by continued subsidence, which broadened the sea-spaces at the expense of the land. In dealing specially with the geographical conditions of the British area, I have sought the assistance of Dr. Callaway, whose studies of the Cambrian rocks are well known, and he has kindly furnished me with an outline of such inferences as he thinks it is safe to draw in the present state of our knowledge. These are as follows: 66 : The ideal reconstruction of the geography of any past epoch is always a rather speculative task, but it becomes especially difficult when we revert to a period so ancient as that of the Lower Cambrian or Longmyndian. We can learn little about tracts which are covered by post-Cambrian deposits, and the deep borings of the south-east of England give us no help. We can, however, determine the approximate position of some of the land masses which rose amidst the waves of the Longmyndian sea. "The groups of volcanoes which in later Archæan times had spread out their lavas and ashes over so large a part of England and Wales were extinct at the commencement of the Cambrian period, but it would seem that several mountain chains, some of them crowned like the Andes or the Rocky Mountains by numerous volcanic cones, stood above the waves. Dry land probably extended over what are now the counties of Shropshire, Staffordshire, Leicestershire, Warwickshire, and Worcestershire, at least as far south as Malvern, but how far this land stretched to the north, east, and south, we have no means of knowing; it can only be said that analogy points to its having had a considerable north-east and south-west extension. 66 The preponderance of coarse sandstones indicates a rapid submergence of the neighbouring land, and the basal conglomerates in Wales indicate the gradual advance of the waves over a sinking shore. The rapid thinning-out of the Cambrian strata toward the west suggests that in the sea which now separates Wales from Ireland, land must then have predominated. Parts of Carnarvon and Anglesey are all that remain of the eastern side of this vanished country. "Beyond this there may have been another sea or gulf if the rocks in the east of Ireland are rightly referred to the Cambrian; but it is a significant fact that no deposits of Cambrian age have been found in any part of central or northern Ireland, or in any part of Scotland, except, perhaps, in Ross and Sutherland. It is rather unlikely, if any considerable part of these areas had been under water in Cambrian times, that every remnant of the deposits then formed should have been since either swept away or covered in. The balance of probability, therefore, is in favour of the conclusion that land masses of considerable size lay to the north and west of what is now Great Britain, and formed a Cambrian Atlantis, of which a part of Norway, the Hebrides, Donegal, and the highlands of Connemara are the worn and inconspicuous remains." The To these remarks it is only necessary to add that at the close of the Cambrian period the sea had gained very largely on the land, that the tract of land indicated as existing over central England was entirely submerged, and that the promontory or island between Wales and Ireland was probably reduced to very narrow limits, though parts of it seem still to have remained above the sea-level. entire absence of Upper Cambrian rocks, so far as is yet known, in northern Ireland and northern Scotland, seems to justify the supposition that these areas remained in the condition of land throughout the Cambrian period. The Torridon sandstone may have been formed in a narrow gulf penetrating into this land, or else, as Sir A. Ramsay has suggested, in an inland freshwater lake. CHAPTER III. ORDOVICIAN PERIOD. 1. Stratigraphical Evidence. GLAND.—The most southern area of Ordovician ENGLE rock in England is a small tract on either side of Dodman Head on the east coast of Cornwall. The rocks exposed consist of brown grits, quartzites, slates, and ashbeds, with some conglomerates, and they contain fossils of Bala age, but as their base is not visible it is impossible to draw any certain inferences as to the neighbourhood of land. It is true that Mr. Godwin-Austen has called the conglomerates "shingle-beds," and infers that they were formed in close proximity to a coast-line, but Professor Sedgwick describes them as trappean and schistose conglomerates, and speaks of their passing into "schaalstein" and "trap shale" (i.e. volcanic ash), so that they hardly seem to warrant the inference drawn by Mr. Austen. 1 At Malvern there is a total absence of Ordovician strata, the Upper Cambrian shales being unconformably covered by Silurian, but as the former dip under the latter, it is possible that Ordovician rocks exist beneath the Silurian area to the westward, and it would not, therefore, be safe to infer that the Malvern area was land throughout Ordovician times, though it will be seen from the sequel that land did then exist over the centre of England, and it is 1 66 Quart. Journ. Geol. Soc.," vol. xii. p. 44. quite possible that the Malvern Hills were part of the southern border of this land. The larger part, if not the whole, of Wales formed part of the Ordovician sea, and in this area the Cambrian sediments are succeeded by a thick series of comparatively deep-water rocks, though they certainly were not formed at any great distance from land. The succession consists of (1) the Arenig series-black slates and flags, from 3,000 to 4,000 feet thick, with the addition of much volcanic material in the north-west of Wales; (2) the Llandeilo series-another group of black slates and flags, with a thin limestone, in South Wales; (3) the Bala series—a more variable group, consisting of shales and sandstones, with two bands of limestone, the whole from 4,000 to 6,000 feet thick. As these Ordovician strata occur in full force on both sides of the Pembrokeshire Archæan axis, it is probable that this was completely submerged and covered by these sediments. In Shropshire the same complete series is found on the western side of the Longmynd ridge, but in the Caradoc district, on the eastern side of that ridge, the Arenig and Llandeilo series are absent, sandstones of Bala age, having a conglomerate at their base, resting on the local representative of the uppermost Cambrian. Fig. 1 is an ideal section, drawn for the purpose of showing the possible underground geology of the midland counties, for it is absolutely necessary to form some conception of this before we can draw any inferences regarding the probable position and extent of the land area which seems to have existed over the centre of England in Ordovician times. From the section it is seen that the Caradoc sandstones are supposed to thin out beneath the Silurians against the pre-Cambrian rocks, and that there is reason to suppose that they were never deposited over that part of |