crystalline character and some fine-grained sediments have become masses of semi-crystalline limestones. There are gradational types between these two extremes.

It thus seems that we have the coming in of the change to the northern type of Lower Carboniferous deposits, already shown in the eastern part of the Lower Craven area. Not only is there a marked change in the character of the limestone, but, as might be expected, there is a change in the fauna. Instead of the abundant Rhynconella (pleurodon, pugnus, acuminata), Spirifera (bisulcata, glabra, lineata, striata), with Amplexus coralloides and Lithostrotion, Cyathophyllum, we get more of those forms which are common to the Pendleside Series, e.g., Athyris ambigua, Orthis Michelini, Productus scabriculus, and P. semireticularis, with an exceedingly abundant development of Syringopora (three species), and of Zaphrentis (Enniskilleni, &c.).

The strata under consideration, with the succeeding formations-Pendleside Series, Millstone Grits, and Lower Coal Measures afford an excellent illustration of Lord Avebury's principle. His experiments in mountain building showed that when a series of formations are thrown into a common mass of folds, the lower strata show most folding, with the sharpest dip, &c. This may well be seen by tracing the beds from the middle of the Clitheroe anticlinal through Pendle to Nelson.

It is similarly seen when passing from Thornton through the Carlton synclinal and the Lothersdale anticlinal to the next synclinal of Reedshaw Moss.

There is thus extensive folding and faulting seen amongst the Sedimentary Limestones of the complex anticlinal (and the same holds for the small, sharp Lothersdale anticlinal), but it seems obvious that there would be some difficulty in accommodating the thickened beds of "massive " shell, &c., deposits to such a system of internal folding. And in this obvious difficulty it seems there may be an explanation of at least one type of reef-knoll. It is important to notice that known faults accompany three out of four of the knoll-masses of Downham in such a way as to suggest that faulting has some connection with the actual form of these knolls.

INGLEBOROUGH.

PART IV.* STRATIGRAPHY AND PALEONTOLOGY OF THE

SILURIAN.

BY T. MCKENNY HUGHES, M.A., F.R.S., F.G.S., WOODWARDIAN PROFESSOR OF GEOLOGY IN THE UNIVERSITY OF CAMBRIDGE.

"Each country ought to be described without any accommodating hypothesis, according to the type after which it has been moulded. But, in comparing the unconnected deposits of remote countries, we must act on an opposite principle; learning to suppress all local phenomena, and to seize on those only which are coextensive with the objects we attempt to classify." Sedgwick.†

I am endeavouring to follow this good advice, and, while offering a tentative correlation of the various subdivisions of the Silurian rocks seen under Ingleborough with those of other areas, to avoid forcing a foreign nomenclature upon our district. In the case of the lowest divisions of the system, we shall see more clearly as we follow the details of the classification that we cannot identify bed for bed even within the limited distance from Ingleborough to the next outcrop of Silurian on the north, where the Rawthey and its tributaries have cut down to and exposed several clear sections in the Silurian rocks, but, if we regard the whole of the series which occurs at the base of the system, we find that it can be correlated with equivalent beds much further afield.

In our examination of the paleontology also, we must bear in mind that a small difference in the character of a deposit often determines what forms of life may be expected. We see this upon any shore. Cockles and mussels and whelks and purpuras, all have their own peculiar habitat. Yet, as accidental occurrence we find a creature that loves the mud drifted away on to the sand, or one that lives on stones carried

an

* Continued from Proc. Yorks. Geol. Soc., vol. xv., p. 371. That part, pp. 351-371, should have been numbered Part iii.

+ Life, vol. i, p. 297, Trans. Geol. Soc., ser. ii, vol. iii.

on to a mud flat. So in the rocks Trinucleus concentricus is found in shale, but Trinucleus seticornis more commonly in sandy beds; yet they are contemporary and are often found together. When the sand was carried over the mudflats where certain Graptolites thrived they had to go or die, but when mud settled there again they or their representatives returned. Some of the common Silurian fossils, such as the smooth Pentameri, have a very short range in time, others, as Cardiola interrupta, run right through the greater part of the Silurian.

Therefore when we are applying the palæontological vernier to our scale of rocks we have various problems to solve before we can trust the first results at which we appear to have arrived, especially when we are making use of negative evidence. When we fail to find the characteristic fossils of an otherwise wellmarked horizon, we must inquire whether the deposits in which those fossils were found have under locally varying conditions of sedimentation gradually changed from coarse to fine or fine to coarse, so that synchronous deposits carrying different forms of life occur in adjacent areas. Sometimes we can explain

the absence of the fossils we expected to find by supposing that, without any discordance, sedimentation ceased over the area under examination, so that the deposit in which those fossils do occur has no equivalent in time in the area under examination, but has thinned out to nothing before it reached it. This of course implies that beds representing it, or part of it, may be expected in the intermediate area.

We should, for the practical working out of the geology of a district, classify by the more obvious petrographical characters, and then identify by paleontology, but petrology and palæontology should each be kept subordinate, as the handmaid, not the mistress, of stratigraphy.

In the preceding part of this paper I discussed the question of what we should take as the base of the Silurian, and gave a general sketch of the whole series, with a view to determining what part is represented under Ingleborough. I now proceed to examine these in greater detail.

Above the irregular and ill-defined deposit which we have spoken of as the basement bed, we found thin variable bands,

not identical in any two sections under Ingleborough, and, though showing points of lithological and paleontological similarity with many of the sections at the same horizon further north, nowhere exactly represented in them.

It may be useful, therefore, to ask what we might, from analogy, have expected to find here before we examine in detail what we do see.

I give two sections, one of which (Plate VI.) shows an almost unbroken succession from the Bala Beds to the bottom of the Coniston Flags, and the other (Fig. 1) illustrates the way in which we generally find these easily crushed and faulted beds cut up and repeated, where they come out from under the Carboniferous Fells north of Ingleborough. When such beds are highly inclined, it is almost impossible to detect sharp plications and displacements, and we must always entertain the idea that the beds we miss may be nipped out by faults.

The first section to which I refer (Plate VI.) is the most complete that I know of in the lower beds of the Silurian and the upper beds of the Bala series of our district. A detailed description of this section, with full lists of the fossils which occur at the various horizons, is given by Messrs. Marr and Nicholson in their valuable paper on the Stockdale Shales.* This section gives a sample of the material which was being laid down on the north side of Ingleborough while the Austwick sediments, which we have been describing, were being spread over the south side. We have found an immense thickness of felspathic slates with subordinate grit in Chapel-le-dale, and inferred that there were more of them exposed to denudation when the lowest beds of the Silurian were being formed. So also in the northern section we see a still greater thickness of felspathic mud such as might be derived from the volcanic series. Moreover, each and all of the finer beds at the base of the Silurian are much thicker in the northern area.

In the Spengill section we have the Lower Coniston Flags passing down into hundreds of feet of Pale Slates. These are generally gray with red bands, and at one horizon yield a beautifully striped rock. The graptolites which have been found in

* Q.J.G.S., vol. xliv., 1888, p. 654.

SECTION IN WHINNY GILL: UPPER PART OF CROSS HAW BECK, SEDBERGH, YORKSHIRE.

this striped band are referred to Monograptus Sedgwickii Portl., while immediately above it Climacograptus scalaris His., and Diplograptus palmeus Barr. occur.

There are some closetextured beds consisting of the finest felspathic material reset into a porcellaneous rock suggestive of the metamorphism of the more susceptible volcanic material. One mass seems to be a dyke, but in the field can hardly be distinguished from some portions of the rock above mentioned. Near the base there is a band of false bedded sandy mudstone which is the only indication of proximity to shore throughout the series. This passes down by the intercalation of bands of shale into the Stockdale Shale, which here consists of a fine sandy shale breaking up into finger-like prismatic fragments. At the base there is a gritty calcareous band with fossils, and at the top there are irregular massive blocky mudstones and black and grey shales with lenticular beds of grey fossiliferous limestone, to which I have referred as the Spengill