phlogopite. New York, in metamorphic limestone (Rammelsberg).

41.96 13.47 .... 2.67 0.34 27.12 9.37 .... F 2.93 2.81. W 0.60.

— — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — —

Margarite. Nexos (Smith).

30.02 49.52 1.65 .... 10.82 0.48 1.25 W 5.55 2.99.

Chlorite. Dauphiny (Marignac).

26.88 17.52 29.76 .... .... 13.84 .... .... W 11.33 2.87.

Rapidolite. Pyrenees (Delesse).

32.10 18.50 .... 0.06 .... 36.70 .... .... W 12.10 2.61.

Talc. Zillerthal (Delesse).

63.00 .... .... trace .... 33.60 .... .... W 3.10 2.78.

— — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — —

THE AMPHIBOLE AND PYROXENE GROUP.

1 2 3 4 5 6 7 8 9 10.

Tremolite. St. Gothard (Rammelsberg)

58.55 .... .... .... 13.90 26.63 .... .... FW 0.34 2.93.

Actinolite. Arendal, in granite (Rammelsberg).

56.77 0.97 .... 5.88 13.56 21.48 .... .... W 2.20 3.02.

Hornblende. Faymont, in diorite (Deville).

41.99 11.66 .... 22.22 9.55 12.59 .... 1.02 W 1.47 3.20.

Hornblende Etna, in volcanic (Waltershausen).

40.91 13.68 .... 17.49 13.44 13.19 .... .... W 0.85 3.01.

— — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — —

Uralite. Ural (Rammelsberg)

50.75 5.65 .... 17.27 11.59 12.28 .... .... W 1.80 3.14.

— — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — —

Augite. Bohemia, in dolerite (Rammelsberg).

51.12 3.38 0.95 8.08 23.54 12.82 .... .... .... 3.35.

Augite. Vesuvius, in lava of 1858 (Rammelsberg).

49.61 4.42 .... 9.08 22.83 14.22 .... .... .... 3.25.

Diallage. Harz, in Gabbro (Rammelsberg).

52.00 3.10 .... 9.36 16.29 18.51 .... .... W 1.10 3.23.

Hypersthene. Labrador, in Labrador-Rock (Damour).

51.36 0.37 .... 22.59 3.09 21.31 .... .... .... 3.39.

— — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — —

THE OLIVINE GROUP.

1 2 3 4 5 6 7 8 9 10

Bronzite. Greenland (V. Kobell).

58.00 1.33 11.14 .... .... 29.66 .... .... .... 3.20.

Olivine. Carlsbad, in basalt (Rammelsberg).

39.34 .... .... 14.85 .... 45.81 .... .... .... 3.40.

Olivine. Mount Somma, in volcanic (Walmstedt).

40.08 0.18 .... 15.74 .... 44.22 .... .... .... 3.33.

— — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — —

The minerals which form the chief constituents of these igneous rocks are few in number. Next to quartz, which is nearly pure silica or silicic acid, the most important are those silicates commonly classed under the several heads of feldspar, mica, hornblende or augite, and olivine. In Table 28.1, in drawing up which I have received the able assistance of Mr. David Forbes, the chemical analysis of these minerals and their varieties is shown, and he has added the specific gravity of the different mineral species, the geological application of which in determining the rocks formed by these minerals will be explained in the sequel.

From Table 28.1 it will be observed that many minerals are omitted which, even if they are of common occurrence, are more to be regarded as accessory than as essential components of the rocks in which they are found. (For analyses of these minerals see the Mineralogies of Dana and Bristow.) Such are, for example, Garnet, Epidote, Tourmaline, Idocrase, Andalusite, Scapolite, the various Zeolites, and several other silicates of somewhat rarer occurrence. Magnetite, Titanoferrite, and Iron-pyrites also occur as normal constituents of various igneous rocks, although in very small amount, as also Apatite, or phosphate of lime. The other salts of lime, including its carbonate or calcite, although often met with, are invariably products of secondary chemical action.

The Zeolites, above mentioned, so named from the manner in which they froth up under the blow-pipe and melt into a glass, differ in their chemical composition from all the other mineral constituents of volcanic rocks, since they are hydrated silicates containing from 10 to 25 per cent of water. They abound in some trappean rocks and ancient lavas, where they fill up vesicular cavities and interstices in the substance of the rocks, but are rarely found in any quantity in recent lavas; in most cases they are to be regarded as secondary products formed by the action of water on the other constituents of the rocks. Among them the species Analcime, Stilbite, Natrolite, and Chabazite may be mentioned as of most common occurrence.

QUARTZ GROUP.

The microscope has shown that pure quartz is oftener present in lavas than was formerly supposed. It had been argued that the quartz in granite having a specific gravity of 2.6, was not of purely igneous origin, because the silica resulting from fusion in the laboratory has only a specific gravity of 2.3. But Mr. David Forbes has ascertained that the free quartz in trachytes, which are known to have flowed as lava, has the same specific gravity as the ordinary quartz of granite; and the recent researches of Von Rath and others prove that the mineral Tridymite, which is crystallised silica of specific gravity 2.3 (see Table 28.1), is of common occurrence in the volcanic rocks of Mexico, Auvergne, the Rhine, and elsewhere, although hitherto entirely overlooked.

FELDSPAR GROUP.

In the Feldspar group (Table 28.1) the five mineral species most commonly met with as rock constituents are: 1. Orthoclase, often called common or potash- feldspar. 2. Albite, or soda-feldspar, a mineral which plays a more subordinate part than was formerly supposed, this name having been given to much which has since been proved to be Oligoclase. 3. Oligoclase, or soda-lime feldspar, in which soda is present in much larger proportion than lime, and of which mineral andesite are andesine, is considered to be a variety. 4. Labradorite, or lime- soda-feldspar, in which the proportions of lime and soda are the reverse to what they are in Oligoclase. 5. Anorthite or lime-feldspar. The two latter feldspars are rarely if ever found to enter into the composition of rocks containing quartz.

In employing such terms as potash-feldspar, etc., it must, however, always be borne in mind that it is only intended to direct attention to the predominant alkali or alkaline earth in the mineral, not to assert the absence of the others, which in most cases will be found to be present in minor quantity. Thus potash-feldspar (orthoclase) almost always contains a little soda, and often traces of lime or magnesia; and in like manner with the others. The terms "glassy" and "compact" feldspars only refer to structure, and not to species or composition; the student should be prepared to meet with any of the above feldspars in either of these conditions: the glassy state being apparently due to quick cooling, and the compact to conditions unfavourable to crystallisation; the so-called "compact feldspar" is also very commonly found to be an admixture of more than one feldspar species, and frequently also contains quartz and other extraneous mineral matter only to be detected by the microscope.

Feldspars when arranged according to their system of crystallisation are MONOCLINIC, having one axis obliquely inclined; or TRICLINIC, having the three axes all obliquely inclined to each other. If arranged with reference to their cleavage they are ORTHOCLASTIC, the fracture taking place always at a right angle; or PLAGIOCLASTIC, in which the cleavages are oblique to one another. Orthoclase is orthoclastic and monoclinic; all the other feldspars are plagioclastic and triclinic.

MINERALS IN METEORITES.

That variety of the Feldspar Group which is called Anorthite has been shown by Rammelsberg to occur in a meteoric stone, and his analysis proves it to be almost identical in its chemical proportions to the same mineral in the lavas of modern volcanoes. So also Bronzite (Enstatite) and Olivine have been met with in meteorites shown by analysis to come remarkably near to these minerals in ordinary rocks.

MICA GROUP.

With regard to the micas, the four principal species (Table 28.1) all contain potash in nearly the same proportion, but differ greatly in the proportion and nature of their other ingredients. Muscovite is often called common or potash mica; Lepidolite is characterised by containing lithia in addition; Biotite contains a large amount of magnesia and oxide of iron; whilst Phlogopite contains still more of the former substance. In rocks containing quartz, muscovite or lepidolite are most common. The mica in recent volcanic rocks, gabbros, and diorites is usually Biotite, while that so common in metamorphic limestones is usually, if not always, Phlogopite.

AMPHIBOLE AND PYROXENE GROUP.

The minerals included in Table 28.1 under the Amphibole and Pyroxene Group differ somewhat in their crystallisation form, though they all belong to the monoclinic system. Amphibole is a general name for all the different varieties of Hornblende, Actinolite, Tremolite, etc., while Pyroxene includes Augite, Diallage, Malacolite, Sahlite, etc. The two divisions are so much allied in chemical composition and crystallographic characters, and blend so completely one into the other in Uralite, that it is perhaps best to unite them in one group.

THEORY OF ISOMORPHISM.

The history of the changes of opinion on this point is curious and instructive. Werner first distinguished augite from hornblende; and his proposal to separate them obtained afterwards the sanction of Hauy, Mohs, and other celebrated mineralogists. It was agreed that the form of the crystals of the two species was different, and also their structure, as shown by CLEAVAGE— that is to say, by breaking or cleaving the mineral with a chisel, or a blow of the hammer, in the direction in which it yields most readily. It was also found by analysis that augite usually contained more lime, less alumina, and no fluoric acid; which last, though not always found in hornblende, often enters into its composition in minute quantity. In addition to these characters, it was remarked as a geological fact, that augite and hornblende are very rarely associated together in the same rock. It was also remarked that in the crystalline slags of furnaces augitic forms were frequent, the hornblendic entirely absent; hence it was conjectured that hornblende might be the result of slow, and augite of rapid cooling. This view was confirmed by the fact that Mitscherlich and Berthier were able to make augite artificially, but could never succeed in forming hornblende. Lastly, Gustavus Rose fused a mass of hornblende in a porcelain furnace, and found that it did not, on cooling, assume its previous shape, but invariably took that of augite. The same mineralogist observed certain crystals called Uralite (see Table 28.1) in rocks from Siberia, which possessed the cleavage and chemical composition of hornblende, while they had the external form of augite.

If, from these data, it is inferred that the same substance may assume the crystalline forms of hornblende or augite indifferently, according to the more or less rapid cooling of the melted mass, it is nevertheless certain that the variety commonly called augite, and recognised by a peculiar crystalline form, has usually more lime in it, and less alumina, than that called hornblende, although the quantities of these elements do not seem to be always the same. Unquestionably the facts and experiments above mentioned show the very near affinity of hornblende and augite; but even the convertibility of one into the other, by melting and recrystallising, does not perhaps demonstrate their absolute identity. For there is often some portion of the materials in a crystal which are not in perfect chemical combination with the rest. Carbonate of lime, for example, sometimes carries with it a considerable quantity of silex into its own form of crystal, the silex being mechanically mixed as sand, and yet not preventing the carbonate of lime from assuming the form proper to it. This is an extreme case, but in many others some one or more of the ingredients in a crystal may be excluded from perfect chemical union; and after fusion, when the mass recrystallises, the same elements may combine perfectly or in new proportions, and thus a new mineral may be produced. Or some one of the gaseous elements of the atmosphere, the oxygen for example, may, when the melted matter reconsolidates, combine with some one of the component elements.

The different quantity of the impurities or the refuse above alluded to, which may occur in all but the most transparent and perfect crystals, may partly explain the discordant results at which experienced chemists have arrived in their analysis of the same mineral. For the reader will often find that crystals of a mineral determined to be the same by physical characters, crystalline form, and optical properties, have been declared by skilful analysers to be composed of distinct elements. This disagreement seemed at first subversive of the atomic theory, or the doctrine that there is a fixed and constant relation between the crystalline form and structure of a mineral and its chemical composition. The apparent anomaly, however, which threatened to throw the whole science of mineralogy into confusion, was reconciled to fixed principles by the discoveries of Professor Mitscherlich at Berlin, who ascertained that the composition of the minerals which had appeared so variable was governed by a general law, to which he gave the name of ISOMORPHISM (from isos, equal, and morphe, form). According to this law, the ingredients of a given species of mineral are not absolutely fixed as to their kind and quality; but one ingredient may be replaced by an equivalent portion of some analogous ingredient. Thus, in augite, the lime may be in part replaced by portions of protoxide of iron, or of manganese, while the form of the crystal, and the angle of its cleavage planes, remain the same. These vicarious substitutions, however, of particular elements can not exceed certain defined limits.

BASALTIC ROCKS.

The two principal families of trappean or volcanic rocks are the basalts and the trachytes, which differ chiefly from each other in the quantity of silica which they contain. The basaltic rocks are comparatively poor in silica, containing less than 50 per cent of that mineral, and none in a pure state or as free quartz, apart from the rest of the matrix. They contain a larger proportion of lime and magnesia than the trachytes, so that they are heavier, independently of the frequent presence of the oxides of iron which in some cases forms more than a fourth part of the whole mass. Abich has, therefore, proposed that we should weigh these rocks, in order to appreciate their composition in cases where it is impossible to separate their component minerals. Thus, basalt from Staffa, containing 47.80 per cent of silica, has a specific gravity of 2.95; whereas trachyte, which has 66 per cent of silica, has a specific gravity of only 2.68; trachytic porphyry, containing 69 per cent of silica, a specific gravity of only 2.58. If we then take a rock of intermediate composition, such as that prevailing in the Peak of Teneriffe, which Abich calls Trachyte-dolerite, its proportion of silica being intermediate, or 58 per cent, it weighs 2.78, or more than trachyte, and less than basalt. (Dr. Daubeny on Volcanoes second edition pages 14, 15.)