the river Wye cut across the anticlinal of the strata, and consequently cut across the same beds in two different directions, i.e. dipping both with and against the underlay of the beds.

That the inclination or dip of the cleavage planes varies infinitely more than the strike is quite certain, and observation shows that, in many cases at least, the cleavage planes on each side of an anticlinal may dip inwards towards the axis of elevation, consequently in opposite directions, so as to present a fanlike arrangement; this was specially noticed by Darwin in South America, and subsequently by other geologists. The late Mr. D. Sharpe attempted to generalise on these facts, and brought forward the hypothesis that the strike of the cleavage, when studied on the large scale, followed a general direction, but that the varying lines of dip formed a series of anticlinal and synclinal "systems of cleavage," extending over vast areas like a series of immense arches. It is to be feared, however, that subsequent researches in the field have not corroborated this hypothesis.

From the direction of the strike of the cleavage in some of the districts first examined it was at once taken for granted that the normal direction of cleavage planes would be exactly east and west of the magnetic north, and this gave rise to numerous attempts to account for this structure by the action of magnetic or electric currents; it required, however, but a little more extended observation in the field to prove this generalisation unfounded. Sedgwick also announced that "the strike of the cleavage is nearly coincident with the strike of the beds," which is, no doubt, often the case in districts in which the direction of the beds is tolerably uniform, but quite at fault when the stratification is of a more complicated and contorted character. Professor Phillips, in 1843, arrived at the conclusion with respect to the slate rocks of Wales, that the cleavage planes are always parallel to the "main direction of the great anticlinal axes, but are not affected by the small undulations or contortions," and this conclusion appears to hold good in the majority of cases; yet exceptions are not wanting, as Professor Phillips has himself pointed out.

In summing up the evidence it can only be stated, that it appears certain that there is a decided tendency in the cleavage planes to follow, more or less, the strike of the strata or axis of elevation. Yet, at the same time, the inclination of the cleavage planes does not appear to have any connection with the dip of the stratification, and it appears most in accordance with facts to assume, that both the strike and dip of the cleavage planes are entirely governed by local circumstances. Since cleavage is not known to occur in normal sedimentary rocks, it is evidently dependent upon the local disturbance of the strata,

and most probably intimately connected with the intrusion of the eruptive rocks of the district.

The causes which are concerned in the production of cleavage structure were for a long period involved in obscurity. Some of the early observers attributed these phenomena to concretionary, crystalline, or chemical action, and others to the effect of magnetism or electricity; as, however, the explanations given were both vague and unsatisfactory, besides not always consistent with the supposition that the expounders of these hypotheses were themselves altogether at home in the study of those forces which they thus invoked to their aid, it is to be suspected that the hypotheses advanced were but another mode of accounting for ignorance.

It was quite evident, however, that crystallisation could not be the cause, since, in all cases of perfect cleavage, no trace of crystallisation can be detected even with the microscope, and in cases where crystalline structure is developed, the cleavage is invariably much more imperfect, or altogether absent.

The late Mr. Daniel Sharpe was, however, the first who pointed out the true road to the solution of this problem, when he announced, as the result of his observations, that cleaved rocks had undergone a compression of their mass in a direction perpendicular to the cleavage planes, and showed that in cleaved rocks the particles were always arranged with their flat sides parallel to the cleavage planes, and that they were usually longer on the line of dip of the cleavage than on the strike. He also confirmed what had previously been noticed by Professor Phillips, that the fossils, when present in cleaved rocks, were always more or less distorted, the distortion being greatest where the angle between the cleavage and stratification is least.

The microscopic examination of sections of cleaved and uncleaved slate rock showed their structure to be very dissimilar. In the latter, as seen in fig. 8 a, Pl. LVII., the grains of sand, clay, mica, &c., which compose the mass of the rock, are seen as if disposed at random, without any trace of definite arrangement, whilst fig. 8 b, which shows a section of cleaved slate (roofing slate), proves the particles to be distinctly arranged and elongated along the line of cleavage of the slate.

Sorby also showed, as figured in fig. 10 a and b, Pl. LVII., that analogous differences in structure occur in the cleaved and uncleaved devonian limestones from Devonshire, the latter being a section of an encrinite joint, in which the cells are arranged without any order, and retain their normal or equiaxial shapes. In the former, however, which is a highly cleaved limestone from Kingskerswell, the structure is seen to be entirely changed,

the cells being in this case flattened and drawn out, so that the longer axis lies in the line of the cleavage.

In the same manner the yellow concretions, which look like so many blotches in the Welsh slates, when found in uncleaved slate rock, are seen to be nearly spherical or ellipsoidal concretions occurring in the lines of stratification, and which, when they are longer in one direction than another, invariably have their major axis lying in the plane of bedding. In the cleaved slate, however, they are found to be considerably changed, as seen in fig. 5 a and b, Pl. LVII., where they are represented as flattened out and compressed sidings; the major axis of these now always follows the direction of the lines of cleavage.

The peculiar effect of cleavage in altering the form of fossils, as has previously been noticed, has proved a puzzle to the palæontologists, who, deceived by appearances, have frequently given different names to distorted specimens of the same fossil. This was the case with the two names Spirifer disjunctus and Sp. giganteus, the latter being subsequently shown only to be the former greatly distorted and expanded by the effects of the cleavage in the slates at Tintagel in Cornwall. In like manner, the shell Euomphalus pentangulatus, from Little Island, Cork, represented by the figures 12 a, b, and c, Pl. LVII. (from Haughton), had frequently (owing to the distortion and consequent change of form from their normal appearance, fig. 12 b) been described as Ellipsolithes, and considered as quite distinct from Euomphalus; and several other quite similar mistakes

could be mentioned.

Another example of how greatly the appearance of the same fossils, when occurring in cleaved rocks, may differ from one another will be seen in the case of the trilobite Asaphus Homfrayi (Salter), shown in fig. 9 a b, Pl. LVII.

A comparison of the normal and distorted fossils proves, without doubt, that they must have not only been subjected to a compressing action, but also that this has at the same time been combined with a sort of sliding movement, which has caused them to become drawn out or elongated along the planes of cleavage in one direction much more than the other. As an experimental illustration of how such an effect can be produced, the changes which a coin, in this case a farthing, will experience when passed through a rolling-mill, are shown in fig. 6 a, b, c, Pl. LVII., and which will not require any explanation. The action of the rolls upon a substance inserted between them is a combination of direct pressure with forward movement, and consequently not only causes the raised impression on the coin to become flattened out, but also at the same time elongates its one axis in the direction of the forward movement, i.e. that in which it had been inserted between the rolls, whilst

the other axis is but little expanded; by this means distorted images like b and c can be produced at pleasure, merely by passing the coin through the rolls in one or other direction.

The idea that an approximate measurement of the amount of compression to which cleaved rocks had been subjected might be deduced from calculations based upon the relative dimensions of normal and distorted fossils, naturally suggested itself, and has been followed up by several observers, particularly by Professor Haughton, whose conclusions are strikingly corroborative of those obtained by other experimenters by totally different modes of procedure, as will subsequently be alluded

to.

The researches of Mr. Sorby, however, not only advanced the previous knowledge of the phenomena of cleavage structure, but proved, experimentally and conclusively, that cleavage was not in any way connected with chemical, crystalline, or electrical agencies, and that it was solely the effect of mechanical forces acting upon the rock so as not only to effect a considerable compression or condensation of its substance, but at the same time to re-arrange or change the position of the particles of which the rock is composed.

One of the most interesting illustrations brought forward by Mr. Sorby to show the effect of such compression, is seen in fig. 11, Pl. LVII., being a vertical section from near Ilfracombe, where a bed of very much contorted coarser arenaceous slate is seen interposed between two beds of fine-grained well-cleaved slate, the stratification lines of which are denoted by the darker bandsa, b. An inspection of this section at once renders it apparent that these beds, originally deposited as horizontal strata, must have been squeezed together from the sides, so as to crumple up the non-yielding arenaceous bed, whilst the particles of the slate, giving way to the compression, were only packed together more closely, and became so much denser than before. To use the author's words, these phenomena are "analogous to what would occur if a strip of paper, for instance, was included in a mass of some soft, plastic material, which would readily change its dimensions. If the whole was then compressed in the direction of the length of the strip of paper, it would be bent and puckered up into contortions, whilst the plastic material would change its dimensions without such being the case; and the difference in distance of the ends of the paper, as measured in direct line or along it, would indicate the change in dimensions of the plastic material."

Numerous observations, some of which have been already referred to, prove that the arrangement of the particles in cleaved rock differs from what it is in normal sedimentary rocks, and that in the former the particles composing them

have, as a rule, their flat sides or long axes parallel to the planes of cleavage, and not to those of stratification, as is the case in unaltered sedimentary deposits. That this effect was due to the same mechanical force which produced the cleavage, was demonstrated by Mr. Sorby by the following simple experiment:

A quantity of scales of oxide of iron was mixed promiscuously, so as to distribute them in all directions throughout pipe-clay; after drying, a section of the mass presented the appearance shown in fig. 7 a, Pl. LVII. Upon submitting this mass to a pressure sufficient to change its dimensions to the same extent that is supposed to have taken place in natural slate rocks, a section at right angles to the diameter of the applied pressure (which, consequently, would correspond to the dip of the cleavage planes in roofing slate) now presented the appearance seen in fig. 7 b, rendering it apparent that the effect of the pressure had caused the scales to re-arrange themselves in nearly parallel lines, perpendicular to the pressure, along which the mass now admitted of easy division into thin plates, whereas it did not do so in other directions.

Ordinary clay, which contains but few flat particles, when submitted to pressure was also found to receive a most distinct laminated structure, but did not cleave perfectly, so that it is probable that the perfection of the cleavage in roofing slates depends, in part at least, upon the presence of the numerous flat particles of mica, chlorite, &c., which, by their re-arrangement, also contribute to determine the easy splitting qualities of the rock. That rocks may receive a comparatively welldeveloped cleavage structure, even in the absence of such particles, is evident, however, from the descriptions given by Sorby of the cleaved limestones of Devonshire; and the subsequent experiments of Tyndall showed also that a most perfect cleavage could be induced in pure white wax, by the application of mere pressure.

The consideration of the above, and numerous similar data connected with the subject, led Mr. Sorby to the conclusion that rocks possessing a perfect cleavage structure, such as the Welsh roofing slate, had experienced a diminution in bulk by compression to about one-half of their original volume (consequently would, after cleavage, increase so much more in density), and to advocate a purely mechanical theory of cleavage, such as explained above.

EXPLANATION OF PLATE LVII.

FIG. 1. Sketch section across Llwyd-Mawr, Carnarvonshire; (a) quartz porphyrite on which rests the purple cambrian slate rocks; (d)