and quartz, in nearly equal proportions. After the absence of mica, the next mineral which usually disappears is the felspar, leaving the compound a mixture of schorl and quartz. . . . Even the compound of quartz, felspar, and mica would seem more variable towards the outskirts of the mass than in the interior, and schorlaceous veins and veins of fine granite appear more frequent under such circumstances."

Generally speaking, De la Beche also notices the very variable character of the Devon and Cornwall granites; and incidentally describes many varieties of elvans or felsites; recording some of the changes that take place in short distances in elvan dykes themselves, alike in colour and in structure: as for example when he says of some of the West Cornwall elvans that they sometimes acquire a more granitic structure in the central parts of the dyke. He remarks: 2

"They may be considered as mere granite dykes, the chemical composition of which bears great analogy to that of the chief granitic masses in the vicinity of which they occur, the mineral structure considerably depending on the conditions for cooling to which they have been exposed."

It was hardly possible when De la Beche wrote to have more clearly indicated the general conditions of the problem. Many years ago the late Mr. Godwin-Austen suggested that the granites and granitic rocks of Devon were of three periods-the ordinary porphyritic granite of Dartmoor, the schorlaceous granite, and the elvans; and for a long period that opinion was generally held. At length it was recognised that the schorlaceous varieties were the product of alteration in the mineral structure of the original rock, and that while subsequent in form they were not really later in origin. The periods were thus reduced to two-that of the formation of the granite, and that of the formation of the elvans or felsites, which are clearly more recent than the mass of the granite, since they intersect it in common with the adjacent rocks. And this view has generally been accepted up to the present day.

As the result of my own investigations I was led in 1888 to suggest that in Dartmoor we had the base of a volcanic region, which at some point or points reached the surface definitely as a volcano, but the superstructure of which had been denuded. I suggested also that in the elvans or felsites, we had a series of rocks intermediate between the plutonic granites, and the sub-aerial lavas which formed part of that superstructure.

The question was then put, and naturally: If Dartmoor was a volcano, where are its volcanic rocks? It is, I think, a sufficient answer to point out that the amount of denudation to which Dartmoor has been exposed would naturally remove this class of rocks, in laying bare the basal granite. As a matter of fact, however, felsitic fragments do occur here and there on Dartmoor, which show flow structure and approximate very nearly to eruptive rocks; and alike in South and North Devon, (as in detritus at Cattedown3 and among the porphyritic pebbles of Rockham Bay1) example of andesites and volcanic grits are found, associated with ordinary granites and felsites, which cannot be traced in Devon in situ, but which such a volcano as that suggested would naturally yield.

Still further light was thrown upon the subject by an inquiry into the origin of the igneous fragments in the breccias at the base of the Trias of South Devon, and specially along the coast from Teignmouth to Exminster. It was found. that these consisted of granites precisely resembling those of Dartmoor; schorlaceous rocks of various kinds, to all appearance of similar origin; a wide variety of felsites, mostly indistinguishable from Dartmoor elvans, while others had a strongly marked rhyolitic or andesitic character; and a number of more or less felspathic traps, ranging from andesites to basalts, most of them closely resembling in situ traps of the Exeter, Crediton, and Tiverton areas.5

Thus sundry gaps between the plutonic granites and their volcanic representatives were filled. Morever it was noteworthy that the granites were much less frequently represented than the felsites and andesites and their allies, which is precisely what we should expect to find at such a comparatively early stage of denudation as that which supplied the Dartmoor materials to these breccias.

A further step was taken when a boss of the Cawsand felspathic trap at Withnoe was found to pass in the course of a few feet from "a rhyolitic pitchstone to a trachytic quartz felsite" that is, from a rock which would unhesitatingly be pronounced volcanic, to one which would unhesitatingly be accepted as a felsite or elvan-a rock that is known to pass absolutely and certainly into granite.

3 Trans. Devon Assoc. xxi. 77-80.

4 Ibid. xxiii. 400-406.

See "The Igneous Constituents of the Triassic Breccias and Conglomerates of South Devon." Quar. Jour. Geo. Soc. xlvi. 69-83.

See "Additional Notes on the Cornish Trias." Trans. Roy. Corn. Geo. Soc. 1891, 338-346.

Since then Mr. Ussher has observed in the vicinity of the Trias of Thurlestone a dyke of quartz-porphyry (Horswell Quarry, North Milton) or elvan, which with equal certainty passes into an andesite. In fact we seem to be getting this sort of evidence on all hands.

The list given will, I think, show in its wide variety and close and intimate gradation that our granites, felsites, and andesites, are not distinct species, but changing conditionsrather pathological than physical-of a common original.

What I have to suggest therefore is this: that instead of our granites and felsites being of three periods as Mr. GodwinAusten held; or of two periods as we have been accustomed to hold of late; that they with the felspathic traps of Central and East Devon-are really of one. I do not of course mean that they are absolutely contemporaneous, but that they are part of one and the same great series of igneous activities. That the elvans and felsites are later than the granites is clear, because they traverse them; but this may have happened at any period subsequent to consolidation, and need not infer any great range in time. What is equally to the purpose is the fact, hitherto too much overlooked, that porphyritic felsites traverse some of the felspathic traps, and must so far therefore be subsequent to them. I have found fragments in the red-rock breccias, showing junctions between porphyritic felsite and andesite; and examples occur in situ.7

There is no doubt that the granite of Dartmoor is later than the surrounding Carboniferous rocks, because it sends veins into them. It is evident also that the granites, the felsites, and the felspathic traps are earlier than the red-rock breccias, because their fragments are found therein.

In my paper on the "Igneous Constituents of the Triassic Breccias," I concluded

"That the elevation of Dartmoor and the associated igneous phenomena, which have been commonly regarded as post-Carboniferous and pre-Triassic, may in all probability be assigned to narrower limits, and be regarded as not earlier than Permian times, and possibly as occupying the Permio-Triassic interval, continuing into the earlier stages of the Trias. Certainly if the eruptions of the 'felspathic traps' of Devon and their associated elvans are related to the great Dartmoor movement (and, as we have seen, these traps are in part of Triassic date), we cannot well give the origin of that movement a higher than Permian antiquity."

7 There are dykes of quartz-porphyry in a cliff quarry of andesite near Withnoe, noted and described by Mr. Bernard Hobson, F.G.S.-Quar. Jour. Geo. Soc. xlviii. 500.

8 Op. cit. Quar. Jour. Geo. Soc.

There seems now, however, to be a growing consensus of opinion that these breccias are Permian. If this should prove so, it will only be needful that the conclusion be modified by restriction; when we should have to assign the great series of Devonian igneous rocks which have the granites at one end and the andesites at the other, to a period between the Lower Carboniferous and the Permian (to some extent inclusive).

It seems a question whether the "gabbros" in the Tavistock and Mary Tavy district, and some at least of our intrusive dolerites, may not be connected with a still earlier period of the same range of igneous activities, but I cannot deal with that now.

The list which follows is arranged alphabetically under the names of places-the only method that seemed to answer the purpose in view. When no definite localities are given it may be assumed that the examples occur either in beds of streams or alluvial deposits. The descriptions are purposely (and necessarily, seeing that over 400 varieties are dealt with) simple and macroscopic. Granite is used in its ordinary acceptation; felsite for all rocks with a felsitic matrix; pegmatite for quartz-felspar rocks; greisen for quartz-mica-; luxulyanite for felspar-schorl-; micas generally are described by their colour, and no attempt is made to discriminate between the felspars. The main object has been to give the general composition and physical characters in a plain and succinct form; and to this other details have been subordinated. Finally it must be borne in mind that the list is one of rocks, not localities, and does not profess to be exhaustive.

Avon Valley (Black Tor). -Bluish-grey close-textured, but developing modified porphyritic characters, both felspathic and quartzose, with a sprinkling of schorl, evidently replacing mica.

Avon Valley (Black Tor).-Pinkish-brown: differs from the last almost wholly in colour, but is somewhat less disposed to the porphyritic phase.

Avon Valley (near Whiteaburrow).-Reddish-brown: fine evengrained slightly schorlaceous pegmatite.

Avon Valley (ditto).-Pink, weathering to cream: fine-grained pegmatite, with porphyritic quartz blebs, occasional porphyritic felspars, and frequent stellate aggregations of schorl.

Avon Valley (ditto).-Reddish-brown fine even-grained granite with black mica, and occasional porphyritic felspars of a lighter hue; some schorl, occasionally segregated and, less frequently, pseudomorphic.

Avon Valley (near Ryders Rings).-Dark-red to reddish-grey: this rock has little resemblance to ordinary granite, the base consisting of a granular aggregate of quartz and ferruginous matter, with occasional flecks of white mica; but it contains porphyritic crystals of felspar and quartz. Some of the ferruginous material has decayed and left the texture of the rock open; probably it replaces an iron mica.

Avon Valley.-Brownish-red: even-grained schorlaceous pegmatite, with veins.

Avon Valley.-Grey, creamy: finely-granular pegmatitic base, with occasional blebs of quartz and porphyritic felspars, and here and there a nest of schorl, or a little needle that may replace mica. Avon Valley.-Red-brown: massive jaspideous quartz-porphyry. Avon Valley.-Red-grey coarse-textured but compact granite, with red felspars, many well-twinned, black and white micas, and schorl.

Avon Valley.-Dark red-brown: uneven-textured, compact granite, with black mica and rich Indian-red felspars.

Auswell Rock.-Grey and brown: a somewhat open-textured (having small irregular cavities) sub-crystalline quartzose rock. Bel Tor.-Pale red crystalline schorlaceous granite; small content of mica, but schorl generally present and developed in prismatic aggregates.

Bickleigh Vale (in situ).-Warm-buff: compact felsite, saccharoid grain with an occasional small pyramid of quartz.

Blackett (near Wrangaton).-Buff: loose-textured medium-grained granite with abundant black mica.

Blackett. Warm-grey: (a) sub-granular felsite approaching an earthy texture, with small porphyritic granules of quartz and needles of schorl; (b) coarser and less compact variety; (c) chalky-weathering ditto.

Blackett. - Buff: fine-grained granitic matrix with partly kaolinised felspar, porphyritic quartz and a little black mica and schorl; associated with pegmatite of similar texture, and large porphyritic felspars.

Blackett, (near)-Creamy-pink: sub-crystalline felspathic rock with casual schorl and porphyritic characters; the felspathic base is graphically mottled red and cream.

Bovey Tracey-Grey; a fine grained felsitic pegmatite, in parts rather porcellanous: the felspar largely kaolinised. Brattor (near Bridestowe)-Pale red a coarse schorlaceous pegmatite; felspar largely predominating; crystalline forms both in quartz and felspar; some of the latter patched greenish-grey. Brattor-Red: moderately coarse-textured pegmatite, with porphyritic blebs of quartz and occasional nests of black mica.