the true relations of the genera, it is undesirable even to attempt to rectify the nomenclature of the groups.

It may be useful to future investigators if I enumerate the chief writings which deal with Rhizodus. They are as follows:

:

HIBBERT, "On the Limestone of Burdiehouse, Edinburgh," Trans. Roy. Soc. Edinburgh, vol. xiii. pt. i. pp. 202–214, t. viii. figs. 1, 2, t. ix. figs. 2, 3, 9, 10, t. x. figs. 1, 2, 3 (1835).

OWEN, Odontography, p. 75, t. xxxv. fig. 2, tt. xxxvi., xxxvii. (1840-45).

SEDGWICK and MCox, British Palæozoic Fossils, p. 611, 612, t. 30. fig. 17 (1855).

HUXLEY, "Essay on the Systematic Arrangement of the Fishes of the Devonian Epoch," Mem. Geol. Survey, Decade x. (1861).

YOUNG, "Note on the Scales of Rhizodus," Quart. Journ. Geol. Soc. vol. xxii. p. 317 (1866).

YOUNG, "Notices of new Genera of Carboniferous Glyptodipterines," ib. p. 596 (1866).

HANCOCK and ATTHEY, "Remains of Reptiles and Fishes from the Northumberland Coal-field," Nat. Hist. Trans. Northumberland and Durham, vol. iii. p. 81 (1869).

TRAQUAIR, "On some Fossil Fishes from the neighbourhood of Edinburgh," Ann. Nat. Hist. ser. 4, vol. xv. p. 258 (1875).

The thanks of those who may be interested in this subject are due to Prof. T. McKenny Hughes, M.A., the Woodwardian Professor of Geology, Cambridge, by whose permission the skull of Rhizodus is now described and to Mr. Thomas Atthey, of Gosforth, Newcastle, who cleared the specimen from the matrix, and developed its features with much care and skill.

46. NOTES on HAYTOR IRON-MINE. BY CLEMENT LE NEVE FOSTER, B.A., D.Sc., F.G.S. (Read June 23, 1875.)

Ir is nearly fifty years since this mine first attracted the attention of mineralogists by the occurrence of haytorite (chalcedony pseudomorphous after datholite). Several papers on the mineral were contributed to the Philosophical Magazine*; and an account of the mine was given by Mr. J. T. Kingstont. His description deals with the beds seen in an open work, and is so full and clear that there would be little left to add were it not for the fact that a new adit-level has intersected the beds and exposed a new section. There are certain points in this section which I think are worthy of being recorded; besides which it is well to call attention to the fact that there is now a good opportunity of studying the mode of occurrence of a fine deposit of magnetite without the necessity of making a journey to Norway or Sweden.

Haytor Mine is situated on the eastern borders of Dartmoor, about three quarters of a mile from the pile of granite rocks from which it derives its name. The iron-ore occurs in the form of thick beds of magnetite interstratified with altered shales and sandstones of Carboniferous age. Near the iron-ore the rock becomes highly charged with hornblende, and is sometimes apparently entirely made up of actinolite. Garnets, too, occur in great abundance in some places; and I also found a thin layer containing iron pyrites and what appears to be axinite.

The section seen near the end of the adit-level is shown in fig. 1.

Fig. 1.-Section in the Adit-level in Haytor Mine.

The exact thicknesses of the various beds were not measured by me, but were taken by the agent of the mine, Captain William Grose.

*The Philosophical Magazine, vol. i. 1827, pp. 38, 40, 43; vol. x. 1831, p. 111. The Edinburgh Journal of Science, vol. vi. pp. 297 & 301. + The Philosophical Magazine, vol. iii. 1828, p. 359.

It appears that there are three beds of iron-ore. No. 1, or the uppermost, is 10 feet thick, with partings of rock giving about 2 feet of waste; No. 2 is 14 feet thick, with about 1 foot of waste; and No. 3 is proved to be at least 6 feet thick, with a similar amount of waste. There is altogether therefore a total thickness of some 26 feet of iron-ore. This is much more than is shown by Mr. Kingston; but then it must be recollected that the adit intersects the iron at a depth of some 20 fathoms below the old open work, and there has been plenty of room for changes in thickness of ore and partings to occur.

The beds strike about E. 25° S., and dip northwards 30°..

A fourth bed of magnetite, about 3 feet thick, is seen cropping out about 300 yards N.E. from the others, and appears to be running parallel to them.

It will be seen by referring to the Geological-Survey Map, Sheet 25 (where the outcrop of the iron is incorrectly marked), that the line of junction of the Granite and Carboniferous rocks is running here about north and south. Consequently, as the Carboniferous rocks strike about W. 25° N. against the granite, it is evident that the granite has broken across the strata here. This same fact is proved in another way. Just above No. 3 bed is a layer of finegrained granite, 6 or 8 inches thick, running parallel to the stratification. At first sight it might be thought to be interbedded; but closer examination shows that it sometimes breaks across the stratification, as is seen in fig. 2. It is consequently an intrusive sheet, and, if it could be traced, would doubtless be found to be connected with the neighbouring mass of granite.

Excepting this granite vein, the ore and enclosing rocks (country) have simply the appearance of altered strata. Beds of iron-ore deposited contemporaneously with shales and sandstones seem to have been subjected to a metamorphic action, probably due to the proximity of the granite. The iron-ore, perhaps originally in the form of beds like the Cleveland ore, has been altered into magnetite; whilst the change undergone by the shales and sandstones consists in an extreme

silicification. It is, however, possible that the apparently stratified magnetite may have been formed by ferruginous emanations, which accompanied or followed the granitic intrusion, and spread out between the planes of bedding of the Carboniferous strata. The presence of actinolite and garnets, which I believe almost invariably accompany deposits of magnetite, may be explained under either hypothesis; these minerals were probably formed by metamorphic action out of the surrounding rocks, the ore furnishing the iron which enters into their composition.

The outcrop of these beds of magnetite may be traced eastwards for a distance of about three quarters of a mile. Near the surface the magnetite has been converted into ochre from the action of atmospheric agencies; and the outcrop has been worked for that substance.

In addition to the actinolite, garnets, axinite (?), and a little ironpyrites, I also found chalcedony and crystallized chalybite occurring with the magnetite. A little fluor-spar seen on one of the rubbish heaps was probably derived from a small vein. I was not fortunate enough to obtain any haytorite.

47. Descriptions of the GRAPTOLITES of the ARENIG and LLANDEILO ROCKS of ST. DAVID'S. By JOHN HOPKINSON, Esq., F.L.S., F.G.S., and CHARLES LAPWORTH, Esq., F.G.S. (Read December 16, 1874.)

[PLATES XXXIII.-XXXVII.]

INTRODUCTION.

In the neighbourhood of St. David's, the Arenig rocks, described by Mr. Hicks in his paper read at the last meeting of the Geological Society, are the earliest in which Graptolites are known to occur; and yet, when once they appear, so diversified are their forms that these Pembrokeshire rocks are only equalled, in the number and variety of the genera they contain, by the Canadian Graptolitebearing rocks of equivalent age known as the Quebec Group. In more ancient deposits two species only, belonging to one of the two great sections into which these fossils are divided, have hitherto been detected, viz. Dictyograptus (Dictyonema) socialis, Salter, and Dendrograptus Hallianus, Prout. The former occurring in the lower portion of the Tremadoc rocks of North Wales, and the latter in the equivalent strata (the Potsdam Sandstone of America), it is impossible to say which genus is the earlier, or whether the group is first represented in Britain or in America.

Before the discovery, in 1872, of the extensive series of Graptolites which characterize the Lower Arenig rocks of Ramsey Island, the Skiddaw Slates of Cumberland were supposed to be our earliest Graptolite-bearing rocks; but it is now known that the lowest rocks of the Arenig Group exposed in the vicinity of St. David's, in which Graptolites abound, are of greater age than any part of the Skiddaw Slates yet described; and it is highly probable that they are also older even than the lowest beds of the Quebec Group known to contain Graptolites, as will presently be shown.

As early as 1841 Graptolites were discovered in the "black slaty rocks of Pembrokeshire" (in the Llandeilo series) by Sir Henry De La Beche; and soon afterwards Professor Ramsay found the wellknown species Didymograptus Murchisoni in the Llandeilo rocks of Abereiddy Bay; but it is not until 1866 that we find the first mention of the occurrence of these fossils in the older rocks of this district. In their 'Second Report' on the St. David's rocks, presented to the British Association in that year (1866), Messrs. Salter and Hicks mention the occurrence of the genera Didymograptus and Dendrograptus in the black slates and flags of Whitesand Bay, which they then recognized as being of Arenig age; and in the same year, in a paper by Mr. Wyatt Edgell, "On the Arenig and Llandeilo Groups "+, the

"On the Succession of the Ancient Rocks in the vicinity of St. David's, Pembrokeshire, with special reference to those of the Arenig and Llandeilo Groups, and their Fossil Contents." Quart. Journ. Geol. Soc. vol. xxxi. p. 167. + Proc. Geologists' Assoc. July 1866. Geological Magazine, vol. iv. p. 113.