ceased to decompose at all; when, on examining it, he noticed that it was coated with a kind of enamel. It at once occurred to him that the process in question might be used to obtain such a coating; but he found, after a few days' exposure of the iron to the atmosphere, that the coating scaled off, and he pursued the matter no farther. The iron employed in this case was rusty; but if it had been new, my father would in all probability have been the accidental author of the process which Professor Barff discovered ten years later. That consists in subjecting iron or steel articles to the action of superheated steam; and, when they are at a temperature sufficiently high, the following chemical change takes place: 3 Fe + 4 (II2 O) = Fe3 0, +8 H. My father thought that what Professor Barff could effect with steam, he might also effect with air; and experiments were made varied both in character and results. On considering the fact that air is oxygen and nitrogen in mechanical combination only, I came to the conclusion, that, to form the lower or magnetic oxide, the quantity of free oxygen, and so of the air employed, must bear some proportion to the surface of the articles exposed to its action, more especially when a comparatively low heat is employed; and it has been found that the quantity of air passed through the retort during most of the unsuccessful experiments was three hundred or four hundred times more than was actually necessary. The mode of action I adopted was to admit a few cubic feet of air into the retort at the commencement of every half-hour, and then leave the iron and air to their own devices; the retort, of course, being tightly closed. During each half-hour a coating of magnetic oxide was formed, and the operation was repeated as often as was considered necessary. This was effective, but costly; both this and the Barff process requiring the external heating of the chamber. Successful experiments were made with air, but open to the same objection in regard to cost. Experiments with carbonic acid, produced by the decomposition of chalk, which should give 3 Fe + 4 (CO2) = Fe, O4 + 4 (CO), gave a coating of light color and easily removed; the film probably being a mixture of Fe O and Fe3O4, or something nearer the metallic state than is magnetic oxide. But, even if successful, the cost of this method would still be too high. I therefore proposed to use a fuel gasproducer, similar in principle to the Siemens generator, but altered to suit other requirements; to burn the combustible gases thus produced, with a slight excess of air over and above that actually required for perfect combustion, and to heat and oxidize the iron articles placed in a suitable brick chamber by these products of combustion. I also arranged a continuous regenerator of fire-clay tubes underneath the furnace; so that the products of combustion, leaving the oxidizing chamber, passed outside the tubes, imparting a portion of the waste heat to them, which was taken up by the in-going cold air passing through their interior on its way to the combustion-chamber. I had hoped in this way to be able to so regulate the excess of air over that required for complete combustion, as to be able to produce magnetic oxide direct, instead of the lower and useless oxide or combination of oxides. I obtained some beautiful results, and some again were unaccountably bad; and I soon found that it was as difficult to regulate the precise amount of oxidation as it first was in the Bessemer process. But I was fortunate enough to hit upon an almost parallel remedy; that is to say, I increased the quantity of free oxygen mixed with the products of combustion, and oxidized the iron articles to excess during a fixed period of generally forty minutes, when magnetic

oxide was found close to the iron, and sesquioxide over all. Then for twenty minutes I closed the airinlet entirely, leaving the gas-valve open, and so reduced the outside coating of sesquioxide to magnetic oxide by the reducing action of the combustible gases alone.

The Barff patents have been purchased by my father. His process is better than ours for wrought iron, and perhaps for polished work of all kinds, as iron commences to decompose steam at a very low temperature, - in fact, much below visible redness. For ordinary cast iron, and especially that quality which contains much, carbon, the Barff process is much too slow in its action; and some specimens that I have treated in England have taken as many as thirty-six hours to coat effectually, which could readily have been finished off in five hours by the Bower process. The main distinction between the two is, that the Bower is much more energetic in its action. The objection to the use of a closed muffle externally heated in the Barff process has been almost entirely overcome by simply putting wrought iron into a Bower furnace previously well heated, then shutting off both the gas and air supplies, and admitting steam into the regenerator tubes. Steel, I consider, can be equally well treated by both processes; except polished steel, which is better treated in a low-temperature Barff furnace. Of the fuel burnt in the gas-producers, a non-caking coal is the best. Virginian splint has suited very well in this country; and of this about one ton every three days is required for a furnace with an oxidizing chamber 13 feet long and 4 feet 3 inches wide and high. When a gas-coal is employed, it should be fed through the charging hoppers just before each deoxidizing operation, when a smoky flame is of great advantage. I have, however, discovered that anthracite coal can be used as well as a gas-coal by simply allowing petroleum to drop, at the rate of one gallon per hour, upon the red-hot surface of the coal in one of the producers. This method has been exclusively used in this country.

These magnetic-oxide processes not only protect from rust, but the coating is of such a beautiful color as to render articles ready for the market directly they are out of the furnace and cooled. One remarkable feature of these is, that there is no more cost (except in the labor of handling them) in treating 2,240 articles each weighing a pound than in coating a cube of the metal weighing a ton; and so penetrating is the process, that every crevice, no matter how intricate the pattern may be, is as effectively coated as the plainest surface. There is absolute certainty that paint used on iron so coated will adhere as well as on wood or stone; and thus iron may be used for construction work in a thousand directions in which it has not up to the present time been possible on account of its liability to rust, no matter what the coating used to protect it has been. Manufacturers appear far more ready to apply the processes here and on the continent of Europe than, up to now, they have been in England; but perhaps the reason has been, that, so far as Professor Barff's process is concerned, it has only just been shown how large masses can be dealt with by the use of the Bower furnace. For ordinary hollow-ware for kitchen or table use, whether of cast or wrought iron, the process is admirably adapted. It is intended to apply the process to cast-iron gas and water pipes; and, as the former have comparatively little pressure to bear, they may be made much lighter if rendered incorrodable: while, for water, there is no reason now why wrought-iron or mild steel pipes should not be used. In the case of railway

sleepers in iron and steel, which are now almost wholly used in Germany, the process is likely to prove of much advantage. For fountains, railings, and all architectural work, the process is invaluable; and iron may now be used in many instances instead of bronze. The cost has been carefully estimated at two dollars per ton; and this may be reduced by giving several furnaces in charge of one workman, and by a better system of taking the articles out than that in use when the estimate was made. Tests have been made as to the effect of the process on the strength of the metals, with the result that no alteration was detected in the strength. Theoretically one would suppose that iron and steel would be somewhat toughened, as the tendency of the process is to anneal, and would, no doubt, if continued long enough, render some classes of cast-iron malleable. A very thin article, if excessively coated, might probably be weakened, due to the fact that the coat of magnetic oxide would form an appreciable percentage of the bulk of the article; but that, of course, is a very extreme case, and one which is not likely to ever occur in practice.

Note on the jacketing of roasting cylinders at Deloro, Canada.

BY PROF. R. P. ROTHWELL OF NEW YORK.

The speaker said, that he merely desired to place on record the fact that he had been using roasting cylinders jacketed, to prevent any one from taking out a patent on the idea. He did not wish to deprive any one of the privilege of using it, but he also did not wish to be deprived of that privilege himself. In the roasting of arsenical sulphurets he had employed what is commonly known as the White and Howell cylinders, of plain boiler-iron, with fire-brick lining and shelves. He used two of them; the ore passing from one to the other through a pipe, without losing its heat. The first cylinder is 30 feet long and 5 feet in diameter, and takes out a large part of the arsenic and sulphur. The second is 24 feet long and a little less than 4 feet in diameter, in which the roast is finished. The two make a complete roast for chlorinating, and give from 94% to 98% of the gold. But these cylinders radiated an immense amount of heat, too much to allow the temperature to be kept sufficiently high to obtain a complete roast. This loss by radiation has been avoided by jacketing. A sheetiron jacket is placed around the cylinder, leaving an air-space of two inches; outside of this is another jacket with a space of two and a half inches, which is filled with mineral or slag wool; this is mixed with plaster of paris, and further covered with roofingpaper bound on with wire. Immediately upon the use of this apparatus there was noticeable a tremendous reduction in the consumption of fuel required, and a remarkable increase in the amount of ore roasted. As thus made, it even resulted in heating the upper portion of the first cylinder too much, and roasting too quickly, not leaving in the ore the sulphur necessary for the treatment in the second cylinder. The trouble was remedied by removing eight feet of the jacket around the upper part of the cylinder. Geological relations of the topography of the South Appalachian plateau.

BY PROF. W. C. KERR OF WASHINGTON.

By aid of a rough black-board sketch of the Blue Ridge and Smoky Mountains, the backbone of the system, the speaker showed from a study of the rivers, that the plateau has been gradually travelling west

ward. A series of spurs are thrown out by the Blue Ridge on the east, making a drainage system of cross valleys; here are the head-waters of the Tennessee river, which force their way through the great escarpment of the plateau, and through the Smoky Mountains, which in some places attain an altitude of 6,000 feet. This is a very remarkable and curious fact. The cañon through which the waters break is 4,000 feet deep, and has rocky sides not easily removed or eroded. A study of the situation shows, that since the establishment of the water-system there has been slow and steady rise of the mountain chain, the waters at the same time cutting their way down. There is another curious feature in this connection: the Tennessee river runs between this chain and the Cumberland ridge, and it would naturally be supposed that there is a rise from the west side of the river to the Cumberland. But observations with the barometer show, that there is really a continuous descent from the top of the Smoky Mountains to the base of the Cumberland chain, and here we have a river running at a higher level than its tributaries. The explanation is simply, that the Cumberland ridge has been gradually sinking since the establishment of the watersystem.

The collection of flue-dust at Ems.

BY DR. T. EGLESTON OF NEW YORK.

In the treatment of silver from lead-ores, this subject is a matter of growing importance in Ems at the works under the charge of Herr Freidenbach, and of some importance here. In 1874 it was found at Ems that there was a considerable loss of product by the dry method, and the wet method was substituted; and still the loss of dust was much greater than had been supposed. There were three difficulties to overcome: to arrest the material carried off by mechanical means, to collect the material which is volatilized, these two problems being comparatively easy of solution; but, when the collection was made, it was another thing to keep the material collected where it was, and prevent its further loss. The works are located on a plateau and hill. They run first down the valley, and then, turning on themselves, up the hill, continuing in a straight line to the top, where there is a chimney. In 1874 the length of the flue was 460 m., and it was furnished with the old style of condensing-chambers. The canal was then lengthened to 2,000 m., and carried to the flue 200 m. above the bed of the river. It was noted at once, that there was an immediate precipitation of flue-dust, much larger than had been anticipated, but still not effecting a sufficient reduction of the loss. An examination of the pipes led to the adoption of iron pipes, with the lower part terminating in zigzags 75 cm. deep, through which, by means of a door and close-fitting tube, the dust could be drawn out of the flue. This dust was rich, and the results of the method were satisfactory until the assays showed that much matter was lost by volatilization. Freidenbach soon found that the oldstyle arched flue was the worst that could be used; for, while its form gave strength to resist pressure from without, it also rendered it weak against pressure from within, and the gases found a comparatively easy means of exit through it. The flues were then made rectangular, bound together with iron, and made as tight as possible to prevent the escape of vapors. This form is now adopted everywhere. In the length of the flue was a series of condensation-chambers, but these were found to give no great results. The flue was now 2,600 m. in length, with an area of 42,650 m., and had cost 255,000 marks. A series of condensation-houses was built beyond the chimney,

and still the results were unsatisfactory. It gradually became apparent that what was wanted was surface, and not volume. The iron pipes before described not having been affected, there were introduced into the flue sheet-iron plates hung vertically. Four of these plates were at first put in; but the results were so immediate and so gratifying, that the number was increased to six, with still better effect. The conclusion was at once jumped at, that the flue would stand all the plates that could be put into it; and accordingly seventeen plates were introduced, having a space of 10 cm. between them. It was then discovered, that nearly all of the material carried off mechanically was thrown down near the furnace, and that volatilized was deposited a little farther on. These results having been reached, the difficulty was to keep these deposits where they were, and to prevent them from being carried off by the immense draught in so long a flue. This last obstacle was surmounted by placing transverse sheets of iron in the bottom. When the deposits reached a certain amount on the vertical plates, they dropped off from their own weight, and fell to the bottom, where the transverse plates retained them. Experiments were made as to the distance from the works at which the deposits were made; and at a short distance away was found nearly all the mechanical dust, that from volatilization being a little farther on. There was no material diminution in the draught occasioned by the introduction of the plates. The dust collected so quickly and to such an extent that it became a serious question as to how to remove it. The flues were constructed with manholes at the top, and the dust was in such fine state that the men would be subjected to the danger of suffocation. The problem was solved by setting fire to the flue and burning the dust, which was found in agglomerations easy to remove, and in just the condition to be put into the furnace. The removal was a matter of little difficulty, the manholes having been changed to the sides of the flue. Next arose the question of temperature, and whether or not the lowering of it had any effect on the collection of the dust. It varied from 300° C. near the chimney to 64° C. at some distance from it; and it was found that the degree of heat made little difference. This led to important conclusions; and the substitution was begun, near the chimney, of pasteboard for the iron plates. They answered the purpose just as well, provided they were of sufficient thickness to sustain themselves, and were also much cheaper. After the success of these experiments, the method of cleansing flues by water will probably be abandoned. They have demonstrated the importance of surface over volume, and of the rectangular against the arched flue. It is doubtful if any method can save the whole of the material carried off by mechanical means or volatilization; but it is proved that there can be saved two or three times more than was believed possible.

President Rothwell said that he had visited these works, and had taken much interest in going over them. By the process, a saving of about four per cent is effected over the old way; and Freidenbach charges a royalty of two per cent, or one-half of what he saves. Since the collection of the dust by burning, the pasteboard surfaces had been dispensed with, as they would be destroyed. He had closely observed the iron plates, and found that they were little affected. The first plates used were those which had been discarded from the screens, and had been lying about the yard, being as likely to be acted upon as any; but they showed no signs of deterioration. He had observed the same effect of surface in the collection of

arsenic dust in the works at Deloro, although at times he had been obliged to use a fan to secure a draught in long flues. The fan, however, needs frequent cleaning. His observations in regard to the ability of the iron to withstand action by the vapors led him to believe that arsenical chambers might be constructed of the same material with advantage. In regard to the flues at Ems, he had the fault to find, that they were built partly beneath the ground, and were apt to become too warm. He was in favor of building them above ground, and on arched supports, which would give the additional advantage that they could be opened without stopping the run.

Lines of weakness in cylinders.

BY PROF. R. H. RICHARDS OF BOSTON.

It has long been known to boiler-makers and to the users of cylindrical pipes of many kinds, that, when a tube is exposed to internal fluid pressure, the resolution of forces is such that the material of the walls of the tube is exposed to twice the stress in the direction tending to produce longitudinal rupture, that it is in the direction to produce circumferential fracture. By longitudinal fracture is meant the fracture by a rent parallel to the axis; by circumferential fracture, fracture by rents running round the cylinder. In consequence of this, makers of boilers always lay the fibre of their metal around the boiler; and the same is true with the makers of gun-barrels. I have never seen any good and simple illustration of this law until I met it in blowing glass. If a thin bubble of glass be blown out in a spherical form, and then exploded, it will be found that the particles tumble into totally irregular shapes, showing no special direction in the molecular structure of the material. If, now, a bubble of glass be blown out, and so manipulated that it will take a cylindrical form, and then be exploded, it will drop into ribbon-shaped pieces from end to end; and the only parts that will be found to differ from this form will be the two hemispherical ends, which will remain whole, having a fringe of ribbons representing the lines of fracture from the cylinder. The main point of difference between this experiment and the accidental explosion of large boilers appears to be, that in a boiler the shell goes at its weakest point, and once the rent is started it tears the boiler to pieces without much regularity of lines: while in the glass cylinder the walls are so nearly of the same strength that it can hardly be said to have a weakest point; when, therefore, it gets to its limit of strength, and is on the verge of exploding, there is no one place to initiate the explosion, and the glass explodes everywhere. This it does as it should do, by tearing into innumerable ribbons parallel to the axis of the cylinder. If Pthe pressure, and D = the diameter of the PD cylinder, then = stress tending to longitudinal = stress tending to circumferential

rupture, and

PD

2

rupture. Professor Richards illustrated his statements by experiments with glass tubing and a blast, with the most complete success.

The shop-treatment of structural steels.

BY MR. A. F. HILL OF NEW YORK.

The speaker urged the importance in the manufacturing-arts of a knowledge of the effects on iron and steel of the various processes to which those metals

are subjected. He took up these processes in their order, and gave the results of a close and careful study into the matter. In the operations of punching and shearing, it is conceded that the effect is to harden the metal to a local extent only; and also that enlargement of the area punched by reaming restores the plate to its original state. But Mr. Hill did not agree with Lieut. Barber, who has announced, as the result of his researches, that the amount of enlargement is a fixed quantity: on the contrary, the amount is dependent upon the carbon percentage and the thickness of the plate. The experiments were made with plates 18 inches wide, 4, 3%, and 1⁄2 inches in thickness, and .30, .40, and .50% carbon. They were cut in the planer, crosswise to the direction of the fibre; and three pieces from each plate were taken - one from the centre, and one from each end-for examination. The result of the experiments led to the conclusion, that the heavier the plate, or the lower the carbon percentage, the greater the effect of punching. Here is a clear indication of the direction which must be given to this line of investigation; but the conclusion is evident, that a restoration of strength is effected by reaming, although the enlargement is not a fixed quantity. In the cases of sheared and hammered open-hearth steel plates, annealing always restores the plate to its original strength. The capacity for welding is in inverse ratio to the carbon percentage, and the metal must not be heated any higher than is absolutely necessary to effect the weld. Annealing should immediately follow the welding, and the metal must be carried to a higher temperature than when it was last worked. It is a most important operation, and its effect varies directly with the carbon percentage. A metal bath gives unsatisfactory results: the best are obtained by annealing with oil. There is no more danger to be apprehended in annealing steel than in performing the same operation on iron; and nearly all trouble can be traced to poor workmanship.

The strength of American woods.
BY PROF. S. P. SHARPLES OF CAMBRIDGE.

When Gen. Walker was put in charge of the Census department, he was authorized to appoint experts to inquire into special industries. Under this act Prof. Charles S. Sargent of Brookline was appointed to gather statistics in relation to forest industries. Soon after his appointment, in 1869, he became convinced that it would be desirable to make an examination of the fuel-value of the various woods of the United States; and this work was placed in my hands. At the same time I made the suggestion, that, while we had the opportunity, it would be well to test also the strength of these woods: the suggestion was at once adopted, and Professor Sargent immediately set his agents at work in various parts of the country to collect specimens of all the trees growing in their localities; employing, as a rule, botanists who were familiar with the flora of the region in which they were at work. The result was the collection of over 1,300 specimens of wood, comprising more than 400 species and varieties, nearly 100 of which had not before been described as trees growing in the United States. The ash and specific gravity of every specimen in this collection have been determined, in most cases in duplicate: there have been about 2,600 ash and 2,800 specific-gravity determinations. About 325 species were further tested for transverse strength and resistance to crushing. In these series about 1,300 specimens were tested; and, as each was tried in three different ways, it made in all about 3,900

tests. There was a total of about 10,600 tests made on the specimens, many of them being of a series that required at least ten entries on the final report. In addition, seventy tests were made of the carbon and hydrogen in a number of the specimens. These tests have already, so far as the results of the ash and specific gravity of the dry wood are concerned, been published (Forestry bull., No. 32); and a bulletin is soon to be published giving the deflections under various loads.

After the wood had become thoroughly seasoned, it was dressed out into rods 4 centimetres square and 11 decimetres long. These were tested on the Watertown machine, the stick being placed in a perpendicular postion, resting on supports that were exactly one metre apart; the deflection being measured by an ordinary Brown and Sharp's scale graduated to millimetres. The force was applied at the centre of the length, by means of an iron bearing with a diameter of 12.5 millimetres. The loads were applied 50 kilogrammes at a time, and the deflection read on the scale after each weight was added. When the weight equalled 200 kilos, the load was taken off, and the set was measured; the load was again put on, the reading taken at 200 kilos, and again at every 50 kilos until the stick was broken, the breakingweight being also noted. In entering the test, a record was made of the direction of the fibre in each piece, i.e., whether the pressure was applied parallel with, or perpendicular to, the annual rings, or quartering them, - but this portion of the test resulted in a failure, the wood seeming to have equal strength in all directions of application of pressure. The stick was also weighed to about half a gramme, from which was calculated the specific gravity. To determine the specific gravity exactly, blocks were taken, carefully dressed out to precisely 11 centimetres in length and 35 millimetres square. They were carefully dried at the temperature of boiling water for a week, and were then measured with a micrometer caliper, and weighed; the specific gravity being calculated from the measurement and weight.

The ash was determined by igniting small blocks, thirty-five millimetres square and a centimetre long, dried in the same way, in a platinum dish in a muffle furnace heated by gas, the heat being applied so carefully that in most cases the ash retained the exact shape of the block: by taking care not to melt the ash, there was avoided a common error resulting from the non-combustion of a portion of the carbon. The ash was perfectly white, except where manganese or iron was present in the wood. It was judged best to report the ash exactly as found, and not to attempt any correction on account of carbon dioxide that might have been lost from the calcic carbonate present. From the results of the specific gravity and ash, the approximate full value was calculated. Count Rumford made experiments from which he came to the conclusion that the same weight of all woods will give the same amount of heat when burned under the same conditions; and Marcus Bull of Philadelphia, in 1826, reached the same result. These are the only attempts known to determine the fuel-value of wood. It is evident, that, if the cellulose in all woods is of equal value, that with the most ash is of the least value for fuel.

In 1848 Liebig made determinations of the carbon and hydrogen in the average composition of European woods; and, singularly enough, all of his experiments were made on hard wood, with one exception, that of fir. I determined the carbon and hydrogen in forty specimens of hard, and twenty-nine specimens of soft, wood. The average results agreed

within one-tenth of one per cent with those of Liebig: in soft woods the hydrogen is almost the same as in hard, but the carbon is from 4 to 5% greater, giving pine a higher fuel-value than hard wood. In these values we find mountain mahogany at the top (on account of its weight); the southern long-leaved pine is next, and at the bottom is poplar; shell-bark hickory is third on the list, these three having 49 to 54% of carbon. The pines are very close together, with over 52% of carbon, while the hard woods average a little under 49% of the average fuel-value by weight for soft wood: burning one kilo gives 4,488 units of heat; hard wood, 3,993.9: by volume, soft, 2,524; hard, 2,776.

In the tests for breaking-strength, the coefficient of elasticity was calculated for all sticks for the first two deflections, i.e., at loads of 50 and 100 kilos, and that at 100 kilos was found in many cases to be larger than that at the lesser load; but the explanation is found in the fact that there is more or less twist in the stick, no matter how carefully it is dressed; and this twist is increased by seasoning. The first load of 50 kilos is just about sufficient to take out the twist, and the second represents the true deflection. The results have shown, that it is by no means necessary to break two sticks to show which is the stronger, provided they are of the same kind of wood: the weak stick will show the largest deflection from the start. The strongest stick found was a piece of common yellow locust, the average of eight or nine specimens giving a breaking-weight of 543 kilos; hickory and southern pine follow closely; ash was found to stand very well up to a certain point, and then it gives way suddenly and without warning, generally shattering badly; California red-wood shatters thoroughly when it breaks, and shows the effect all over, rendering the entire stick worthless; white oak is inferior to several other oaks and to southern pine, the average breaking-weight of 40 specimens being 386 kilos, while the average of 8 specimens of the southern low oak was 528 kilos; 27 specimens of southern pine gave 490 kilos; 36 specimens of the Douglas fir from the Pacific coast, 374 kilos; 6 specimens of western larch, 523 kilos; 13 specimens of white pine, 274 kilos; 11 specimens of beech, 454 kilos; 16 specimens of large nut shell-bark hickory, 464 kilos; 20 specimens of white hickory, 512 kilos; 24 specimens of white ash, 378 kilos; 8 specimens of locust, 543 kilos.

The next series of tests were made on specimens of the same-sized square as before, and 32 centimetres long, compressing them in the direction of their fibres. Nine specimens of locust stood an average weight of 11,206 kilos; 5 specimens of western larch, 10,660 kilos; 35 specimens of white oak, 8,183 kilos; 24 specimens of southern pine, 10,498 kilos. The effect of the pressure on the specimens was very curious. Professor Sharples exhibited a number of specimens thus treated, which showed curious changes under the pressure.

The third series of tests was to find the force necessary to indent the wood at right angles to the grain. These are not yet finished, and I can give only a few general results. The load was noted at every onehundredth of an inch of indentation, and it was found that the first one-hundredth was the hardest to make. After that the amount of force necessary diminished with each one-hundredth, until, at one-tenth of an inch indentation, it was found that the force required was only twice that at one one-hundredth. specimens were often destroyed, however, before reaching the greater depth. In closing this paper, I wish to express my public thanks to Col. Laidley for

The

many valuable suggestions made during the work, and to Mr. Howard for his careful aid in bringing the tests to a successful issue.

The eozoic and lower paleozoic in South Wales, and their comparison with their Appalachian analogues.

BY DR. PERSIFOR FRAZER OF PHILADELPHIA.

This paper embodied the observations of the author at St. David's, South Wales, during a visit at the invitation of Prof. Archibald Geikie, director-general of the geological surveys of Great Britain and Ireland, and Mr. B. N. Peach, geologist in charge of the survey of Scotland. The occasion offered a rare opportunity for studying those classic rocks, - the Cambrian; but there were other series of rocks exposed of the greatest interest to the student of Appalachian geology, not only from their points of resemblance to other rocks met with frequently on the Atlantic border of the United States, but from the similar relations which they seemed to bear to the measures in contact with them. At Roch's Castle is an area of Llandeilo flags, resembling what Dr. Frazer has often designated as argillaceous shale; and, in specimens where the decomposition into clay had proceeded very far, there was almost invariably the same disposition to split into prisms of unequally large pairs of parallel planes, no two of which were perpendicular to each other, giving them a remote resemblance to some of the indefinitely numerous varieties of triclinic crystals. Like similar argillaceous shales and slates near the town of York, Penn., and elsewhere in America, the slabs split up into almost any desired degree of thinness. The rock on which the castle is built is a silicious, greenish rock, showing everywhere included crystals of more or less definite outline, and generally of about the size of a buckshot, and containing a whitish or yellowish feldspar. The analogy between this rock and the jaspers' of Rogers, of which Dr. T. Sterry Hunt was the first to point out the real character, is striking. In the porphyry of Roch's Castle, the feldspar is oftener yellowish-green than in the orthofelsite porphyries of the South Mountain and of the eastern United States, as there is much of the Welsh orthofelsite which shows flesh-colored feldspar, and much of that of the South Mountain which exhibits green and other colors. The lamination and flaggy structure, when it was apparent, seemed to be entirely due to the arrangement of the cleavage surfaces of numbers of small crystals in the same plane; because a large part of the rocks defied all attempts to define sedimentary structure. Similiar exhibitions of orthofelsite are found in quantity on the eastern slope of the South Mountain in Pennsylvania, from Dilsburg to Monterey. In the latter regions, however, the beds, which are generally in contact with them, have a more chloritic and a more schistose character than the Llandeilo flags. They are marked, too, in America, for a part of their extent, by an horizon of copper ores, of which no trace was observed in South Wales. To the west and north of the beds of intrusive rock which seem to underlie St. David's, and in the harbor of Porth Ceri, there occurs a thick series of greenish, arenaceous beds, showing numerous streaks of chlorite. They are of very great interest, because they are unmistakably hydro-mica schists of light greenish or grayish color, very finely laminated, and resembling the rocks of parts of the South Valley Hill, and of parts of Fulton and Manor townships on the Susquehanna river. Similar schists, which (according to the writer's theory of structure, based on