the opinion expressed by Sir R. Murchison, that a large part of the Arctic region was land in Lower Silurian times, because no rocks of that age had been discovered there. Nor have any been found since this opinion was expressed, though (Upper) Silurian rocks cover large areas in the north-western part of the Arctic Archipelago.

It seems probable, therefore, that the great northern continent of early Cambrian times was partially broken up by the submergence which ensued, and that an open sea extending from North America to Scotland and Scandinavia was established for the first time at the beginning of the Ordovician period; that this sea divided the Cambrian continent into two parts, the larger part lying over Greenland and the northern regions, while a smaller mass lay to the south, and stretched eastward through the north of Ireland and into the south of Scotland, so as to separate the northern sea from that which lay over the European area. The thick conglomerates of Llandeilo and Bala age which occur in the Girvan district may indicate a part of its southern border, but of its northern shore we have absolutely no evidence at present. Mr. J. E. Marr takes a somewhat similar view,' and has suggested the existence of a ridge of pre-Cambrian land across the centre of Scotland in Arenig times, separating the sea in which the Orthoceras and Durness limestones were formed from that of the Girvan area. It is possible, therefore, that the land barrier had a still further eastward extension, but until more is known of the geology of the central Highlands this must remain uncertain, and the form given to the land in the accompanying map is of course merely a suggestion based upon the preceding considerations. The map is in fact only a pictorial representation of the theory suggested by the facts which are known to us, and may require considerable modification as our knowledge increases, and the same may 1 "Classif. of Camb. and Sil. Rocks," p. 68.

be said of the succeeding restorations of Paleozoic geographies, so little do we really know of these rocks and their extension beneath the British Islands.

During the deposition of the Llandeilo and Bala rocks depression continued, and the whole of the British area seems to have been submerged, with the exception of a portion of the central English island; the dimensions of this island were of course considerably diminished, as testified by the extension of the Caradoc sandstones to the east of the Longmynd, but these do not extend into Staffordshire or Warwickshire. Llandeilo and Bala rocks seem to have been laid down over the whole of Ireland, and if any part of the land barrier above mentioned remained above water it must have been outside the present limits of Ireland.

Numerous volcanic islands, however, came into existence during this period, and portions of the lava-streams which they emitted are interbedded with the sedimentary rocks of Llandeilo and Bala age. In Llandeilo times a line of such islands seems to have stretched across Ireland and the north of England in a south-west and north-east direction, while in Bala times great eruptions took place from a group of volcanoes in Wales and the east of Ireland. The period was brought to a close by an upheaval which brought up the submerged ground, but gave it a different outline and extension from that which it possessed in Arenig times.

CHAPTER IV.

SILURIAN PERIOD.

1. Stratigraphical Evidence.

OMMENCING with what is usually regarded as the

typical Silurian district, it may be stated that the Silurian rocks form a continuous escarpment through Shropshire and the east of Wales, and that they rise to the surface at intervals in the counties of Gloucester, Monmouth, Hereford, and Stafford.

The lowest beds in this district are sandstones, shales, and conglomerates; this (Llandovery) division is thickest in the western part of the area, and is there divisible into two groups, with a combined thickness of nearly 2,000 feet, but the upper stage frequently overlaps the lower, and it is the only one found in the more eastern outcrops, where they include beds of conglomerate; these basal conglomerates vary much in thickness, and sometimes overlap the Ordovician strata, as in the Longmynd and Malvern areas, where they rest upon the Cambrian rocks.

The Llandovery sandstones are succeeded by a great series of dark grey shales, in which limestones are developed at intervals; some of these limestones attain a thickness of 100 feet, but they are all lenticular deposits, and it is often difficult to identify the beds seen in one detached area with those which occur in the others. Only three of these limestone bands have received distinctive names,

these being respectively, in upward succession, the Woolhope, the Wenlock, and the Aymestry limestones, but at Tortworth, May Hill, and Malvern, there is a fourth limestone between the Wenlock and the representative of the Aymestry, which might be regarded as a distinct horizon. This great shale and limestone series is from 2,000 to 3,000 feet thick.

The highest Silurian beds in this region consist of yellow sandstones and reddish shales, the latter passing up into red laminated micaceous sandstones and marls, some of which contain marine fossils; these beds form a passage from Silurian to Old Red Sandstone, and were evidently formed in shallower water than the dark mudstones and limestones of the Wenlock series.

As regards the eastward extension of the Silurian rocks very little is yet known; the most easterly outcrop is near Walsall and Barr in South Staffordshire, but Wenlock shales have been found below the cretaceous rocks in a deep boring at Ware, near Hertford, so it is probable that they have a wide subterranean extension beneath the south and east of England.

area.

The subdivisions which have been established in the limited area where the Silurian rocks were first described, will not apply to the rocks of the same age outside that When traced in any westward direction, a great change takes place in the lithological character of the strata: all the limestones thin out and disappear, and the soft mudstones change into a series of hard shales, flags, and grits. In central and northern Wales these alternations of shales and sandstones attain a thickness of from 5,000 to 7,000 feet, and this change suggests the existence of a considerable mass of land to the westward.

The Silurian rocks of the Lake District are similar to those of North Wales-they have a conglomerate at the base which lies unconformably upon Ordovician rocks, and

to this succeeds an immense series of shales, grits, and flags, which are no less than 13,000 feet thick. Of this total only 250 feet belong to the lower or Valentian group, the great mass of the beds (11,000 feet) representing the Salopian shale and limestone group, while the upper 2,000 feet represent the red passage beds of the Welsh border. Although no Upper Silurian beds are found in Cumberland north of their Coniston and Windermere outcrop, yet they set in again on the north side of the Solway, and, as the late Mr. J. C. Ward writes,1 "there is every reason to believe that the whole series once extended over the now exposed volcanic rocks, for there is nothing in the deposits themselves to indicate a land margin near their present outcrop. Such a thickness of beds as that just described implies a continued subsidence of the sea-bed throughout the whole period of deposition."

The prevalence of sandstones and flagstones, however, shows that the water was never very deep or very far from land, and we are guided to the direction in which some of this land lay by considering the lithological changes which the lowest group exhibits when traced in different directions. In Westmoreland it consists entirely of shale, with a local conglomerate at the base; to the south-east, near Settle, it is represented by a band of calcareous mudstone resting on a few inches of grey shale, while to the north and north-west it thickens into a great series of shales and sandstones, which are spread over large areas in the Southern Uplands of Scotland.

In the Moffat district (Dumfries) the Birkhill and Gala groups must be several thousand feet thick, and consist largely of hard sandy and micaceous flagstones, but without any actual shore-beds or conglomerates. In Ayrshire, however, near Girvan, they include thick beds of conglomerate; the basal bed of this district is a boulder conglome

1 "Geol. Mag.," 1879, Dec. 2, vol. vi. p. 55.