doubt that they were soon deeply submerged and covered by the Silurian sediments.

The evidence for the existence of this Hiberno-Cambrian island is not at present very strong, but it furnishes us with a means of accounting for the great thickness of the Silurian deposits on its eastern side, for if we suppose that the tidal currents came from the south-westward, a reference to the map (Pl. I.) will show that they would naturally sweep the débris gained from the erosion of its southern and western shores round to the more sheltered side, and that great banks of sand and mud would be likely to accumulate along the eastern shore and beyond the north-eastern extremity; the sea itself may originally have been of considerable depth, and the continued submergence would prevent it from being entirely silted up.

There is better evidence for the existence of two smaller tracts of land to the eastward, which may have been separate islands, or may have been promontories from a larger tract lying to the north-east of them. One of these was formed by the rocks of the Longmynd and Shelve district; the Llandovery conglomerate wraps round the edges of this tract, which doubtless had a farther extension to the north-east in Silurian times beyond the range of the Wrekin. The coasts of this island were probably rather steep, and the sea, on the eastern side at any rate, was deep, or limestones would not have been formed in such near proximity.

The other island was probably rather larger, and was the remnant of that which seems to have existed over the midland area in Ordovician times, but only the higher ridges between Malvern and Charnwood Forest were left, and it is impossible to say whether these remained above water throughout the Silurian period.

It is probable, indeed, that both this and the Longmynd island were surrounded by coral reefs which were built up

around them as they sank beneath the Silurian sea, till after existing for a time in the condition of atolls, they were eventually submerged and buried under the higher Silurian shales and mudstones.

It is just possible that another island existed to the south-west of England, and included part of Cornwall, for no rocks of true Silurian age have yet been found in that county, and it is believed that the oldest Devonians rest on the Ordovician rocks of Mevagissey and Veryan Bay. Further examination of this district is required, however, before any definite conclusions can be drawn.

Basal Silurians exist in Brittany, but nothing comparable to our Wenlock and Ludlow series is found there, and French geologists believe that during the latter part of the Silurian period the Armorican and Cotentin region formed an island the limits of which were very nearly those which are now presented by the pre-Devonian rocks of this district.1

1 Lapparent," Traité de Géologie," second edition, p. 781.

CHAPTER V.

OLD RED SANDSTONE AND DEVONIAN.

'HERE is much uncertainty about the classification

THE

and correlation of the rocks which intervene between the Silurian and the Carboniferous systems, and this uncertainty mainly arises from the fact that there is no locality in Britain where an unbroken series of these intervening rocks can be studied. In several areas the Silurian is seen to pass upward into a series of red and purple sandstones (Old Red Sandstone), but in every case there is a break before the Carboniferous series is reached. Again, there is one area (North Devon) where Carboniferous rocks rest conformably on a series of strata containing a peculiar fauna which stamps them as of intermediate age between Silurian and Carboniferous, but the base of these (Devonian) rocks is not seen, so that we do not know what they rest upon, and it is possible that a considerable thickness of rocks intervenes between the Lower Devonian Sandstones and the Silurians of Upper Ludlow type, if such beds exist beneath North Devon.

The Old Red Sandstone is divided into a Lower and an Upper Series, with a gap between them. The greater part of the Devonian system was doubtless deposited during the period which this gap represents, though its lowest portion may have been coeval with part of the Lower Old Red, and its uppermost portion is generally acknowledged as the equivalent of the Upper Old Red.

E

Since, however, we cannot be sure that the strata which conformably succeed the Silurian are represented to any great extent in the Lower Devonian group, and as it is possible the mass of them may be older, the safest plan will be to treat the rock-groups above mentioned as three separate series, viz.: (1) Lower Old Red Sandstone, and its homotaxial equivalents in Scotland and Ireland; (2) Devonian rocks; (3) Upper Old Red Sandstone. This is the plan adopted by Mr. H. B. Woodward in his new edition of the "Geology of England and Wales."

1. Stratigraphical Evidence.

1. Lower Old Red Series.-In England this series is only found on the borders of Wales, in the counties of Brecknock, Monmouth, Hereford, and Shropshire, where it covers a considerable area. The rocks consist of sandstones, flagstones, and marls, the sandstones generally red, the marls red, grey, or green, and including lenticular masses or nodules of limestone, which are locally known as "cornstones;" hence the whole series is sometimes called the "Cornstone Series," and its thickness is supposed to be from 2,000 to 2,500 feet. It is succeeded by the Middle Old Red Sandstone, or "Brownstone Series," but Mr. Symonds believes that there is a break in the succession at the summit of the Cornstone group.1

The limits of this group are not yet accurately known, but its western outcrop is believed to extend from the head-waters of the river Usk in Brecknock to Much Wenlock in Shropshire, and as it also flanks the Silurians of Woolhope and Malvern, &c., it would appear to lie in a broad synclinal which has a general north-east to southwest strike. The area becomes narrower towards the north

1 "Records of the Rocks," p. 234.