eres PDE se ee lee en A ce oe er ee en a ye veers ' tt> rat a= : -+ wt ‘ rf ’ : 1 ' ' R «tt * ' to at = . rate ’ ‘ ns 28's ‘ els caren oe ‘ about this time committed himself in part to both hypoth- esis, though evidently disposed to favor that of transport. “The coal itself is admitted to be of vegetable origin and the state of the plants and the beautiful preservation of their leaves in the accompanying shales precludes the idea of their having been floated from great distances. As the species were evidently terrestrial, we must conclude that some dry land was not far distant; and this opinion is confirmed by the shells found in some strata of the New- castle and Shropshire coal-fields.” The alternation of marine lime- stone with strata containing coal beds may be due to alternate rising and sinking of large tracts, which were first laid dry and then sub- merged again. He is clearly inclined to agree with the suggestion made by Sternberg and Ad. Brongniart, that the beds of mineral detritus were derived from waste of small islands arranged in rows and he thinks that the suggestion is supported by the observation that the Coal Measures flora is of insular type. At a later period, Lyell accepted the autochthonous origin of the coal beds, as appears in the “ Travels in America.” Buckland,”* in 1836, accepted the theory of transport. “The most early stage to which we may carry back its origin was among the swamps and primeval forests, where it flourished in the form of gigantic Calamites and stately Lepidodendra and Sigillarig. From their native bed, these plants were torn away, by the storms and _inundations of a hot and humid climate and transported in some adjacent Lake or Estuary or Sea. Here they floated on the waters, until they sank saturated to the bottom, and being buried in the detritus of adjacent lands, became transferred to a new estate %C. Lyell, “ Principles of Geology,” 5th ed., 1st Amer. ed., Philadelphia, 1837, Vol. L, p. 134. **W. Buckland, “Geology and Mineralogy considered with Reference to Natural Theology, Philadelphia, 1837, pp. 362, 353, 354. PROC. AMER. PHIL. SOC. L. 198B, PRINTED APRIL 24, IQII. 17 18 STEVENSON—FORMATION OF COAL BEDS. [April 21. among the members of the mineral kingdom. A long interment fol- lowed, during which a course of Chemical changes, and new combi- nations of their vegetable elements have converted them to the min- eral condition of Coal.” On an earlier page, Buckland referred to the existence of erect stems in the Coal measures rocks: he was convinced that none of those recorded, aside from some near Glasgow, could have grown where they were found. From this date onward the discussion respecting erect stems, became increasingly important. The facts and the conclusions are alike contradictory. It is better to pass by this matter for the pres- ent and to treat it apart. . Sternberg’ did not accept the hypothesis that coal was formed from peat. He thought that one should conceive of a forest in the ancient time, when neither man nor plant-eating animals existed; that this forest grew for an indefinitely long period in a warm, humid climate; that the offal of buds, leaves, seeds, fruits and decayed stems accumulated on the ground; many generations of plants grew, one on the other, and so a mass, consisting of mold from wood, fruits, seeds, leaves, with complete examples of smaller plants, would be produced, whose surface would be covered with still living vegetation. Conceive now of a cataclysm, when a hurri- cane casts down the living plants and is followed by a flood, loaded ‘with sand and mud—thus one has a true picture of the mode in which the overlying deposits of the stone coal are formed. Cases are rare where one finds erect stems of trees between two coal beds, losing themselves above and below in the coals. The water-cover would hold the mold in place, would bring about decompositions and changes in the different materials and would cover the whole with clay and sand. It is unnecessary to borrow carbon from the air or water in order to get a coal forma- tion, since in this interval, as well in the dry as in the wet way, humus and other acids, bitumen and coal itself have been produced, as occurs even to-day in peat bogs. The material existed in abun- *K. Sternberg, “Versuch einer geognostisch-botanischen Darstellung der Flora der Vorwelt,” Siebenstes und achtes Heft, Prag, 1838, p. 88. 18 _gtt.J STEVENSON—FORMATION OF COAL BEDS. 19 dance and fermentation necessarily followed under the covering of water and sediment. It is unimportant to determine whether the _ water was fresh or salt. In this way, he sees no difficulty in accounting for accumula- _ tion of stone coal deposits, even those of Saint-Etienne, which are 60 fathoms thick. He emphasizes the fact that the particular vege- _ tation of the stone coal period produced colossal stems. Link?* was the first to study the texture relations of coals. He observes that two theories had been offered to account for the origin of coal beds; that of driftage does not commend itself to him, but that referring the coal beds to ancient peat bogs appears more rea- sonable. After summarizing the opinions of v. Beroldingen, de _ Luc, Steffens, Hutton and Leonhardt, he presents the results of his own investigations. Von Buch, feeling perplexed by some recent _ publications, had given him some specimens of coal from Bogota _ and had asked that he study them microscopically. The composition of one of those coals so resembled that of peat that he was led to a _wide study of coals and peats from several horizons and regions. In all peats, whether loose or compact, cell tissues form the _ body of the mass; the difference in quality of the peats being due probably to difference in the plants; the stone coals resemble peat in structure, some recalling the comparatively loose Linum peat used as fuel in Berlin, while others are more like the dense, almost _ wood-like peat from Pomerania; the Mesozoic coals vary, one from _ the Muschelkalk closely resembles peat, but the Liassic coals appear to be composed largely of woody fiber; the brown coal of Green- land is like the Linum peat, while that of Meissner in Saxony is similar to the dense Pomeranian material. __ Link observes two quarto plates illustrating the vegetable struc- tures observed in each of the peats and coals examined. _ Logan’s?* notable memoir on underclays appeared in 1841. He =H. Link, “ Uber den Ursprung der Steinkohlen und Braunkohlen nach mikroskopischen Untersuchungen,” Abhandlungen d. k. Akad. d. Wiss. Berlin, 1838, pp. 33-44. *=W. E. Logan, “On the Character of the Beds of Clay Lying Immedi- ately below the Coal Seams of South Wales,” Proc. Geol. Soc. London, Vol. III., pp. 275, 276. 19 20 STEVENSON—FORMATION OF COAL BEDS. [Agee ae had found almost one hundred coal beds in the South Wales coal- field and, with rare exceptions, each overlies a clay bed from six inches to ten feet thick. The clay varies much in composition but it is a persistent deposit, so that coal beds which have thinned out in the workings have been found again by following the clay. Ordi- narily, Stigmaria occurs abundantly in the clay and Logan thinks that plant was the source of most of the coal. Soon after the field work of the Virginia and Pennsylvania sur- veys was completed, H. D. Rogers*®® gathered the salient facts bear- ing upon the origin of coal beds and presented them in a paper which has become classical. It bears the impress of the time, but it was based on broad observations by the author and his equally celebrated brother, William B. Rogers, aided by a corps of able assistants; the studies, lasting six years, were in detail for an area of somewhat more than 20,000 square miles, but in addition less detailed studies had been made in Ohio and Kentucky, so that the region under consideration was not far from 40,000 square miles. The discussion was the first serious attempt to account for the origin of the Coal Measures, which was based on actual study of a vast area. At the outset, Rogers pronounced against any theory of delta formation, as according to his belief the Appalachian ocean deep- ened toward the west and northwest.*t The deposits are traceable coastwise for 900 miles, so that it seems improbable that fluviatile currents could have assembled them. The sandstones decrease in thickness and coarseness as they re- cede from the ancient shoreline at the east; the shales increase in that direction for a time and then decrease, while the limestones, wholly wanting near the shore line, increase in thickness and purity so as to become imposing before the Ohio River has been reached. The animal remains found in the limestones are marine. There % TH. D. Rogers, “An Inquiry into the Origin of the Appalachian Coal Strata, Bituminous and Anthracitic,” Reps. of Amer. Assoc. of Geologists and Naturalists, Boston, 1843, pp. 434, 459, 463-467. Tt should be noted here that when Rogers wrote the conditions on the west side of the Appalachian basin were not known; but does not affect the general argument. 20 ee, ee Oe ee ee a 1911] STEVENSON—FORMATION OF COAL BEDS. 21 were many alternate periods of movement and of total or compara- tive rest. Limestones indicate periods of comparative tranquillity. Some of the coal beds are of great extent. The Pittsburgh bed had been traced around an area of 14,000 square miles and there are isolated basins holding that bed far southeast from the main area, so that the Pittsburgh coal must have covered a surface of not less 30,000 square miles. The uniformity in thickness and the absence of abrupt variations are as remarkable as the area. These features “‘seem strongly adverse to the theory which ascribes the formation of such deposits to any species of drifting action.” The alternation of laminz of bright and dull coal; the lenticular form of the bright layers; the predominance of mineral charcoal in the dull laminae seem to be almost conclusive arguments in favor of belief that the vegetable matter grew where it was deposited. He finds it difficult to understand why the coal does not consist principal- ly of the larger parts of trees if any drifting agency brought the materials together. The leaves and smaller parts would be detached before the trunks could become waterlogged. But the beds have subordinate divisions, coal, clay, impure coal, so persistent in great areas that miners can recognize their bed at great distance from their own locality ; only one method of accumu- lation can explain this. “I cannot conceive any state of the surface, but that in which the margin of the sea was occupied by vast marine savannahs of some peat-creating plant, growing half immersed on a perfectly horizontal plain, and this fringed and interspersed with forests of trees, shedding their offal of leaves upon the marsh. Such are the only circumstances, under which I can imagine that these regularly parallel, thin and widely extended sheets of carbonaceous matter could have been accumulated.” The purity of the coal is inconsistent with any notion of drifting of the vegetable matter, “which according to any conceivable mode of transportation, would be accompanied by a large amount of earthy matter, such as abounds in all delta deposits and even mingles with the wood as in the raft of the Atchafalaya.” The underclay, irregular in structure, accompanies nearly every coal bed in the Appalachian basin and usually contains Stigmaria 21 22 STEVENSON—FORMATION OF COAL BEDS. [April 21. ficoides with its fibrous processes. The roof contrasts with the underclay and is, normally, a laminated shale due to more or less rapid current and it contains vast numbers of plant impressions. When the roof is sandstone there is evidence of tempestuous cur- rents and the vegetable fragments are trunks and stems of large plants. Occasionally limestone forms either roof or sole of the coal bed but there is usually a very thin layer of calcareous shale parting them. No hypothesis, thus far presented seemed satisfactory to Rogers, and he presented his own to account for origin of the Coal Measures. He imagined extensive flats bordering a continent, the shore of ocean or bays, beyond which was open sea. The whole period of the Coal Measures was characterized by a general slow subsidence of the coasts, interrupted by pauses and gradual upward movements of less frequency and duration, and these merely statical conditions alternated with great paroxysmal displacements of the land. During gentle depression, the coast was fringed by marshes while arborescent plants were on the land side. The meadows would give pulpy peat; leaves blown in or moved by higher tide would rest on the peat ; some would be buried and become pulpy, or, in some cases, by rapid re- moval of volatile constituents would remain as mineral charcoal. An earthquake comes. Water is drained from the swamps and their tributaries; muddy water draws from swamp and swampy forests leaves and the rest to distribute them with the mud over the bog. This is the laminated shale. The sea returns, rolls over the swamp to the dry land; withdrawing, it brings uprooted trees, and washed off soil, strewing the land stuff in a coarse promiscuous stratum. — Repeated waves would add to the mass. The disturbance ends; coarse materials sink, then the less coarse and last of all the finest sediment, light vegetable matter and the buoyant stems of Stigmaria, would sink together. A new marsh would be made and once more the savannahs would be clad with vegetation. This he terms the paroxysmal theory. Petzholdt *? found two questions involved in the problem; were coal beds formed during a brief period and were they formed im situ A Petzholdt, “ Geologie,” Zweite Auflage, Leipzig, 1845, pp. 413-417. 22 lis Re eC ee as 1911.] STEVENSON—FORMATION OF COAL BEDS. 23 or from transported vegetable material. The answer to the first question is certain—a great period of time was required for forma- tion of the coal beds and their associated strata; but the second question is more complex and he is inclined to believe that both methods are possible, though there may be difficulty in determining which prevailed at a given locality. Vertical stems are not decisive, for they are found at times in rocks formed by transport, while prostrate stems occur in deposits clearly made im situ. He believed that there were no continental areas during Carbonif- erous times, that the dry land consisted only of islands. For this reason, it is impossible to accept the hypothesis that coal was formed in great lakes or at the mouths of rivers. The only method of formation by transport would be the driving of great masses of vegetable matter against an island, which would collect in the quiet eddy on the opposite side, where, becoming waterlogged, they would sink and be covered with mud. He clearly prefers the doctrine of origin in situ. An island, heavily forested for an indefinitely long period, be- comes covered by a mass of bark, wood, etc., and similar remains of small plants. If the island be flooded by the outburst of granite and consequent elevation of the sea-level, the vegetation will be pro- strated. By frequent outbursts the sea-level will be raised perma- nently and the island remains submerged. Deposits of sand and mud bring the island again to the surface of the water ; a new forest rises on the grave of the old one. He thinks the alternation of strata and the formation of coal in situ can be explained very simply in this Way. Murchison,** after his study of the Donetz field in Russia, was convinced that the doctrine of transport alone could explain the conditions. The sections in southern Russia show “that the hypo- thesis of the formation of coal beds by masses of vegetation which there grew having subsided in situ (the truth of the application of which to some basins we do not deny) cannot be applied to the cases in question any more than to the pure marine coal beds of the north- ern districts, Northumberland and the northwestern parts of York- “R. I. Murchison, “The Geology of Russia in Europe and the Ural Mountains,” London, 1845, Vol. I, pp. 112-114. 23 24 STEVENSON—FORMATION OF COAL BEDS. [April 21. shire, etc.” Limestones with marine fossils are found at various horizons in the Donetz section. The presence of an underclay proves nothing—even though Stigmaria ficoides be the only plant present for a confused assemblage of plants is seen above and below the coal beds and the fossil beds are exclusively marine. The fine underclay indicates only that the sea bottom was covered with detritus of plants washed in by floods; the heavier earthy matters, accompanying the detritus, sank to the bottom, while the plants floated and formed the upper stratum. Those plants, thus left on the muddy slime, were covered afterwards by other sediment. Much of the coal, in strata alternating with marine sediments, may have come from the wash- ing away and sinking into the sea of floating masses of matted earth and plants. At a later date,** he discussed the question more broadly. He refers to the terrestrial conditions exhibited in the Upper Carbonif- erous of England and to the lack of a physical break there between the Lower and the Upper Measures, such as appears in Germany and France. In those countries, the later accumulations may well be accounted for by depressions of low woodlands and jungles beneath freshwater, followed by elevations and depressions. There is no physical break in Britain, but there is the same passage from marine to terrestrial conditions, of which the coal beds offer posi- tive evidence; for the roots of Sigillaria are found in the underclay, which was the soil of a primeval marsh or jungle. The view, which supposes many and successive subsidences of vast swampy jungles beneath the level of the waters, best explains how the different organic masses became so covered with beds of sand and mud, as to form the sandstone and shale of such coal fields. But this theory of oscillations . . . can scarcely have an application to those other seams of coal, which, as before mentioned, are interstratified with beds containing marine shells, the animals of which, such as Producti and Spirifers, must have lived in comparatively deep water.” He conceived that the latter class is to be explained only by the supposition that great rivers, flowing through lowlands, trans- ported vast quantities of trees, etc., entangled in earth, and de- *“ Siluria,” 3d ed., London, 1859, pp. 315-317. 24 rg1t.] STEVENSON—FORMATION OF COAL BEDS. 25 3 _ posited them on the bottom of the estuaries, or that vast heaps of Ss organic matter were carried as floating masses to the sea. The _ Northumberland deposits, large tracts of Scotland, as well as the _ Donetz field in Russia offer fine proofs of these conditions. There - were at least two modes in which coal measures were formed, one terrestrial, the other subaqueous. ____ Goeppert®* in his elaborate work on the formation of coal beds gave the results of many years of study in the Silesian coal fields. _ A large part of the volume is devoted-to determination of the _ materials forming coal; it will be considered in another connection. The chapter on the formation of coal beds is supplemented by a mass of illustrations drawn from the coal fields of Silesia, the whole discussion being so compact, so free from unnecessary detail that _ to make a just synopsis is difficult. The standpoint in Goeppert’s _ work differs much from that in the discussion by Rogers, the only _ preceding study with which it can be compared. Rogers knew 4 little about the intimate structure of coal itself and reasoned wholly from stratigraphical conditions ; Goepert was a skilfull paleobotanist _ as well as stratigrapher. © Be The important question for Goeppert is, were the coal beds _ formed of plants growing in place or of plants brought in from other localities. There were many islands, mountains, valleys, rivers, etc., in the Coal Measures time. The organic matter was deposited on plains which were covered with sand, clay or mud. The extent of the deposits, their occurrence as plains or as basins show that they were laid down on the sea-bed, on slowly changing coasts or in enclosed sea or lake basins. The few marine products found in coal beds do not favor the opinion that the coal-forming material was collected from distant places and deposited in the depths of bays; everything indicates the utmost quiet; the vegetation covered *H. R. Goeppert, “ Abhandlung eingesandt als Antwort auf die Preis- frage— Man suche durch genaue Untersuchungen darzuthun, ob die Stein- kohlenlager aus Pflanzen entstanden sein, welche an den Stellen, wo jene gefunden werden, wachsen; oder ob diese Pflanzen an anderen Orten lebten, und nach den Stellen, wo sich die Steinkohlenlager befinden, hingefiihrt wurden?” Amsterdam, 1848, pp. 119-131, 136-139, 141-160, 278, 270. 25 26 STEVENSON—FORMATION OF COAL BEDS. [April 21, the low-lying horizontal sea-strand. Changes of level, elevation and subsidence, led to burial of the plants under the ocean; sand and clay were deposited on the plant covered surface; dunes were formed, on which plants grew to run the same course. Through repetition of this process, the different beds were formed, separated by sand and clay. The conditions were like those of the present day, for submerged bogs and forests have been observed at many places along the coasts of Europe and America. Well preserved stems are wanting because the plants lacked a dense interior structure. Filled stems are rare in Tertiary deposits because the trees were dicotyledonous; whereas they abound in the Coal Measures because the loose interior structure decayed quickly. Plants grew in these hollowed stumps; Goeppert found Lepido- dendron, Calamites and ferns in decayed Sigillaria; in the stump of Lepidodendron he found the stem of a new genus, two feet long and vertical. If the coal had become compact or if -the quiet were undis- turbed, the boundary between coal and the succeeding deposit is sharply defined; at most one finds only impressions of stems lying upon the upper surface. This latter condition occurs frequently in Upper Silesia, where the coal is composed chiefly of Sigillaria. It is quite true that filled stems occur even within the coal itself; Goeppert found them. He explains their presence by supposing that clay and sand were brought down by floods before con- solidation of the coal, before the spaces between the stems had been obliterated by compression. In the same way he accounts for Brandschiefer or bituminous shale; the influx of muddy waters caused the alternation of laminae of bright coal, containing 2 per- cent of ash, with dull layers, containing much mineral charcoal and 20 to 25 percent of ash. The overlying beds were deposited after complete formation of the coal bed and the time-interval between the two deposits is as variable as the intervals interrupting the formation of a coal bed itself. Partings in coal beds show how the time required for dif- ferent types of deposits may vary. A-parting, ten inches thick, may be equivalent in time to a sandstone deposit elsewhere, many fathoms 26 re is ee. ggrt.] STEVENSON—FORMATION OF COAL BEDS. 27 thick. Perhaps one may regard layers containing great abundance of plants as equivalent to deposits in which the plants do not P 3 form beds, because in the latter case the plants were brought in con- termporaneously with the sand and mud masses. He is convinced that the coal and the enclosing sandstone or shale beds are wholly independent deposits. And this belief is strengthened by the fact _ that the material, filling stems in coal, clay or sandstone, differs from that which surrounds them—an additional evidence of the extreme quiet prevailing during deposition. Goeppert was the first to recognize that the coating of the filled stems is the converted bark. The roots of Sigillaria ‘and Lepidodendron were feeble, as are those of related plants to-day, and the trees were overthrown easily ; and thus it happens that the stems, as in Upper Silesia, con- tribute to the formation of the coal. When overthrown, their cel- lular interior was squeezed out and converted into coal, as is seen near Dombrowa. All the phenomena indicate that the coal deposits © _ were made during conditions of quiet, which would be impossible unless the plants grew where the coal is found. The vast extent and constancy in structure exhibited by coal beds is important. He cannot think that such a mass could be floated in _ at once, yet how could it be deposited so regularly by any other means? He agrees with Lindley and Hutton and with Burat that _ the mass is too great for transport. He is unable to believe that _ the coal was the product of forests, because the amount is so vast; _ but the evidence satisfies him that the plants have not come from a distance. He prefers to accept the opinions presented by v. Berold- ingen, De Luc, Ad. Brongniart, Link, and to believe that, if not all coal beds, at least the thickest originated as peat bogs—the more so because of the resemblance which a buried peat bog has to a coal bed. He conceives that on the damp floor there grew lycopods, cala- a mites, ferns, stigmaria and other plants, corresponding to the crypto- _ _ gams and monocotyledons of present day bogs. Tree-like Sigillarie id Lepidodendra grew on the borders of the bog and at times were uprooted by floods. He laid great stress on the preservation of the plants, as precluding the possibility of transportation; he finds the mode of decay of tree stems equally important, for the conditions observed in Calamites are the same with those found in his experi- 27 28 STEVENSON—FORMATION OF COAL BEDS. [April 21, ments with Arum. The presence of vertical stems is noteworthy, because they are so numerous. It is possible for floods to carry away whole trees and to deposit them in vertical position; that oc- curred in the great débacle near Martigny in Switzerland. This explanation would suffice for an isolated instance ; but the number of such stems in the Coal Measures is too great; the analogy is in submerged forests of our own day. The distribution of plants, both vertically and horizontally, has an important bearing on the subject. At one locality the flora may consist almost wholly of one species and at another, almost wholly of another species. There is a group-like distribution, so to speak, a social occurrence. In Upper Silesia, the coal may be termed Sigillaria coal, while in Lower Silesia it is Stigmaria coal. He asserts that an observer, in viewing the coal bed, involuntarily thinks of a peat bog. Lyell’s volumes on his second visit to the United States appeared at this time and had material influence in moulding public opinion. They will be cited in another connection. Naumann*® recognized the distinction between deposits formed on the sea border and those in fresh-water lakes, as had been done by Elie-de-Beaumont and Burat. The former contain, especially in their lower portions, rock layers with organic remains correspond- ing to the marine mode of formation, while the latter, less extensive, have no traces of marine fossils or anything else to show co-working of the sea. These types he terms paralisch and limnisch. These terms are equivalent to pelagic and mediterranean of Elie-de-Beau- mont, to terrains houillers de haute mer and terrains houillers des lacs of Burat. The coal deposits of Great Britain, Belgium, West- phalia, Russia and America are paralisch or pelagic ; those of central France, Saxony and Bohemia are limnisch or lacustrian. The prevailing rocks of the Carboniferous are conglomerate, standstone and clay shale, which occur in paralisch and limnisch alike. They are derived mostly from destruction of other rocks and their materials were transported. The land consisted not of small low-lying islands but mainly of great islands and continents * C. F. Naumann, “ Lehrbuch der Geognosie,” Leipzig, 1854, Vol. II., pp. 451, 452, 571-580. 28 rgtt.] STEVENSON—FORMATION OF COAL BEDS. 29 - with mighty rivers, along whose coasts and in basin shaped depres- _ sions was deposited the vast system of sand and mud strata: This at length became marshland, offering the ground for the luxuriant vegetation of the first coal bed. In the Appalachian region, there may have been the flat coast of a land extending far to the east, from which great rivers carried sand and mud into the shallow sea at the west, in which, farther away, limestone was forming. Proc- esses such as those now seen in the Nile, Mississippi, Hoangho and other rivers, continuing for many thousands of years, would raise the sea bed until it reached the water surface as a wide-spread marshland. Similar operations were going on in freshwater basins of the dry land leading to the formation of morasses, supporting Calamites, Sigillarie and other Carboniferous plants, which would give a deposit of peat. The alternation of a great number of coal beds with thick masses of sandstone and shale is not so easily explained as is the origin of the first coal bed. The causes in paralisch areas are different from those in limnisch basins. Lyell, Lindley and others held the opinion that seacoasts, on which paralisch deposits were formed, underwent slow subsidence during Carboniferous time. If one suppose that this subsidence was interrupted periodically, we have a mechanism by which the forma- tion of successive coal beds could be explained. A similar result would be secured by occasional elevations of the sea-bottom, ac- cording to Petzholdt’s conception. There is necessary in each case a general rise of the sea-level to cover the plant deposit with the sandstone and shale needed to give another swampy surface. This alternating subsidence and stability of the sea-bottom explains why _ the shale, covering coal beds, encloses a mass of plant remains and also why paralisch territories may have many but thin coal beds. This explanation is not wholly satisfactory for limnisch areas, since one can hardly suppose that all of those could have suffered the repeated subsidence. One must conceive that between longer periods of stability there were epochs in which increased fall of inflowing streams or a diversion of flow occurred. The greater carrying power of the streams would bring the plant deposit and 29 ~ 30 STEVENSON—FORMATION OF COAL BEDS. [April 21. at length form a new surface on which vegetation would begin once more. This would give a smaller number of beds. The, at times, great thickness and the frequent irregularity of coal beds in lim- nisch areas may be explained in part by supposing that they were not formed wholly as peat deposits, but received masses of uptorn vegetation, swept out by floods, and this leads to the question of the formation of a particular coal bed. There are two theories, transport (Anschwemmung) and in situ (an Ort und Stelle). Both may be correct. The great beds, beyond doubt, are of im situ origin, but there are many deposits boas can be explained only by transport of plant masses. It is known that streams bring down astonishing quantities of plant material; that ocean currents carry driftwood far and that it accumulates in vast masses on shores. Currents of the olden time must have been similar. If the widespread masses were buried under sediments, they would be transformed into coal beds. Neu- mann thinks that repetition of this process at mouths of streams in lakes or on the sea-coast would give a system of strata like the present series of coal beds with intervening sandstones and shales. Such drift masses are irregular in extent and thickness, often as blocklike masses. Such transported material would give conditions like those observed in coal beds of some limnisch areas, great irregu- larity and variation in thickness, breaking up into separate benches, some of them excessvely thick. He thinks that under especially favorable conditions a coal bed might be formed in this way which would resemble one formed in situ. He considers also that this theory of transport explains many regular coal beds, such as those between limestones or other strata distinctly marine, as well as beds resting directly on granite, limestone, etc., without an underclay. He agrees with Murchison that in some cases the transport sie has value. But for the greater part of the coal beds, the in situ theory must be accepted; their material was produced by vegetation an Ort und Stelle. All beds continuous over great areas, with regular and not too great thickness and with a stigmaria-filled underclay are to be explained in this way. But one must not think that there were real forests, which were thrown down in place, compressed by in- 30 rr.) STEVENSON—FORMATION OF COAL BEDS. ea : ‘coming sediments and changed into layers of plant material. The = Carboniferous was not a tree and forest flora; it was morass and strand vegetation, developed on great emerging plains of marshland. a The prevailing forms suggest that formation of the widely extended a coal beds was analogous to the formation of peat bogs. 2 _ The purity of coal substance, the continuity of the beds, their _ regular thickness, the arrangement in benches due to clay layers 4 _ produced by inconsiderable inundations, the upright plant stems and all the remaining relations of most coal beds appear to find sufficient explanation only in this or a similar conception of the mode of their 4 formation. When at length a permanent elevation of the sea-level comes, the bog is buried under sand and mud, in whose first layers, just as in the last conditions of peat vegetation, a great mass of = plant remains is found, torn from the neighboring land; so that it is _ clear that the roof shale of a coal bed encloses as a rule a large _ number of isolated plant remains. ie. Newberry’s® attention was attracted to the cannels and semi- _ cannels of Ohio at the beginning of his studies. Observations _ made in peat bogs of this country and Europe led him to believe that cannel was formed in lagoons, where completely macerated vegetable tissue, probably parenchyma for the most part, accumulated as vege- table mud. Among other arguments favoring his hypothesis, he urges that cannel is more nearly homogeneous than cubical coal ; that it contains mere volatile matter, with more hydrogen, and must have been deposited .n a hy drogenous medium which prevented oxidation ; that it contains aquatic animals, so abundant at times, as to prove that they inhabited pools in which cannel was a sediment; that the plant remains in cannel are usually skeletonized; and that in open water lagoons of modern peat marshes, fine carbonaceous mud ac- cumulates, which when dried is very like cannel. Le Conte** compared the peat bog and estuary theories. The arguments in favor of the peat bog theory are, the purity of the coal, the fine preservation of the tender and more delicate parts * J. S. Newberry, Amer. Journ. Sci., 1857. A synopsis of this paper. with _ some additional notes was given by him in Geol. Survey of Ohio, Vol. II., 1874, p. 125. _ ™ Joseph Le Conte, “Lectures on Coal,” Ann. Rep. Smithsonian Inst. for 1857, Washington, 1858, pp. 131-137. : 31 32 STEVENSON—FORMATION OF COAL BEDS. [April 21, of plants, the position of these plants in the roof shale, the com- pletely disorganized condition of materials in the coal, the presence of the underclay, with roots and the occurrence of vertical stems rooted in the underclay. The chief objection to the theory is the repeated alternation, in the same locality, of coal seams with marine and freshwater strata. There being as many as one hundred coal seams, it would appear as though the same spot has been raised above water level and had been depressed below it at least one hundred times. The estuary theory was proposed to avoid this difficulty. As an estuary at the mouth of a great river is occupied now by salt- and again by fresh-water, it should contain alternating deposits of marine and fresh-water origin. In seasons of freshet, the salt water is pushed out and the river water, loaded with mineral detritus and timber rafts, makes its deposits; during low water, the sea returns and marine deposits follow. Le Conte finds insuperable objections to the latter. He thinks that coal beds were formed as peat bogs at the mouths of large rivers. The analogy is to be sought, not in the bogs of Ireland, but in those of the Mississippi delta. He supposes a vast delta, with spaces protected by fringes of plants from influx of river muds. There pure vegetable matter would accumulate until during some violent flood the barrier would be broken down and the whole space covered by mud. The delta, like that of the Mississippi, subsided slowly and the covering of mineral detritus eventually became ground for a new marsh. If the subsidence were more rapid than the river deposits could overcome, the sea would take possession and limestone would be formed. There is no necessity for con- ceiving repeated upheavals and depressions. ‘‘ Coal has almost cer- tainly accumulated in situ in extensive peat swamps at the mouths of large rivers, upon ground which was slowly subsiding during the whole period.” - Lesquereux,®® after long study of peat bogs in Europe, came to the United States, where as palaeobotanist to several official sur- ° TL. Lesquereux, Paleontological report on fossil flora of the Coal Measures, Third Ann. Rep. Geol. Survey of Kentucky, Frankfort, 1857, pp. 505-522. 32 — rtrd STEVENSON—FORMATION OF COAL BEDS. 33 _ yeys, he examined coal beds within a large part of the Appalachian 4g and Mississippi coal fields. His first report upon the work in Ken- a tucky is prefaced by discussion of matters relating to the origin of _ coal beds as illustrated by conditions in the Appalachian basin. . Bog plants are partially immersed and ordinarily are woody. _ The trees are mostly resinous and are such as can thrive only in bog conditions. The Coal Measures plants are ferns, clubmosses, _horsetails, reeds and rushes, in character much resembling the forms prevailing in modern bogs. The peat of the Great Dismal and Alli- _ gator swamps rests on white sand and fills the depressions, while 4g its surface is covered by canes, reeds and shrubs; where there is 4g a cover of water, the soft black mud supports cypress and magnolia, 4 and a great mass of material is added each year. Some ponds were q once covered with vegetation, now sunken, as in Lake Drummond, _ which has at its bottom a forest, probably carried down by its own _ weight. He found similar phenomena in Sweden, Denmark and g Switzerland. The water, to permit formation of peat, must have a _ constant level and be stagnant. The clayey bottom of bogs was _ made by fresh-water mollusks and infusoria or by Characee and - Conferve. Peat always has this mud. Comparing these conditions with those prevailing in the Coal _ Measures, Lesquereux finds: (1) The fireclay varies in thickness, - color, composition and in the quantity of Stigmaria; sometimes no coal rests on it—the soil was ready but conditions did not favor accumulation. Yet fireclay, without coal at one place, is likely to bear coal elsewhere. (2) The coal varies abruptly in physical and chemical features, just as peat varies in all directions, horizontal and vertical; and these variations depend largely on the plants con- cerned as well as on the amount of foreign matter introduced. (3) The roof shales, usually very fine, are evidence of slow subsidence, sometimes without marine invasion, as shown by plant remains; sometimes with marine invasion, as where the shales contain shells of brackish water type. (4) The limestones, equivalent to or con- tinuation of the shales, need quiet deep seawater. Influence of the sea is very distinct in erosions due to currents. (5) The sandstones were due in many cases to turbulent waters, as appears from the PROC. AMER. PHIL. SOC., L, I98C, PRINTED APRIL 24, IQII. 33 34 STEVENSON—FORMATION OF COAL BEDS. [April 21. erosions and the mighty erect trees. The sand may have been derived possibly from dunes such as those on the Rhine or Elbe. Lesquereux knows of no peat composed of fucoids and marine plants. Jukes’s*® contribution to the discussion is not less important than those by Rogers and Goeppert, as it is the first presentation of the transport theory based on careful observation in an extended area. It covered the ground so thoroughly that little aside from detail or local coloring has been added since its publication. Two opinions exist respecting the origin of coal beds; the first is that trees and plants were drifted into lakes, estuaries and shallow seas, where, becoming waterlogged, they sank to the: bottom and became covered by the other accumulations; the second is that the plants were not drifted but grew and perished on the spot where they have formed the coal, just as our peat bogs would form coal if long buried under a great mass of earthy matters. While he does not purpose to range himself as an advocate of either opinion, he finds difficulties in the way of the latter which make him hesitate to accept it exclusively. These, observed in the South Staffordshire coal-field, he gives in detail. 1. The “rolls,” “swells”? or “horsebacks,” which are ridge-like accumulations of clay rising sometimes eight feet above the floor, cannot be explained if the coal were formed at or above the level of the water; but if coal and “swell” alike were formed under water no difficulty exists. 2. The “rock faults’ in the Thick coal. These are of two kinds. One, which he has not seen, is due to erosion of the coal after deposit, the hollow being filled with the material deposited on the coal. The other comes from contemporaneous deposition of silt or sand with the coal, so that they alternate at short intervals. The coal encloses cakes, layers or masses of sandstone, more or less inter- mingled with it. One such “ fault” seen by Jukes, was 286 yards wide and it had been followed 400 yards without reaching the end. The upper part of the coal bed passes over the sandstone. At the “Jj. B. Jukes, Memoirs, Geol. Survey of Great Britain. “The South Staffordshire Coal-field,’” 2d ed., London, 1859, pp. 34-42, 44-49, 201-206. The writer has not seen the first edition, published at least ten years earlier. 34 i — ns x91] STEVENSON—FORMATION OF COAL BEDS. 35 - lateral border, both coal and sandstone split up so as to interlace. The condition is precisely similar to a cake of sandstone in clay. ukes asks, if the sandstone was deposited in water, why not the coal also, for they are interstratified. The partings of sand in coal beds are of the same type. The laminz of coal are obviously laminz of deposition; their arrangement and their alternation with films of shale or with thicker partings of clay or sand would all be __ explained by the gradual deposition of laminz and strata of dif- 2 ferent kinds of substances and by different degrees of mingling at the bottom of some body of water. a 3. The extreme bifurcation of some coal beds; and here are phenomena extremely perplexing from the standpoint of the in _ situ theory. The great bed near Dudley, known as the Thick coal, E is composed of numerous benches, each with its own persistent _ peculiarities. At two miles north from Dudley there are eleven : _ benches, with 36 feet 6 inches of coal and 2 feet 11 inches of part- ’ _ ings; while at one mile east from Dudley, there are thirteen benches _ with 28 feet 7 inches of coal and 1 foot 9 inches of partings. But at two miles east of north from Dudley, the upper two benches, _ there known as the Flying Reed coal, are at 84 feet above the Thick - coal; at two miles farther, the interval has increased to 204 feet, hile an intercalation of 10 feet appears midway in the Thick coal below. The benches retain their distinctive features throughout. imilar conditions prevail toward the west, where the interval be- tween the Flying Reed and the other portion of the Thick coal increases from almost nothing to 128 feet within barely three miles. There is a higher bed known as the Brooch coal. It is 95 feet above the Flying Reed, where that bed is 10 feet 6 inches above the Thick; but where the latter interval becomes 115 feet, the former is only 30 feet. Thus, while the Brooch and Thick are rudely parallel, e Flying Reed is oblique between them. ; The normal section persists in the central southern part of the _ field to some distance south from Dudley ; but toward the southwest the Thick coal breaks up, loses its structure and becomes worthless ; oward the southeast, the bed thins out, has little good coal and is troubled by “ rock faults” or “cakes of sandstone.” 35 36 STEVENSON—FORMATION OF COAL BEDS. [April 21. An additional difficulty is found in the expansion of the Thick and other coal beds toward the north. The expansion of the whole series and the splitting of the beds in that direction seem incom- patible with the idea that the coal beds were formed at or above the surface of the water, while the intervening strata were deposited under it. Of the intervening rocks, those of coarse material are heaped up usually and thin out rapidly in all directions, while those of fine material have a greater area. This is true of superimposed beds forming a group; when material is fine, the disappearance of a bed is gradual. This law of area and thickness means only that fine materials were spread over a larger area “in consequence of their comparatively light specific gravity, or at least of their being more easily and therefore more widely transported by water, and being more generally diffused through it before finally coming to rest at the bottom. It was pointed out before, too, that beds of coal so far from forming any exception to this general rule, are its most marked example at the one extreme, while coarse sandstones and conglomerates form the most striking example at the other. . I wish merely to say as the result of an experience of a good many years, confirmed by the particular instance under examination, that the phenomena of the lamination and stratification of beds of coal, and their interstratification and association with other stratified rocks are explicable solely by the relation of the specific gravity of their materials to the action of moving water, and the consequent diffu- sion of their materials through the mass of that water.” The materials of the clays and sandstones were most largely deposited on the northern side of the coal field and sometimes failed to reach the southern part of the area, whereas the coal beds “were diffused equally, or at least more equally, over the whole area.” He finds in the Bottom coal bed a notable illustration of these conditions—and it is only one of many. One “cannot fail to be struck with the obvious ‘ delta-like’ or ‘ bank-like’ form which the Coal Measures of South Straffordshire must have originally possessed, and the perfect resemblance they must have had to an undisturbed subaqueous accumulation. It seems to me then impos- sible to suppose otherwise than that the whole series of the Coal 36 1g1t.] STEVENSON—FORMATION OF COAL BEDS. 37 Measures, coals included, were deposited by one connected operation of the same forces acting in obedience to the same physical laws on similar but slightly differing materials, through an indefinite but immensely long period of time.” _ Dawson spent many years in investigation of the Acadian coal fields, but devoted his attention especially to the South Joggins region where exposures are almost complete in a section of more _ than 11,000 feet thickness. He visited that locality with Lyell in F : 1852 and 1853 and afterwards made detailed study of each coal bed as well as of each ancient soil, subjecting samples from all to careful macroscopic and microscopic examination. The results of his 3 studies were given in several memoirs and the details were pub- a lished in the second edition of “ Acadian Geology.” In his first _ elaborate memoir* he called attention to the gradual passage from coal to the roof shales through lamine composed of coaled leaves and flattened trunks, separated by clay. He expresses the opinion that erect forests explain to some extent the accumulation in situ. The sandstone casts of Sigillarie are enclosed in bark converted into caking bituminous coal, while remains of the woody matter remain as mineral charcoal at the bottom of the cast. A series of such stumps with flattened bark and prostrate trunks may consti- tute a rudimentary bed of coal, of which many occur in the South Joggins section. He is convinced that the structure of the coal accords with the view that it accumulated by growth and not by 5 driftage and that accumulation was very slow. He regards Sigil- _ laria and Calamites as the chief contributors to the formation of coal. The woody matter remains mostly as mineral charcoal, while the cortex and such other portions as were submerged gave the compact coal. This memoir is concerned, for the most part, with _ the origin of coal. 7 In a later memoir,*? he considered especially the subject of accu- _ mulation. After describing the formations and the physical condi- “J. W. Dawson, “On the Vegetable Structures in Coal,” Q. J. G. S., Vol. XV., 1859, pp. 638, 640. ““On Conditions of the Deposition of Coal, more Especially as Illus- trated by the Coal Formations of Nova Scotia and New Brunswick,” Q. J. G. S., Vol. XXII., 1866, pp. 95-104. 37 38 STEVENSON—FORMATION OF COAL BEDS. [April 21. tions observed in the numerous coal beds, he presents these con- clusions: (1) The occurrence of Stigmaria under nearly every bed of coal proves accumulation in situ; the sediments between the beds prove transport of mud and other materials, the conditions being those observed in swampy deltas. (2) True coal consists mostly of bark of Sigillarid and other trees, leaves of ferns and Cordiates with other débris, fragments of mineral charcoal, all grown and accu- mulated where they are found. (3) Microscopic structure and chemical composition of cannel and earthy bitumen as well as of the more highly bituminous and carbonaceous shales prove that they were fine vegetable mud as in the ponds and lakes of modern swamps. (4) A few underclays consist of this vegetable mud, but most of them are bleached by drainage. They contain not sulphide but car- bonate of iron; rain, not seawater, percolated through them. (5) Most of the erect and prostrate trees had become hollow shells of bark before final embedding and their wood had been broken into cubical pieces of mineral charcoal; land snails, galley worms and reptiles were caught in them. There is much mineral charcoal on surfaces in all the larger coal beds. (6) Sigillaria roots have much resemblance to rhizomas of certain aquatic plants, but structurally are identical with cycad roots, which the stems resemble. Sigillarie grew on soils supporting conifers, Lepidodendra, Cordaites and ferns, which could not grow in water. There is remarkable absence of aquatic vegetation. (7) The occurrence of marine or brackish water forms is no evidence of sub-aqueous formation. The same condition is observed in the case of submerged forests. The channels, sand or gravel ridges, inequalities of floor observed in coal beds are familiar features of modern swamps. The lamina- tion of coal is not aqueous lamination; it is the superposition of suc- “cessive generations of more or less decayed trunks and beds of leaves. It is very different from the lamination observed in cannels and in the carbonaceous shales. The doctrine that coal is composed of the débris of land plants, though maintained by nearly all students, did not pass unchallenged. As far back as 1815, Parrott suggested that seaweeds had contrib- 38 Per 918.) STEVENSON—FORMATION OF COAL BEDS. 39 uted materially to the formation of coal and, at a later date, Bischoff conceived that the Sargasso sea might yield a coal bed. Mohr,** in 1866, presented this view with great energy, and his opinions received more or less support from some eminent students. _ Mohr contrasts stone- and browncoal, the one being fusible the other infusible. Land plants with much woody fiber yield charcoal, which soon decays when exposed to air and moisture. But sea- weeds, river and lake alge, having no fiber, decay to slime, which _ hardens through loss of CO, and CH, the original composition being q that of starch and the allied substances. He combats Bischoft’s assertion that Calamites and other land plants were concerned in forming coal, for the mass of the coal is amorphous and no treat- ment gives trace of plant skeleton. Evidently, everything with rec- _ognizable structure is a foreign body. Coal did not originate from land plants but from water plants, whose growth is protected from air and decay. Only one of these water plants, a grass of wide distribution, is a phaenerogam ; the genera and species of the others are very numer- ous and their mass is inconceivable. The Sargasso sea alone has an area seven times as great as that of Germany and none of its material can escape. Here is ample material; contributions from land plants are only accessory. The ash from sea weeds contains clay; that from coal, lignite and peat consists of silicates not _ belonging to plants and contains clay. This material is derived from land detritus. The ammonia in distillates from coal is of animal origin; no accumulations in landlocked basins could have animals enough to supply this ammonia, but Darwin and Meyen have described the vast numbers of mollusks and other forms attached to seaweeds. _ Sea plants are swept away, decay and sink or are distributed by currents. They are heaped up to great thickness, there being 338 feet of coal in the Saarbruck basin. Darwin saw immense masses _ of seaweed, floating, so constant in position that they are mapped _as reefs and sand banks. If a layer of leaf coal occur, it is evi- dence only of material brought in from the land. The absence of *“F. Mohr, “Geschichte der Erde,” Bonn, 1866, pp. 82-100, 130, 137. 39 40 STEVENSON—FORMATION OF COAL BEDS. [April 21. animal remains in stone coal is due to the solvent power of carbon dioxide coming from the decomposing seaweeds. Muck** came strongly to support of Mohr’s doctrine in the first edition of his work. The essential objections to the theory are: (1) That great accumulations of seaweed are not likely to reach the bottom; (2) that remains of seaweeds have not been found in dredgings, which bring up only inorganic materials and animal remains; (3) the poverty of earthy materials in stone coal; (4) the absence of sea plants, and (5) the rare occurrence of sea shells in stone coal. The answers to these objections are: That the first is based on supposition, originating in lack of knowledge ; that, for the second, it may be well to wait for its invali- dation by opposing facts; as for the third, it stands in close connec- tion with the second and so may be of narrowly conclusive value, but it is to be remarked that the ash-poor glance coal alternates with the often very ash-rich matt- and cannel coal, whose ash does not proceed from beds intervening between the coals, but is so intimately mixed with the coal stuff that it can be due in only small degree to later infiltration ; as for the fourth, absence of sea plants is explained by the fact that those plants, in dead or torn condition, with or without access of air, undergo decay very quickly, becoming, within a few weeks or months, a structureless mass, in which organic remains cannot be recognized; the fifth is answered very easily, for animal remains are calcareous and are removed by carbon dioxide which originates during the coal making process. In the second edition of his work, Muck, though no longer urging the theory, argued that sea plants, embraced under the trivial term “Tang,” offered and do offer enough material for stone coal forma- tion. The disappearance of organic structure in ‘stone coal is ex- plained as easily for seaweeds as for land plants by a kind of peaty fermentation. The morphological differences between seaweeds and the land plants corresponds to chemical differences in composition. Petzholdt*® gave a more than halting adhesion to this doctrine “F. Muck, “ Die Chemie der Steinkohle,”’ Leipzig, 1st ed., 1880; 2d ed., 1891. The citations are from the second edition, pp. 162-165, 168. * A. Petzholdt, “ Beitrag zur Kenntniss der Steinkohlenbildung,” Leipzig, 1882, pp. 25, 26, 27. 40 tgtt.] STEVENSON—FORMATION OF COAL BEDS. 41 _ though without mentioning Mohr in connection with it. Referring to the current opinion that the material for formation of coal may be wholly or at least in great part derived from land plants, he says that this is evidently pure hypothesis, for remains of undoubted land plants occur in coal only under exceptional conditions. As, at the time when stone coal and anthracite were formed, the land was _ sunken, it is doubtful if the then production of land plants could yield the vast quantity required for the coal beds, he is led to look elsewhere for suitable material—and that, the sea plants appear to have produced. Remembering that the fauna of the Coal Measure _ time was marine and that, for these vast numbers of genera and _ species, the nourishment could come only from alge, he asks with Bischoff, “where are the remains of the vast masses of sea plants, which since the Plant Kingdom first appeared on earth, have grown and then perished?” He replies that they have been consumed in _ forming coal and anthracite beds; and he is compelled to admit the conclusion that algz, not land plants, produced the chief material for coal-making. At the same time, he is careful to state that this is only hypothesis, without direct proof, since remains of alge are as rare in coal as are those of land plants. Mietzsch** devoted much space to discussion of this hypothesis, which he regarded as baseless. His objections are those tabulated by _ Muck in the work just cited. In the concluding part of his argu- ment, he points out that the Challenger expedition crossed the ocean _ along several lines and that the results of dredging leave uncertain _ whether seaweeds, after death, reach the bottom, become decomposed at the surface or become covered with animal remains. The Chal- lenger expedition found no seaweed on the way to coal, though, several times it crossed the area, where, if ever, such deposits might be expected. Not only the petrography of coal but also the palzon- tology opposes the hypothesis. Seaweeds have not been discovered and the forms known in earlier days as fucoids have proved to be land plants. : Lesquereux** referred to Mohr’s hypothesis only to reject it. OH. Mietzsch, “ Geologie der Kohlenlager,” Leipzig, 1875, p. 244. “TL. Lesquereux, Ann. Rep. 2d Geol. Survey of Penn. for 1885, p. 104. Same for 1886, p. 465. 41 42 STEVENSON—FORMATION OF COAL BEDS. [April 21. Seaweeds have cellular structure alone. They decompose quickly whether exposed to atmospheric oxygen or protected from it. They are soon transformed into a fluid, black material which penetrates the sands along the seashore. He thinks it possible that remains of marine alge may have been thrown casually on swamps and that ~ their decomposition products, added to those of the decomposing materials, may have enriched them and may have given cannel. J. Geikie*® sees in the alternation of coal and limestone, evidence of prevailing subsidence, while the coal seams indicate frequent re- currence of land surfaces. The cannels and iron-stones show that many wide lakes and lagoons existed. He finds lines and ribs of cannel associated with splint and even ordinary coals, while the can- nel itself passes into common coal or black shale or even into black- band ironstone. The varying conditions are due to the mode of accu- mulation. Cannel was formed under water, for it contains fresh or brackish water fossils. The expanse of fresh water was surrounded by wooded flats; slimy vegetable mud, with, in places, ferruginous matter, was deposited where the streams entered. Along the shores were marsh plants, while farther back were trees and fern under- growth. The last gave ordinary coal, the marshy plants were con- verted into splint, while the slime became cannel, oil-shale or even iron-stone. Stevenson,*® as the result of studies in the Upper Coal Measures (Monongahela) of Ohio and West Virginia, came to the conclusion that at the close of the Lower Barren Measures (Conemaugh) the northern part of the Appalachian area basin was a half-filled trough separated from the western coal areas by the Cincinnati fold. He accepted the in situ doctrine without reserve. The conditions ob- served in the Upper Coal Measures prove a succession of gradual subsidences interrupted by intervals of repose, during each of which a lid of coal was formed over all or part of the basin. The sub- sidence could not have been paroxysmal, for, as the shore line sank, ** J. Geikie, “On the Geological Features of the Coal and Iron-stone- bearing Strata of the West of Scotland,” Journ. Iron and Steel Inst., Vol. III., 1872, pp. 13, 14. ; “J. J. Stevenson, “ The Upper Coal Measures West of the Alleghany Mountains,” Ann. Lyc. Nat. Hist., N. Y., Vol. X., 1873, pp. 226-252. 42 ae STEVENSON—FORMATION OF COAL BEDS. 43 the great marsh, which became the Pittsburg coal bed, crept up the shore—and this perhaps to the very close of the epoch. Thus it is, that though giving origin to many subordinate seams, the great bed diminishes westward. The Pittsburg coal bed began at the east _ and advanced westwardly. There is evidence in the distribution of sandstones and shales that a delta formation at the east pushed out into the basin, so that conditions favorable to coal-making existed first on the east side of the great basin. His summary is: (1) The great bituminous trough west from the Alleghanies does not owe its basin-shape primarily to the Appalachian revolution. (2) The coal measures of this basin were not united to those of Indiana and Illinois at any time posterior to the Lower Coal Meas- ures (Allegheny) and probably were always distinct. (3) The __. Upper Coal Measures (Monongahela) extended as far west as the _ Muskingum river in Ohio. (4) Throughout the Upper,Coal Meas- __ ures epoch, the general condition was that of subsidence, interrupted . by longer or shorter intervals of repose. During subsidence, the Pittsburgh marsh crept up the shore, and in each of the longer intervals of repose it pushed out upon the advancing land, thus giving rise to the successive beds of the Upper Coal csceamcasa bine (5) The Pittsburgh marsh had its origin at the east. Two years later,®° after further studies in West Virginia, he offered additional arguments in favor of his suggestions and ex- tended the scope of his hypothesis. The Appalachian basin at the beginning of the Upper Coal Measures was closely landlocked, com- ‘municating with the ocean at the southwest by a comparatively nar- row outlet. At the east and southeast, rivers brought in their loads of detritus to be spread over the bottom of the basin; on the opposite side, few and sluggish streams flowed from the low Cincinnati fold. During periods of repose, deltas were formed and the marsh ad- _ vanced on the newly formed land. If the period of repose were long enough to permit the filling of the bay, the marsh would extend across if begun on one side, or to the middle if passing out from all sides. The basin in West Virginia was never so filled with detritus as to permit coal beds to cross it. The Appalachian basin was never *”*“On the Alleged Parallelism of Coal Beds,” Proc. Amer. Phil. Soc., -Vol. XIV., 1875, pp. 283-205. 43 Ad STEVENSON—FORMATION OF COAL BEDS. [April 21. united with those at the west, anywhere north from Kentucky and he leaves to others to decide if there was at any time a connection farther south. Still later,®* after very detailed studies in southwest Pennsylvania, he discussed the question Are coal beds continuous? He describes the Pittsburgh, Waynesburgh, Waynesburgh A and Washington coal beds as practically continuous in the northern portion of the Appa- lachian basin within Ohio, Pennsylvania and West Virginia—that is to say, that they are almost invariably present wherever their horizon is reached. But that is not true of the intermediate beds, which fre- quently are wanting in considerable areas; yet they are constant in many great spaces of from 100 to 1,000 square miles: he cannot resist the conviction that these beds are not in isolated patches but that for the most part these apparently separate areas are merely parts of a connected whole. The barren spaces mark localities which did not present conditions favorable to accumulation of coal. Res- pecting coal beds older than those of the Upper Coal Measures, he is convinced by the evidence of borings, that all, with possible excep- tion of two, merely fringed the border of the basin. Andrews,*? in rendering the final report upon his work in south- eastern Ohio, presented the conclusions respecting formation of coal to which he had been led by his many years study of the Ohio measures. The Lower Carboniferous detrital rocks were deposited in shallow water; the sandstones show ripple marks, strie and branches of marine plants (the indefinite Spirophyton). Some conglomerate appears in the early part of the Coal Measures, but it is confined to the shore side of the basin and disappears eastwardly [toward the center of the basin]. Rocks exhibit rapid variations laterally ; sand- stones pass into shales ; limestones into shales and sandstones. Some marine limestones, formed in shallow water, indicate, as do the coal beds, pauses in the almost continuous subsidence ; but the great lime- stones of the Upper Measures [Monongahela] are to be considered merely as calcareous muds, for they vary as do the other mud rocks. *t 2d Geol. Surv. Penn. Rep. KKK, Harrisburg, 1878, pp. 283-295, 301-303. FE, B. Andrews, Geol. Survey of Ohio, Vol. I., Part I., Columbus, 1873, PP. 345, 347-351, 354, 357, 358. 44 I aE eet 1911.) STEVENSON—FORMATION OF COAL BEDS. 45 They were deposited in shallow water, for they are close to coal beds and show the shrinkage cracks due to drying. Andrews adheres to the doctrine of accumulation in situ, assert- ing that his studies leave no room for any other conclusion. The vegetation grew on marshy plains skirting the ocean or perhaps making low islands near the shore. Slates as coal partings are of great geographical extent, holding the same stratigraphical position throughout, thus implying a temporary overflow of the marsh by ocean waters, with an even distribution of the sediment. Some beds contain evidence of tidal flows, for beachworn sticks, replaced by pyrite, lie in the coal as they were drifted upon the marsh. After complete submergence of the bog, trees growing on the surface were overthrown by turbulent waters; thousands of trunks of Pecopteris arborescens are seen in the roof of the Pomeroy coal bed, bent or broken down by the sediment-carrying water; and with them are great trunks of Sigillaria and Lepidodendron; while, in sandstone, drifted and buried trees from upland areas are not rare. The con- tinuity’of coal seams was often interrupted, as should be expected in great areas. Andrews’s studies were confined chiefly to southeastern Ohio and the adjacent portions of West Virginia, where the coal area ap- proaches the central part of the basin, the original western border having been many miles beyond the present western limit of the Coal Measures. The irregularities of deposit are comparatively insigni- ficant and the important members show a remarkable parallelism. He was led by the phenomena of his region to deny the possibility of notable variations in thickness of intervals between coal beds and he refused to accept as correct the great variations reported from the anthracite areas. There are many evidences of erosion and planation during deposi- tion of sandstones. The great bed on Sunday creek shows erosion from one foot to entire thickness of the bed, the overlying sandstone filling the trench and resting unconformably on the eroded edges of the coal. The eroded surface is smooth, there being no traces of rough work such as one should expect to find, if the coal were still soft and unconsolidated at the time of removal. 45 46 STEVENSON—FORMATION OF COAL BEDS. [April 21. Andrews thought that cannel was originally vegetable mud. He emphasizes the abundance of Stigmaria, saying that they fairly reveled in this ooze. They, with their rootlets, abound throughout ; their existence in these beds for hundreds of miles almost necessitates the conclusion that they are in situ. If they are roots of Sigillaria, those trees must have grown in the wettest portions of the marsh, which, in that case, could not have been lagoons. The Stigmaria are evenly distributed. *If they had been drifted in, he thinks they ought to have gone to muck with the rest. Newberry,** in the following year, discussed the origin of the various deposits composing the Coal Measures. The coarse rock underlying the series contains rounded pebbles of quartz, igneous and metamorphic rocks, with rounded and angular sand of the same material as well as cherty pebbles from the Lower Carboniferous. The pebbles for the most part must have come from Archean areas at the east and north; but he finds difficulty in explaining how ma- terial from those areas could be distributed in sheets at hundreds of ~ miles from the only possible sources of supply. It is difficult to con- ceive of rivers as the transporting agency and he is inclined to find the explanation in the drift deposits of the Mississippi valley, ice being the transporting agent. Where the rock is coarse, fragments of the tree trunks, of Calamites and of roots are present, all broken, and sometimes heaped in masses covering several rods. Fruits, like Trigonocarpum, occur in hollow calamites and the mass is Ri driftwood, everything broken and battered. The fireclays sometimes contain stumps of Sigillaria and Lepido- dendron in unbroken connection with Stigmaria roots. Coal is seldom wanting above fireclay, though at times it has been removed by erosion. Coal beds were formed in situ. Fine sediment accumu- lated in pools and these were invaded by vegetable growth, to be filled up finally by bitumenized remains of generations of plants. Aquatic plants remove alkalies, phosphorus, sulphur and silica from the soils, as is seen in peat bogs, where the underclays are often fireclays. The varying deposits are explained by alternate eleva- tions and depressions of the surface. Limestones were formed in arms of the sea and their presence is proof of unequal subsidence. J. S. Newberry, Geol. Survey of Ohio, Vol. II., part I., Columbus, 1874, pp. 104-115, 118. 46 19t1.] STEVENSON—FORMATION OF COAL BEDS. 47 Newberry opposed the doctrine that spores of cryptogamous plants are important constituents of coal. Sporangia and spores are common enough in American coals but they are an inconsiderable part of the whole. Dana,** reasoning from chemical analyses, objected to Dawson’s suggestion that coal was derived largely from bark or material of that nature. Though nearer coal in composition than is true wood, bark resists alteration longer and is less easily converted into coal. The occurrence of stumps and stems outside of the coal beds, “ while __ proof that the interior wood of the plants was loose in texture and a very easily decayed, is no evidence that those trees contributed only _ their cortical portion to the beds of vegetable debris. Moreover, 3 the cortical part of Lepidodendrids (under which group the Sigil- larids are included by the best authorities) and of Ferns also, is made of the bases of the fallen leaves, and is not like ordinary bark _ in constitution ; and Equiset@ have nothing that even looks like bark. This cortical part was the firmest part of the wood; and for this reason it could continue to stand after the interior had decayed away —an event hardly possible in the case of a bark-covered conifer, how- ever decomposable the wood might be. Further, trunks of conifers are often found in the later geological formations, changed through- out the interior completely to Brown coal or lignite.” He appears to be convinced that the whole plant material contributed to formation _ of the coal, which he regards as the product of marsh accumulation. Dawson* returned to the discussion in view of Huxley’s asser- tion that spores are an important constituent of the coal-forming mass. Referring to his study of more than eighty coal beds in Nova Scotia and Cape Breton, he asserts that the trunks of Sigillaria and _ similar trees constitute the great part of the densest portion of the coal and that cortical tissues, rather than wood, predominate. Spores and spore cases, though often present abundantly, constitute only an infinitesimal part of the great coal beds. Sporangites or __ bodies resembling them are present in most coals, but they are acci- “J. D. Dana, “ Manual of Geology,” 2d ed., New York, 1874, pp. 361, 362, 366. . *J. W. Dawson, Amer. Journ. Sci., 1874. Supplement to 2d ed. of “ Acadian Geology,” 1878, pp. 65. 47 48 STEVENSON—FORMATION OF COAL BEDS. [April 21. dental rather than essential constituents, more likely to be found in cannel and shales, deposited in ponds near lycopod forests, than in the swampy or peaty deposits, whence the coal beds proceed. While giving credit to Huxley and his predecessors for calling atten- tion to the importance of spores in coal, he is compelled to maintain that they have generalized on insufficient basis, that sporangitic beds are exceptional among coals and that cortical and woody matters are most abundant. The purest layers of coal are composed of flattened trunks; other coals are made up of finely comminuted par- ticles, mostly epidermal tissues—not only from fruits and spore cases but also from leaves and stems. Mietzsch** attempted to answer the question, how did the vege- table material accumulate in great beds? Was it brought down by rivers from forest covered areas or did the plants grow where the coal is now found? The mode of occurrence can be explained measurably by either supposition ; at times one process may act alone, at time it may be permissible to regard both as contributing. He describes the heaping up of driftwood along streams as well as on coasts, whither it has been carried by currents; and he thinks that in this way may have originated some tertiary deposits of lignite, composed almost wholly of stems stripped of their bark. But many deposits of lignite and brown coal contain stems with bark, twigs, leaves and fruit preserved. The Suterbrander lignite of Iceland was formerly supposed to be driftwood, because of the present conditions in that land; but Heer discovered well-preserved buds, leaves and twigs of the plants, represented by the stems, which still retain their bark. The same criterion must be applied to the black coals. Many deposits of these and the greater number of brown coals have numerous tokens, rendering improbable, in part impossible, the supposition that they were made of transported plant masses. It is difficult to understand the regularity and vast extent of coal beds on the theory of transport, for driftwood accumulations are irregular and of small superficial extent. The composition of coal tells against the theory of transport, for in most beds the ash is very small—surprisingly small, for in the process of coalification no part of the mineral content of the plants °H. Mietzsch, “ Geologie der Kohlenlager,”’ Leipzig, 1875, pp. 244-257. 48 — x9tt.] STEVENSON—FORMATION OF COAL BEDS. 49 _ disappears, aside from soluble alkaline compounds. Some have found proof of transport in the composition of ash from stone coal, since it is quite similar to clay‘shale. But Mietzsch points out that living Lycopodiacee contain from 22 to 26 per cent. of clayey earth in the ash and asks why one should suppose that the older types were different. But if the coal contain an abnormal proportion of ash, there is reason to recognize influx of fine mud. . x The fineness of the materials, clay and sand, in contact with the _ coal, proves a long period of quiet; and the same may be said of _ the plant deposits themselves. Such a period can hardly be ac- a cepted for rivers or for currents along coasts: The conditions of _ the underclay ; the resemblance of the clay in many cases, as Steffens _ showed, to vegetable mould; the interlacing of S: tigmaria roots like a wicker work; and the occurrence of erect trunks are all opposed 4 to the doctrine of transport. In most cases the conditions can be explained only by the doctrine that coal beds owe their origin to _ plants which grew where their remains are now found. He accepts _ the peat bog theory as advanced by v. Beroldingen and presents _ many facts as additional evidence in its support. The advance of _ bogs into lakes is proved by the discovery of pile constructions in 3 Swiss peat bogs; along the seashore, algae form dense floating felts on which bog plants grow and the mass sinks to the bottom. Zee- - land was once cut by bays much longer than now and part of the " former sea-area is filled with peat. He strengthens his argument by _ many references to phenomena observed in the great swamps of _ Europe and North America. In order to explain the origin of coal-bearing strata, holding a number of coal beds, one must distinguish between those formed along a coast and those formed along rivers or in the interior of an land or continent. Those of the first type are explained by the subsidence of coasts bordering on the North Sea. The preliminary work for drainage of the Zuyder Zee, as well as similar work else- _ where, has proved the existence of peat bogs in. extended areas of shallow sea ; anchor flukes have brought up peat from depths of 200 _ meters on the English coast. Such bogs become covered by river sediments and in case of long-continued slow sinking, the shallow sea area is filled, so that a number of bogs may be formed suc- PROC. AMER. PHIL. SOC., L, I98D, PRINTED APRIL 24, IQII. 49 + 50 STEVENSON—FORMATION OF COAL BEDS. [April 21. cessively. Among other illustrations, he refers to the discovery at — Rotterdam of two bogs, 5 and 6 meters thick, separated by 4 meters of clay; to the presence of erect trees which, despite the long period which has passed since they sank below the water surface, are still standing on the sea bottom, partly surrounded by sediments; such trees on the coast of the islands of Sylt and Romo are of types which : disappeared from that region many hundreds of years ago. Changes in grade of rivers, caused by damming or by crustal movements, would lead to covering of bogs with sand or mud and to the accumulation of rock masses. He finds confirmation of this view in Livingstone’s statements respecting the floods of African rivers and in the observations of others elsewhere. : Lesley®’ in prefaces to reports by geologists of the Pennsylvania survey, made frequent references to hypotheses respecting formation of coal beds. Ordinarily, he preferred to present the matter, as it were judicially, giving the difficulties in the way of accepting the hypotheses and leaving the decision to the reader. But in two of the prefaces he offers some important suggestions. W. G. Platt described a little basin, barely a mile and a hale across, in which three sections of Coal bed D were obtained. In all of them, the bottom bench is 2 feet 7 inches thick and composed of brilliant coal; but the upper part is a dull cannel or cannel shale, measuring I foot 3 inches, 8 feet 3 inches and 1 foot 2 inches, while between the last two the dip is about 8 degrees compared with about one degree elsewhere. A noteworthy feature is that while the ash in the cannel is from 21 to 25 per cent. and that in the pure coal below is only 1.6 per cent., yet the ratio of volatile matter to fixed carbon is practically the same throughout. Lesley felt convinced that the petty basins, in which canvlal was deposited, were waterways or pools and that more of them existed at once in certain horizons than in others. They were not due to erosion for the underlying coal bed is not cut out, it is merely de- pressed. There is no evidence of currents, for the mud is fine, the lamination perfect and the roof soft. The pools were almost stag- nant. How could a depression come about to give, as here, a dip * J. P. Lesley, Second Geol. Survey of Pennsylvania, Indiana County, 1878, pp. xiv-xviii; Lawrence County, pp. xix—xx. 50 1911.] STEVENSON—FORMATION OF COAL BEDS. 51 of 5 to 10 degrees to an almost dead level bituminous coal bed? There is no room for suggestion of crustal movement as the area is too small; equally the cavern theory is excluded for no limestone underlies the horizon except at vast depth. He can see no explana- tion for most of the localities except in the subsidence of a floating bog, such as Lesquereux has described. On this the fine muds ac- cumulated and the pool was filled. _ He was led in this connection to consider the sequence of coal beds. If the Carboniferous plain consisted of a low area with shallow ponds, the coal forming vegetation would conform to the dimpled surface and there would be but one coal bed, intersected by river channels. This plain, if continuous, would be not less than 1,000 miles long by 300 miles wide [this refers to the Appalachian basin]. It is very difficult to account for the submergence of this continental plain to a depth of 50 feet below sealevel in order to give opportunity for formation of a second bed. Yet this “slow de- ‘pression theory” may not be rejected easily, for without it, one cannot conceive how 20,000 to 40,000 ‘feet of palaeozoic sediments could have been deposited; the more so, since many of the strata give every evidence of deposition in very shallow water. As a partial alternative, he suggests that the relative sea level may have been changed by the filling of basins. The effect of deposits by great rivers and that of glaciation are discussed but no conclusion is reached. In the preface to the Lawrence report, he attempts to explain the origin of underclays. A peat bog and even a lake invaded by sphagnous growth must-have some water circulation due to percola- tion from the surrounding land and to evaporation from its own sur- face—but the movement would be very feeble and it could transfer only the finest mud, though in course of time the result would be important. Dry grounds are largely fine gravel with rounded quartz and feldspar grains; the feldspar is soluble, it follows the indraught and settles beneath the evaporating surface with its floating peat. If the peat area be surrounded by clayey land, the percolation would be at a minimum; the water supply would be from the surface and less muddy, so that the underclay would be less in quantity. It ‘would appear, then, that when the margin was a tight clay, deposits 51 52 STEVENSON—FORMATION OF COAL BEDS. [April 21. of calcareous type show that limestone must have been exposed within the drainage area. The thickest underclays should belong to beds next or near above the great sandrocks and it is a fact that our great clay beds are near the base of the Lower Productive Coal Measures | Alle- gheny] and that the few important clay deposits high in the series have coarse grained sandrocks not far below them. A logical con- sequence of such conditions is that sandrocks geologically close to such great underclays should be purer, more open sands and gravels than others which had not been robbed of so large quantity of interstitial clay. If the surrounding land contained iron in its gravel, there should be ball ore in the fireclay—as is seen in the New England ponds surrounded by drift. Davis®* described a cannel deposit in Yorkshire, somewhat resem- bling that discussed by Lesley. The bed is thickest in the center and thins away in each direction, meantime becoming less pure and passing into bituminous shale at the circumference. The condition is due to in-floating of plant remains, which sank to the bottom of the pond. The marked interlamination of shales and their marked increase toward the border resulted from more rapid subsidence of the muds. In some places the pond was filled up; there the under- clay has abundance of Stigmaria and the plants growing in such places were converted into ordinary coal. Afterwards the whole mass was submerged and covered with black mud. The cannel is fine, close-grained, homogeneous, with conchoidal fracture, without planes of deposition and everywhere yields beautiful specimens of fishes. Reinsch®® undertook the microscopic study of coal. He pre- pared a great number of sections, subjected them to close examina- tion and published his results in an elaborate volume with 95 plates. These exhibited the structure of the coal as well as numerous forms which seemed to be organized. Reinsch maintained that the coal J. W. Davis, “On the Fish Remains found in the Cannel Coal of the Middle Coal Measures of the West Riding of Yorkshire,” Q. J. G. S., Vol. XXXVI., 1880, p. 56. °*P. F. Reinsch, “ Neue Untersuchungen uber die Mikrostructur der Steinkohle des Carbon, der Dyas und Trias,” Leipzig, 1881. 52 Ler -, ‘2S =a “4 44 ah ees ° os Ea npn > Ls “ ‘ Shi ee IE ee te es oe a) eer ee) NOt eee eo Te GeO eat 1911.] STEVENSON—FORMATION OF COAL BEDS. 53 substance originated, mainly, from marine plants of such peculiar form that they cannot be assigned to any group of known types. He created a new group for their reception, Protophyte, of which he made seven divisions. Remains of land plants are of very rare occurrence. This hypothesis differs from that of Mohr in that the plants are microscopic. Petzholdt® at once made a fierce critique of Reinsch himself, his methods and his results. Of the seven divisions of Protophyte two are decomposition products, three are certainly inorganic, one consists of fragments of land plants and one is based on minute fragments of coal. The decomposition products, mistaken for organic bodies, are termed bitumen by Petzholdt, who thinks them the same with those discovered fifty years before by Hutton in his study of the Newcastle coals. ‘ Fischer and Rust," following Reinsch’s method, found not only yellow and reddish resin-like bodies in black coal, such as make up the great part of the Scotch boghead, but also small grains, showing wood structure, in anthracite. In the black coal, they observed spindle-shaped or serpent-shaped bodies, whose relations they could not determine. The English cannel from Lancashire is very rich in little resinous cylinders and, as far as richness in resinous matter is concerned, is intermediate between the Bogheads and the ordinary coals. These studies have an important bearing on investigations which have attracted much attention in more recent years. Green™ says that it is not easy to see how light material, such as dead wood, could be spread out evenly over tracts of hundreds of square miles, so evenly that the deposit shows comparatively little variation in thickness; and it is equally difficult to understand how, in case the coal be composed of drifted materials, it could be so pure as we often find it. The water bringing the vegetable matter @ A. Petzholdt, “Beitrag zur Kenntniss der Steinkohlenbildung,” Leip- . zig, 1882, pp. 23 et seq. * H. Fischer and D. Rust, “ Ueber d. mikroskopische Verhalten verschie- dener Kohlenwasserstoffe, Harze und Kohlen,” Groth Zettschrift f. Kryst., Vol. VIL, pp. 209-243. This has not been seen by the writer. Cited by Petzholdt and v. Giimbel. “A. H. Green, “Geology,” Part I., Physical Geology, London, 1882, pp. 257-262. 53 54 STEVENSON—FORMATION OF COAL BEDS. [April 21. _ would certainly carry also mineral matter, The coal and its ash may, both of them, be of vegetable origin. Logan’s discovery of the underclay or Seatstone under nearly every coal bed was the first great step in the right direction toward solving the problem. Bin- ney’s study of an erect stump discovered by Hawkshaw near Man- chester was the next, for there a Sigillaria with Stigmaria roots was rooted in a seat clay, while the stem was surrounded by rock. Many similar cases were discovered. The underclay was the old soil sup- porting plants which produced a layer of nearly pure vegetable matter. When the surface was lowered beneath the water, sand and clay were laid on top and the band of dead plants was converted by pressure and chemical change into a seam of coal. When sinking ceased, the shallow water was filled up and a swampy plain was made. Vegetation spread out from the land and a second coal bed began to accumulate. This process repeated many times over gave a succession of sandstone and shale with coal beds at intervals. The great swampy expanses in the delta of the Ganges and Brahmapootra must bear close resemblance to the marshy flats in which the coal was formed. The nearest approach, however, is in the accumulations on the coast of Patagonia, described by Lady Brassey in “ A Voyage in the Sunbeam”; “To penetrate far inland was not easy owing to the denseness of the vegetation. Large trees had fallen and, rotting where they lay, had become the birthplace of thousands of other trees, shrubs, plants, mosses and lichens. In fact in some places, we might almost be said to be walking on tops of the trees, and first one and then another of the party found his feet slipping through into unknown depths.” There are, however, deposits of subaqueous coal, derived from driftwood carried down and buried amid mechanical deposits, but they are irregular and are apt to be impure. It is probable that the patches of cannel coal mark sites of pools or lakes in which vege- table matter lay “until it was macerated into a pulp. This passes gradually by increase of earthy admixture into well-stratified carbo- naceous shale. Green had already presented the same suggestions, though briefly, in his work on the Yorkshire coal-field published in 1878. 54 t911.] STEVENSON—FORMATION OF COAL BEDS. 55 Grand’ Eury,® in the first section of his notable memoir, gives the grounds on which his theory of transport is based. When one makes minute examination of coal, he discovers that the plants have been broken up and the parts scattered; fruits and leaves are apart from the stems; the layers of the bark are sepa- rated and dispersed; the interior parts of the stems have disap- peared and the flattened cortex alone remains. . The woody portions of the stems have been dispersed as fusain [mineral charcoal]. Stems are split and torn, Cordaites leaves are imperfect, everything, bark or leaf, is broken up. Hethinks that a great part of the tissues was transformed into a kind of vegetable pulp, which makes up most of certain coal beds. That this was not wholly fluid or homogeneous is evident, for one may distinguish some traces of organization with the microscope or even with a magnifying glass. The disintegration of the plant organs occurred after death and its character puts aside all suggestion of violent action. All the evidence contradicts the supposition that the forests were ravaged. by inundations ; everything points to quiet, peaceable flow of water. Most of the material was decomposed in place and carried away piecemeal. The vegetable matter was not deposited in deltas within either the north or the center of France. The preservation of stems reduced to their bark is not surprising, for there was little wood in trees of the Carboniferous; but the min- eral charcoal is not so easily accounted for. Itseems to be fossilized buried wood, dried in the air and not changed into coal. It did not originate through maceration, though after formation it may have been subjected to moisture, as is indicated by lack of sharpness in outline. The vegetable disaggregation was rapid, mostly in air, and was completed in swamps before removal. The conversion into detritus. and the quasi-dissolution were sometimes pushed very far at the base of damp forests and at the bottom of swamps. The Car- boniferous forests were marshy and aquatic. The plants grew quickly, reached maturity and soon died. Growth had to be ener- getic in order to carbonize the bark so as to make the contraction ; “G. Grand’ Eury, “ Memoire sur la formation de la houille,” Ann. des Mines., Ser. 8, T. i., Paris, 1882, pp. 1o1—122. 55 56 STEVENSON—FORMATION OF COAL BEDS. [April 21. small in coalification. That the air was damp and warm is proved by the aerial roots of Psaronius and Calamodendron; and the heat of the climate appears from the dense resinous bark, which often dominated the wood. Strong light, great heat, excessive humidity, great marshes in which plants grew quickly and died, explain condi- tions not easily explained by conditions of the present time. The residues falling into the marshy bottom of the forest, underwent aqueous rotting; they were then transported to the areas of deposit, which preserved them from complete destruction. Grand’ Eury published much relating to this subject and in 1900% he summarized all the results of his long studies in a memoir presented to the geological congress. He describes fossil forests in situ, which show that the Car- boniferous plants, though arborescent, were forms of marsh-habit like those of the Dismal Swamp, the foot and adventive roots in the water, but the stocks and rhizomas creeping on the bottom. The ‘ forests were very local. Growing in stagnant water and fixed by few roots to the ground, they were destroyed by slight causes and the roots alone remained. This would give a “soil of vegetation” as described by Dawson—a feature as familiar at Saint-Etienne as in Canada. Coal is stratified, evidently deposited under water. There is no evidence that roots ever traversed the parallel laminz of which it is composed. The stocks and roots, descending in the roof, spread out on the coal but never penetrated it. This condition is constant and is due to the circumstance that slowly deposited vegetable mat- ter, undergoing fermentation, is opposed to the introduction of roots, which, being unable to live in it, instinctively refuse to pierce it. Similarly there is no relation between stocks and overlying coal. Their roots are often enclosed in coarse twisted coal composed of overturned stems, with leaves, branches, which, however, is con- tinuous with overlying laminated coal. The elements are the same in both and they are identical with those in the adjacent shales, so that transportation from a distance is impossible. There is then in some coal beds evidence of formation in place or almost in place. *C. Grand’ Eury, “Du bassin de la Loire,’ Compt. Rendus VIII™é Congres Geol. Intern., Paris, 1901, pp. 521-538. 56 is dug Smtr Ue abe an ahaa ei “d .- 4 a ‘Z > r: a 3 : 1911.] STEVENSON—FORMATION OF COAL BEDS. 57 But most of the material forming the beds was transported; yet all coals resemble that found almost in place and the parts, cer- tainly transported, are identical with similar parts of the rooted stems. The materials were derived from marshy forests on borders of the basin, which doubtless succeeded those temporarily installed in the basin of the deposit which afterwards became a lake. At the foot of this forest was elaborated, as in peat bogs, the humus or fundamental material of the coal. The basin of deposit was much like the bottom of a morass, for the mud of coal beds often resem- bles the clay underlying peat bogs. The debris of plants falling into water on the borders of the marsh became stratified in its depths. Grand’ Eury was convinced that by this hypothesis he had reconciled the opposing theories, for he has shown that certain coal beds were formed by concurrence of both processes, as in the sub-" aquatic parts of some swamps. The permanent swamps, where primitive peat was elaborated, were not exposed to deposit of mineral sediments, they remained uncovered and disappeared; so that very little of the coal formed in place remains. The researches of Renault and C. E. Bertrand on cannel and the fundamental matter of coal show that coal was not always deposited on lake bottoms under moving waters, but’that it may have been formed in stagnant or quiet waters of swamps. The coal was deposited slowly, not continuously and there may have been long periods of arrested growth. The concentration of fossil forests and soils of vegetation in and near coal beds proves for the thick beds a very long period. Additional evidence in this direction is found in the advanced decomposition of the rocks form- ing the roof, their new chemical combinations, their impregnation with carbon, showing that they had been long in contact with the swamp before being transported and deposited on the coal bed. The basin of the Loire was subjected to orogenic movements. The fossil forests have irregular distribution both vertically and horizontally; great sterile deposits break up the continuity. The basin was deepening throughout the period of formation, but each important coal bed corresponds to an interval of stability. That the mineral materials were brought in by streams is shown by their 57 58 STEVENSON—FORMATION OF COAL BEDS. [April 21. distribution. The granitic rocks of the northern portion thin toward the south and their rooted stems lean toward the south and south- east; but the micaceous rocks of the southern portion thin toward the north and the rooted stems lean in the same direction, sometimes strongly. These mineral deposits interlock as wedges. But the coal beds pass from one type of rock to the other, preserving well their distance and parallelism. Grand’ Eury finds no evidence to support the delta-theory of accumulation in deep basins; every feature leads | to the belief that the mass of rocks could accumulate only by means of a subsidence, equal and progressive from the clay bottom. In a still later paper,®* Grand’ Eury shows that coals of all kinds are practically alike in origin. Coal beds are deposits of allochthonous peats formed by an «exuberant vegetation, loving water, whose detritus was carried from shores to interior of immense marshy lagoons, where barks, cuticles and the rest were stratified with ulmic substances under the water. Stipites or dry coals of the Secondary in France are clearly the same in origin with the coals. Mineral charcoal is so abundant in one of the Upper Cretaceous coals as to give a finely-stratified structure to the bed. The brown coals of the Tertiary resemble coal completely in mode of occurrence; they are composed of marsh plants, leaves of dry land plants being in small proportion. Lignite is wood-like in appearance though formed of red humus from plants; they show much variation, but the mass of the material is derived from marsh forms. The peats of lowland areas or marshy plains are allochthonous—they resemble almost all deposits of mineral coal. Gruner® notes the ancient forest in the quarry of Treuil, which had been described by Alex. Brongniart many years before. At 100 meters lower and almost directly under the quarry, Gruner found in the Treuil mine twelve great trunks in a space of less than 10 meters square; their roots spread out over the coal but did not pene- trate into it. He cannot accept the doctrine that coal consists of transported material. The continuity and uniformity of coal beds make a serious’ *C. Grand’ Eury, “Sur la formation des couches de houille de stipite, de brownkohle et de lignite,” Autun, 1902, pp. 123-132. *T. Gruner, “ Bassin houiller de la-Loire,” Paris, 1882, pp. 160-170. 58 27s a “ n Se ae St ee sty Rr ee ee ee ey pee ye eee 1911.) STEVENSON—FORMATION OF COAL BEDS. 59 objection. Inthe little basins of Saint-Etienne, beds can be followed 5 to 10 kilometers in one direction and 2 to 4 in another with little change. He thinks that a current capable of uprooting trees would tear away the soil and pebbles also, so as to give a mingling of trees | and detrital matter. METS. ee eT i PoUn awe, een ee As large streams carry much mineral material there should be an alternation of vegetable elements and mud—and this is found in coal beds where shale appears in thin layers between benches of coal. These shales or the nerfs of fine sandstone could be pro- duced only by water-currents, by inundations of brief duration cov- ering the debris on the surface or invading shallow basins in which leaves, etc., were deposited slowly. The two modes of accumulation went on simultaneously in the coal period as they do now in peat bogs. He does not assert that coal was the peat of palaeozoic times ; ’ the flora and the climate were different; but the mode of formation was the same. The plants of the coal epoch grew where their remains are found. He cannot accept Grand’ Eury’s theory, which opposes the doctrine of in situ accumulation because stumps and trees are wanting in the coal beds themselves. Grand’ Eury main- tains that the vegetable matter was transferred from the place of growth to the basin where the coal is found, but the distance was small. Gruner maintains that the current would have brought more than leaves and stems and that it would have distributed its load unequally ; he thinks it preferable to conceive of a marshland extensive enough to admit of a thick cover of vegetable débris over an area of several thousands of square kilometers—as one finds in the Nord basin. Grand’ Eury emphasizes the absence of stumps and roots passing from coal beds to the mur. But at Saint-Etienne itself, Lyell and Gruner saw rootlets passing from the coal into the underclay and Gruner saw the same condition in the Batardes coal bed, where Stigmaria abounds in the mur. The absence of stumps in the coal is to be expected, because the soft tissues would be crushed quickly under pressure and all traces would be effaced; moreover, in the nature of the case, stumps would be only a small portion of the mass. A negative result of study does not prove that the plants 59 60 STEVENSON—FORMATION OF COAL BEDS. [April 21. did not grow sur place. Since the rapid current, which piled sand around the forests of Treuil, did not uproot the trees, one finds difficulty in understanding how the waters so slightly agitated as to be able to draw off only leaves and twigs did not leave in place the stumps whose roots are seen to-day in the underclays. The preservation of the underclay proves that the stumps were not torn out before deposit of the plant debris forming the coal, bed. The clay shows no signs of erosive action such as are seen so often in the roof. The deposit of the clay is itself a proof that then had begun the long period of tranquillity, which continued during forma- tion of the coal. He is convinced that it must be admitted as almost proved that the coal beds have come from a vigorous local vegeta- tion, whose debris accumulated at the bottom of shallow stagnant water and probably, quite as often, on a damp but not flooded _ surface. The intervening rocks are, in character, wholly similar to part- ings in the coal beds, but they were formed not by petty inundations but by strong currents of prolonged duration. The existence of these is proved by erosions as well as by the sands which covered the coal forests. The surface subsided at intervals, as shown by phenomena connected with the faults in the Loire basin. But the flora was not destroyed, for one finds forests or isolated trees in place, in sandstones at all horizons, their bark preserved as coal. The sands are evidence that the agitated water prevented quiet depo- sition of vegetable debris. That was destroyed or scattered afar. Meanwhile, the sunken surface was leveled up and the depres- ~ | sion was filled. A second marsh was formed above the first, now buried under a thick bed of sand or mud. If the deposit of sand, etc., did not exceed 30 meters, the conditions under which the new bed was formed might not differ from those of the earlier bed. But — when the sterile interval attains great thickness, 100 to 800 meters, the period of depression was very long and before its close the flora had undergone modification. Thus it is that one finds successive appearance of varied types, so that classification of the Coal Meas- ures by their flora becomes possible. Subsidence of the type here conceived has been observed in rocks of all epochs. Lament and 60 a yell eee Lie eee ERENT ee ah ae ae ET A eh ee ee _ eu - Perio ee ee, ees t91t.] STEVENSON—FORMATION OF COAL BEDS. 61 Degousie, in sinking artesian wells at Venice, found beds of lignite and carbonaceous clays at 40, 60, 100 and 120 meters from the surface. Von Giimbel,®* perplexed by the contradictory results presented in memoirs, undertook a series of systematic studies, covering all phases of the subject. His study did not concern itself with chem- ical or technical matters and had little reference to botanical rela- tions. At the outset, it deals only with questions relating to the constitution of coals; it begins with examination of peat-like sub- stances and advances, step by step, to anthracite and graphite; it ends with a discussion of the mode in which coal beds accumulated. In breadth of scope, this study excelled that by any predecessor ; in compactness and precision of statement the memoir has rarely been excelled. Much of the earlier portions bear directly on questions respecting the transformation of vegetable matter into coal, a sub- ject to be considered in a later part of this work; but some of his observations are so closely connected with the final part of his dis- cussion that they cannot be neglected. The method of investigation by means of thin sections did not commend itself to v. Giimbel, who preferred the method proposed by Franz Schultze. The broken coal was treated first with potas- ‘sium chlorate and strong nitric acid, and afterward with ammonia, in order to separate the particles and to make the transparent por- tions more readily available. Absolute alcohol completed the prepa- ration by removing coloring matters. He gives specific directions as to the use of the reagents and warns against the possibilities of error in the study. This investigation led him to recognize that the whole series from peat to anthracite is continuous and of similar origin. All of the members are made up of combustible materials. “Stone coal con- sists, apart from the earthy admixtures, of parts of plants, which, changed into a coaly substance, have taken up into their empty spaces, as well as into the intervals between the plant débris, a humin-like or ulmin-like substance (carbohumin) which was origi- = C. W. v. Gumbel, “ Beitrage zur Kenntniss der Texturverhaltnisse der Mineralkohlen,” Sitzungs. Berichten der k. bayer. Akad. d. Wissenschaften. Math.-Phys. Klasse, 1883, pp. 113 et seq. The citations are from pp. 190-212. 61 62 STEVENSON—FORMATION OF COAL BEDS. [April 21. nally soluble, but became insoluble, so that the whole is amorphous and apparently structureless.” The taking up of this material is the Inkohlungsprozess. Adjacent rocks, containing plant remains, may have contributed to this coalification by means of circulating waters. It is self- evident that this soluble material might be deposited by itself apart from any remains of plants, not merely as layers of a coal bed but also in cracks and fissures; but such layers of structureless coal could have contributed in only subordinate manner to the formation of coal beds. The several types of coal, Glanz-, Matt-, Faser-, Cannelkohle and the rest cannot have originated under similar conditions. In considering these he takes the most complicated condition—where several varieties occur in the same bed: Three modes of explana- tion are suggested by the investigations: (1) Original differences in kinds and parts of plants; (2) differing conditions, chemical and mechanical, in which the plants came to contribute toward making the coal; (3) heterogeneous external conditions under Wiha the transformation was completed. Difference in material in the several types of coal appeared con- stantly during the study; bark, and woody parts along with leaves in Glanzkohle ; abundance of leaf organs, especially of the epidermis layers and less abundance of hard parts in Mattkohle; constant recurrence of little balls, membranes, the spores of authors, in aston- ishing abundance with algae-like clumps in cannel-like layers; all proving a certain dependence of constitution on the character of the plant remains. It is clear that the condition under which the plant material was accumulated was of great importance. This is evident from the great amount of Faserkohle [fusain, mineral char- coal]. If this material result from decay in free air, as would occur in the occasional drying of the surface in peat bogs, one must con- cede that this process was of vast extent during the coal-making time. It is unnecessary to suppose that the great supply was swept in; it could have been produced as readily on the bog surface. Simi- larly the dismembered parts of plants, clods or flocks, and the rest belong to a stadium anterior to formation of the coal. The pres- 62 3 4 3 ; ie f ha 3 4 tort) STEVENSON—FORMATION OF COAL BEDS. 63 ence of plant remains in soil, in every peat bog, justifies us in tracing back in some degree, certain relations of coal formation to similar origin. Accumulation of cannel-like coaly substances can- not be explained otherwise. The tertiary gas coal of Falkenau, pyropissite and Lebertorf all consist of a similar wholly broken up mass of plant parts. External relations had much to do with the conditions. If ‘inflowing water bring much mineral matter into a bog, the borders are impure while the main portion is pure. Soa coal may be impure on the borders and pure in deeper portions of the basin. Even the character of the overlying rock may be important. Passing from the composition of the coals, he considers the mode of accumulation ; first of all, rejecting absolutely as without founda- tion, the doctrine that coal could have been formed in the open sea and from seaweeds. Coal beds consist of alternating, mostly very thin layers, like beds of sedimentary matter; this, with the fact that they are asso- ciated in series with undoubted sediments, seems to afford proof for the opinion that coal beds originate as do other sedimentary strata, in contradiction of the so-called peat theory, which accepts the idea of an origin in place after the manner of peat bogs. If one confine his attention solely to this layer-like accumulation and make no further inquiry, the conditions appear so completely ex- plained by the former doctrine that facts favoring the latter have no value. V. Gumbel thinks that the presence of upright stems is of comparatively little importance as a proof of autochthonous origin, since their presence is exceptional and it can be explained in several ways—by drifting, by advance of waters into swamp forests or by plant growths floating on the water. A careful examination of the query as to whether or not the lamination of coal can be explained by anything except deposit of suspended matter, leads to surprising results, when extended to the newer coal accumulations. The Quaternary brown coal offers an instructive illustration of the mode in which the lamination origi- nated. These have absolutely the same structure as that of stone coal beds. It is known positively that they owe their origin to peat- like swamps and that the clayey, sandy partings, which accompany 63 64 STEVENSON—FORMATION OF COAL BEDS. [April 21. them, proceeded from occasional overflows. Coming down a step farther to the coal making of our own time and ignoring for the present the various local modifications of peat, one can recognize two distinct modifications ; Autochthonous, that forming or originat- ing in place, and Allochthonous, the sedimentary, due to deposit of plant detritus in pent up waters. The latter shows, of course, evi- dence of sedimentary origin, is more or less dense and homogeneous, contains much earthy matter and the plant remains are notably advanced in change. Often it shows lamination only on drying. All kinds of peat have the lamination. In Moortorf there are often alternating layers, differing in color, density and composition ; in Specktorf the structure is especially distinct. Peat then is not an unstratified mass and one cannot say that the lamination of coal — places it out of comparison with peat. Close investigation shows so many similarities between the peat layers and those of some coals, that this kind of structure favors rather than opposes comparison of coal-making with peat-making. This lamination appears in the autochthonous peat, in the diluvial brown coal originating in peat and in the whole range of the brown coal formation. But one must remember that the coals were not all formed on the same model; that comparison with peat is only tentative, as modern peat is made from moss and swamp grasses, while in the coal time the deposits came from a wholly different moor and swamp vegetation. 2 The stone coal formation for the most part is to be regarded as an inland formation, originating in widespread leveling and sub- siding of the land, in many cases on swampy lowland along the sea- coast, over which floods distributed materials, such as shale and sandstone. On the extensive but not high land of the Carboniferous time, waters were penned in great areas and became converted into morasses, where a luxuriant vegetation flourished. It is very prob- able that in occasional drying of the swamp followed by renewal of the flooding, one may find explanation of the alternating bright and dull coal. This does not exclude influx of broken and shattered plant stuff from the higher surrounding region; that might even have predominated in some localities and have been the basis for cannel and boghead. Even from the swamp vegetation itself, decay- 64 oe) eee eg ome BU 2 S Moe eis ee eT ae oe ap eC TT ae ee ee 191t.] STEVENSON—FORMATION OF COAL BEDS. 65 ing material might float away to deep water within the swamp, so as to be heaped into peculiar massive layers like cannel. Flooding of the plain and deposit of mineral matter checked formation of coal; but the swamp would be re-established and a second formation be made; or possibly for a long period only rock material might be deposited. _ How far variation in the water niveau may affect the question is considered only so far by v. Gumbel as to let him warn against the conception that basins, now filled by a thick series of coal bearing deposits, were filled with water in like manner at the be- ginning. These bowls were filled very gradually; they must be thought of as filled temporarily by a relatively shallow pond of water, which little by little reached a higher level. At times, marine remains occur in strata between the coal beds, a condition which seems opposed to the explanation offered. But this occurrence is due to the fact that the low swamp land was spread out near the sea and was exposed to invasions, so that remains of marine animals might be enclosed in materials originating on the land. Marine or brackish water forms might be enclosed in the coal deposit itself, if it were formed alongside an arm of the sea. In general, coal beds are an autochthonous product of dead, broken and disintegrated plant fragments with only local and petty contribution of transported material of the same character. Wethered® called attention to the fact that coal seams are not single beds, but are separated by partings into benches which may differ in quality as well as structure. Sometimes Stigmaria are present in the partings. The Cannock Chase or Shallow seam, near Edinburgh, has in its upper bench, 1 foot 10 inches thick, the brownish layers composed of macrospores and microspores, while the bright layers, containing some woody tissue, are composed mostly of a structureless material which he terms “hydrocarbon” in preference to “bitumen.” Whence this comes he does not know, but wood tissue may con- tribute to it. The middle division of the bed is very different, consisting almost wholly of “hydrocarbon” with very few spores. ®* E. Wethered, “On the Structure and Formation of Coal,” Q. J. G. S., Vol. XI., 1884, Proceed., pp. 50, 60. PROC, AMER, PHIL, SOC., L, 198 E, PRINTED APRIL 25 IQII. 65 66 STEVENSON—FORMATION OF COAL BEDS. [April 21. It is possible that spores may have been there and that they may have been decomposed, but spores are much more resistant than is woody material. The main division has a great accumulation of spores but also a fair proportion of the “hydrocarbon.” He con- cludes that some coals are made up practically of spores, others are not; the differences in benches of a coal bed are of this character. Harker, reasoning from the ornamentation of the spores, suggested that they may have come from a plant related somewhat to /soetus. In the discussion of this paper, Carruthers took exception to con- clusions based on markings seen on spores. He knew of no reason for referring those spores to Jsoetus or any other form of sub- meged vegetation. Spores in coal were discovered first by Morris; they are associated with Sigillaria and Lepidodendron; the coal was the soil for the vegetation, penetrated by Stigmaria roots of the plants. A Sigillaria stem, at the Leeds museum, filled with white sand, penetrated far into the coal in which it grew. Coal seams are remains of forests which grew on swampy ground. The macro- spores were not composed originally of brown substance, they are merely filled with it. E. T. Newton stated that some coals are certainly made up of macrospores and microspores. Dull coal contains spotted tissue ; in- termediate coal contains both forms of spores; bright coal is a brown substance, usually structureless, but in one case, known to him, it consists wholly of spores. Dawkins had never found sporangia in coal though both macro- spores and microspores are abundant. Coal consists of carbon and resin, the latter giving the property of blazing, which Huxley would — attribute chiefly to the spores. With this conclusion, Dawkins agrees only in part. The carbon comes from decomposition of woody portions, but the resin from cell concretions in the living plant. Carboniferous forests grew on level alluvial tracts but little above the water level. Dawkins,®® discussing the geographical conditions in Great Britain during Carboniferous time described the mode in which the coal beds accumulated. ®W. B. Dawkins, “On the Geography of Britain in the Carboniferous Period,” Trans. Manchester Geol. Soc., Vol XIX,, 1887, pp. 45-47. 66 Se Oe eT ig Me ae ‘ STEVENSON—FORMATION OF COAL BEDS. 67 Oscillations of level still continued as the north, but the land Beonstantly encroached on the shallowing sea, the mud encroaching on the Carboniferous limestone and the sandbanks following the -_- closely. Meanwhile “the terrestrial vegetation was spreading _ from the old Lower Carboniferous land areas over the new Upper Carboniferous marsh lands, from the mountains of Wales and — ¢ d : : 4 “from the other Lower Carboniferous islands, now uplands. These forests contributed in their decay, through many generations, the ' accumulation which now, compacted by pressure and subjected to earth heat, is familiar to us as a coal seam. Each coal seam repre- : sents a land surface, just as the sandbanks and mudbanks (sand- g bi : 4 stones and shales) above it point to submergence. The fact too that the coal seams in a given section are parallel to each other or nearly so, implies that the forests grew on horizontal tracts of land, just as the associated sandbanks and mudbanks, with marine or freshwater shells, prove that these horizontal tracts were near the sea level or within reach of the waters of a mighty river. We may , . 2 y q © Yearn also from the study of the isolated coal fields that this great horizontal tract of forest clad alluvia occupied nearly the whole area of the British isles in the Upper Carboniferous age, from the Scotch Highlands southward, the dead flat being broken only by the higher lands, the old islands of the Lower Carboniferous sea, which T have already described. It was indeed the delta of a mighty river, analogous in every particular to that of the Mississippi—a delta in which from time to time the forest growths became depressed beneath the water until the whole thickness (7,200 feet in Lancashire) was accumulated of coal seams and associated sandstones and shales. After each depression the forest spread again over the bare expanse : of sand and mud piled up in the depression.” q 4 : The great northern and western land, termed by Dawkins, _ Archaia, whence came this mass of mineral deposits, occupied the North Atlantic sea, stretching from the west coast of Ireland and the Scottish Highlands to the American continent. To this great land may be traced the pebbles and groups of pebbles found in the Lancashire coal seams, mostly quartzites, which probably were brought down in flood time in roots of trees from the shingle beach. 67 . 68 STEVENSON—FORMATION OF COAL BEDS. [April 21. Williamson,” in disctissing the characteristics of the great fossil in the Owens college museum, remarked that that specimen had removed finally all doubts respecting the relations of Stigmaria by showing that plant to be the root of Sigillaria. The roots divide only once and after division extend indefinitely. The stigmata are lacking near the stem because the roots increased by exogenous growth and the superficial portion with its rootlets was thrown off. The trees grew in swampy ground as the swamp cypress does in American swamps. The gymnospermous plants grew on drier ground. The particular tree under consideration must have been at least 100 feet high. When it died, decay continued downward to the point shown and then was checked probably because the lower portion was buried in sediment and protected from air. Thence decay proceeded very slowly until the woody tissue of even the root disappeared. Meanwhile, the surrounding rock had hardened and had taken a cast of the stem and roots. The surface sank beneath the water and soft sand filled the cavity; thus the roots have their original form. Fayol, after spending many years in study of the basin of Com- mentry, published his results in a remarkable work, which is un- excelled as a record of detailed observation. This work presented the grounds on which, several years before, its author had based his theory respecting the formation of coal beds. The positive posi- — tion taken in favor of the transport theory and the clearness, with which the observations were offered, caused a notable reaction in favor of the doctrine that coal beds are formed of transported vege- table matter. A year after publication of the work, Fayol gave a “summary of the delta theory, as he termed it, at the summer meeting of the Geological Society, when several members of the society com- mented on the theory. This resume, being the later presentation, is the basis of the present synopsis.” pA Ee Ps See ee ae eae oe LV AVL er a rans RE Smet tee es The theory is based on the laws of sedimentation, as observed in . "W. C. Williamson, “On the Fossil Trees of the Coal Measures,” Trans. Manchester Geol. Soc., Vol. X1X., 1888, pp. 381-387. “H. Fayol, “Etudes sur le terrain houille de Commentry,” I"*. partie. “Lithologie et stratigraphie.” Bull. Soc. Min. Ind. St-Etienne, 2™° Ser., a XV., Liv., III, IV., 1887; “Résumé de la theorie des deltas et histoire du _ bassin de Commentry,” Bull. Geol. Soc. France, 3™° Ser., XVI., pp. 968-078. 68 ts eer a i eu ee “A Ree m git] STEVENSON—FORMATION OF COAL BEDS. 69 _ deltas. Mingled detritus brought in by streams forms a stratified a deposit in the basin, where the beds may be composed of a single _ substance or of several. Those beds are inclined, irregular and of small extent in tranquil waters but less inclined and of wider % extent in agitated water. The inclination may vary from 0 to 45 a degrees ; different portions of a bed may vary much in age, while _ beds at different levels may be contemporaneous. The total thick- ness of a deposit has no necessary relation to the sum of thicknesses of the beds which compose it, for a basin, 100 meters deep, may be _ filled with inclined beds which may have a total thickness of 1,000 _ meters; he gives illustrations of these conditions. The little basin of Commentry is one of several isolated areas in a.synclinal which is about 60 kilometers long. These are separated by granite and gneiss and the evidence shows that they were always separate. That of Commentry, 9 by 3 kilometers, contains only Car- boniferous rocks, except at the northwest, where some Permian re- mains. The rocks are not disposed at hazard, but there are definite zones or areas, each with its own type of rock, and these areas, as it were, interlock laterally. Each contains detritus derived from a _ single locality, though there is a greater or less intermingling where the deposits interlock as overlapping wedges. The history of the basin is thus interpreted by Fayol. : A lake, 9 by 3 kilometers in area and 800 meters deep, was sur- 3 rounded by steep mountains. Rainwater ate away the surface, 4 _ digged valleys, carried to the lake pebbles, sands, clay and plant _ materials, by which at length the lake was filled. This was one of 7 numerous lakes, depressions and alpine elevations on the central : plateau of France. Sediment brought in by the streams was heaped up at mouths and formed deltas. The main stream at the northwest, _ the Bourrus, cut through the mica schist and reached the granite, the _ latter being found in the upper part of the delta. This delta has the '- steep slope, with pebbles, blocks, sand, clay and plant debris, all dis- _ posed in accord with the laws of delta deposit. A somewhat smaller 7 stream, the Colombier, at the east, flowing over anthracitiferous beds 4 and afterwards cutting back to crystalline rocks, formed another _ delta of similar type; while petty streams from the north. formed 69 4 i ian in od 70 STEVENSON—FORMATION OF COAL BEDS, [April 21. small intermediate deltas. Apparently nothing came from the south, where the waters found their outlet. As the deltas increased in size and approached each other, their elements intermingled. i The lighter materials, clay and plant, floated into a bay in the southeast corner, where they formed some beds of shale and coal, while in less degree, similar materials floated off on the other side of the Bourrus delta into the bay at the west, where, in like manner, deposits of shale and coal accumulated. Eventually the Bourrus delta divided the lake into two small ponds and in the larger were formed thin irregular lenticular beds of impure coal. At length the _ lake was filled up and streams began to destroy the coal formation. 4 Disturbances set in afterward but they were not serious, for the . Permian deposits are almost horizontal. | : The facts to support this explanation of the origin of the beds, both mineral and vegetable, are presented abundantly in the great excavations. The walls show local faultings, thinning of faisceaux 4 of beds, pebbles of coal are seen in several strata, a great lenticular parting, in part very coarsely conglomerate, occurs in the Grande _ Couche. This remarkable coal bed is only a few centimeters thick at the southeast outcrop, but it swells thence to 10 to 12 meters and retains that thickness along the outcrop for about 2 kilometers and a half, beyond which it becomes thinner and at length disappears. Fol- lowed down the dip, it decreases in thickness and disappears toward __ the depth of 350 meters. The outcrop resembles an open C and the 4 interval from the outcrop to the old rock is 500 to 800 meters. : Before disappearing at the west, the bed breaks up into six diverging _ branches. Two other beds, the Gres noirs and the Pourrats, are in q contact with the great bed at the southeast but they diverge west- a ward. Some lenticular deposits of anthracite occur at the base of the ; series in both bays. a Fayol made careful calculation of the quantity of vegetation 4 which could be produced on the whole drainage area of the lake and © — asserted that enough be produced to give ten times the coal present— __ and this within the period of 17,000 years. This period is a maxi- mum, corresponding to a very slow filling and to the minimum trans- ’ portation of vegetable material. On the hypothesis of formation in 70 ” 1911.] STEVENSON—FORMATION OF COAL BEDS. 71 situ after the manner of swamps, he thinks a period of 800,000 years would be required. Fayol’s delta theory, then, is that the deep lake was filled gradually with material carried down by the streams; that this material was deposited according to its gravity, fine clay and vegetable matter being regarded as equivalents; the arrangement being that observed in deltas. It differs from the theory offered by Jukes by adding the suggestion of great original depth of the basin, a conception against which v. Gumbel had argued a number of years before. ; The record of the summer meeting of the Geological Society was issued as a separate” and it contains the discussions by several mem- bers. The doctrine as enunciated by Fayol was regarded by Busquet as applicable to the basin of Decize, by Nougarede as supported by much observed in the basin of Epinac, and by Bergeron as explaining the conditions observed at Grassesac and Decazeville. Renevier™* was not prepared to give assent to the doctrine and he suggested some grounds for hesitation. Vegetable materials in sus- pension are equivalent to fine mineral débris. If the coal beds were formed, as Fayol thinks, by the sweeping off of vegetable débris from the land and its deposition on the surface of the delta, that débris should accumulate on the border of the dejection cone, in the more tranquil waters, so that the deposit should have only a gentle original slope. But the great bed of Commentry has an extreme dip of 50 degrees, the same with that of the beds which accompany it. He regards these dips as impossible in a cone of dejection and suggests other modes of accounting for them. He maintained that the phe- nomena indicate, in part at least, the agency of marshy or semi- aquatic vegetation. Even the great thickness of the Grande Couche seems to him an argument in favor of vegetation in place, receiving increment brought in from the neighboring forests. Delafond** was inclined to question the applicability of the doctrine without modification to. the basins of the Saone-et-Loire (those of Autun, Blanzy and Creusot). Fayol conceived the exist- ence, before the coal deposition, of a deep depression transformed ™“ Réunion extraordinaire dans l’Allier,” Bull. Soc. Geol. de France, 3™* Ser., XVI., 1890. FE. Renevier, “ Réunion, etc.,” pp. 77, 78. “F. Delafond, “Réunion, etc.,” pp. 73-78. 71 a ee Se Seistuaeth, aia ty 72 STEVENSON—FORMATION OF COAL BEDS. [April 21. into a lake, in which would be deposited, in form of a delta, the vari- ous elements which constitute the Coal Measures; the plants, giving the coal beds, would have been furnished principally by the luxuriant forests which grew on the alluvial plains of the deltas. During and after the formation of the Coal Measures, the movements of the crust were so unimportant as to leave no apparent trace, so that to- day one can easily find all the circumstances accompanying the for- mation of the deposit. But these were not the conditions in either the basin of Autun or in that of Blanzy and Creusot. There were important movements of the crust during and after the Carbonifer- ous and the Permian. In Autun the successive stages overlap in such fashion as to be explained only by admitting, during the process of deposition, the existence of crustal movements which modified profoundly the shape of the basin. Further, it would be difficult to explain by this doctrine why in Autun the important coal beds are in only the lowest part of the formation, at the time when the alluvial plains of the deltas were small; whereas, in the later part of the formation when those plains should have acquired great extent and could support immense forests, there were formed only some insignificant deposits in the Upper Coal Measures. Similarly'in the other basins of the Saone-et- Loire, there were movements during the formation of the Coal Measures and of the Permian, which caused the overlapping of deposits. Delafond recognizes that the process of delta formation explains the manner of deposit, the separation of the various materials, coal, shale, sandstone; but the intervention of movements of the crust is indispensable. De Launay®™ remarked that it would not be incompatible with the theory of deltas to believe that movements of the crust occurred during the period of the Coal Measures and that they had given progressively the great depth observed. to-day. Almost at once after the appearance of Fayol’s first publication, de Lapparent*® gave his adhesion to the new doctrine. His first *L. De Launay, “ Réunion, etc.,” p. 102, footnote. *° A. de Lapparent, “L’Origine de la houille,’ Assoc. France. Avance. Science. Conferences de Paris, 1892. The same in Rev. des quest. scien- tifiques, Juillet, 1892. 72 r911.] STEVENSON—FORMATION .OF COAL BEDS. 73 publication was in 1887; in 1892 he presented his views in vigorous fashion. The statements are made with that clearness and precision which characterized his writings, so that it is well to give the synopsis in detail. : bo Sn ge a ek ho as aN ee Ts a aa ee CES a I ah a a hl a na oe I a ae a a oi aa SS te aa Ce The early observers regarded coal as due to transported vegetable _materials but the fascination of actual conditions, as exposed by Lyell, led men to abandon that explanation and to see in the vast peat bogs of this day the modern representative of coal beds. De Lappa- rent gives a synoptical statement of the peat bog theory. He thinks this doctrine deserving of a double reproach—it draws no argument from the nature of the coal itself"? and it does not consider suffic- iently the topographical conditions of each bed. De Lapparent says that coal, especially in the great maritime basins, has wholly mineral aspect, laminated, with conchoidal fract- ure and showing no sign of organization; even thin sections show only amorphous material with rare indications of cellular structure. In most cases, chemical and microscopical examination must be com- _ bined, but sometimes the former is unnecessary. Fayol discovered at Commentry, in 1883, lenticular brilliant zones which proved to be flattened stems. Grand’ Eury, in 1876, asserted that the coal of the Loire basin was formed of vegetable remains laid flat in a position uniform enough to suggest a liquid in repose. Several beds at Saint- _ Etienne consist wholly of Cordaites bark and the Grande Couche at Decazeville is composed of bark of Calamodendron. This determina- tion, first made by Grand’ Eury, is interesting as showing that the leaves, barks, etc., play in the coal the same part that vegetable im- _ prints do in the shale. The ulmic matter, resulting from maceration of vegetable detritus, formed the sediment in which the recognizable - remains were buried. To explain the origin of this amorphous material, he quotes : Saporta, who relates graphically the conditions existing in the dense forests of the hot, humid Carboniferous time. The rapidly accumu- lating mass of leaves, loose internal material from tree trunks, was “Tt is well to remark in passing that de Lapparent’s statement was made _ $4 years after Link’s investigations, 33 years after Dawson’s publications in the Q. J. G. S. and 9 years after publication of v. Giimbel’s elaborate researches. 73 74 STEVENSON—FORMATION OF COAL BEDS. [April 21. converted into ulmic material, the lower part of the deposit becoming a blackish paste. Detached heaps of leaves, peripheral sheaths of ferns, cortex of Sigillaria, Cordaites, etc., obstructed places at foot of slopes and awaited only the passage of waters in order to abandon to them the great mass of material in various stages of decomposi- tion. This vegetable pulp is the amorphous gangue in which one finds the barks and leaves. But it is no longer in place. It shows evidence of having been suspended in water; the condition of the fragments shows that they have been subjected to frequent and energetic friction. By what mechanism was this transport effected? Grand’ Eury thought that the waters of great rains sweeping down the slopes drew the vegetable detritus into lagoons—such waters were limpid. At other times the streams carried muddy water with sand and clay giving sandstone and shale. Thus was explained the alternation of coal with other rocks. But de Lapparent cannot understand this selective process—the conditions are unlike those of the present day. The delta theory of Fayol is preferable and it applies perfectly to the lacustrian basins of central France. It is no mere hypothesis, but the result of long, painstaking observa- tion in the great open quarries of Commentry. More, Fayol made experiments which proved that the conditions were such as must be due to delta formation. The cause was gained and it remained only to answer objections offered by adherents to the old theory. The presence of vertical trunks was shown to be not only not inconsistent but rather consistent with the theory. And this was the most important objection. The presence of Stigmaria in the underclay is no objection. Those are rhizomas capable of giving origin to Sigillaria; when swept by tor- rential currents, they were drawn into the deltas, where being heavier they would pass to the bottom of the mass which was to become coal. The delta theory is full of important consequences. There is no further need of numerous and complicated movements of the crust. The beds have been deposited one on the other as sediments on the surface of a submerged dejection cone. If complete stability of the surface be one of the conditions of the phenomenon, there is at least no a priori reason to put it in doubt; as the beds had to be deposited 74 2 4 1911] STEVENSON—FORMATION OF COAL BEDS. 75 with a certain inclination, there is no need of calling in, for lake basins, dislocations to explain phenomena which may very well be primordial. The time required for the deposits is vastly shortened. Not only a complete coal bed, whatever its thickness, but also a por- tion of the underlying clay and sandstone, becomes before our eyes the product of a single flood. Fayol has shown also the rapidity with which vegetable matter is transformed into coal. The coal of pebbles in the rocks is coal, so that when a portion of the delta was exposed by a change in equilibrium of the surface, its coal suffered erosion as did the other rocks. De Lapparent finds in the study of Commentry some important matters bearing on the origin of-the coal itself, which will be considered in another connection. The coal of the maritime basins of France is a vegetable allu- vium deposited in a delta; but the material has been brought from a greater distance and by the action of the waves it has been spread out over a greater area. In the central plateau the vegetable paquets descended violently from the neighboring steep slopes to be deposited en bloc with pebbles of the torrent, thus producing some thick but very localized masses of coal. In the Nord area, there must have been, far above the mouth, wide river sheets in time of flood, many kilometers broad, like the Amazon and Orinoco, on whose surface the vegetable matter was spread. In subsiding, the ulmic materials, which formed the chief mass, separated themselves from the fine clays. This explains the constancy of the floor, while the roof may consist of any material. As the unmacerated vegetable matters, fronds and barks, had to float on the surface of the ulmic materials, one can understand why they are so abundant in the roof. The mouth of rivers changed their position, which explains the invasion of brackish waters. Thus is understood easily the filling of the old arm of the sea. Why is it that a theory, so luminous, has not gained the adhesion of any but Frenchmen? De Lapparent thinks the hesitation due to lack of confidence in anything novel which comes from outside, _ and tends to overthrow notions so long accepted that they seem to be part of a national patrimony. Foreign doctrines are subjected to quarantine as foreign goods at a custom house. It is possible that 75 Se ee ree a en ee ees ae eke 76 STEVENSON—FORMATION OF COAL BEDS. [April 21. the hesitation is due to imperfect exposition of the doctrine at the outset, when Fayol declined to accept crustal movements as having had any influence ; but that error was corrected afterward by Fayol. De Lapparent considers that to deny all influence of orogenic move- ments upon even the lacustrian areas would be excessive. Coal basins are depressions, feeble lines of the earth’s crust, are land- marks of fractures whose equilibrium has been disturbed frequently. Malherbe’® notes that, though the explicit statement is not made, Fayol evidently regarded his doctrine as of universal application. But Malherbe asserts that, while it may suffice for Commentry, it cannot suffice for other basins. He utilizes the Liege basin as test- ing ground. That basin has an area of 40 by 15 kilometers, with 50 coal beds and numerous petty seams. The northerly border is but slightly disturbed; away from that the disturbances become serious and some of the faults extend through the formation, which is 1,200 to 1,500 meters thick. This is very different from Commentry, which is small in surface and depth, enclosing an insignificant num- ber of beds. If the Commentry strata are in the original position, those of the Liege basin must be the same; but everything proves the contrary—the enormous displacements of the beds, the presence of Cardium in horizontal and inclined beds alike; all show original horizontal deposit. The waters from the Liege basin carry salt and Roget-Laloy has proved the same for the coal formation of the north of France, concluding therefrom that that is the sea water of the coal time imprisoned in the rocks. The deposit is not lacustrian but fluvio-marine. Fayol’s capital objection to theories other than his own is the apparent impossibility. of periodicity ime deluges due to terrestrial oscillations. Malherbe thinks it equally difficult to explain by Fayol’s hypothesis the transport of a mineral formation, 1,500 meters thick and enclosing 50 coal beds from 0.45 meter upwards on an area comparable with that of modern seas—for the elevations break- ing the area into basins came after the coal time. Oscillations are known in the present time, they are probable for other times. If one recognize that subsidences necessary for formation of beds *R. Malherbe, “ Géologie de la houille,’” Ann. Soc. Geol. de Belgique. T. XVII., 1890, Memoirs, pp. 25-40. 76 1911.] STEVENSON—FORMATION OF COAL BEDS. 77 occurred only during accumulation of the great beds and that the overflows, bringing about the deposition of sterile rocks, led to transportation of vegetable matter intercalated in the intervals as veinettes, the number of overflows would be greatly reduced. Mal- herbe discusses Fayol’s doctrine in detail and at the close expresses much doubt respecting its competence to explain even the phenomena of Commentry. Renault*® says that coal beds are intercalated among beds of sandstone and shale and, like those, they have all the features of deposits made in water. In sandstone, the fragments are inorganic and preserve the chemical as well as the mineralogical characters of the rocks whence they came; in coal, they are derived from plants and conserve the anatomical, at times, also the chemical characters of the plant organs. The fragmentary condition of these organs, the small proportion which they form of the mass, consisting chiefly of a blackish vegetable powder as gangue, show that the plants had been subjected to repeated energetic friction before their burial. So one cannot admit that coal beds were formed solely by accumula- tion, sur place, of debris from an exceptional vegetation, spreading over marshes, lowlands, lagoons, etc., near lakes or the sea; that the surface, subject to elevation and depression, saw, checked and again restored, that great vegetation of which innumerable genera- tions would be represented by successive coal beds. The fragments of wood and bark are very small. If the vege- table materials had been changed into coal and buried where their debris is found, it is certain that, in place of these reduced frag- ments, there would be entire trunks, branches and complete leaves as principal constituents ofthe mass. More, taking into considera- tion the diminution of volume, which vegetable tissues experienced in becoming coal, it is evident that numerous forests of high trees growing successively on the same place, would form hardly a few centimeters of compact coal—even though one suppose that, at the foot of the trees, there grew a mass of herbaceous plants. Further, the thick coal beds are separated by great deposits of sand- stone or shale; as those deposits were formed slowly after the ™B. Renault, “ Etudes sur le terrain houiller de Commentry,” Livr. 2™° Flore fossile, Saint-Etienne, 1890, pp. 704-712. 77 oo ty ee ace wt A Pl eet ee rsa Sm a te aces 78 STEVENSON—FORMATION OF COAL BEDS. [April 21, manner of sediments, one must assign, if he admit this succession, an extraordinary duration to the coal epoch. | Renault accepts the explanation offered by Fayol and commends especially the shortness of the time which it requires. During the Carboniferous time, the air held more moisture than now, as no ice cap covered the polar regions; the rains were frequent and abundant; depressions occupied by lakes were filled rapidly. If one consider the strength of the torrents, greater than now, and the vigorous growth of vegetation, surpassing that of the present tropical regions, he will recognize that the formation in the Basin of Com- mentary could have been deposited in even less time than is re- quired by the Fayol hypothesis. The selection, so distinct in deposi- tion in inorganic materials, would take place with equal readiness in the plant materials. Coarse fragments, such as trunks, branches, would be dropped with the sand and clays, while the lighter, finer materials would be carried beyond into deeper parts of the basin. Erect stems have little bearing, upon the question at issue. Many of them are merely in-floated fragments, while those, which are in situ, do not penetrate the coal beds and have no relation to them. Spring®® undertook investigation along a new line. His study, though bearing largely on the question of transformation, finds place here because the results have an important bearing on the manner of accumulation. The homogeneity, the structure and com- position of coal beds all seem to favor the doctrine of transport; but the stratification within coal beds does not exclude the doctrine of in situ origin, for with rare exceptions modern peat bogs show a structure resembling that of coal. It is clear that a definite conelu- sion respecting mode of formation cannot be reached by study of the coal bed alone; he determined to investigate the shales of mur and toit. The mur of a bed formed by transport wou!d be impregnated with vegetable matter to some distance below the coal while the toit should contain little. In the Belgian terrane, the shales of the toit, °”W. Spring, “Détermination du carbone et de l’hydrogéne dans les schistes houillers,” Ann. Soc. Geol. de Belgique, XIV., 1888, “ Memoires,” pp. 131-154. 78 ESR ee ne ao he Pm gee ag PE A jms in i a A, c= or 7 = 1911.) STEVENSON—FORMATION OF COAL BEDS. 79 _when broken up by atmospheric agencies, yield a hard rather plastic material, resisting plant growth, yet they are as black as those of the mur. The theory of origin from peat would require that, in the mur, the quantity of carbon increase as it approaches the coal, as it must contain roots of plants; while in the toit the carbon should decrease gradually as one recedes from the coal. There is no abrupt change from coal to shale in the roof, so that the latter should be richer in carbon than the mut.. It is necessary to see how transformation of vegetable matter into coal is explained by each theory. This necessity is felt by de- fenders of the transport theory, because the flowing water furnishes only a mass of wood, bark, leaves whereas according to the theory of peat bog origin, the change of vegetable matter into peat is associated with the deposition. In passing from vegetable matter to coal, there is great loss in hydrogen and great enrichment in carbon. Either the plant ma- terials were changed into peat, lignite and the rest successively, or a the organic matter was converted at once into its present state with- out passing through the intermediate stages. The latter explanation rests chiefly on Fremy’s experiments, which showed that vegetable matter, subjected to high pressure and a temperature of 200° to 300° C. for a long time, becomes converted into a material very closely resembling bituminous coal. A fundamental objection to this theory is that no evidence exists suggesting that any such e temperature prevailed, and nothing is less established than the con- ception that time could compensate for deficiency in heat. However this may be, it is evident that, according to the doctrine of transport, the change going on in materials between the shales _ requires that specimens of shale collected at equal distances in re- _ ceding from the coal, should show the carbon and hydrogen varying in a determinate manner; in proportion as one recedes from the coal the shale should have less of carbon and more of hydrogen as the more volatile hydrocarbons would go farther. But the doc- _ trine of peatbog origin leads to a contrary condition. A determination of the carbon and hydrogen in shales near coal beds may aid in answering the question as to whether the 79 80 STEVENSON—FORMATION OF COAL BEDS. [April 21, hydrocarbons are impregnations from the forming coal or are due, og as in the coal itself, to transformation in place of vegetable debris, 5 imprisoned when the shales were deposited. These determinations | would tell us if one should prefer the doctrine of transport to that of formation in situ, and whether the transformation of vegetable matter into coal has been accomplished by a kind of distillation or has been caused by a special kind of fermentation. i In the course of his studies, Spring discovered an unexpected _ condition—that the shales, containing organic matter, were the @ seat of slow oxidation, depriving them of hydrogen. The shales — not only protected the coal from erosion but also from oxygen, as ° gas or in solution, the action of the oxygen being exhausted in the shales. As the encasing rocks are not the same everywhere, the character of the coal should differ in the same bed and in different beds. Usually, meager coals are on the peripheral parts of a basin while fat coals prevail in the middle portions. May this be be- cause the latter have been better protected against the action oe oxygen? ? The shale samples studied were from the Saint-Gilles mine near Liege, eight of them, with one from the coal bed. Five were taken from the toit and three from the mur, each representing a vertical space of a half meter.. They are marked “a,” “b,” “¢" "d@ ang “e” for the toit and 1, 2 and 3 for the mur. The material was dried and analyzed with these results ; . Coa!. 1a? a6 aya? bind bs “eget ‘hig I 2 Fe MEU oes o5ch'6 5 3 86.61 7.54 3.35 221 1.20 070 0.00 003 O80) Hydrogen ........ 4.65 0.79 062 054 056 0.59 084 053 0.58 Sh OR Anes 1.84 98.33 92.05 93.86 92.00 94.08 95.16 93.50 93.20 Oxygen, sulphur \ ha Difereich 480 3.34 3.08 410 6.24 463 38 5.04 5.42, The carbon varies greatly but regularly, decreasing as one re- cedes from the coal. No conclusions can be drawn from conditions in the mur as the quantity is very small, but the variation in the toit is a logarithmic curve, the cause producing the variation is in inverse relation to distance from the coal. This seems to show that Fremy’s conclusions are right and that the shales were impreg- nated with carbonaceous materials at expense of the coal, the com- 80 git.) STEVENSON—FORMATION OF COAL BEDS. 81 pounds less rich in carbon going farther. But the relations lead ‘to a chemical impossibility ; “a” gives C,H,,, while the coal gives ‘practically CyeHyo. _ The reason is that not all of the water of hydration goes off at 20°. To escape this error, Spring employed hydrofluoric acid and ‘continued the solution until the ash was about ro per cent., the same with that of many coals. Analysis of the residues gave these ratios; Be ! “a” “p” “or ~ «g” “e” a > § & SS SR eee 24.80 30.45 36 ? ? 19.80 ? ? _- The results for “d” and “e” are uncertain as are also those for 2 and 3, the hydrogen being present in such small quantity. De- termining the absolute relation of hydrogen he has 3 Coal er gel bee ‘ort mast Sy ee. > 2 3 oe 88.61 7.54 335 221 1.20 070 099 093 080 Hydrogen ....... 465 030 O11 0.06 ? ? 0.05 ? ? eae 19.09 24.28 30.45 36.00 ? ? 19.80 ? ? The relation of carbon and hydrogen in the mur is very nearly the same as in the coal; it contains particles of coal little altered. But the toit results are remarkable; the hydrogen diminishes in relation to the carbon and in “d” and “e” it is no longer in appre- ciable quantity. Evidently the roof shales are not impregnated by -yolatile materials coming from the coal, as required by Fremy’s theory. The transformation of the vegetable matter is rather by ‘ulmic fermentation. Within the primitive marshy mass the plant ‘substances have yielded ulmic materials while becoming richer in carbon. These have impregnated the whole and have been modified _by external agencies. __ The doctrine of transport seems to be out of harmony with the results as by it one would have difficulty in explaining the richness in carbon characterizing the toit. The alluvium, because of its physical nature, could not support a sufficient vegetation. If one Suggest that the alluvium at its origin was mingled with much vege- table debris, it may not be superfluous to ask if the plants ‘could _Tremain on slopes, denuded and torn up by the flood which had swept way the most thoroughly rooted plants. Everything speaks of Origin in situ. But returning to the analyses. PROC. AMER. PHIL. SOC., L. 198 F, PRINTED APRIL 25, I9II. 81 a Rana SEP PTH FREY Ie 82 . STEVENSON—FORMATION OF COAL BEDS. (Apri ats If the alluvium covering the peat bogs came gradually it would be mingled with a greater or less quantity of vegetables, which had to undergo the same changes as the underlying mass in order to become coal. One ought to find in the alluvium the same propor- tion of carbon and hydrogen as in the coal itself, or at least near so. If this relation do not exist, evidently some external influence has been exerted. And the relation does not exist; the variation increases as one recedes from the coal; this irregularity must be _ due to some slow action becoming appreciable through lapse of time. Everything seems to indicate that slow oxidation went on in the shales, acting chiefly on the hydrogen, for which oxygen has the greater affinity, so that it has converted the vegetable matter into anthracite in the more distant part of the shales. i According to this conception, coal with abundant gas could have been formed only when the material was protected against atmos- pheric agencies. The many varieties of coal owe their origin rather to unequal degrees of protection; the fattest coals give off the — most abundant grisou—evidence that the enclosing rocks are im- — permeable. Wild*® in describing the Lancashire coal-field, referred ‘to fs “bullions ’’ which are characteristic of the Mountain-Four-foot coal bed. These, embedded in the coal, are ferro-calcareous “ concre- tions” more or less pyritous, frequently enclosing mineralized wood, “showing the woody and cellular structure of the plants which — have produced the seams of coal from which the concretions are — extracted.” Shells are absent, the nodules being for the most part fossil wood in varying degrees of preservation. The coal bed is persistent and its roof shale contains concretions, known as “ baum- — pots,” which at times are embedded partly in the coal. These are ironstone or calcareous, sometimes weigh 40 pounds and contain marine shells but rarely any wood. After a review of all the coal beds he considered the question ‘of their. formation. The generally accepted theory that coal comes from growth in situ seems to be a natural conclusion, for the roots in the underclay pass through several layers. It is true that under- *G. Wild, “Lower Coal Measures of Lancashire,” Trans, Manchester Geol. Soc., Vol. XXI., 1892, pp. 364 et seq. 82 11.] STEVENSON—FORMATION OF COAL BEDS. . 83 clay is not essential for vegetable growth, but more than three rths of the coal beds have it. The “ bullions,’ composed of fossil ood, occasionally show rootlets working their way through the decaying wood, separating the fibers which now surround them. The fossil wood is often parallel to the bedding of the coal, a condi- _ tion familiar in prostrate forests and in peat accumulations. Erect trunks and stems are unusual both in coal and peat. The underclay was the land surface which supported vegetation like the forests . swamps where warmth and moisture prevail. _ If coal is to be considered as derived from drifted material, he is puzzled to discover what has become of the shells and fishes, which must have abounded in the tracts of water in witich the deposits were laid down. To float some of the large trees either vertically or horizontally, with their outspread roots having a radius of 15 to 20 feet, would certainly require enough water to accommo- date fishes and mollusks. Remains of fishes are not necessarily de- stroyed by embedding them in coal-forming material, and shells are as capable of resisting destruction as fish spines are. Shells and fish remains occur often in impure cannel. The “bullions” have yielded no shells, and fish remains are very rare in pure coal. That the trees were forest growth is proved by the splendid specimens in the Manchester and other museums. _ Estuarial swamps with intermittent subsidence, permitting de- _ position of sand and mud, would explain alternations of coal and other strata, whichever theory of coal accumulation be accepted. Marine conditions frequently followed directly upon formation of coal bed ; fishes of shark-like types are in shales directly overlying coal at many horizons. But shells and fish are unknown in the underclay. __ Orton,*? in his description of the coal-fields of Ohio, considers the various theories of formation; some of them appear to be based on merely local conditions, others are extravagant and only a very iall proportion of the explanations seems to have been the result of careful observation in extensive areas. _ Ina coal-field, one finds a system which can be explained only by subsidence. Limestone is found above and below coal beds and “E. Orton, Geol. Survey of Ohio, Vol. VII., Antioch, 1893, pp. 256-262. 83 84 STEVENSON—FORMATION OF COAL BEDS. [April 21, is accompanied by iron ore. The coal beds, though variable, are wonderfully persistent and are always associated with fireclay There is no haphazard mode of occurrence. Coal is product of land life; limestone is of marine origin; the ore depends on life for con-— centration; sandstone, occupying the intervals between other rocks, is due to inorganic forces and it may be about equivalent to the others. Orton’s conclusions based on more than ke years of study in much of the Appalachian basin, are: (1) The Ohio coal-field, at the beginning of the Carboniferous, was an arm of the sea with the Cincinnati arch as the western boun- dary. (2) Marginal swamps of varying width became the earliest coal seams by long continued growth and subsequent fossilization. (3) While the swamps were submerged, in succession, and covered | by shale, sandstone or limestone, in turn covered by other swamps, the continental nucleus grew slowly at the south and the Cincinnati arch united with it by like advance eastward, expelling the waters of the gulf and converting the earlier formed portions of the coal — formation into dry land. (4) Every coal swamp had a narrower area than its predecessor. (5) As all coal seams were formed at sea level, so all were raised by continental growth to an approximate — equality, which their outermost outliers still retain. (6) To look for the earlier formed seams in the center of the basin would be to look for the living among the dead. (7) In the formation of one seam, in particular, the floor of the gulf, around which the swamps were growing, seems to have been raised nearly to sea level at mani points, and coal appears to have been formed in island-like masses over much wider areas than any single marginal swamp woul account for. pe Bolton®* describes a peculiar deposit of coal in Irelgae Th Jarrow coal bed appears to be a great cake, attaining a maximum thickness of 16 feet and thinning in all directions except towar the west, in which direction no tests have been made. Underclay absent at almost all localities. The lower part of the deposit is smutty anthracite with slaty structure and containing abundance 0 * H. Bolton, “ Notes on the Plant and Fish remains from the Jarrow Col- liery, Co. Kilkenny,” Trans. Manchester Geol. Soc., Vol. XXIL., 1804. 84 STEVENSON—FORMATION OF COAL BEDS. 85 Lepidodendron stems. The upper part is a pure typical anthracite. _ Fish remains, Gyranthus, Megalichthys, etc., occur throughout. The ‘plant remains are Halonia, in the form of crushed cylinders of wood. This condition and the mingling of fish remains led Bolton to conceive that the deposit was due to the bursting of a lagoon-like swamp and to the discharge of vegetable debris, consisting of bottom accumulations as well as of the twigs, etc., on the surface. He refers, for illustration, to the bursting of Solway moss in 1771, which spread over a square mile of ground, giving a mass of vege- table matter, 30 to 40 feet deep, demolishing houses, overturning trees and so contaminating the Esk that no salmon ventured into the river during that year. _ Kuntze** took up the discussion from a botanist’s standpoint and advanced a wholly new theory. He antagonized v. Gumbel’s con- clusions which he maintains are wholly at variance with that observer's facts. His own studies from 1879 to 1883 had shown that the Carboniferous flora was sylvo-marine, a floating vegetation. e objection that marine forms are wanting does not hold good: the forms, described by v. Gumbel as resembling algae, are chitinous bryozoans related to Aulopora. These, as stated by v. Gimbel, occur abundantly in cannel and make up a great part of the boghead coals. Carboniferous coals contain much sodium chloride, one fourth to one half kilogram per ton; Tertiary coals contain none. It is certain that the Carboniferous coals are not allochthonous; the flora must have been marine. He contends that students have failed to interpret Stigmaria rightly, for the appendices, regarded as rootlets, are water leaves. ‘The Stigmaria, with intertwining rhizomas and hollow stems rising above the water, formed floating islands. When overloaded, they sank to the bottom and through the mud until checked by some harder rock. He agrees with Potonié’s conclusion that they are not allochthonous but he cannot concede that the underclay or clay shale is a petrified humus, for the clay is no more a soil than are the - granite and other silicious rocks with which coal beds are often in p=" ©. Kuntze, “ Geboerieliactie Beitrage,” Leipzig, 1805, pp. 42-77. Sind Carbonkohlen autochthon, allochthon oder pelagochthon? 7 85 86 STEVENSON—FORMATION OF COAL BEDS. _ [April2 contact. The thickness and extent of some coal deposits are serious — objections to growth in situ. Richthofen describes a bed in China, 20 to 30 feet thick and having an area of 600 German square miles. This would require at least 400 feet of plant remains. The bottom three feet might have been a soil in which Stigmaria rhizomas could have grown, but the sturdiest defender of autochthony would be at loss to find a soil for the remaining 397 feet. Such a deposit could have been made only by a matt of sylvo-marine vegetation. : All allochthonous and land basin theories are untenable because transportation yields no undisturbed sedimentation; there is no transportation of organic detritus without contemporaneous trans- portation of inorganic material—the transportation of purely plant detritus is a superstition; subsiding land basins giving 7,000 me-_ ters of Carboniferous rocks, while neighboring basins subside at different rates, would be a marvel, for in order to account for the thick mineral beds the process of coal making would have to” be intermitted a hundred times; there are no basins so great as those of coal sedimentation. The four great deltas do not equal the Pennsylvania coal-field alone; Richthofen’s southeast Shansi field — would require a basin sixteen times as large as the Caspian sea. The great basins must have been sea basins and a sylvo-marine forest alone explains the intermittent deposit of coal, the clays being due to influence of streams. Kuntze classifies the theories as Autochthony, the irregular daa posit of the coal-producing substance directly on the place of vegeta- tion; Allochthony, the irregular deposit of coarse coal-producin substance on a distant place; only the powdery substance is depos ited after the manner of sediments. . Pelagochthony, the sedimentary deposit of coarse substance in water of the sea directly under the vegetation; a secondary product is the powdery detritus sometimes floated away from the coal magma and deposited elsewhere as anthracite. Autochthonous types are found in tropical or subtropical brow1 coal from wood-covered bogs, without sphagnum; newer peats cooler regions with sphagnum; shore swamps and some others. Allochthonous types are drift woods; sedimentary peats; s 86 — -rgrt.] STEVENSON—FORMATION OF COAL BEDS. 87 _ peat; paper peat, which is a bituminous clay with infusoria ; Blatter- kohle, a marly clay with a little sedimentary peat. _ Pelagochthonous types are: (1) Normal Carboniferous coal fields. ; The coal beds have originated from floating forests and remains of rooted trees occur in very limited localities. Naumann’s paralic _ coal-fields belong here; they are found in America, China, etc. (2) © Sea basin deposits, consisting of limited but often very thick beds, the coal frequently thinning seaward; these contain, besides sylvo- _ marine remains, abundant remains of trees rooted in clay. Best _ seen in France. Here, in part, Naumann’s limnic basins. (3) _ Amorphous anthracite, consisting of the finest detritus and forming irregular deposits ; does not include Faser-, Staub- or Koksanthracite coal. Penhallow® has given the results obtained by study of cannel- like coal from the lower Mesozoic of British Columbia. All the samples are composed of rod-like bodies more or less closely com- pressed, which resemble dark amber and are embedded in a cement- ing material. The rods show tubules within, many of them branch- _ ing, which are very suggestive of Mycelium; granulations are com- mon and often form zones around hyaloid areas. The features revealed by the microscope are: : (1) Absence of structure, (2) tubular ramuli of diverse dimen- sions, (3) rounded cavities, (4) large proportion of material in angular fragments and resembling that of the rods, (5) an amor- phous substance, associated with (4), occurring as distinct flakes or _ as cement to unite the rods. _ Appearance of structure was observed in only one rod and in that case it is evidently due to shrinkage; he thinks the spore-like _ aggregations are of chemical rather than of organic origin. The general character of the ramuli at once suggest Mycelium, but the intimate features and the arrangement forbid reference to vegetable structure. They rather resemble effects of internal shrinkage, fol- _ lowing hardening of the outer layer, such as one sees in amber and _ other resins. The material occupying spaces between the rods and — *D. P. Penhallow, “A Preliminary Examination of So-called Cannel Coal from the Kootanie of British Columbia,” Amer. Geologist, X., 1892, pp. 331-339. 87 88 STEVENSON—FORMATION OF COAL BEDS. _ [April at. apparently cementing them “consists of an amorphous and irregular — mass full of rounded holes, thereby giving it a spongy character.” It contains fragments of perhaps broken rods, the material in both being the same. The source of the amorphous material is not — certain. Penhallow offers no positive hypothesis respecting the origin of these coals, though he is inclined to think that it must be “sought elsewhere than in modified vegetable structure.” At the same time, he feels that the evidence is not sufficient to justify the assertion that they did not originate in vegetable structure. In 1892 and 1893 there appeared papers by Bertrand and Renault describing Bogheads and related types. Afterwards those observers published their results independently. The later studies of Renault concern the matter in hand only indirectly and they will receive consideration in another portion of this work. It is necessary, how- ever, to make detailed reference to Bertrand’s contributions, for, though they consider similar topics, the conclusions have a notable bearing on the formation of coal beds; and in this connection, the stratigraphical relations of the several types must be given. With- out that one cannot appreciate the full bearing of the studies. The joint study by Bertrand and Renault*® was of boghead obtained from Permian beds at Autun, France. This deposit occupies an area of 7 kilometers by 150 to 450 meters. The chief constituent is a thallophyte, Pila bibractensis, which makes up about three fourths of the mass; the remaining fourth being the “ fundamental mate- rial” with some clay. Vegetable débris is wanting, but poe of Cordaites and remains of fishes are present. These observers recognized the bodies of yellow, red and other tints, which had been mentioned by earlier students, but their study proved that “certain resin-like bodies represent the organic gelose and even entire organisms. A great proportion of the yellow and red bodies enclosed in coals are in this category and M. P. F. Reinsch has the great merit of making this known.” The inferior gelatinous plants have been preserved in this way when buried in “ C. Eg. Bertrand et B. Renault, “ Pila bibractensis et le boghead d’Autun,” Bull. Soc. d’Hist. Nat. d’Autun, V., 1892, Separate, pp. 95, pl. 2. 88 : ro1t.J STEVENSON—FORMATION OF COAL BEDS. 89 _ ulmic materials. The Autun boghead, 24 to 25 cm. thick, is not an ' accumulation of resinous pellets due to injection of hydrocarbons : q into plant débris, but it consists of 1,600 to 1,800 beds of algz, which sank to the bottom along with grains of pollen and the fundamental material as well as the detritus. The fundamental material is brown, rather flocculent and feebly colored. It is a precipitated brown _ substance analogous to the ulmic matters which color the Amazon and certain of its affluents. It contains particles of a darker mate- _ rial, thelotite, an infiltration which penetrates the thalli. The Pilas were alge of very low type. Their isolation in the fundamental material, their accumulation in beds, with traces of pressure on the under surfaces, suggest that they were floating algze _ like the fleurs d’eau. The pollen grains, usually reduced to their _ coats, were a powder resting on the water with the fleurs d'eau. The accumulation, which may have been very rapid, was only an incident in the formation of bituminous shale. It was made in quiet waters, with little or no current, and so rapidly that putrefac- tion could not begin in the mass. The deposit was laid down prob- ably in shallow brown waters, like those of the Amazon region, _ whose acidity is unfavorable to development of many bacteria. _ Nearby, were forests of Cordaites, which furnished the pollen. The second paper by the same authors’ gives results of study of the so-called kerosene shale of New South Wales, which had been utilized as a source of gas and illuminating oil. This shale, known as Hartley mineral, Wollogongite and, in some reports as -Torbanite, is of uncertain occurrence. Mackenzie** says that the ‘deposits are very irregular, there being no guide to discovery except the presence of fragments at or below the outcrop. Toward the border of a mass, the rich mineral becomes deteriorated and grad- ually passes into indurated clay, bituminous or non-bituminous shale, coal or ironstone. It occurs at two horizons in the Permo-Carbo- niferous of New South Wales, the most notable deposits being in the Upper Coal Measures, including the well-known areas of Hart- “C. Eg. Bertrand et B. Renault, “Reinschia australis et premiéres re- _ Marques sur le kerosene shale de la Nouvelle-Galles du Sud,” Bull. Soc. Hist. Nat. d’Autun, VI., 1893. Separate, pp. 105, pl. 7. * J. Mackenzie, Ann. Rep. Dept. of Mines for 1896, p. 100. 89 90 STEVENSON—FORMATION OF COAL BEDS. [April 2x. ley, Joadja creek and Wollongong at the south and Murrurundi. (Doughboy hollow) at the north. The only important deposit in — the Lower Coal Measures is at Greta near Newcastle in ngeiieat e port of the province. < Long ago, Clarke*® recognized the close resemblance of this : mineral to the boghead or Torbanite of Scotland. He thought it due to local decomposition of some resinous wood and believed that — the lens-form of the deposits and their passage laterally into shale — could be explained easily by supposing the mineral to be due to — drifted resinous trees, undergoing changes in shallow pools sur- rounded by material changing into ordinary coal. The quartzose — constituents are merely sand carried by wind into the pool. The — thickness of the deposit depended only on the supply of drift timber. — Wilkinson” says that the kerosene shale occurs in irregular — lenses, sometimes in actual contact with layers of coal as at Joadja creek, sometimes wholly unassociated with layers of coal, as at Hartley, or even as forming part of a great coal bed, as at Greta. . At the last locality, the boghead is a great lens in the coal, but there are many petty lenses of the same material scattered through the | coal benches. At Joadja, one finds small irregular patches of bright jet-like material, plant remains lying horizontally and numerous vertical stems of Vertebraria, whose lustrous bright jet substance is — in contrast with the dull luster of the shale. ao David" found-the shale in one place at the bottom of a great coal bed; Mackenzie®? found it at the top in another; while in still another David found a mass of alternating coal, clay and “ shale,” five beds of the boghead and four of bituminous coal. At the last locality the whole mass thinned out in one direction, the several — layers disappearing in succession until the last layer of boghead — passed into bituminous shale. There he saw many stems of Verte- — braria, both vertical and prostrate; in one tunnel, some of them four — _ W. B. Clarke, “ Mines and Mineral Statistics of New South Wales,” Sydney, 1875, pp. 179-180. ” C. S. Wilkinson, “ Mines and Min. Stat., 1875,” p. 131; Ann. Rep. Dept. Mines, 1884, pp. 149, 156; 1890, p. 208. “T. W. E. David, Ann. Rep. Dept. Mines, 1888, p. 170; 1890, pp. 221-224; 1892, pp. 150-163. be @ J. Mackenzie, Rep. 1895, p. 104. 90 1911.) STEVENSON—FORMATION OF COAL BEDS. 91 inches in diameter were converted into coal. Mineral charcoal is abundant in a mine in Camden County, while at Murrurundi the boghead “contains numerous fragments of mother-of-coal and small fragments of what appears to be coniferous wood like Araucaria, together with coniferous fruit.” In the pages already cited, David gives ten analyses by Mungaye, __ which show that at Murrurundi the ash varies from 17 to 68 per : cent. and the fuel ratio from 0.11 to 0.24; while at Ketoomba eight z analyses show ash from 10.7 to 78.1 and the fuel ratio from 0.13 Sm to 1.10. A specimen from Joadja Creek had 77 per cent. of silica _ + in the ash. ; The material studied by Bertrand and Renault consisted of two great blocks, one in Paris and the other in Brussels, each more than one meter thick, apparently the full thickness of the deposit. Like the Autun mineral, the kerosene shale consists of a fundamental brown, flocculent material, holding alge and remains of dead plant tissues. The algae are assigned to the genus Reinschia, now ex- tinct, but belonging to a group which was spread widely during Permo-Carboniferous times. The alge are all separate, though, at times, owing to paucity of the fundamental matter, they are in con- tact, they are still independent. They were free, floating on the s surface of absolutely tranquil brown water, and they rained down % upon the bottom, while at the same time, under the influence of cal- careous waters, an ulmic jelly was precipitated to form the funda- mental material. The great specimen in the Paris Museum shows _ 36,000 beds of these alge, but the proportion of algae varies in the _ several layers from 0.019 to 0.900 of the whole mass. At Joadja creek the mineral is often beautiful, with a satin-like homogeneous _ surface, and it consists almost wholly of the alge. _ Infiltrations are here as at Autun. The most important is red- brown, in strings or sheets, and shows fluidal structure; it is harder than the fundamental material; it often impregnates leaves and wood ; some plants have the property of absorbing this to a notable extent. its mode of occurrence and its tendency to penetrate the substance of plant remains suggest great resemblance to the thelotite of Autun. The authors make no attempt to decide respecting the 91 92 STEVENSON—FORMATION OF COAL BEDS. [April 21, source of this infiltration; they are convinced that it penetrated the © deposit, if not contemporaneously, at least very soon after its forma- tion and they suggest that it may be a kind of asphaltum, like that of lake Brea in Trinidad. The kerosene shale contains no animals except at Murrurundi, where some coprolites have been discovered. _ It is a charbon produced by unaltered gelosic organisms. Bertrand’s® later studies were published in a series of papers, _ his conclusions being summed up in a memoir presented to the — Geological Congress at Paris in 1900. The bogheads, typified by deposits at Autun of France, the Tor- _ banite of Scotland and the kerosene shale of New South Wales are . charbons gelosiques of Bertrand, accumulations of fresh water algae in a humic jelly,.their fossilization being in the presence of bitumen. The basal material of all is a clear brown fundamental jelly, the dull part of the bogheads and the same as the basal material of v. Giimbel’s Mattkohle. Spores and pollen have undergone macera- tion, but they did not liquefy. They gave two kinds of yellow bodies and they condensed bitumen strongly. When they abound, the coal, though dull, is brighter than mattkohle. Débris of vege- table matter, also a contribution by the wind, is distributed irregu- larly. The hardened tissues are usually brilliant, prismatic like vy. Giimbel’s Glanzkohle. Wood and barks can be found as brilliant coal, but this depends less on their organic nature than on the extent of alteration and their capacity to imbibe bitumen. Vegetation along river banks yielded tree-trunks, which, after imbibing bitu- — : men, were converted into bright coal. The alge were fleurs d’eau. They consisted of gelose and a little protoplasm, which, when humefied, would condense bitumen. _ They descended in sheets with other accidental bodies; in times of low water, the descent would be very slow, being impeded by the — *C. Eg. Bertrand, 1, “ Nouvelles remarques sur le kerosene shale de Nouv, Galles du Sud.,” Bull. Soc. d’Hist. Nat. d’Autun, 1X., 1896; 2, 3, “ Con- férences sur les charbons de terre,” Bull. Soc. Belge de Géol., etc., VII., 1894; — XI., 1808; (4) Caractéristiques du kerosene shale,” Assoc. Franc. pour l'avancem. des Sci., 1897; (5) “Les charbons humiques et les charbons de purins,” Trav. et Mem. de Univ. de Lille, VI., 1808; (6) C. R. du Congrés. es. Int. de Géol., Paris, 1900, pp. 458-497. 3 92 1911-] STEVENSON—FORMATION OF COAL BEDS. 93 fundamental jelly. Each ball of gelose yielded a little mass of glassy, transparent gold-yellow hydrocarbon. _ The bituminous matter found in all is wholly different from the fundamental material. There is proof of its intervention, for it _ follows clefts made by contraction of the fundamental material, which it does not color. The coalified stems of Vertebraria on Joadja creek are humefied vegetable material charged with bitumen. There is no evidence that this bituminous enrichment was due to condensation of resinous matter held in suspension by the funda- mental material; nor is there any evidence that the fundamental material originated from alteration of the enclosed bodies. The accumulation could be made with remarkable rapidity. A few good days with low water would suffice. All the accidental ‘bodies, enveloped in a humic coagulum, make a raft on the abso- lutely tranquil water. A very slight cause, colder weather, more water, would hinder formation of gelose and cause descent. The precipitation of brown matter was continuous but formation of _ gelosic matter was fortuitous; with check of algic growth, the deposit passes over to a humic coal or organic shale. The vegeto-humic _ deposit was fixed at once and remained unaltered. The fossiliza- _ tion was in the presence of bitumen, which became altered so as to be insoluble in the ordinary solvents of asphaltum.** : Bertrand’s charbons humiques differ from the charbons gelo- _ siques in that the fundamental matter is not diluted with foreign bodies. They are typified by the Broxburn shales of Scotland, con- taining, according to Cadell, about 75 per cent. of ash. Accidental bodies, such as alge, spores, pollen, vegetable débris are in small proportion. Bitumen penetrated through the fundamental jelly and enriched the shale. Bertrand finds no evidence that this bitumen is a leakage or exudation from a fermenting vegetable mass; he a believes that it was in the water and that it penetrated the accidental _ bodies only with difficulty. “ After the memoir was read in the Paris congress, de Lapparent asked _ what is to be understood by the term “bitumen.” Bertrand replied that “the term bitumen implied for him the idea of a substance charged with carbon and hydrogen, intervening wholly formed in the rock.” 93 94 STEVENSON—FORMATION OF COAL BEDS. _ [April2n — Gresley® called attention to the persistence of slate partings in the Pittsburgh coal bed as having an important bearing on the origin of coal beds. Two of them, one fourth to one half inch thick — and separated by 3 to 4 inches of coal, are present in an area of 15,000 square miles. Under the lower one is a coal bench somewhat more than 2 feet thick, while above the upper one is a bench varying 5 from 3 to 5 feet. The clay of the thin binders or slate partings is extremely fine grained, mottled, non-plastic, contains macrospores a and indefinite plant remains, but no Stigmaria. er Accepting in full the doctrine of transport, he assumes that, at — the close of deposition of the lowest bench, that mass of vegetable matter lay practically level on the bottom of a vast lake or inland a sea. Such being the condition he finds difficulty in explaining the overlying shale as due to fine material brought in by currents; the __ shale is uniform in thickness and composition over a great area, sO that the supply of material must have been uniform throughout ; there could have been no changes in currents or offshore conditions during the period of deposition. The quantity is not less than 100 tons per acre. He finds equal difficulty in the suggestions that the shale consists of wind-blown dust, that it is a precipitate from solu- tion, that it is concretionary. The supposition that these shales are substitution or replacement formations or that there was a segrega- tion of inorganic substances during solidification or the process of coal-forming involves serious difficulties. “To suppose that such shale bands were originally thin films of chalky mud, since chem- — ically converted into silica, alumina, iron, etc., would, I think, be ie exceedingly unsafe.” At the same time, he suggests that the globi- gerina ooze, widespread “over the bottom of the Atlantic, where deepest and farthest from land would seem to furnish us with about the only way (as to physical conditions) in which our shale binders _ in the ‘ Pittsburg’ coal bed can be imagined to have accumulated.” 2 If the lower slate binder was really deposited as silt by aqueous ~ transportation, the interesting query presents itself, How could the — succeeding 4 inches of coal be formed in situ? *®W.S. Gresley, “ The Slate Binders of the Pittsburg Coal Bed,” Amer. Geologist, X1V., 1894, pp. 356-395. 94 grr.] STEVENSON—FORMATION OF COAL BEDS. 95 _ The Pittsburgh coal bed thickens toward the southeast and the ‘slate partings, as well, thicken in that direction. The evidence favors the assumption that the organic as well as the inorganic materials came from the land surface in that direction. The absence of tigmaria casts reasonable doubt upon the hypothesis of formation situ, and this doubt is increased by the discovery of an aquatic fauna in the underclay of the bed, which Gresley has found to be a calcareous shale. _ The extraordinary uniformity of the Pittsburgh coal bed in _ purity and structure, the evenness and geographical extent of its several divisions make it the most remarkable known. In explana- | ‘tion of its phenomena, about all that can be said safely is “that, everything being horizontally stratified, every part of it was most likely accumulated under water. I have therefore come to the con- clusion that this coal is the accumulated remains on the bottom of a lake or sea of vegetable growth of aquatic forms (though much of it did not necessarily grow in the water) living afloat and dying and decaying, falling through the water.” All the familiar phe- nomena can only be explained by an aqueous origin for the coal. The problem of coal accumulation attracted Potonié’s attention ‘in 1886 but he published no results of direct study until 1895.°° In that year he had opportunity to study a core obtained in the Upper ilesian coal field. This core, 750 meters long, one to 2 decimeters diameter, begins in Saarbruck beds and ends in the Upper Ostrau ‘deposits. As submitted to Potonié, it was complete and it was studied by him in company with C. Gaebler of Breslau. The core shows not less than 27 coal beds, each of which is in direct contact ith a Stigmaria underclay; in most of them, remains of Sigillaria are present and some contain Lepidodendron—particularly in the accompanying carboniferous shale. Ochsenius, who urged the allochthonous origin of coal beds, explained cases, such as are present in the core, as due to local ubsidences and thought them of rare occurrence. But Potonié, as an outgrowth of broad observation, asserts that these cases are “H. Potonie, “Ueber Autochthonie von Carbonkohlen-Flétzen und des Senftenberger Braunkohlen-Flétzes, Jahrb. d. k. preuss. geolog. Landesanstalt fiir 1895, pp. 31, pl. 2. 95 96 STEVENSON—FORMATION OF COAL BEDS. [April 21. merely illustrations of the ordinary conditions. ‘“ The allochthonous formation of fossil humus beds is not the normal, as Ochsenius maintains, but autochthony is the normal, exactly as in the co sponding beds of the present day.” But this does not exclude con- tributions from other localities. He cites the abandoned ox-bow of the Mississippi, into which drift wood is thrown at high water, but which are filled eventually with autochonous peat in which the driftwood is enclosed. The existence of Stigmaria in interveni beds is a normal thing and to be expected, as appears from condi- — tions in cypress swamps of North America. Its existence in the coal itself is explained by autochthony, for, on that hypothesis, the old decaying vegetation becomes soil for the new. Indeed, the only difference between deposits of the several geological peri is in character of the vegetation, there is none in the mode | accumulation. : He finds a fossil swamp of the American type in the Miocen deposits of brown coal at Gr. Raschen near Senftenberg, which con tains, among other plants, Taxrodiwm distichum. The brown coal is 10 meters thick and shows several generations of forests, one above the other, the stumps remaining rooted in the brown coal. — Every feature of recent swamps is reproduced there except that the humus has become brown coal. Many of the stems are holl containing more or less of Schweelkohle. It is worthy of note tha an old peat bog exists on the clay overlying the brown coal, and tha’ -in the humose sand covering the peat, there are trunks of Pin silvestris: the conditions favoring accumulation of humus continued — there until diluvial time. The Schweelkohle is due to resinous — exudations from broken parts of the tree—the familiar process o closing wounds. Absence of stumps in no wise proves allochthonous formati ; If the fossil moor had borne only non-resinous dicotyledons, the Gr. Raschen condition could not have come about. The fact t# Stigmarie are often filled with sand is no evidence of allochthon for hollow alder stumps in West Prussia swamps, exposed to high water, are filled with sand even to the roots, so that they must be cleaned out before the axe is applied. 96 J STEVENSON—FORMATION OF COAL BEDS. 97 In a later paper,” Potonié says that having supported the cause autochthony, he must describe a deposit of allochthonous type. distinctions are simple; in plants of autochthonous origin, the 2 tender parts are preserved but they are practically wanting those of allochthonous origin. In connection with coal beds one 3 to do chiefly with autochthonous plants; but in Culm localities e has to do with “ Haecksel,” shreds of plants, which are ‘char- istic of allochthony. These fragments are at time large zh to show by their arrangement the direction of the transport- jag current. _Allochthonous deposits of carbonaceous material have few botanically recognizable plants; many stems and branches, often coal coated but with surface sculpture so obscured that determina- tion is impossible ; stems of the Knorria type are of frequent occur- ‘rence ; while Stigmaria is almost wholly absent, those which do occur _ being imperfect. Sub-surface organs can be carried away only after having been washed out from their place: other portions of plants must be the essential material of a transported mass. _ He presents the following contrasts. aS Autochthony. Allochthony. 1. Coal beds common. . Coal beds are. . Haecksel deposits absent or 2. Plant remains prevailingly insignificant. Haecksel. Determinable plants numer- 3. Few determinable plants; if ous, especially in roof. coal bed, Haecksel in the Few indeterminable casts. roof. . Knorria rare. 4. Indeterminate casts abundant. . Abundant Stigmaria in the 5. Knorria abundant. liegend. With their appen- 6. Stigmaria absent or rare; they dices. are without appendices. . Excellent preservation of Lael Potonié presented a brief systematic discussion of the whole _ “H. Potonié, “ Die Merkmale allochthoner palaeozoischer Pflanzen-Abla- angen,” Naturwiss. Wochenschrift, XIV., 1890, pp. 81, 82. PROC. AMER. PHIL. SOC., L, 198G, PRINTED APRIL 26, III. 97 98 STEVENSON—FORMATION OF COAL BEDS. _ [April2 subject in 1905;°* since that time, in successive editions, he has” widened the scope of his inquiries until, in the fifth, the presentation covers every phase with abundant illustration from German areas and references to those of other lands. Only certain portions of the work can be referred to in this place but, farther on, many citations will be made. He approaches the subject from the double stand- point of stratigraphy and palaeobotany. ‘ The coals and allied substances are termed Kaustobiolithe, be- cause they are combustible rocks of organic origin. He groups” them into— | ae Sapropel deposits, originally “stinking muds” composed of aquatic animals and plants, ' Humus deposits, derived from land plants. The former include the cannels, the oil shales and, as a derived product, petroleum; the latter include the ordinary brown and stone coals. The difference in origin of the two groups is evident from ag the physical composition shown by the microscope as well as by the chemical composition, the Sapropels yielding compounds of the para- ffine group while humus deposits yield compounds of the benzol | group. The Sapropels are formed in quiet, almost or wholly stag nant water and are of limited extent; whereas the humus deposits were formed as are the moors of to-day and are of vast extent. He illustrates the modes of origin by description of a great bog in northern Germany, which exhibits the passage from sapropel muds _ at its shore, to the Flachmoor, well wooded; thence by the Zwisch- enmoor, with changing type of trees, to the Hochmoor, hour-glass i form, which is treeless except alongside of rivulets. He comparé the conditions with those existing in sapropel and humus deposits o the older periods. The existence of both autochthonous and alloch thonous deposits is recognized, but he asserts that the former have been the prevailing type throughout and that, in every age, the latter have played an insignificant part. 4 . Potonié finds a strong argument for autochthony in the surpris ing resemblances, chemical and physical, existing between beds o * H. Potonié, “ Die Enstehung der Steinkohle,” Naturwiss. Wochenschrift, IV., 1905, pp. I-12; the latest edition is “ Die Enstehung der Steinkohle un der Kaustobiolithe ttberhaupt,” funfste Aufl., Berlin, 1910, pp. 225. 98 STEVENSON—FORMATION OF COAL BEDS. 99 brown and black coal on one side and the modern Flachmoor on the other. The laminated structure is frequently present in peat. The st extent of some beds of comparatively pure coal cannot be paralleled in recent autochthonous deposits, but the latter are of great extent in some regions, whereas no extensive areas of alloch- thonous carbon deposits are known to exist anywhere. He lays great stress upon the occurrence of sub-surface parts of fossil plants in the soils where they grew, the so-called petrified humus-soils. He ‘emphasizes especially the mode in which the Stigmaria rhizomas and their appendices penetrate the underclay of coal beds, spreading out and interlacing in such a manner that transport is inconceivable. hey must be in place. Equally conclusive are the modes in which _ roots of Calamariacee rhizomas occur in the clays. This almost universal underclay was the soil in which were rooted trees introduc- ing the moor-formation. _ The occurrence of forest beds in stone- and brown-coal forma- _ tion is not infrequent. He notes that at White Inch near Glasgow, _ Scotland, and that near Senftenberg. Sometimes the profile is shown in the roof; sometimes there are successive forests embedded as at Senftenberg, where erect stumps are associated with prostrate trunks. These are conditions familiar to students of modern ramps. The mode in which the Stigmarie have developed indi- _ cates, in some localities, even the prevailing direction of the wind at the time the trees grew. The growth of reeds in banks and the parallel arrangement of their roots are the same in Mesozoic, Ceno- zoic and recent deposits. Potonié carefully distinguishes the features of autochthonous de- sits as contrasted with those of allochthonous origin, elaborating ie discussion given in the paper just cited. He states that Stig- aria is not rare in the Commentry basin and that his search there for that plant was rewarded abundantly. He discovered a fine itochthonous stump, with spreading stigmarian rhizomas, still re- perine the delicate appendices, the whole occupying a space of 6 diameter. He found there also a fern tree, almost com- ¥ preserved and with a frond attached to the stem. He con- es that the condition must have been that of great quiet to ermit so nearly complete preservation. 99 100 STEVENSON—FORMATION OF COAL BEDS. Potonié’s description of secondary-allochthonous formations, to erosion and transport of materials from beds already ¢€ will find place in another connection, as will also his arg drawn from the testimony of the fossil plants, rep which authority is unquestioned. ds Ochsenius®” ‘published many noteworthy papers but that of 1 is especially important in this connection. The author recognizes force of the objection to allochthony—that, as running water ca organic and inorganic materials together, the deposit should indiscriminate mass of both kinds but his recent study of coal in the Lahn country has convinced him that phenomena observ the Frische Haff present a true explanation and destroy the force the objection. The history of the Frische Haff is comple 1510. 100 The Vistula, a stream laden with everything that can x di from a rich lowland province, gives off an arm at Peickel, the | which flows northeast to the long narrow Frische Haff, sey from the Gulf of Dantzig by the Frisch Nehrung or lowland, communicating with the gulf by the Pillauer Tiefs at its ae end. For convenience of discussion, he confines his attention t Nogat, ignoring the old Vistula and the rivers which enter from | east. Under the supposed conditions, the sea having contrelees Haff being filled with salt water, a marine bed is deposited floor. Such beds occur locally at the bottom and higher up in- series of coal deposits. Phase 1 of coal formation is brought abi through sanding up of Pillauer Tiefs by wave action, and the quent conversion of the Haff into a fresh-water basin by influ the Nogat. The débris brought down by that river, an indiscrim mass of organic and inorganic material, will be deposited on bottom. If now the sea cut a shallow passage through the lowla1 floating stems and twigs would form a “rake” at the head of | ~"C. Ochsenius, “Die Bildung der Kohlenflétze,” Verhana— Erste Halfte, pp. 224-230. The Frische Haff is a great sound on the border of the Gulf of Da about 60 miles long by 5 to 7 miles wide. It may be compared as to sup ficial area and position to Lake Ponchartrain on the Mississippi delta. 100 STEVENSON—FORMATION OF COAL BEDS. 101 passage. . If the water-surface of inflowing stream be lowered, a barricade” of wood would accumulate in curves and narrows of Nogat, to become compacted in time—a familiar phenomenon in ‘this day. This “barricade” would prevent passage of large wood and only fine material, ““Spulgut”’ would go over to be deposited as yer in the coal basin, t. e., the Haff. Brushwood would be caught the “rake” beyond. Thus a bituminous shale with plant impres- and paper-like laminz of bright coal would accumulate. Phase 2 comes with moderate height of the water. Now there _ goes with the “ Spulgut ” also the “ Sperrgut,” stems, rooted stumps, branches and the rest, which the Vistula pushes over into the Nogat; the “ rake’ does not permit its escape to the sea, it circles round the basin, finally sinks and forms pure coal. So much of the mud does not pass through the “ rake” will accumulate on the borders be mingled with the coal magma, as clay is in the globigerina ooze ; ‘it may form bands in coal beds. Repeated sinkings of waterlevel the feeding stream, the Nogat in this case, would give a clay shale e the floor as roof, the roof of the coal. No sand or gravel could ss the “barricade” but it would be heaped up there. Phase 3 comes with a high flood, which overthrows the “barricade”’ and es all into the coal basin. The sand and gravel form sandstones d conglomerates as roof of coal beds and formation of coal ceases. [he woody portions become at most only isolated stems buried in the ‘Rollgut”’; by repeated pressure they may perhaps be pushed into an oblique position. If the “rake” be torn away, the seawater again enters the basin and lays down a marine bed. _ This is the characteristic succession in coal bed formation. All lepends on the condition of the water-level. Changes in that cause tnation of clay, shale, coal and psammite, and effect the sharp hanical separation of those substances by the easily explained ‘mation of “ Rakes” and “ Barricades.’’ The elevation of import- mountain ranges in Carboniferous and Tertiary times afforded indant material for widespread lowlands, approximating sea-level. e advanced seaward and their luxuriant forest growth yielded tial for the stone-and brown coals. Networks of rivers must cut through the lowlands and must have deposited their loads 101 102 STEVENSON—FORMATION OF COAL BEDS. _ [April in huge depressions. It is clear that many channels fed the basin, but all worked after the same fashion. The process thro out was that which Ochsenius terms “ Barrenwirkungen” or bart cade action. ae The memoir discusses many details, rock fragments in co; stumps filled with sandstone, the occurrence of gypsum, the pre sence: of land shells, all of which are explained very readily by the yeu The conditions at Senftenberg, described by Potonié, are clearly to this barricade-action. ny The numerous coal beds of the Carboniferous were Neto quietly, but they are rarely more than 15 meters thick; whereas brown coal beds are comparatively few in number, show irregu deposit and at times attain a thickness of 50 meters. The explanatio: is simple. The soft plants of the Carboniferous had, at most, diameter of one meter and a height of 40 meters, so that they float easily in a few meters of water over the “barricade”; whereas t Tertiary and Quaternary giant trees had a diameter of 10 meters. a height of 170 meters, so that they needed a depth of, say, 15 me to float them over the “barricade.” Clearly a depth of one m would sink more quickly to some centimeters so as to permit on “Spulgut” to pass than would a depth of 15 meters—whence _ more frequent interruption of coal deposit in the Carboniferous the great constancy of formation in Neozoic time. Aug Almost all our mighty coal deposits are freshwater formnati 1s which came into existence through the factor of “ Barrenwirkunget Autochthony holds in their formation an exceedingly limited 2 in comparison with that of allochthony. Schmitz! has contributed a series of important papers to th literature of the subject. a In 1894, he regarded the in situ doctrine as merely a hye The presence of transported pebbles in the coal itself rather fave the doctrine that the coal is composed of transported materia ™ G. Schmitz, “ A propos des cailloux roulés du houiller,” Ann. Soc. Ge de Belgique, XXI., 1804, pp. Ixxi-lxxv; “La signification géogénique de: Stigmaria oan des couches d’houille,” Ann. Soc. Sient. de Bruxelles, X 1897, 6 pp.; “ Formation sur place de la houille,” Rev. des. Quest. Scientifique : Avril, 1906, 35 pp., 9 pl. 102 rgtt.] STEVENSON—FORMATION OF COAL BEDS. 103 These pebbles, he had discovered, are much more numerous than had been supposed. They are covered with a carbonaceous patine and are found mostly in the lower portion of the beds; he never saw any in the upper portions. The patine suggests that the pebbles may have made a long journey in fermenting pulp, and he thinks that their _ presence with this coating is confirmatory of Renault’s opinions re- _ specting the conditions of deposition of materials composing the coal. _ At the same time, the “ French theory” of the origin of coal, though probable for the ensemble of the coal formation, does not explain the underclay. As for Belgium, the special, the constant facies of the mur is evidence of formation in place. The convincing fact is the presence of Stigmaria in the mur with interlacing of the rootlets. Stigmaria remains in the roof are fragmentary. In 1897, Schmitz reviewed Potonié’s paper on Autochthonie; he recognizes that the mur is autochthonous but is not satisfied that that necessarily involves the conclusion that the coal itself is autochthon- ous also. A mur without coal is evidence of erosion, that its vege- table cover has been washed away. If there be coal without mur, it is allochthonous. A thick bed may be autochthonous below and allochthonous above. While recognizing the valency of many of the _ arguments presented by Potonié, he is not convinced that they are final. In 1906, he reviewed the whole subject. His own position in 1896 was that of uncertainty between the old doctrine of autochthony and the new forms of allochthony presented by Fayol and Grand’ Eury. Many phenomena observed in the Belgian basins seemed to support Fayol’s hypothesis, but the mur, with Stigmaria, clearly in loco natali, is a fact which cannot be ignored. Autochthony found its chief sup- port in conditions observed in recent swamps, but the knowledge of those was too imperfect to make the argument wholly satisfactory ; so that Schmitz, at that time, was inclined to hold an intermediate position and to think that both doctrines might be true. But Potonié’s later publication, based on the study of swamps in a great area, goes far toward removing objections. Schmitz sum- ‘marizes the processes described as occurring in the formation of “Die Enstehung der Steinkohle.” 103 104 STEVENSON—FORMATION OF COAL BEDS. [April 21. — Sapropel ; the gradations from peat to coal; and asserts that all point — toward autochthony. He antagonizes conclusions drawn from the presence of vegetable matter in material obtained by deep-sea dredg- ing in the Gulf of Mexico, for that is mingled with ooze and proves nothing for transport. He maintains that vegetable pulp cannot be transported far without notable loss and he urges that black waters from swamps soon lose their color through oxidation, as appears — from conditions in the Congo, Rio Negro and other rivers. De Lap-— parent has protested against the “ fascination of present causes” and Schmitz admits willingly that it is an error to seek in the present an ‘ : absolute representative of the past; but he asserts that it is equally an error to disregard the present in the study of the past. : Schmitz presents an elaborate argument. He traces the for-— mation of Sapropel in an arm of the sea, the encroachment of vege-_ tation, the formation of a bog covered by trees—the tourbiére boisée, the loss of moisture and the destruction of the forest, the for mation of the moss bog with Sphagnum, Scheuchzeria, ete.—the tourbiére bombée or hochmoor, which may continue to rise until it reach the heath stage—that of final decrepitude. He shows how this — normal development is often interrupted, that a newer stage may 4 return to an older stage or may originate without existence of pre- vious stages. a The wooded bogs are modern representatives of the Carbonifer ous type. They show conditions observed in the coal beds; peaty maceration disintegrates the most resistant plants so that one rarely recognizes the parts. The mode of growth in bog plants resembl that of the coal plants; the root is radial not tap. He describes an extensive bog in Hanover, in which the peat had been burned, leaving exposed great tree-trunks, the luxurious crown existing when the bog was wooded; if that bog had been covered with sediment during th life of those trees, there would have been a legion of autochthonous tree-trunks. . The immensity of the great coal areas, to be compared with the immensity of modern bogs, must not be disregarded. One canno think of the great Westphalian-Belgian-English basin as a mere lagoon to be filled by rivers; and Schmitz asks how vast must have 104 1918] STEVENSON—FORMATION OF COAL BEDS. 105 been the low country to yield humic material for the coal beds of that basin. He thinks that to accept the land conditions necessary would require too great draft on one’s credulity. But the case is wholly different with the peat bog theory. Schmitz concludes that the coal systems consist of allochthonous rocks and autochthonous coal beds. The underclay is not a special _ sediment; it is a sediment modified by the establishment of vegeta- tion. There must have been some allochthonous deposits of carbon- aceous matter, but they were merely local. The accumulation as a _ whole was autochthonous, after the manner of the forested swamps. Sterzel’® thinks that very probably no theory of formation is of universal application, the conditions being unlike in different regions, even in different parts of the same region. In studying the Zwickau region area, he became convinced that plants embedded in shales accompanying the coal are not in their original place, for they are _broken, they are in different stages of decomposition, their remains are mostly parallel to the stratification, and they show distinct evi- dence of sorting due to currents of water. Plants im situ occur only locally. Some features favor belief in the autochthony of coal; the narrow variations in thickness of important beds within great areas; the small proportion of ash in many beds; the localization of Stigmaria in the Liegenden; the occurrence of erect stems in the Hangenden. _ But there are others equally favoring allochthony ; the distinct lami- nation of the coal; the mineral matter, often forming a considerable _ part of the bed, is mostly clay, the same with that of the roof and floor, and it tells of quiet deposition ; Stigmaria occurs abundantly in the roof of coal beds ; erect stems are of exceptional occurrence. __ The greater number of phenomena favor allochthonous origin of the Zwickau coal beds. They were deposited in a lake basin sur- rounded by forested swamps. The gently inflowing waters carried little mineral matter and the plant material accumulated long time on the bottom, where it was converted slowly into coal. When the *$T. Sterzel, “ Palaeontologische Character der Steinkohlenformation und des Rothliegenden von Zwickau in den Erlauterung zur geologischen _ Specialkarte,” Section Zwickau, 1891, pp. 87-142; “ Mittheil. aus d. Naturw. - Sammlung d. Stadt-Chemnitz,” 1903, 22 pp. 105 106 STEVENSON—FORMATION OF COAL BEDS, [Aprilan. water-courses swelled, a great quantity of material, inorganic detri- tus, was brought down to form the intervening bed, on which, when quiet was restored, the plant material was deposited anew. Period-— ical changes, slight crustal movements, variation in fall of rivers, lead to deposit of a great mass of rock over the coal bed; the thick- ness of this intervening rock depending on the extent and contin- uance of those changes. When quiet returns, the forested swamp again expands. Many localities with particular species of plants” had been destroyed wholly and those forms do not reappear in later beds—an explanation of the irregular occurrence of plant forms in the series. Z The lake was comparatively deep, for the Zwickau measures are _ about 400 meters thick. By accepting this hypothesis of a lake, one finds explanation also of the origin of the great salt-content char- acterizing the Zwickau deposits—in 1854, 400,000 kilos of sodium > chloride and 15,000 kilos of calcium chloride were obtained from mine waters of the Tufen Planitzer beds. In 1903, Sterzel qualified a statement made on p. 90 of the pre- ceding paper, which refers to the value of erect stems as evidence. The only stems of that sort, observed by him, were “ Sargdeckel,” at the “coal-pipes”’ of English miners. One Sigillaria stump, exam- ined by him, was completely cut off at the base, with no trace of — Stigmaria. It had been torn from its place by running water, robbed 2 of basal branches and then deposited in the roof of the bed, where — ‘its softened bottom was flattened under pressure. He notes that — the overlying rock is sharply defined, that there is no passage of plants from the coal, such as would be the case if the place of plant- growth were flooded by masses of rock material. | Lemiere™* presented a memoir to the Geological Congress of 1900, which discussed the conversion of vegetable matter into coal. — In 1904, he returned to the subject and considered in addition the manner in which coal beds accumulated. The discussion is based largely on the assumption and conclusions of Fayol that the coal *T Lemiere, “Sur la transformation des végétaux en combustible fos- — siles,” C. R. Congrés Géol. Intern., Paris, 1901, pp. 500-520; “ Formation et recherches comparies des divers combustibles fossiles,” Bull. Soc. de VInd. — Min., 4™°. ser., 1V., V. Published separately, 1905. Citations from pp. 70-142. 106 grt] STEVENSON—FORMATION OF COAL BEDS. 107 beds were formed of transported vegetable matter deposited in basins of deep water. In this later memoir, he discusses the laws _ governing deposition of inorganic materials of varying density and shape, on lake bottoms in tranquil water, on beds of streams and on ___ shores exposed to the action of waves. This completed, he applies a the ascertained principles to explain the formation of coal deposits. ce The basins in which those deposits were laid down were ordi- -narily gaping faults, very long except where divided transversely by uplifted granite, and, in many cases, the fault is still apparent. Streams began to flow into the basins at once. Where the fault valley was divided transversely by uplifted granite, lake basins were formed like Commentry, Montvicq, etc., in which the beds are irreg- ular. At other times the fracture valley retained its length and was _ wide enough to be a strait or estuary, common to several rivers and ___ bordering on seas extensive enough to be affected by tides and waves. __ Respecting the latter he makes the frank remark: “It is hardly pos- sible to admit that the areas of coal deposit were in direct commu- nication with the high sea, because high-level floods are little com- patible with free access of this [the ocean] ; now, the floods are a condition, sine qua non, of vegetable contributions; it is necessary, q then, to admit that the areas of deposition were lagoons, sheltered _ from the ordinary tides, fronted by vast low plains, themselves _ above the tides and furnishing few coarse elements to the river : a =. load.” ‘ Other basins retaining their length, were less affected by marine _ conditions, possibly because of the narrowness or because of varia- tions in level. Ofsuch is the great syncline extending from Moulins to Decazeville. The deposits are lacustrian. The form of the _ depression affects the speed of currents and therefore the type of deposits ; if broad, the rivers from different points form deltas, but if narrow, the speed along the middle is such as to sweep away such deposits. The contrasting conditions are shown by the Saint-Etienne and Rive-de-Gier divisions of the Loire coal basin. The vegetable matter, to form coal beds, was brought in mostly during floods; some of it remained afloat; some was held in sus- pension; while some, which had undergone thorough maceration, 107 108 STEVENSON—FORMATION OF COAL BEDS, sank immediately. But all alike were deposited at last on the soaked — talus of the delta. The lake basin, in which the deposition was made, is conceived to have been quite deep, for Lemiére’s diagram shows curves to a depth of 350 meters and the last is still at consid-— erable distance from the bottom; it is supposed also to be large in comparison with the breadth of the tributary streams. The impor : tant source of plant material is the space along the streams between the average low water line and that reached by high floods; but is still higher portions of the drainage area, being exposed to rain ak wind, would contribute. ots During a long period of low water, little aside from inorganic matter would be carried to the basin; but when that was followed by a period of heavy rains, the forested area was invaded, the vege- table contributions were increased, while inorganic contributions. were decreased. The forest soil was covered with humus, which had been accumulating without cessation. The soil, thus ppl became increasingly unfavorable to vegetation, whose roots — — Eury says, hate to penetrate it deeply. Lemieére thinks at a “peremptory argument against formation sur place of coal beds — formed by aerial plants very different from those which have formed peat bogs. That the forest might continue and might renew itself after destruction, it was necessary that the soil be cleared away at — intervals by winds, rains and especially by floods.” | 3 The humus, already macerated and denser than living plants, was swept off first; afterwards, the living plants would be uprooted and broken. The macerated humus, being denser, was deposited on the convex surfaces of the delta, while the living plants had to become watersoaked before sinking, so that they were superimposed upon the other plant material. They would come to rest more abundantly in the bays between deltas, so that one should find more of volatile matters in coal laid down within the bays than in that deposited on the delta slopes, along the axes of the currents. The volatile should increase as one departs from those axes but it should — decrease with the depth at which the vegetable matter was deposited. Floating islands are possible, since a flood might tear off bodily 3 part of a forest, which, carried down, might float for a while and 108 1g11.] STEVENSON—FORMATION OF COAL BEDS. 109 then sink to give the appearance of growth im situ. If the storm continue long enough, it would wash off the soil itself, which would become an intercalation in the bed. If the flood return, attaining a higher stage than before, another area of forest region would be torn off to form a new bench of coal, possibly directly on the other. When the flood subsides, the superficial currents would find only _ inorganic materials on which to act, and the first deposit would be mud to form the roof of the coal bed, after which would follow some sandstone and conglomerate. Between floods the vegetation is restored and the area is increased by encroachment on the lake. During this long interval, the flora might be changed. _ Lemiére is convinced that, by his hypothesis, he has succeeded in explaining converging beds, parallel formations and floating islets. All are allochthonous; aerial plants have formed no autochthonous beds, for no erect stem has been found in the coal; in fact, the plants could not thrive in a humus not nitrefied. Peat cannot become coal, as its tannic acid checks the process of conversion. He applies his doctrine with great ingenuity to several basins in France and - finds it confirmed in all. Lemiére?” has published several papers in more recent years and he presented a résumé of his opinions in 1910. In that he expresses _ surprise that in recent congresses the dominant opinion was that coal beds are ancient wooded-bogs buried by successive subsidences, because this opinion involves the supposition that the coal beds were not formed in the same way as the sterile beds which enclose and, at times, penetrate them. This opinion is based upon palaeobotan- ical evidence, which is often untrustworthy, providing two-edged weapons, available equally for defenders of each theory. It is nec- essary to discover some criterion which will be conclusive. In an earlier paper, he had demonstrated finally in geometric form that the ‘peat bog theory leads to arrangement of beds unknown in nature. In this he proposes to restudy the conditions after the same method, avoiding palzontological discoveries, and availing himself of dis- coveries which have the character of certitude. He describes three types of structure observed in areas of the coal formation. **©L. Lemiere, “Résumé des théories sur la formation de la houille,” Bull. C. R. mensuels. Soc. Ind. Min., Sept., 1910, separate, 19 pp. 109 110 STEVENSON—FORMATION OF COAL BEDS. [April as. The first type is that of the Hainaut coal basin in Belgium, a small area, 15 kilometers wide and separated from the Campine basin at the northwest by 60 kilometers of older rocks, while on the south- east it is bounded by a fault. The fossils show that, at times, this basin communicated with the sea. The deposits are thin at the north, where the beds have remained unaffected by subsequent dis- turbance; but they thicken to 3 kilometers toward the southerly — border of the basin, where the disturbance increases as the fault is _ approached, the downthrow having caused close folding. The hinge | é of movement was near the southeast bounding fault. If the peat bogs were formed at the unvarying sea-level, the first of them should have had, when the basin was filled, an inclination of 25 cm. per meter and the last should be almost at sea-level, while the inter- — mediary beds should converge toward the shore line at the north- — west. The conditions being absent, it is evident from this mathe- matical demonstration that the coal beds are not buried peat bogs. The warning against the dangers of dependence on palaeontology is repeated, and the necessity for the warning is proved by the dis- covery of the Bernissart iguanodons in rocks other than those to which the animals belonged, as well as by the possibility that some day remains of fossil man may be discovered under a landslide from — a chalk cliff. : The second illustration is that of an area, increasing in extent as it deepens. There, convergence of the beds toward the hinge — of movement would not be a criterion. The upper beds should be @ of greater extent than the lower. This is to explain conditions . existing in the Appalachian basin, where one thick coal bed, the Pittsburgh, has an area equivalent to not less than 400 kilometers square. It is difficult to understand how materials from the anti- — clinal borders could reach the central parts of such a synclinal to a give parallel beds there. In the central parts of the basin are great : masses of red shale and beds of limestone and the coal beds are not — rigorously parallel. He is inclined to think that the materials within — the central parts are due to precipitation (from solution) without — mechanical transportation from the borders. One cannot assert " 110 . pnt] STEVENSON—FORMATION OF COAL BEDS. 111 _isoclinal formation in an isoclinal valley, bounded on one side by a fault. The reference to this area is brief. Lemiére states that the phenomena of the faisceaux at the north and the dips in the basin suggest, a priori, that here one has a case of peat bog formation. But he plots the conditions in a diagram and states that, as shown thus, they are evidently due to influence of the fault. He concludes that the French coals as well as those of the _ Franco-Belgian basin are not old peat bogs but are of alluvial origin ~ and that the same conclusion is probable for the coal beds of North _America. These conclusions do not proscribe the theory of peat _ bogs; on the contrary they appropriate those conditions and their results. All that is insisted on is that, at present, we can find no e trace of successive deepenings of feeble amplitude and repeated for _ each bed ; but there are evidences of many subsidences, important or _at distant intervals, corresponding to the faisceaux of beds. ___Lemiére, feeling himself no longer in danger of being paralyzed 3 by the question, Is coal formed im situ or as alluvium?, proceeds to _ show wherein his doctrine differs from other forms of the transport theory. As the distinction depends in great measure on his con- E ception of the mode in which vegetable matter was converted into coal, the details have no place here. This extended reference to Lemiére’s publications is justified by the fact that he has presented the characteristic of the transport theory more fully than most of his predecessors and has attempted _ to explain all the conditions as far as they are known to him. Stainier,?°* whose numerous contributions will find consideration in another connection, believes that formation of coal beds is essen- E tially a geological problem and he maintains that geologists have ~ been negligent in that they have left the discussion too long to the palzobotanists. Fayol and Grand’ Eury, by studying the matter “The diagram, illustrating the structure in this second case, shows a _ bounding fault on one side, such as limits the little basins in France. ™ X. Stainier, “De la formation des gisements houillers,” Bull. Soc. Belge de Géol., XX., 1906, p. V., pp. 112-114. 111 112 STEVENSON—FORMATION OF COAL BEDS. [April 2 geologically, have succeeded in solving the problem for the bas of central France. He hopes by following their methods to solve the problem in the great basins of northwestern Europe. an If one study not the coal beds alone but also the whole ser of deposits in those coal basins, he finds that their strata differ ir no wise from those of terranes, whose marine origin is recognized — by all. No feature of coal beds suggests a different origin for them. On the contrary, when one endeavors to explain the formation of — coal beds by the im situ doctrine, he find himself, at each step, con- tradicting the best established laws of geology. These contradic tions, naturally not apparent to the botanists, ought long ago to have spurred geologists to make investigations for themselves. The have led Stainier to believe that coal beds, like the encasing rocks are of purely sedimentary origin. ae For him, the coal plants grew on continents, bordering great depressions, into which meteoric agencies carried the vegetable - débris along with materials torn from the land by erosion. These Z materials, vegetable and inorganic, were mingled intimately while the water was in agitation; but in proportion as the condition o calm was re-established, they were thrown down to the bottom in a — well defined order, determined by density of the materials. In cases where the succession is complete, there was formed, first, a bed of Es sand, ultimately becoming a bed of sandstone; then a peculiar irregular rock, which constitutes the mur and contains the denser parts of the vegetables, i. ¢., the sub-surface organs; then the remain- ing portion of the vegetable débris was deposited to form a coal bed; and finally, the impalpable elements, fine clays, reached the bottom : giving tender fine shales, the roof of the coal bed. The reasoning on which the conclusions are based is to be given in a memoir not yet published. Ashley?®® has offered suggestions which are not without interest here. Adopting the doctrine of autochthony, he ignores in his cal- culations the cannels as well as other merely local deposits, which are allochthonous and therefore outside of the discussion. He finds — 8G. H. Ashley, “ Maximum Deposition of Coal in the Appalachian Coal Field,” Econ. Geology, I., 1906, pp. 788-793; II., pp. 34-47; “ Significant Time Breaks in Coal Deposition,” Science, N. S., XXX., 1900, p. 129. 9 112 STEVENSON—FORMATION OF COAL BEDS. 113 that under exceedingly favorable conditions a peat bog has gained e foot of thickness in five years but that in one case this increase appeared to be only one foot in two hundred years. With the con- ditions normal, the rate of increase seems to be not far from one foot in ten years. Reasoning from the approximately ascertained ratio of volume of peat and the resulting coal, he conceives that 300 years would be required for the formation of one foot of coal, thus giving a period of about 4,000 years for accumulation of the Pitts- burgh coal bed in western Maryland. The minimum period to be assigned for formation of the 300 feet of coal in the Appalachian basin is not far from 100,000 years. In his later paper, seeking to ascertain whether or not a coal bed may be utilized as a time measure, he indicates some complexities of the problem, one of which is important. A coal bed, 18 inches thick at one locality may be 15 feet at another, the latter thickness requiring for accumulation 4,000 years more than the other. As the rocks accompanying the thinner bed show no compensating differ- es, the 18 inches is all that was formed while the 15 feet was accumulating elsewhere. There was either slow growth or a time- break, that is a period of no deposition, before or after deposition of the thin bed. “Smooth-partings ” are evidences of time-breaks and represent locally nonconformity between the under- and the overlying beds: a “ smooth-parting” at one place may be equivalent to 40 feet of shale at another; an inch or two of cannel may have similar equivalence. Slow growth and temporary cessation of deposition are important elements of the problem. Dannenberg’” finds strong arguments in favor of autochthonous rmation in the vast extent of some coal areas, the presence of the tenderest plant-parts in coal inclusions, the abundant occurrence of roots directly under the coal, and the identity of coal-forming ant species with those found in the enclosing shale rocks. Not all localities show these features with equal clearness, for in some cases ere are variations along dip and strike like those in delta deposits, ™ A. Dannenberg, “Geologie der Steinkohlenlager, Berlin, 1909; Erster Teil, 197 pp. The citations are from pp. 18-27. PROC. AMER. PHIL. SOC., L. I98H, PRINTED APRIL 25, IQII. 113 114 STEVENSON—FORMATION OF COAL BEDS. _ [Aprilar such as appear in the basins of central France, which Fayol has proved to be allochthonous. if The deposits must have been made in shallow water; Grand’ Eury has shown that the autochthonous flora of the Loire basin a - could not have grown in water more than 10 to 15 meters deep. There must have been a special combination of circumstances, since the deposits, in spite of the shallowness of the water, have in some basins a thickness of some thousands of meters. This can be under- stood if one accept a constant though variable subsidence through- — out the period of deposition. A certain instability of coast line in paralic basins is proved by repeated inroads of the sea. If the sedi- _ ments be laid down less rapidly than the-surface sinks, marine con- ditions prevail. Periods of rest, possibly of some elevation, would be favorable to development of swamp vegetation, which, when sub- sidence began again, would be buried under muddy and sandy depos- _ its, until a new swampy area was produced, on which vegetation began de novo. These movements can be followed with great clear- ness in the Saarbruck and Loire basins. Similar movements in the period of man can be recognized alo many coasts. Dannenberg regards the Tertiary and Quatern history of the Netherlands as especially instructive. This he gives in detail, showing that there have been successive advances and — retreats of the shore line, so that the section of Tertiary and Quater- nary beds consists of sandstones, conglomerates, shales, marine beds and peat deposits, wholly similar to the succession observed in the Coal Measures. The filled river valleys observed in the Coal Meas- ures, have their counterparts in these newer deposits. And it must not be forgotten that, in the Carboniferous time, great orogen movements occurred, so that there was abundant material for filling the basins. Stevenson," after studying the area, found himself unable to accept Fayol’s conclusions respecting the mode in which the coal beds were formed in the basin of Commentry. He agreed fully with Fayol as to the process by which the inorganic deposits were laid “° J. J. Stevenson, “ The Coal Basin of Commentry in Central France? Ann. N. Y. Acad. Sci., X1X., 1910, pp. 161-204, 6 pl.; “ The Coal Basin of Decazeville, France,” the same, XX., 1911, pp. 243-204, 2 pl. 114 1911.] STEVENSON—FORMATION OF COAL BEDS. 115 down, seeing there the conditions of delta formation as long recog- _ nized by geologists in American coal fields ; but he could discover no reason for supposing that the coal beds were formed of plant ma- terials washed in from the drainage area. That hypothesis, as pre- sented for this region, seems to be self-contradictory. The supposed surface conditions at the beginning of the history were such that dense vegetable cover seems in the last degree improbable; but the vegetation required by the hypothesis was so dense, that it would have been its own protection against any but a long-continued series of the most terrific cloud-bursts; in case of such a débacle, only a small part of the vegetable matter could be deposited as a coal bed, for the trees, supposed to have composed one half of the whole vegetation, would be loaded by material around their roots, would be snags in the mass of detritus and would be buried in the sands; even the twigs and underbrush would be entangled in the mass, for there could be no sorting action in the short course of the little tor- rent and all would be dropped when the flood’s velocity was checked on the comparatively broad delta surface, supposed to exist when formation of the Grande Couche began. Only the finest material, mineral, or vegetable, could find its way to the bottom of the basin— yet it is certain that trees make up a very considerable part of the Grande Couche. The objections presented by this writer will be con- sidered in another connection. He thinks that the structure of the x ‘Grande Couche shows that its vegetation accumulated in situ and that there is no evidence to favor the suggestion that Lake Com- mentry was a deep water basin at the time when coal accumulation Study of the Decazeville basin led him to similar conclusions respecting that area. The conditions there are very different from those in the Commentry basin, so different that any doctrine of transport formulated to account for the conditions at Commentry could not be applicable at Decazeville. _ Study of investigations by v. Giimbel and Potonié led Gothan** _ to study the coal area near Fiinfkirchen. The economic importance the Liassic coals within that area had been known for more than ™W. Gothan, “Untersuchungen iiber die Entstehung der Lias-Stein- kohlenflétze bei Fiinfkirchen (Pecs, Ungarn),” Sitzungsber. d. k. preus. kad., VIII., 1910, pp. 129-143. 115 116 STEVENSON—FORMATION OF COAL BEDS. [April ar 100 years and the relations of the beds had been described by several geologists ; but nothing was known which showed the mode in which the coals had accumulated. The section contains about 100 coal beds, of which fully 25 attain workable thickness in much of the area. Gothan had already discovered underclays with roots asso- ciated with Mesozoic coal beds on the Yorkshire coast of England, — and it seemed probable that search for similar clays at Fiinfkirchen — would be successful. He was not disappointed, though he found the difficulties in the - way of study greater than anticipated. Under the coal bed, no. 7, there is a well-marked underclay with irregular branching coaly markings, varying in diameter and in every — respect resembling roots; and, at one locality, a rhizoma with its — rootlets was complete, enabling him to determine the relations of the other forms. “Through such horizontal rhizomas, the analogy of this Mesozoic underclay with the Carboniferous Stigmaria-beds and the recent or sub-recent reed-beds is the more marked.” A four-inch layer of carbonaceous shale lies between the underclay and the coal, but one cannot trace the roots in it; they cannot be dis- tinguished in the dark material, which is so crossed by cleavage planes that none but irregular angular fragments can be obtained. The planes do not coincide with the direction of the rootlets. Roots are seldom observed in the freshly exposed rock within the mines, but they are distinct enough where the rock is somewhat weathered. Gothan exposed the outcrop for several meters at dif- ferent horizons and in the course of a day’s excursion, he found well-marked underclays, with roots, associated with 8 coal bed The analogy with Tertiary and Quaternary underclays is complet His conclusions are that the underclay, associated in more than a dozen instances, with the Fiinfkirchen coal beds, shows that th are, for the most part, of autochthonous origin, as are, predomi- nantly, the younger and older humus deposits of the present time as well as those of the Tertiary and Palzzoic. The failure to secut proof of this origin for all the Fiinfkirchen beds is due merely ' the unfavorable conditions to which reference has been made. a footnote he notes his discovery of typical underclay, with root just. below a Wealden coal bed in a neighboring district. 116 THE TRANSPIRATION OF AIR THROUGH A PARTI- TION OF WATER. By C. BARUS. (Read April 21, 1911.) 1. Molecular Transpiration of a Gas—Ever since 1895 I have observed that the Cartesian diver, used in my lectures, grew regularly eavier from year to year. The possibility of such an occurrence is at hand; for the imprisoned air is under a slight pressure-excess as compared with the external atmospheric air. But this pressure gradient is apparently so insignificant as compared with the long column of water through which the flow must take place, that oppor- tunities of obtaining quantitative evidence in favor of such trans- piration seem remote. If, however, this evidence is here actually forthcoming, then the experiment is of unusual interest, as it will probably indicate the nature of the passage of a gas, molecularly, through the intermolecular pores of a liquid. It should be possible for instance to obtain comparisons between the dimensions of the molecules transferred and the channels of transfer involved. _ 2. Apparatus—Hence on February 27, 1890, I made a series of definite experiments’ sufficiently sensitive that in the lapse of years one might expect to obtain an issue. The swimmer was a small light balloon-shaped glass vessel, vd, Fig. 1, unfortunately with a very narrow mouth, 2 mm. in diameter, at d, in the long column of water A. The small opening however gave assurance that the air would not be accidentally spilled in the intervening years. For this Teason it was temporarily retained, the purpose being that of getting safe estimate of the conditions under which flow takes place. In Fig. 1 ab is a rubber hose filled with water, terminating in the receiver R. Here the lower level of water may be read off. More- over R is provided with an open hose C, through which pressure or suction may be applied by the mouth, for the purpose of raising or _ * Am. Journ. of Sci., 1X., 1900, pp. 397-400. 117 118 BARUS—THE TRANSPIRATION OF AIR [April lowering the swimmer, vd, in the column A. In this way cons of temperature is secured throughout the column. oe 3. Barometer.—The apparatus is obviously useful for o ordi barometric purposes, and provided the temperature, ¢, of the. -—~a ‘a es | | | ha 2 | | ee | | 4 2 A Fic. 1. Swimmer and appurtenances. changes of temperature, so that slow manipulation is es These and other precautions were pointed out in the nein (1. ¢.) Ld 4. Equations. Manipulation—Let h be the difference of af of the imprisoned water and the free surface in the reservoir Then it follows easily that p Rm a h+ fm nas - Pp, gM (1 + m/M) — p,/p, where H is the corrected height of the barometer (from which mercury head equivalent to the vapor pressure of water is to r91t.] THROUGH A PARTITION OF ‘WATER. 119 4 deducted), pm, pw; pg, the densities of mercury (0° C.), water (f° C.), and glass, respectively, m the mass of the imprisoned air at v, R its gas constant, and r==t + 273° its absolute temperature. M is the mass of the glass of the swimmer and g the acceleration of gravity. _ The equilibrium position of the swimmer is unstable. To find it, R may be raised and lowered for a fixed level of the swimmer ; or R may be clamped and the proper level of the swimmer determined by suction and release at C. The dropping of the swimmer throughout Fic. 2. Cylindrical Swimmer. the column of water may occasion adiabatic change of temperature of .23°. It was my practice to use the latter method and to indicate the equilibrium position of the swimmer by an elastic steel ring, circling A. In this way the correct level may be found to about mm., and afterwards read off on the cathetometer. _ After making the observations, the hose ab is to be separated at a, so that the swimmer falls to a support some distance above the bottom, admitting of free passage for diffusion. Clearly this dif- fusion is due to the difference of level, h”, between the water level in wv and at the free surface of the liquid, f (see Fig. 2). Increase of rometric pressure has no differential effect. A large head h” how- ever means a longer column for diffusion. 5. Data.—tIn the following summary a few of the data made in 1900 are inserted, chosen at random. 120 BARUS—THE TRANSPIRATION OF AIR [April 2 In the intermediate time, I did not return to the measurements — until quite recently (January, 1911), when a second series of obser- vations was made. As much as one fourth of the air contained in 1900 had now, however, escaped, in consequence of which the above method had to be modified and all heads measured im terms of mercury. Hence if H denotes the height of the barometer diminished by the head equivalent to the vapor pressure of water, and if m/M_ be neglected in comparison with 1 (about .06 per cent.) the equation becomes Ss =P Evita 1 /Py) @ in which the first factor of the right-hand member is constant. If the observations are made at the instant the swimmer sinks from the free surface in A, Fig. 2, H must be increased by the mercury equivalent of the height h’ of v. The table contains all the data reduced to mercury heads. A—Mgpm/R. Consequently 1,842 x 10° grams of the imprisoned air escaped in the intervening 10.92. years; i. e., .265 of the original mass of air. In other words 168.7 X 10°° grams per year, .462 X 10° grams per day, or 5.35 x : 10°?” grams of dry air per second. | 6. Conditions of Flow.—It is now necessary to analyze the above : experiment preparatory to the computation of constants. The mouth — of the swimmer had an area of but .0314 cm.? When sunk the head of water above the surface v was h” ==24 cm. The column of water between v and d was h’”==8 cm. Hence the length of column within which transpiration took place was 24 +2 8== 40 cm. The right section of this column is taken as .0314 cm.” through- out. Naturally such an assumption, accepted in the absence of a better one, is somewhat precarious; but it may be admitted, inas- much as the pressure of the gas sinks in the same proportion in which the breadth of the channel enlarges. Thus there must be at least an approximate compensation. In more definite experiments a cylindrical swimmer whose internal area is the same as the annular area without will obviate this difficulty (see Fig. 2). The pressure difference urging the flow of air from v is 1gtt.] THROUGH A PARTITION OF WATER. 121 Ap = 24 X .997 X 981 = 23470 dynes/cm.? Hence per dyne/cm.? per sec. 10°” x 5.346 = 10x 2.28 IO X 2.347 grams of air escape from the swimmer. | A few comparisons with a case of viscous flow may here be inter- esting. Using Poiseuille’s law in the form given by O. E. Meyer and Schumann’s data for the viscosity of air, it would follow that but .194 X 10-* cm.? of the .0314 cm.” of right section at d is open _ to intermolecular transpiration. The assumption of capillary trans- _ piration is of course unwarrantable and the comparison is made | merely to show that relatively enormous resistances are in question. Again the coefficient of viscosity n i 2.7 1+ 4€/r~ m 16 Re” —?) _ may be determined directly. In this equation m is the number of grams of air transpiring in ¢ seconds through the section zr* and in virtue of the pressure gradient (P—p)/1, when 7 is the viscosity and é the slip of the gas. Hence the value »/(1 + 4£/r) —4.8 X 10° would have to obtain, a resistance, which would still be enormously large relative to the viscosity of air (y==180 X Io), even if the part of the section of the channel which is open to capillary tran- _ spiration is a very small fraction. 7. The Coefficients of Transpiration—To compute the constants under which flow takes place the concentration gradient dc/dl ‘may be replaced either by a density gradient dp/dl or a pressure _ gradient dp/dl. If the coefficients in question be k, and kp respec- tively _ where a is the section taken equal to the area of the mouth of the _ swimmer, R is the absolute gas constant, r the absolute temperature of the gas, and m the loss of imprisoned air in grams per second. _ If v= mRr/p is the corresponding loss of volume at + and p 122 BARUS—THE TRANSPIRATION OF AIR [April 21, i k, pi : % #, Ra ™ aRe apf . oy If in equation (3) the ay value of m is inserted and t denotes cur _ rent time, or m==m/t; i : MPS dl” hk" + 2h” where pw is the density of water, h” and h’” the difference of level (see Fig. 2) of the surface in v below the free surface in A and above the mouth at d, the relations are es m 21+ 2h" [h" A ae (le he), = kr. mates (5) a The acceleration of gravity, g, has dropped from both equations; &, is independent of Rr. The coefficient kp, however, is more per- spicuous. a If h’” is made very small in comparison with h” (care being taken to avoid loss of air during manipulation) h’” will also vanish; or for h’ =o . by = ee (1h — 1/0), “6 and similarly for h” =o » i I t - hh" m= k apg. Thus the apparatus is most sensitive if a is as large as possible and h'”/h"” as small as possible and the length of the column in 4 is eventually without influence on the result. Hence if for a cylindrical swimmer the internal right section is equal to the area of the annular space between the outer wall of the swimmer and the inner wall of the vessel A, if the column of water above the swimmer is removed during the prolonged intervals of time between observations, the : Torr.) — THROUGH A PARTITION OF WATER. 123 section a through which capillary transpiration takes place is defi- —nitely given. It is obvious that the swimmer must be suspended, for _ instance by fine cross wires, above the bottom of the tank A. Reference is finally to be made to convection and to temperature. The manipulation during observation necessarily stirs up the water and distorts the regular pressure gradient. Hence observations are to be made rarely. Again to obviate convection in general the vessel must be kept in a room of nearly constant temperature. a 8. Values of the Coefficients —lf the data of the above summary ___ be inserted in the equations for kp and Rep, c mRr 5.35 x 107” x 2.87 x 10° x 298 é k, = wie adp|dl 10314 x 23470/40 == 250 xX 10°, , = &, | Rr =.29 x 10-™. Hence for a gradient of 1 dyne per cm., 2.9 X I0°** grams of air _ flow between opposed faces of a cu. cm. of water per second. This may be put roughly as about 2.4 * 10°*° cu. cm. of air per second. _ The speed of migration of individual air molecules intermolecularly _ through a wall of water is thus 2.4 X 10-?° cm./sec. for a dyne/cm. gradient. Since the gradient is the energy expended when the cu. cm. is transferred 1 cm. along the channel and if the number of air mole- cules per cu. cm. be taken as N60 X 1078, the force acting per molecule to give it the velocity just specified is 1/(60 X 10**) dynes. Hence the force or drag per molecule if its speed is to be 1 cm. per I I I - 24810 "760% 10" 144% 1 I s {6.9 107” dynes, if v= cm. /sec. oF dynes __. This may be compared with the force necessary to move a small _ sphere through a very viscous liquid of viscosity ». This force is f =Omrv. If v1 cm./sec., 27—10-° X 2 cm. the diameter of the sphere of influence of a molecule, and f—6.9 X 107 dynes the value just found, 124 BARUS—THE TRANSPIRATION OF AIR. [April 21, 6.9 x 107"! 1 64x 107° - ; In other words the molecule moves through a liquid about twice as viscous as the air itself. 9. Conclusion—The above ‘data are subject to the different bya eses stated; but it has been shown that the results may be obtained by the method described free from ulterior assumption. It seems_ to me that detailed investigations of the above kind carried on with — reference to both the chemical and the physical properties of the — liquid, i. e., with different liquids and different gases at different - temperatures and pressures, cannot but lead to results of importance : bearing on the molecular physics involved. Hence experiments of — this kind have been begun in this laboratory and I hope to report the — results from time to time. Obviously in a doubly closed water ma- nometer (U-tube) the unequal heads of the two columns of liquid — must in a way similar to the above vanish in the lapse of time. This — method seems particularly well adapted to obviate convection. Finally hydrogen shows a measurable amount of molecular tran-_ spiration in the daily march of results already obtained, and with this gas a new and direct method for obtaining the molecular diameter is foreshadowed. = 366 x 10~%, Brown UNIVERSITY, ProvIDENCE, R. I. ELLIPTIC INTERFERENCE WITH REFLECTING GRATING. By C. BARUS. (Read April 21, 1911.) 1. First Method—There are two or three typical cases in the use of reflecting gratings for the production of interferences in the spectrum, each of which shows peculiarly interesting features. _ The of these is given in Fig. 1 and corresponds closely to the method described for transmission gratings in a preceding paper. If L is the source of light and M a glass plate grating, it was shown that Fic. 1. Diagram, showing positions of mirror, M, and grating, G. plane mirrors in the positions G» and Gn, each reflecting a spectrum from M, produce elliptical interference whenever the rays returned after passing M by transmission and reflection, respectively, are made to overlap in the spectrum, under suitable conditions. _ The present method is the converse of this, since the gratings and the opaque mirrors change places. Parallel rays from L strike e plate of glass M and the component rays reach identical reflecting 125 126 BARUS—ELLIPTIC INTERFERENCE [April 21 gratings Gm and Gn, placed symmetrically with respect to M at an — angle 1 to the E and L directions. The undeviated rays pass off — eccentrically at R and are not seen in the telescope at E. They may, Z however, be seen in an auxiliary telescope pointed in the line R and they then facilitate the adjustments. Rays diffracted at the angle 2i, however, are respectively transmitted and reflected by M and ~ interfere in the telescope in the line E. Similarly rays diffracted at | an angle 6’ >1 interfere in the line D. ; To make the adjustment it is sufficient to bring the Fraunhofer lines in the two spectra seen at E into complete coincidence, hori- zontally and vertically. Coincidence of slit images at R (at least — vertically) aids in the same result. It is also necessary that the — rulings on G,, and G, and the slit should be parallel, or that the images of slit and spectra shall lie between the same horizontals. One of the gratings, Gn, may now be moved parallel to itself by the — micrometer screw until the elliptic interferences appear. If the © plate M is not half silvered there are three groups of these as — described in the preceding paper. Each group passes from the initial _ degree of extreme fineness, through maximum size, to a final degree, — for a play of the screw of about 1 mm. There is the usual radial — motion of the fringes, together with the drift through the spectrum ~ as a whole. To bring out the set of solitary ellipses, the silvered — surface of M should be towards the light and remote from the eye. a As a rule the adjustment is difficult, as an extra condition is imposed i in the parallelism of the slit and the rulings of the gratings. The a ellipses are liable to be coarse with their axes oblique, clearer in some parts of the spectrum than in others, unless means are provided for : placing the rulings accurately parallel. Even when well adjusted « they are rather polygonal than rounded in their contours. They are about as strong with non-silvered glass M as with half-silvered glass ; but in view of the multiple spectra the adjustment is much more difficult in the former case. It has been suggested that the white slit images must appear eccentrically in the direction R. Hence if a special telescope is — directed in this line, the final adjustment is reached on coincidence — of the proper slit images, provided the rulings of the gratings and the slit are parallel. torr.) WITH REFLECTING GRATING. 127 For & >i the second series of interference spectra occurring at D, eccentrically, are broader, but only on perfect adjustment do they occur simultaneously with the other set. In fact, since for the pre- ceding case 1= 0, or 2sint=A/D and in the present case, aa sin & —sini—2A/D, _ therefore 3 sin # = 3 sint==3 sin 6. i There is also an available set in the second order to the left of E. In the gratings used above D lies in front of ee being nearer the E than the L direction. ‘ 2. Inversion of the Method.—The occurrence of the caleviated _ tay R suggests another method. For if the white ray R is reversed, i. e., comes from an eccentric collimator, slit images will be seen in telescopes at L and E, whereas overlapping spectra will appear in _the direction D’ eccentrically and in the lines R and R’. One of the _ latter may be lost in the collimator. The former occurs for the _ Same angle 6’ so that t sin & = 3 sini. Moreover, if 45° is the angle of incidence of L upon M when sodium light is taken, so that 6 —26° 14’, i—8° 28’, the R, D, D' rays make angles 21, # +-1, 6’ —1, respectively, with the E direction; _ or the sum of the angles at D and D’ with the E line is 26’, their | é difference 2i, and the rays D, R, D’ intersect at a common centre on Gm. Hence if we place the plane of G» at the centre of the spherom- eter and arrange M and Gy, eccentrically, the angles may be meas- ured as before. __-3. Resolution of the Slit Image—lf the sharp white images of _ the slit in a Michelson apparatus for the case in which the incident ; light consists of parallel white rays from a collimator, be accurately superimposed and the opaque mirrors be set at the proper distances _ from the semi-transparent mirror by the micrometer, the slit image _ may itself be viewed through a grating and will then show elliptic ‘interferences in all the spectra. The apparatus is here eccentric, 128 BARUS—ELLIPTIC INTERFERENCE while the grating (either transmitting or reflecting) must be at the center of the spectrometer, if angles are to be measured. The same is true for any of the other superimposed white slit images in the above or the earlier experiments and may even be repeated with successive transmitting gratings. It is interesting to note that the position of the center of ellipses is at the same wave length in all the spectra though the form of ellipses may differ enormously. The same phenomenon may thus be seen by a number of observers at the same time, each looking through his own telescope. Fic. 2. Diagrams showing position of gratings, g, g’. 4. Third Method. Parallel Gratings—In this case the two halves of the grating are treated displaced parallel to themselves, from their original coplanar position in the grating, from whi ' they are cut. Their mounting is thus something like the case of t two black plates of Fresnel’s mirror apparatus, one of the plates being adapted for displacement parallel to itself. In Fig. 2 g and g’ show the two halves of the grating cut alon the plane S, normal to the plates and parallel to the rulings. The half g’ is provided with a micrometer screw, so that it may be su cessively moved from the position g’ in Fig. 2 to the position g’ in Fig. 3, through all intermediate positions, while the half g remains stationary. Each of the halves g and g’ is controlled by three a justment screws (horizontal and vertical axes of rotation), to sec complete parallelism of the faces of the grating. Each, moreover, WITH REFLECTING GRATING. 129 ay be rotated around a horizontal axis to place the lines parallel the slit of the collimator. The duplex grating is mounted on a etrometer as is usual for reflection. Finally each half may be >d and lowered and moved horizontally to and fro, parallel to f, so that the half gratings when coplanar may approximately oduce the original grating. _ After each of the spectra are clear as to Fraunhofer lines, the iterferences here in question are produced by bringing these lines “(the D lines for instance) into perfect coincidence, horizontally and ertically. Under these circumstances if the distance apart, e, is ‘ably chosen, the interference fringes will appear throughout the trum. These consist of an approximately equidistant series of nes parallel to the slit, i. ¢., vertical lines, which are finer, cet. par., the breadth of the crack at S between the gratings is larger. hey may be increased from the extreme fineness as they enter the ge of visibility to a maximum coarseness (in the above experi- nts) of about three to five minutes per fringe, after which they ish. They cannot, in practice, be passed through infinite size; ler can they be produced symmetrically on the two sides of the iustment for infinite size. They cannot in other words be changed m the positive to the negative condition of appearance. The occurrences are in fact as follows: if as in Fig 2, 1>8, rallel white rays coming from L and L’, R and R’ being reflected, ) and D’ diffracted rays for the normal n), the grating g’ must be n advance or forward of g. If now the airspace e is reduced micro- trically, g’ retreating, the lines travel in a given direction (from to right) through the spectrum, while at the same time they row continually larger until for a minimum value of e still positive, hey vanish as a whole. The period of indistinctness before evan- ence is not marked. On the other hand if 6 >i as in Fig. 3, the grating g’ must be to > rear of g and the air space e is throughout negative. If this is decreased numerically the lines travel through the spectrum the opposite direction to the preceding case, while at the same ime they coarsen until they vanish as a whole as before. The ra ting g’ is still behind g when this occurs. PROC. AMER. PHIL. SOC. L. 198 I, PRINTED APRIL 27, IQII. 10h BARUS—ELLIPTIC INTERFERENCE Finally if for any suitable value of e the grating g’ is mi V tance apart is very peculiar, as is also the fact that they ¢ f not passed through infinite size or appear symmetrically on both si of this adjustment. To investigate this case I provided both collimator and the telescope with slits so that the parts of the ing g and g’, from which the interfering pencils come, a investigated. Fe If-a single vertical slit about 1 mm. wide is passed fron to left toward the objective of the telescope, a black, line” »& Fic. 4. Diagram. across the field of the spectrum, which line is merely the imag the crack at S. In the diagram Fig. 4, the G rays, for instance, ¢ from the edge of both gratings g and g’, whereas the R rays ai the ’ rays come from but a single grating. Now when the spac —-tgrr.] WITH REFLECTING GRATING. 131 is diminished, the black band at G gradually vanishes and in its place appear the coarsest fringes producible. When the slit F is removed these coarse fringes disappear. The fringes visible through the slit have however both an inferior and superior limit of angular size. When e is diminished to zero they vanish and when ¢ is sufficiently increased they again vanish, though they now ‘appear when the slit is either removed or widened. From this it follows that the coarsest fringes come from the edges of the crack 'S of the gratings, and that the remainder of the grating will not pro- duce coarse fringes. By moving the slit the fringes may be made to appear in any other part of the spectrum. ___ The same fact may be proved by putting the vertical slit F over the lens of the collimator and allowing the white light L to fall on the edges of the grating at S. Coarse fringes limited as to range and size are then seen throughout the spectrum at g. . _ Whenever the slit or vertical stop is used, the fringes are ex- -ceptionally sharp and easily controlled for micrometry. It is not even necessary to adjust the two spectra horizontally with the same care as when no slit is used, but the vertical coincidence of spectrum lines must be sharp. Naturally the use of the slit has one draw- back, as the resolving power of the grating is decreased and the ‘spectrum lines are only just visible. The adjustment, however, may be made before the slit is added. A few examples may be given. For a slit 1 mm. wide over the telescope or collimator, only the immediate edges at the crack S, about .5 mm. each in breadth, are active. A narrow range of large fringes aré seen in the field easily ‘controlled by the micrometer screw. With a slit 3 mm. in width the lower limit is much increased the upper diminished, to a size __ of about 3 inches per fringe. In the absence of the slit the field is free from fringes. With a slit 6 mm. wide, the upper limit is again decreased the lower much increased ; nevertheless the finest fringes appear only after the slit is removed. Using double slits over the collimator, each 1 mm. wide and 3 mm. apart, fringes of medium Size limited at both ends appear; 3 mm. slits 6 mm. apart show only the very fine fringes. but both sizes are still limited. Finally when al but about .5 mm. of the edge of the crack of the grating g’ is 132 BARUS—ELLIPTIC INTERFERENCE screened off, whereas the whole grating g (about one half inch | square) is without a screen, all the fringes from the maximum size to complete evanescence beyond the range of visibility are pro- ducible. Naturally if the edge of g’ is quite dark everything vanishes. It follows therefore that pairs of corresponding rays are always” in question. These corresponding rays are at a definite ND, apart where D is the grating space and N the number of lines per cm. of the grating in question. This distance ND is greater as the fringes are smaller and may be of the order of a cm. when the fringes pass" beyond the range of visibility. Again ND is equal to the width of the crack when the largest fringes vanish. Finally when ND is zero, as in the original unbroken grating. the size of the fringes is infinite. | It has been stated that the use of the slit or a laterally limited objective is advantageous because all the lines are much sharper. Inert or harmful illumination is cut off. If the slit is over the objective of the telescope only a small part of the field of view shows” the lines; if placed over the objective of the collimator. the fringes are of extreme clearness throughout the spectrum. It may be ulti- mately of advantage to use the edge near the crack g’ only, together - with the whole of g. For if a small strip of g’ at the crack S is used with the whole of g, the smaller fringes are weakened or wiped out. Thus the inner edge of the nearer grating with successive parts of the further grating is chiefly effective in the production of these interferences. . To bring the two edges quite together was not possible in my work, as they were rough and the apparatus improvised. . 7. Data——Some measurements were attempted, with the view only of checking the equations presently to be deduced. The adjust: ment on an ordinary spectrometer is not firm enough and the fringes being very fine (a few minutes of ae) are difficult to follow unless” quite stationary. ‘Table I., however, gives both the values of de/dn, displacemen per fringe, foe different angles of incidence 7 and of diffraction 6, — and d6/dn, the angular deviation per fringe at the D line. In meas- 1911.) WITH REFLECTING GRATING. 133 uring the latter it was necessary to count the fringes between the C and D lines and divide their angular distance apart by these num- rs. As e cannot be measured, its successive increments Ae from the first position are given. These are presently to be associated with the corresponding increments of dn/dé. TABLE I. VaLuEs oF d@/dn, ETC. i==53° 15°. D200 X 10° cm. Observed. Computed. a Meinocs. Are | aco | Mean | *esion- Saeee | ee | ee Aa Be cata | dull |Adald® | ¥20° | 30°27’ | 1’ 17”| 3080 1130 025 1260 | 1028 | x140 | Between € 75 28° 14’ | 1/46”) 1950 and D lines Diff. by seas & 1 90 Oo 7x. 28° 14’ 46” |} 4438 -025 | 1116 | 1027 | 1072 | Near C line : 55 Diff. | 1 07%} 3438f| 1000 | .o50 ; 55’ |17327| 2250 | 1188 | .075 je 5 ° , , o” 8 : ay 29° 49 ve jot 1328 | .025 | 1259 | 1196 | 1228 | Near D line 44’ a ta gi ial Ri ad la ata siliaas a eee ea aes 8. Equations—tIn Fig. 5, L and L’ represent a pair of corre- sponding white rays, reflected into R and R’ and diffracted into D and D’ at angles i and 6, respectively. The half gratings g and g’ are separated along the crack S, and g’ is movable parallel to itself by a micrometer screw normal to g’. Let the normal distance apart of the gratings be e. The incident rays L, L’ strike the originally coplanar grating at points N rulings apart, or ND cm. apart, if D is the grating space. In the separated grating let these points be at a distance c apart. Let d be the incident wave front and h the corre- sponding diffracted wave front and call the angle between c and d, y. When there is reinforcement the path difference of the rays L d L’ from the incident (d) to the diffracted (h) wave front, may be written nA=b—a, where b and a are the distances of h and d from the points of inci- dence of L and L’ on the grating g and g’ respectively. If finally f 134 BARUS—ELLIPTIC INTERFERENCE is the length of the prolongation of L’ between the gratings write in succession (1) d==ND cosi, (2) f=sesecs, (3) a=ND sini—eseci, (Ay. tany=a/d, (5). C=ND cosisecy, (6) | bb an (i 44): Fic. 5. Diagram. To these should be added (7) dN/de=tani/D. Hence after removing y and arranging ni = ND {cosi sin (t+ 0) —sinicos (i+ 6) —sin1} +e seei(t +008 (i+ which reduces to nk = ND (sin 6@—sini) + eseci(1 + cos (1+ 8@)), or since sini — sin@=A/D, 11.J WITH REFLECTING GRATING. 135 1+cos(¢+@) 2ecos’ (¢ + 8)/2 cos z oF cos z : This must therefore be regarded as the fundamental equation of the phenomenon. Equation (7), however, leads on integration to (9) N=etani/D+N,, where N,D is the width of the crack. If the value of N from (9) is put into (8) together with the equivalent of A/D, it appears after reduction that (8) (n+ N)A=e r+ 6 (x + N,)A = e(cos z+ cos 0) = 2¢ cos cos The case of N 0, e > 0 would correspond to the equation (10) #w=e[1+cos(¢+ 0) |cosi-= 2¢ cost *° /cos which is only a part of the complete equation (8). For +> 8, one active half, kh, is necessarily partly behind the other half, k’h’, and therefore not adapted to bring out the phenomenon as explained, unless eo. g. Differential Equations. Displacement per Fringe, de/dn.— To test equation (8) or (10) increments must be compared. The latter gives at once since N is constant relative to e like 1, 6, and A, de Xr oe x =) ae Ee me ae io dn cost+cos@ 2cos eS which is the interferometer equation when the fringes pass a given spectrum line, like either D line, which is sharp and stationary in the field. Equations (7) and (11), moreover, give after reduction 76 (12) dN/dn = tan itan———. Table I. contains values of de/dn computed from (11), made under widely different conditions (i > 6, 1< 6, first and second order). ‘The agreement is as good as the small fringes and the difficulty of getting the grating normal to the micrometer screw in my impro- 136 BARUS—ELLIPTIC INTERFERENCE vised apparatus admit. If this adjustment i is not perio with e. From equation (12), moreover, - bets aN _ dN dd _dN,d0_ aN, adn ad0dn dO dn an’ since N, is constant only relative to e when 6 varies. 5 10. Deviation per Fringe, etc., d0/du, d0/de.—These mez ments are still more difficult in the absence of special appara since ¢ is not determinable and the counting of fine flickering frin is unsatisfactory; but the order of results may be corroborated observing the number of fringes between two Fraunhofer lines, e the C, D and other lines used. Differentiating equations (8) an (10) for variable n, A, 6, and N (since dN /dé@ is equal to dN. /d6, equation (12’)) and inserting —D cos 6- d0/in- a it Son after arranging that ! dd 1+dN/[dn _® I eee (ta) dn eD1+cos(é+0) eDcosi{cosi+cos@) or ‘ : et, € de r i—6 mW .tan adn ecost 2 Combining this with (11) ae x sin 7 — sin 8 (14) dn eDcosi ecosz Since, in equation (13), ¢ is not determinable it is necessary to ce pare increments Adn/d@ in terms of the corresponding increme Ae, whence me :(15) — A(dn/d0) = (cos i/X tan fn “| Ae. Table I. also contains data of this kind computed separately for the Fraunhofer D, C, etc., employed and their mean values. To find the mean width of fringes between these lines, their angular devia- tions were divided by the number of fringes counted between them at different values of e. The results agree as closely as the difficulty of the observations warrants. One may note that without remoy- gtr.] WITH REFLECTING GRATING. 137 N, the corresponding coefficients would be Ad(n + N)/d6, and these are much more in error, here and in the preceding cases. If from d@/dn, ¢ is eliminated in terms of (n+ N) the equation is a@ Xx I dn D(n+N, cos 2’ so that for a given value of i, 6, No, they decrease 1 in size with n. f n=0 they reach the limiting size db % dn DN, cost’ f the crack N,D should be made infinitely small, they would be infinitely large. To pass through infinity, N, must be negative, which has no meaning for i > 6 or would place one effective edge of the crack S behind the other. These inferences agree with the observations as above detailed. If, however, i < 6, a negative value »f N, restores equation (16) for no to ——— (17), as was actually observed (Figs. 2 and 3). Finally equation (14) might be used for observation in the incre- i me A(de/d8) = ~~ Ae; but I did not succeed with it. One loses track of the run of a I. Colored Slit Images and Disc Colors of Coronas.—In the above experiment the fringes were but a few minutes apart. It is _ obvious, however, that if N, is sufficiently small the fringes will grow with decreasing n, in angular magnitude, until there are but a few _ black bands in the spectrum. Under these circumstances the unde- _viated image of the superimposed slits must appear colored, particu- ery so if an effect equivalent to N, is present throughout the ‘grating. This phenomenon of colored slits is apparently of interest _ in its bearing on the theory of coronas, where there is also an inter- _ ference phenomenon superimposed upon a diffraction phenomenon, © as is evidenced by the brilliant disc colors. For instance suppose 138 _ BARUS—ELLIPTIC INTERFERENCE. a corona were produced by a sufficient number of tag ) tributed throughout a plane normal to the undeviated rear of their original position and let the distance betw : planes be small relatively to the wave length of bie I tion of the disc colors of coronas. Brown University, ProvivENcE, R. I. ON THE TOTALITY OF THE SUBSTITUTIONS ON n LETTERS WHICH ARE COMMUTATIVE WITH EVERY SUBSTITUTION OF A GIVEN GROUP ON THE SAME LETTERS. By G. A. MILLER. (Read April 20, 1911.) §1. InrropuctTIon. The problem to determine all the substitutions on m letters which are commutative with every substitution of a regular group on the ‘same letters was first solved explicitly by Jordan in his thesis. It was found that with every regular group there is associated a group _ which is conjugate with this regular group, such that each is com- posed of all the substitutions which are commutative with every substitution of the other.1 These two regular groups were called : _ conjoints by Jordan and it is evident that they have a common _ holomorph and that their group of isomorphisms is the quotient group of this holomorph with respect to either of these two regular groups. The more general problem to determine all the substitutions on n letters which are commutative with every substitution of any transitive group on the same letters seems to have been solved for the first time by Kuhn in his thesis.* He found that with each transitive group G of degree m there is associated a group’K on the same # letters which is composed of regular substitutions on these -n letters, in addition té the identity. The order of K is a, the degree _of the subgroup composed of all the substitutions of G which omit a = given letter being n—a. Hence a necessary and sufficient condi- 7 _ tion that K be transitive is that G be regular, and the number of the ‘ 3 ‘systems of intransitivity of K is always equal to n/a. Whena 2, and determining t totality of the substitutions which are commutative with every si stitution of the intransitive group G thus formed. It is evident this totality of substitutions constitutes a group K which is similar G. That is, G and K are two conjugate intransitive substituti groups each being composed of all the substitutions on these 7 letters which are commutative with every substitution of the other. — The existence of the two amicable intransitive groups G and of the preceding paragraph may also be established as follows Consider the n* m-sets' of the symmetric group of degree m regards the symmetric group of degree » —1. On multiplying th n? m-sets on the right by all the substitutions of this symmet group the 2? m-sets are permuted according to a group G’ similar G, and by multiplying them on the left they are permuted accor ) to a similar group K’. From the fact that multiplication is asso tive it results that every substitution of G’ is commutative with eve substitution of K’.. Moreover as every substitution on these n? letter which is commutative with every substitution of G’ must perm some of its systems, it is evident that K’ is composed of all | substitutions on these letters which are commutative with ev substitution of G’, and vice versa; that is, G’ and K’ are in fact t amicable intransitive groups for every value of u. The gro *If H is any subgroup of a group G, the total number of distinct sets operators of the form SeHSg, where Sa and Sg are operators of G, are kno as the m sets of G as regards H. MILLER—SUBSTITUTIONS ON n LETTERS. 143 : : 1gt1.] 4 generated by G’ and K’ is clearly imprimitive and of order (m!)’. __._ The existence of amicable intransitive groups which are not included in the preceding infinite system can be easily proved by the following examples: Let G be the dihedral group of order 8 and H any one of its non-invariant subgroups of order 2. With respect to _ H there are 8 m-sets of G since H is transformed into itself by 4 of the operators of G. Hence these eight m-sets are permuted accord- ing to a group which is simply isomorphic with G and has two 4 transitive constituents both by right and also by left multiplication. P _ Each of the two substitution groups obtained in this way is clearly - composed of all the substitutions on these eight letters which are _ commutative with every substitution of the other and hence these are two amicable intransitive groups whose transitive constituents _ are not symmetric. a The substitutions which are commutative with every substitution of an intransitive group G either interchange systems of intransi- tivity, or they are composed of constituents which are separately _ commutative with the various transitive constituents of G, The 3 latter have been considered in the preceding section. Hence we shall, for the present, confine our attention to those substitutions __ which are commutative with every substitution of G and interchange _ its systems of intransitivity. It is evident that these systems of _ intransitivity are transformed by all the substitutions which are - commutative with every substitution of G according to a substitu- tion group, and that those transitive constituents of G which are transformed transitively among themselves must be simply isomor- phic in G. These constituents are clearly transformed according to a symmetric group by all the substitutions which are commutative with every substitution of G. Hence the theorem: Jf an intransitive _ group G is one of a pair of amicable intransitive groups, and if the _ transitive constituents of G are such that no substitution on the __ letters of the separate constituents is commutative with every sub- _ Stitution of the constituent, then must the consituents of G be _ symmetric groups. a It is clear that G may have more than one set of transitive con- _ Stituents such that all those of a set are conjugate under the totality 144 MILLER—SUBSTITUTIONS ON n LETTERS. ‘of the substitutions K which are commutative with every substi tion of G. In other words, the substitution group according to which the transitive constituents of G are transformed may be intransiti When this condition is satisfied K is the direct product of two or more symmetric groups. This suggests a more general infinite system of pairs of amicable intransitive groups than the one men- tioned above: viz., Let G be the direct product of the p gro formed by establishing simple isomorphisms between n, symmetric groups of degree n,, m, symmetric groups of degree m,, ---, ny, symmetric groups of degree n, (,, m, «++, m, being distinct numbers greater than 2), it is clear from what was proved above that K is similar to G and hence G and K are ami- cable intransitive groups. It should be observed that G and are always amicable whenever they are similar but that the conver: of this theorem is not always true. This more general system amicable intransitive groups may clearly be constructed by formin the direct product of the p symmetric groups of degrees m,, ms, -*-, n, Tespectively and forming the m-sets as regards the subgroups H obtained by forming the direct product of p symmetric groups degrees m,—1, m,—TI, «++, m, —1 respectively, one being take from each of the given symmetric groups, in order. If the m-se ‘thus obtained are multiplied on the right and on the left by all operators of these sets there clearly results the two sya c amicable intransitive groups in question. To obtain a still more general infinite system of atnicabie intra; sitive groups it should be first observed that the intransitive group formed by establishing a simple isomorphism between m, symmet: groups of degree m,, written on m, distinct sets of letters, is ami with the one obtained by establishing a simple isomorphism betwee n, symmetric groups of degree m,, written on m, distinct sets of letters, where ,, m, > 2. Hence it results that the direct produ formed by multiplying p intransitive groups of degrees n,m,, n4i -,n, m, respectively, each being formed in the former of the tw ways mentioned above, is amicable with the direct product forme by multiplying the p groups of the same degrees respectively, but constructed by establishing a simple isomorphism between m, sym= ge . 5 3 Pi Ag a Bi q oa i fh tg1t.] MILLER—SUBSTITUTIONS ON n LETTERS. 145 metric groups of degree m,, n, of degree m,, ---, m, of degree m, respectively. Moreover, it results from the given theorem that these direct products include all the possible sets of amicable groups in which each of the two groups is intransitive and each of the transi- tive constituents is not commutative with any substitution besides the identity on the letters of the constituent. The above therefore completes the determination of amicable groups when both groups are intransitive, and the transitive constit- uents are such as to involve subgroups whose degrees are just one less than the degrees of the respective constituents. The cases in which at least one of the two amicable groups is transitive were considered in the introduction. It may be observed that whenever an intransitive group is formed by establishing a simple isomorphism between more: than two symmetric groups it is one of a pair of amicable groups. The second group is transitive when each of these symmetric groups is of degree 2, when this condition is not satisfied the second group is also a simple isomorphism between symmetric groups. The group obtained by establishing a simple isomorphism between two symmetric groups is evidently never one of a pair of amicable groups unless the two symmetric groups are of order 2. We may express this result in the form of a theorem as follows: The intransitive group G formed by establishing a simple 1somor- phism between three or more symmetric groups, written on distinct sets of letters, is one of a pair of amicable groups, the second group K being also such an intransitive group whenever the degree of the given symmetric groups exceeds 2. The intransitive group formed by establishing a simple isomorphism between two symmetric groups is one of two amicable groups only in the special case when these symmetric groups are of degree 2. By means of the given results it is not difficult to complete the determination of all possible pairs of amicable intransitive groups. Suppose that G is constructed by establishing a simple isomorphism between any number of conjugate transitive groups written on dis- tinct letters, each constituent being one of a pair of amicable groups. If these constituents are not symmetric and not regular it is clear that G is one of a pair of amicable groups and that the number of the 146 MILLER—SUBSTITUTIONS ON n LETTERS. [April 20, transitive constituents of K is equal to the number of transitive con- stituents in the amicable group corresponding to a transitive constit-_ uent of G. Moreover, G is evidently not one of a pair of amicable groups when its constituents do not have this property. Hence there results the theorem: Two necessary and sufficient conditions that a given intransitive group be one of a pair of amicable groups are: 1) that it be the direct product of transitive constituents which belong to pairs of amicable groups, or of sets of simply isomorphic transitive constituents of this kind, or 2) that the number of simply isomorphic constituents -be greater than two whenever they are symmetric but not regular. From the Introduction it results that the second group of this pair is also intransitive except in the case when the intransitive group is composed of simply isomorphic regular groups. It reduces to the identity whenever the given intransitive group is the direct product of symmetric groups whose degrees exceed 2. The pair of amicable groups are conjugate whenever one is the direct product of regular groups, of sets of m simply isomor-_ phic non-regular symmetric groups of degree n if the m’s and n’s may be put into (1, 1) correspondence such that the corresponding pairs are equal, or of sets of m simply isomorphic non-symmetric ‘ transitive groups of degree n (nm —a being the degree of a subgroup © composed of all the substitutions of the constituent which omit a letter) if the a’s, m’s and n/a’s may be put into (1, 1) correspon- dence such that the corresponding triplets may be a, n/a, ” _ 2 every set of values a, m, n. ise eae UNIVERSITY OF ILLINOIS. PROCEEDINGS AMERICAN PHILOSOPHICAL SOCIETY HELD AT PHILADELPHIA FOR PROMOTING USEFUL KNOWLEDGE 7oOL:: L May—AucustT, 1911 | No. 199 NOTES ON CANNON—FOURTEENTH AND FIFTEENTH CENTURIES. By CHARLES E. DANA. (Read April 20, 191.) There can be seen to-day in the fair city of Florence, on the Arno, an old yellow parchment, upon which is transcribed an edict dated February 11, 1326. This, expressed in the monkish Latin of the _ day, gives authority to the Gonfalonier, the Priors, and twelve _ good men,” to superintend the manufacture of “ palloctas ferras et canones de metallo,” balls of iron and cannon of metal, which _ may possibly, in this case mean brass. What these cannon for the _ defence of Florence were like, or what they did, we shall never ___ know, but with them the real history of ordnance begins; these little 7 pop-guns are the ancestors of the 14- and 15-inch B. L. R. (breech- 4 3 loading rifle) of today; the fathers threw a wee projectile a hun- = dred or two yards; the degenerate offspring throw a shell weighing about a ton, fifteen or twenty miles. That the Florentine guns were | a the very first no one would assert, but with our present information, only legend lies back of them. ; be Of course the credit for the invention is given to the Chinese. ___ There is not time here to do more than state that the Institutes of 5 Timur, about the middle of the fourteenth century, although they PROC. AMER. PHIL. SOC., L. I99 J, PRINTED JUNE 26, IQII. : 147 148 DANA—NOTES ON CANNON. [April 20, Cannon of 1390-1400. One of the earliest representations of a Fire-arm. Royal Library, Munich.) (From German Codex. DANA—NOTES ON CANNON. 149 os 1.J give full details of the equipment of his troops, do not mention either cannon or gunpowder. The “Wuh-li-Siao,’ published 1630, says “gunpowder came from the outer barbarians.” 4 The mention of an explosive in the Sukranita, a Hindu work 4 said to ante-date everything Chinese, is admitted by experts, I under- _ stand, to be a modern interpolation. q The “Liber Ignium a Marco Graeco Descriptus,”’ dating back of the eleventh century, gives some 22 to 35 recipes for the so-called ““ Greek-Fire”’ etc. No Greek or Moslem writer ever uses the term _ “Greek-Fire.” Col. Hime, an authority on this subject* concludes that the earlier recipes in the “ Liber Ignium,” were translated from _ the Arabic by a Spaniard. The first four recipes are for the com- pounding of “‘sea-fire,” or, as there described, mixtures which will ignite “when rain falls on them.” Quicklime was the cause of _ ignition; to it was added (C. 1300) sulfur, oil, gum Arabic; (C. _ 1350) sulfur and turpentine; (C. 1405) sulfur, petroleum, wax. None of these were true explosives. Berthold Schwartz, of Freiburg, in Breisgau, the favorite Ger- ‘man discoverer of gunpowder, made his discovery about 1320 to _ 1330, at the time the Florentines were popping off their.‘ canones de _ metallo.” Schwartz is said to have preceded the Florentines in the making of cannon but this claim has not as yet been established. Lieut. Col. Hime undertakes to translate the “ Epistola de secre- tis,” of the liberal minded Friar Bacon (1214?-1294). This letter is probably earlier than 1249. It is written according to some eryptic method, a bad habit both famous Bacons indulged in, and if the secret of the over-cautious Friar has been guessed with even ‘partial success, we have a right to suppose that while the good Friar was “ experimenting with some incendiary compositions . . . the mixture exploded and shattered all the chemical apparatus near _ it” (Hime. 161). After this smash-up, Bacon could not fail to be convinced that saltpeter, sulfur and charcoal, when mixed in right proportions, had a distinctly explosive tendency—but he never seems to have advanced the next step and discovered the projectile force of the compound. * Lieut.-Col. Henry W. L. Hime, “ Gunpowder and Ammunition,” London 1904. ? 150 DANA—NOTES ON CANNON. [April 20, A-few more words on gunpowder. In early days saltpeter was most difficult to procure; it was collected from cellars and caves; later, depots were established for its reception, while strict laws were passed to ensure its purification and baking. Costing much, saltpeter was very sparingly used, much to the detriment of the — 4 gunpowder of course. The proportions differ greatly according to — the kinds of powder—whether, for cannon, priming, hand gun, etc., from equal parts of each; to, saltpeter, 3, sulfur, 1 %, charcoal, 1 4% to 4: 1:1. and6:2:1. The formula of today being about 6: 1: 1. The price of gunpowder in the fourteenth century seems to have — been almost prohibitive. Assuming that my figures are correct, which is more than doubtful, for there is no real standard of value, _ the price was, in money of to-day, rarely as low as twenty-five dol- lars, and quite possibly, occasionally, as high as fifty dollars a pound; : now it costs a quarter of a dollar or less a pound. These prices | ; rapidly decreased with the systematic collecting of saltpeter. ‘a In a campaign the ingredients for making powder were carried — separately, and mixed only when need came. Here is a note or two © from an authority of about 1465. Keep the three ingredients sepa- _ rate, as the niter and sulfur if mixed soon spoil. Better carry the — willow wood unburned, as charcoal absorbs the damp. A secret process for the preservation of powder: Take clear and very strong vinegar, make the powder into a paste; form cakes of four to eight — livres (livre, about a pound), dry in the shade, sun, or even in an oven. We are getting close to granulated powder. As the usual — powder was in the form of a very fine dust, the ignition must have . been slow, and much of it was, in all probability, blown out at the — muzzle. . The next mention of cannon is in an “indenture” of 1338, be- tween John Starylyng, former keeper of the “ King’s vessels” (Ed- ward IIT.) and Hemyng Leget. “Ij [ij] canons de ferr sanz estuff “y, presumably, without ammunition. Also “un canon de ferr ove ii . chambres, un autre de bras ove une chambre.” The cannon with — two chambers was the form of breech-loader often used even for large bombards until the early part of the next century, and for — smaller iron and brass cannon until the art of casting iron guns was — well understood (in England not until c. 1545), and even into the — ee ——— oti 1911.) DANA—NOTES ON CANNON. 151 en tory Meaeitt Loading Cannon of 1390-1400. Cutting off wooden Plug for Wad. (From German Codex. Royal Library, Munich. ) 152 DANA—NOTES ON CANNON. [April 20, seventeenth century, especially for guns used in China. An iron tube was fastened by bands of iron to a strong stock, with a space hollowed out at the breech to hold an iron box, the “ chamber,” sometimes called “a pot.” Later this hollow became part of the gun, a sort of basket, or cradle. The chamber contained the powder charge, upon which a round section of wood was tightly driven as a wad. The projectile was placed at the breech end of the tube, with a straw, felt or rag wad in front to hold it; the chamber was wedged against it, then primed, and touched off with a heated iron’ rod. Needless to say that a goodly portion of gas escaped between chamber and tube. These were the earliest quick-firing guns, and in the sixteenth century were used on the upper works of the ships, very much as we use quicker firing guns today; the ancient ones, when there were four chambers, say, could be discharged about every two minutes. Later they will be called “murtherers”; and well they earned the name when the projectiles consisted of rusty nails, bullets and ines ‘ The chambers, or ‘ canones,” for the huge bombards, which w shall meet with later, were held in place by heavy timbers. By or before 1400 the heavy, wrought-iron powder chamber was welded or screwed onto the chase of the bombard. In the year 1338 appeared, in the Arsenal of Rouen, a terri engine of destruction, called by its proud keepers, a “pot de f a traire garros,” an iron pot for throwing arrows. These arro much like cross-bow bolts, were tipped with iron and winged w brass, the latter metal obtained from kitchen utensils, cut up melted for the purpose. ' The projectile was wound with leather make it fit snug in the barrel. The powder charge for this dr engine of war was about seven tenths of an ounce of the ill-propo tioned powder of that day. When all was prepared, and fire wz applied, the bolt of destruction no doubt emerged, but certainly wit! considerable reluctance. A recipe of a few years later enables us to approximately fig out the cost of cannon of that day. Five cannon of wrought-iron weighing 25 Ibs. each, and five “canon de metal,” presumably bre cost three hundred dollars of to-day, say $30 each. All of which is” submitted with considerable hesitation. ‘UOLJOUL [ROAVA PUL [vJUOZIIOY JO SURO BuIMOYS MOTA-oplg ‘aptoefo41g MOYS 0} poAouios oaseYyD jo Jud foovjd Ul paspom szoquivysy ‘“Burjutod JojJ [ey ‘aseYyD JO YooIig puv o[pery ‘aBivyo s9pMod poy 0} ,“JOq ,, 10 Joquuivysy ‘OOZT "O-QELI ‘ad01q-saquieyy 153 CANNON. ? f. DANA—NOTES ON 1911r.] 154 , DANA—NOTES ON CANNON. [April 20, The first mention of cannon by Froissart, who is as “ faithful as an eye witness,” is in the year 1340 at the siege of Quesnoy, on ‘ the northeast border of France, not very far from Valenciennes. “Those of Quesnoy let them hear their cannons and bombards, — which flung large iron es in such a manner as made the French ey, afraid of their horses.” The earliest ae use of this word bombard given by Dr. Murray, is in a quotation from John Lydgate, 1430. The noun has left us, but the verb, to bombard still lingers. It is usually asserted that the first field-guns were used at the 4 ‘battle of Crécy, August 26, 1346. The Florentine chronicler, Gio- vanni Villani, remarks, in the somewhat florid manner of the time, “the bombards of the English made balls of iron to leap with fire, to frighten and drive off the horses of the French. . . . That the roar — of the bombards made such a trembling of the earth, such a noise, _ that it seemed as if God thundered, with great slaughter of men and beating down of horses.” This terrible slaughter must have been produced by three small toys somewhat like blunderbusses, the charge for each of which was an ounce, more or less, of very bad powder. Cause and effect do | seem disproportionate. e “Grandes Chroniques de St. Denis” assert that it was the three cannons of the English that spread panic amongst the Genoese ‘ cross-bowmen and made them indulge in the singular antics by which they sought to frighten the English archers. In only one known copy of Froissart is there any mention of cannon at Crécy; this happens to be that in the library of the city of Amiens, not far from the battlefield; there is some reason to believe that the words — in question are an interpolation; when one remembers that from Falkirk (1298) to Flodden (1513)—Bannockburn excepted—the English archer, firing ten or more armor-piercing projectiles a min- ute, with an effective range of 250 yards, was always victorious, it does appear possible that French writers, with more patriotism than truth, introduced these terrible cannon into their accounts of the battle as an excuse for the crushing disaster to their arms. Edward III. took with him several cannon when he entered ? Chap. IV., Book L, p. 40. grt] DANA—NOTES ON CANNON. 155 _ France, July, 1346, the month before Crécy. We have too the _ King’s Privy Wardrobe Accounts, as they were termed, giving lists _ of guns, ammunition, gunners and other details of the ordnance sent _ from the Tower of London to be used at the siege of Calais, which followed the battle of Crécy. _ At this siege of Calais only leaden projectiles are mentioned, and _ from the very moderate amount of ammunition required for their = propulsion, the guns although called “great’”’ must have been ex- 4 a ceeding small. The main reliance of besiegers and besieged contin- ued to be in the huge engines for hurling masses of rock and other ___unpleasantnesses. q The derivation of the word gun is not without interest. Consult- _ ing both Murray and Skeat, we find that Gunnhildr was an Ice- ’ landic, female, proper name, once applied to war engines. As Gunn (Icel. Gunnr) signified war, and hildr a battle, it was certainly ap- = ‘propriate. An account of the munitions in Windsor Castle, 1300/o1, mentions a large ballista named “Domina Gunilda.” As there does : 3 : not seem to have been any great lady, famous or infamous, so called a in the fifteenth century, this is quite probably a survival of the old i Scandinavian name. The M. E. word gunne, is, of course, but a shortened pet name for the fearsome lady. Here is one of the early tragedies connected with cannon. In 1346, the year of Crécy, Peter of Bruges had established a high reputation for the making of “connoiles.’” The word may come from “tonnoiles,” which in its turn, may have come from “tuyaux _ de tonnoire,” or tubes of thunder. In September of that year the consuls of the city of Tournay hearing that connoiles were useful to be let off in a good town when besieged, desired the aforesaid _ Peter to make them one as a sample, and if it proved satisfactory _ they would give him an order for more. Peter, the thrifty burgher, did make one and then proceeded to show the worthy consuls what _ it would do. The connoile was placed with great care, outside the _ gate “ Noire aux Champs.” Peter states in his own account that he _ loaded the connoile with a quarrel, meaning in this case a heavy bolt, not an altercation. To the quarrel Peter affixed two pounds of lead. From the subsequent happenings there is reason to suppose _ that he did not omit powder. Peter “laid” the connoile so that it 156 DANA—NOTES ON CANNON. [April pointed against a door and wall. The spectators heard a “ce noise,” but the antics of the connoile remain a mystery. In quite another part of the city of Tournay, an industrious fuller was busily at work; when lo, along came the erratic quar with its two pounds of lead,—and the guild of Fullers gave deceased brother one of those picturesque funerals for which good town is so celebrated. When Peter of Bruges heard of this mishap, he fled into sanctuary and gave himself up for lost. Then followed a solemn session of the consuls. Contributory negligence could not be charged against the Fuller, for if ever bolt came “ from the blue” it was this one. After a long discussion the conclu arrived at was: Peter of Bruges fired the connoile at the order the consuls; he was not known to have harbored any ill feel against the fuller;—they might have added that neither ill feeli nor skill in aiming would have enabled Peter to hit the fa fuller. The consuls thereupon held Peter blameless, merely rem é ing that the event was a misfortune and a sad pity. A curious point is brought out by the list, dated 1347, of artille in its broader sense, for the defense of the castle of Brioul in Fran At the fag end of the list we are told that one man managed cannons, and that the efficiency of their projectiles, and of ste thrown from the towers by hand, was considered about equal. _ Before glancing at the great bombard of Caen, 1375, wh marks a considerable step in advance, let me say that during the fifty years we have glanced at, there have been cannon of wro iron, occasionally of brass. The largest of the former did not we over 120 lbs. Breech-loaders were common, and the projectiles w bolts, or balls of lead—iron balls are referred to, but never sto March 20, 1375, an order was received at Caen, in Norman from Jehan Le Mercier, one of the King of France’s councill for the building of “a great cannon of iron.” March 21 w began by erecting three forges in the market place, and surrounding them with a wooden paling to keep the curious at a proper distan e. March 22, the four smiths with their eight helpers began to d wages. Fifteen men worked for six weeks, sometimes at night April 3, Jehan Nicolle, a master smith, said to have been the bes in Normandy, arrived from “Sap.” 2,110 pounds of wrought ire I i fH 2: Basel Arsenal. ho nat, Grosse. Built-Up Wrought-Iron Bombard, 1420-1430. Length 8 ft. 6 in. Caliber 13% in. Win Y Y WY P Gy Y Vib ad I oS ee CE tas 5 ay oN aro to RE apt) ey eg ot anal at ON in ee 5 ie ear rr ve wan wanseuine ss UY Yl UY Uy WY wn wy we ee y ; UUMLDOYOUUYYGDLUUBD Ym ‘ fies x * ' A 2 sy we a a tg Me ee a a a eee es art ing ae ‘ . , ' H ” : . : . —_— ee eed ee Se) 4) ee | 158 DANA—NOTES ON CANNON. April 50, and 200 pounds of steel were used. The inner tube was formed © by longitudinal bars ; encircling these were tight-fitting rings of iron, driven on, one touching the other, till they formed an unbroken sur-_ face. 400 pounds of the iron was Spanish, presumably a better quality. The “cuve” for which it was used, may have been that part of the breech which enclosed the powder chamber. 200 pounds of steel were needed; could the chamber have been of that “metal? The chamber seems to have formed a permanent part of the bom-— bard, as the vent is specially mentioned with its large projecting — apron of iron. : . After the metal part was finished ninety pounds of rope was _ wound about the gun, for what purpose we are not told. Over this — was sewn a cover of hide to prevent the rope rotting or the metal rusting if exposed to rain. The manner of attaching the monster to its heavy wooden bed and braces, is fully and confusedly described. General Favé thinks — its carriage was a kind of cage, somewhat like that used by black smiths in France for shoeing unwilling animals. Four stone balls, — size not given, were provided, at a cost of two sous six deniers each | ($1.50? today). Two of these were used in the proof rounds. After this date (1377) the size of cannon rapidly increased. — Froissart mentions 140 cannon used at Odruik or Outherwyck, by — the Duke of Burgundy, in 1377, which threw balls of 200 pounds. A work (name not given) professing to quote contemporaneous authority, mentions a cannon of the Duke of Burgundy, 1377, throwing a shot of 450 pounds, which would require a calibre of say, ts 21 inches. a 1382, at the siege of Oudenarde by Philip van Arteveld, the — Flemings made use of a “marvellously great bombard,” so they — said, at least. They added, that when this bombard was fired, by — day it could easily be heard a distance of five leagues, and by night ten. It made such a terrible din (French “noise”) that to those who listened, it seemed as if all the devils in hell were rushing on. The rather imaginative old chronicler says that this monster had “53 pouces de bec” (mouth). Englished, a trifle over 58 inches. Either we must credit him with having measured the circumference, —rather an unusual manner of classifying artillery, making the real 1911] DANA—NOTES ON CANNON. 159 caliber only about 18 inches, or else—but the alternative is too painful. The accounts of the fighting about Chioggia, 1380, between the Genoese under Pietro Doria, and the Venetians under their beloved _ Vittore Pisano, are well authenticated, and give a vivid picture of the power of these old bombards. January 22, the great bombard, a two-hundred pounder, was fired by the Venetians at the campanile _ of Brondolo; it knocked out a large piece of wall, and some of the flying stones struck and killed Pietro Doria, the Genoese commander, together with his nephew. The next day the same bombard brought down a still greater piece of the same campanile, killing 22 men; so that as an implement of slaughter, the clumsy thing was a success and endeared itself proportionately to the Venetians. = Before leaving the fourteenth century a few short notes might be added. : 4 The castle of Tannenberg, in Germany, was captured 1399. A = huge bombard belonging to the city of Frankfort a/M., was loaned _ to the besiegers. Tremendous difficulties were met and overcome in getting the gun into position, very close to the castle. The first pro- jectile stuck in the wall; the second passed through, and soon the defences were in ruins. These were never rebuilt. Excavations were made in 1849 and many stone balls were found. They varied in diameter from three inches to 3114 inches, the latter weighing 825 pounds, and igre taeem aad one of the shot for the “ Frank- furter Buechse.” Napoleon gives an jeventory of the Artillery of Bologna, 1381/97, in which stone balls of 1,000 pounds for bombards and mortars, together with iron balls of 1, 2, 3 and 6 pounds are men- tioned. A word about field-guns. Froissart,* speaking of the capture of the castle of la Roche sur Yon (1369) by the Black Prince, men- tions “several cannons and springalls with which the army was provided, and from long custom had always carried with them.” In the year 1382 the bumptious burghers of Bruges were engaged in one of their usual wars with their equally bumptious neighbors *“Die Burg beige ait und ihre Ausgrabung,” Hefner und Wolf, Frank- furt, a/M., 1850. “Chap. 268, Vol. 1. 160 DANA—NOTES ON CANNON. of Ghent, who took the field 5,000 strong with 200 “ ribaudequins.’ The latter were heavy built push-carts—Napoleon calls them “ wheel-barrows,” bearing in front two or three, sometimes mor of the small cannon of the day, with an ugly fringe of bristling lances projecting beyond. These disagreeable field-pieces we trundled along in front of the line of battle. The effect of two lin of “ribaudequins ” meeting and neutralizing each other must have given rise to some curious tactics in battle. In this case the 5,000 of sai formed themselves into a dense mass and with “ ribaude- quins ” in front, drove off 40,000 men of Bruges. At the battle of Roosebeke, November 27, 1382, where the Flem- ings were cut to pieces by the French and their leader Philip van Arteveld killed, Froissart states that the battle began by “a Bis. nonade with bars of iron and quarrels headed with brass.” This battle did not end the war, and a curious picture of the : ineffectiveness of the smaller cannon of the day is given by Lieut. Gen. Sir Henry Brackenbury, in his account of the siege of Ypres by the English and Flemings. The siege lasted from the eighth ¢ June, 1383, to the eighth of August. During that time a steady cannonade was maintained, but apart from interfering with the sleep of the good burghers of Ypres, not a soul was one whit worse. Two guns were advantageously posted in front of one the gates, and kept up a steady fire, in all 450 shots. When siege was raised those of Ypres were forced to admit that the g in question was in need of immediate repairs. Much danger to inhabitants was avoided by a thoughtful device; the besiegers c siderately heralded by a trumpet blast each discharge; this enah promenaders to step aside and avoid any possible annoyance 2 intruding cannon balls. | Another curious picture of by-gone days is given us in “Tssue Roll of the Exchequer for 1384,” in which the amount payments for the hire of cannon and cannoniers is given, making plain that private individuals often owned one or more can which they hired out like cabs. | Viollet le Duc mentions this same custom on the continent ; says that during the middle ages the engines of war were made by non-military workmen, and the same rule prevailed after the intro- 191.) DANA—NOTES ON CANNON. 161 _ duction of cannon. Not only did ordinary mechanics make the new artillery, they also served it; letting their cannon for hire as one lets carts and drivers ; and it was not until the death of Charles VII, _ 1461, that they formed companies of bombardiers and culveriniers, _ heavy and light artillery, like the companies of cross-bowmen and archers, gave to them military organization, and placed them under the command of the grand master of artillery. The fifteenth century was one of development, very important but less startling than its predecessor. The most marked advance was in cast bronze and iron guns. Pretty much any date after 1400 may be taken as the beginning of that phase of the smelter’s art. Erfurt claims 1377 as her begin- ning; it seems needlessly early, but no one can say her nay. There are two cast-iron guns in the Leipzig Museum, one between 1400 and 1420; another, less archaic, 1420 to 1430. _ Francis Grose says: = ‘It seems extremely strange, that none of our workmen attempted to - cast them, [cannon] till the reign of King Henry VIII. when in 1521, accord- ing to Stowe, or 1535 [Camden says], great brass ordnance, as canon (sic) and culverins, were first cast in England, by one John Owen, they formerly having been made in other countries; . . . 1543. . . [Stowe] . . . the King minding wars with France, made great preparations and provisions, as well of amunitions and artillery as also of brass ordnance; amongst which at that time, one Peter Bawd, a Frenchman born, a gun-founder, or maker of great ordnance, and one other alien, called Peter Van Collen, a gun-smith, both the King’s freedmen, conferred together, devised and caused to be made, certain mortar pieces, being at the mouth, from eleven inches up to nineteen inches wide. .. . and after the King’s return from Bullen [Boulogne], the said Peter Bawd by himself in the first year of Edward VI. [1547] . . . did also make certain ordnance of cast yron of diverse sorts and forms, as fawconets, falcons, minions, sakers, and other pieces. Chamber’d pieces for throwing stones, called cannon-perriers, port-pieces, stock-fowlers, sling-pieces, port- ingale-bases, and murtherers, were about this time much used in small forts and on shipboard.* _ Of course all these guns were cast hollow ; that is a core covered with clay, was suspended in the center of the mould while the metal ‘Was poured in. Despite all precautions it was very nearly impossible with the imperfect means then in use, to keep this core in place and true; cavities formed in the metal about it, and the scoria did not * Francis Grose, “ Military Antiquities,” London, 1788, IT., p. 383. 162 DANA—NOTES ON CANNON. [April 20, rise freely; certainly too much cannot be said in praise of the — founders who could cast such a gun as the serpentine of Charles — the Bold (say 1476), in the arsenal at Neuveville, near Bern; a cast- iron field gun some fifty-two inches long, and 2-inch bore. Machinery for boring cannon is said to have been invented by Lew, in Switzerland. It was introduced into France 1740/44, by Jean Maritz, born in Bern (1711-1790), who, after accepting office _ under the French was naturalized. Maritz seems to have been the first who thought of placing the gun horizontal and making it, not q the drill, revolve. a The Great Bombard, the characteristic gun of the latter half of the fourteenth century and the greater part of the fifteenth, was often, in its early days but a huge tube—“tuyeau de tonnerre.” It is possible that after the frequent burstings, the occasional sur- vivor noticed that these annoying accidents usually had their origin just in front of the chamber, about where the great stone ball was placed. The gun-maker would naturally strengthen this portion with much thicker bands, and doubtless he would soon deduce the fact that the strain decreased from the bursting point to the muzzle, then he would shape his gun to suit. The early gunners suffered terribly from the bursting of their guns. James II., King of Scot- land, was killed at the siege of Roxborough Castle, 1460, in this manner; 1470, a bombard near Paris burst, killing 14 men_ and wounding as many more. p It was long before the early gunner discovered (the figures are for a 4.25 inch caliber) that the proportional pressure on the bore increased alarmingly with the weight of the ball; 3.6 per square inch for stone; 10 for iron; 10.9 for bronze; 14.5 for lead. For the same caliber ; the cost for one round; 4-inch ball, charge 1-9 wt. of ball:— In money of to-day: with a stone ball, $1.25; iron, $4.75; ich $6.25; bronze, $9.00. s Stone balls had two bad defects—they were apt to shatter te pieces when used for breaching purposes against heavy masonry; and their rough surface greatly damaged the interior of the bom-_ bard; it was sought to correct these defects; the first by bands of | iron about the ball; the second by enclosing the ball in an envelop of : soft lead. tort.) DANA—NOTES ON CANNON. 163 The difficulties in the construction of these big bombards were greatly lessened by the system of forging the very heavy breech piece so that it could be screwed to the chase—and unscrewed, when de- sired ; the square holes for the levers, that worked like capstan bars on board ship, are conspicuous in such guns, usually in the rear ring of the chase, and at the rear end of the breech. In some bom- _ bards there were three divisions, greatly adding to convenience in ___ transportation. Here are the dimensions of the largest bombard that has come down to us, the one on the Place du Marché at Ghent. The lady is _ called “Dulle Grete.” which they tell me can be translated “ Mad a _ Meg.” Her caliber is 26” ; interior of chase 10’ 4”, or five calibers. Chamber caliber 10.23”, length 4’-6.16” ; five to six calibers. The exterior length 16’ 6”. The gun is built up of 32 longitudinal iron bars, 2.17” wide, 1.2” thick; these are soldered together. Over them are 41 iron rings, welded together and diminishing in thick- ness from the junction with the breech to the muzzle, except the three which form the muzzle moulding or swell. In addition, there are 20 bands, called “rondelles,” in two of which, the one at the extreme end of the breech and the one at the end of the chase, are holes for the levers used in unscrewing breech from the chase. - Curiously enough, the breech is not exactly in line with the chase, inclining slightly to the left—might possibly be a trifle trying for the right-hand side:of the chase after a few shots. Meg’s weight is 36,080 pounds, but painters of that day represent Flemish women of her class as distinctly heavy. The ball weighed 748 Ibs., and the _ powder charge was 88 lIbs., between 4 and % the weight of the pro- _ jectile. The range was about 3,000 yards, at least Meg claimed that, though her effective range could not have been more than three hun- dred yards. But, it is only fair to remember that in Nelson’s day _six hundred yards was long fighting range. The date of this huge, but rather useless engine of construction —destruction, I fear, would be gross flattery—is rather uncertain. The Flemings took it to the siege of Oudenarde, in 1452; a coat-of arms I found near the vent is that of the father of Charles the _ Bold, called “Philip the Good,” because he was bad. He warred from 1422 until 1467. 164 DANA—NOTES ON CANNON. [April 205, At the siege of Caen, in 1450—this is all quoted from an account in very bad Latin—the town was ringed about with twenty-four bombards, horrible to behold, for they were of such immense size that a man could sit in any one of them without bending his head! Possibly this old chronicler’s account was intended to fall into the the hands of the besieged—though of course,.cannon of that size would make very comfortable quarters for at least 48 men. ; + The preceding is but a dip into the doings of bombards; we have not touched on that most interesting one, “ Mons Meg,” at Edin- borough Castle; nor the “ Michelettes,” at Mont St. Michel, which - are of unique value; the “ Faule Mette,” was a cast bombard, a resi- i dent of Brunswick, but alas she has disappeared. | Wooden guns are of great interest, not only on account of the! frequent tendency to burst which must have been theirs, but also — on account of the very unreliable descriptions of them that we possess. i About 1450 trunnions appear (not the first, by any niearia) on some of the Burgundian guns, adding, of course, greatly to their efficiency, both in permitting more exact aiming and as a help fo resist the recoil. Omitting all other technical details it may be of - interest to take a look at the Burgundian, and afterwards at the — Freneh artillery. The town of Neuss, near the Rhine, and not far from Duessel dorf, was unsuccessfully besieged by Charles the Bold, 1474/5. The Burgundian artillery, then the best in Europe, was well represented there, and more or less careful accounts of it have come down to us. The following list is taken from Napoleon’s “Etudes”; un-— fortunately the calibers are omitted. Nine large bombards. Eight bronze bombards, 8 to 11 ft. long these had lions’ heads at the muzzle. Ten courtaux, 4 feet long, on wheels. These were a little like the carronade of just before 1800, to forty or fifty years after; there are accounts of courtaux which carried 60-Ib. balls and were used as siege pieces. 115 Ser. pentines, one of which was 13 feet long. Six serpentines of bronz with dragon heads at the muzzle, one of these guns was 8 feet long. Sixty-six serpentines 6 to 9 feet long. Fifteen others of the sam caliber weighing 4,000 pounds. 1911.] DANA—NOTES ON CANNON. 165 LONGUE SERPENTINE EN FER FORGE (Calibce > poures} (Calibre 2*") Fig 2 Echelle des Fig 1 203 t 2 2 > 2 5 ancien P de Pr Field-Guns of Charles the Bold, c. 1476. Fic. 1. More ancient build; wrought-iron bars, banded. : you know how to use it, and do not hesitate to use it. failure, though a magnificent one. Fic. 2. Cast- _ iron; modern looking gun and carriage. Fic. 3. “Portable bombard”; _ throws incendiary bomb or stone ball. : Like many other weapons, artillery is not of much service unless Charles the Bold was the last of the knights-errant, unless we include the chival- ric Don Quixote. Cannon had changed all that and Charles was a 166 DANA—NOTES ON CANNON. [April 20, March 2, 1476, was fought the battle of Granson; Charles of Burgundy had 20,000 men and his splendid train of field-artillery ; both these he proceeded to post as badly as he conveniently could. The Swiss always attacked in solid squares, impervious to cavalry — but just the food for cannon to devour. Nine thousand men, and absolute silence, save the word of command; instant death to who- ever faltered. A few shots took effect on that solid human mass, as it moved slowly towards the guns, each ball mowing down ten — or a dozen men; then a dip in the ground protected them, and the — balls passed over their heads. Now was the time for Charles to — have concentrated his artillery fire on the square and rent it to pieces, _ for his cavalry to drive off the field. Instead, time and again he — launched his magnificent gendarmerie against that bristling wall of : steel, those 16-foot spears held by sturdy mountaineers who knew ; not fear. Every attack failed, panic followed, and that splendid Burgundian artillery now adds interest to a score of Swiss museums. Napoleon III. and General Favé consider the artillery of Charles VIII. the beginning of that arm of the French service. Of course guns of earlier days still lingered on, but the newer ones took on ~ almost the form they were to retain for three hundred years. 38 Guns changed but little from 1500 until the astounding develop- — ment of today. Drake fought the Spaniard with almost the same guns that Nelson used at Trafalgar. e A better organization, and an improvement in tactics was made by Charles of France, before his great Italian campaign of 1495. On the other hand he was opposed by very different foes from the — heroes who defeated Charles of Burgundy at Granson, Morat and ~ Nancy. One may safely say that France easily, almost pityingly. scattered before her powerful guns the very worst troops the world contained at that time. Burgundy, on the contrary, had faced the bravest and best fighters history tells of. Swiss tactics, the old phalanx of Greece, steadily adhered to, soon became obsolete, and the system of rushing the guns with such unwieldy squares, received — its death blow on the days of Marignano, 1515; when, cannon to a the right of them, cannon to the left of them, cannon in front of si them, volleyed and thundered. Two days of carnage failed to shake the Swiss; but when Francis I., massed his artillery, and the — ¢ 1911.] DANA—NOTES ON CANNON. 167 Swiss attack was exposed to a cross fire that tore their squares to shreds,—just what Charles of Burgundy should have done at Gran- son, they sullenly fell back and the rule of the cannon began; alas, that its end is not yet in sight. Ten years later this same Francis was routed and captured at Pavia; one reason for it was that he stupidly masked his own _ guns by advancing his troops in front of them; another, that many _ of the Swiss of Marignano were then fighting on his side; but those _ days of Marignano and the slaughter were not forgiven ; so when the _erisis came, the Swiss, despite the despair and entreaties of their _ officers, threw down their arms and pretended to be cowards,—for a s Swiss it could only be pretence. AUTHORITIES. “ Ancient Cannon in Europe” (fourteenth century), Lt. Henry Brackenbury (Later Lt.-Genl.), Proceedings of the Royal Artillery Institution, 1865, - Woolwich, Eng. “Etudes sur l’Artillerie,’ Prince Louis Napoleon, Paris, 1846. “Etudes sur l’Artillerie,” General Favé, Paris, 1862. “Quellen sur Geschichte der Feurwaffen,” Leipzig, 1877. _ “Military Antiquities Respecting a History of the English Army,” Francis Grose, London, 1812. “Principles and Practice of Modern Artillery,” Lt.-Col. C. H. Owen, Lon- don, 1873. “Gunpowder and Ammunition,” Lt.-Col. H. C. L. Hime, London, 1904. _ “Naval Gunnery,” Capt. H. Garbett, London, 1897. _ “Catalog of the Musée d’Artillerie,” Paris. “Encyclopedia Britannica.” “Encyclopedie Larousse.” * Dict. du Mobilier Frangais,” Viollet le Duc. _ “Biographie Nouvelle.” Author’s notes, etc. MOREAU pe SAINT MERY AND HIS FRENCH FRIENDS. IN THE AMERICAN PHILOSOPHICAL SOCIETY. — By JOSEPH G. ROSENGARTEN. (Read April 20, rort.) Born at Fort Royal, Island of Martinique in 1750, dying at Paris in 1819, Moreau de St. Mery had a career characteristic of the storm, period through which he passed. Of a good family of Poitou, hi: father’s early death left him with little means. At nineteen he came to Paris, became a King’s gendarme, studied law, letters and mathe- matics. Returning to Martinique he became a lawyer at Cap Frangais, and in 1780 a member of the Upper Council of Saint Domingo. He classified the laws of the French Colonies of the Antilles; discov- ered and restored the tomb of Columbus, and sent many scientific papers and many curious archeological articles to the American Philosophical Society, and was elected a member in 1789. Returning to Paris as a member of the Constituent Assembl from Martinique, he was warmly welcomed by the scientific world in recognition of his frequent contributions to scientific societies. When the French Revolution broke out, he was elected President of the Electors of Paris, twice addressed Louis XVI. on their behalf, and was fond of boasting that for three days he had been King of Paris, and helped to secure for Lafayette the command of National Guard. : Elected Deputy from Martinique in 1790, he brought many colo- nial matters before the Constituent Assembly, and in 1791 became a member of the Judicial Council. Wounded in an attack by a maddened crowd, he took refuge in a country village in Normandy, escaped the guillotine and came to the United States. After a short stay in New York, he settled in Phil delphia in 1793, opened a book store at Front and Walnut Streets and became active in the Philosophical Society, attended its meet- ings regularly, contributing papers, making gifts to its collections, 168 i91t.] ROSENGARTEN—MOREAU pe SAINT MERY. 169 introducing many of his fellow exiles, some of them soon elected to the Society. Returning to France in 1799 and making use of his distant relationship to Josephine, wife of Bonaparte, he was em- ployed by Bonaparte in the preparation of a Maritime Code. Ap- pointed to the Council of State in 1800, he was sent in 1801 to Parma as Administrator of the Duchy of Parma, fulfilling his duties with moderation, but showing a lack of firmness and energy that cost him his position, and the enmity of Napoleon, who sent Junot. to replace him, and to end a threatened revolution by fire and sword. When he lost his place in the Council of State, he told Napoleon that his honesty need not be feared, for it was not contagious in that body. The Empress Josephine helped him, and afterwards he be- came historiographer of the Marine Department. He sold to the French government, for a pension from Louis XVIII., his large collection of historical papers, documents, maps, etc., often mentioned by recent historians. One unkind critic, who worked at making a calendar of his papers, says he sold to the gov- ernment not only the copies he had made, but many originals which he had taken from the files in his care. His printed works include one in six volumes, on “ The Laws and Constitutions of the French Colonies in the West Indies from 1550 to 1785.” Louis XVI. ordered a copy to be placed in each French colony in America. His “History of Saint Domingo” was translated by William Cobbett, then living in Philadelphia, and his list of subscribers included many notable Americans then in office and a large number of French exiles in the United States. He translated and published a pamphlet on “The Prisons of Philadelphia,” by Rochefoucauld Liancourt, reprinted in Paris and in Holland, and in one of Rochefoucauld Liancourt’s bulky six vol- umes of his “ Travels in the United States.” He had the honor of an eloge by Fournier Pescay printed in Paris in 1819 and the bio- graphical dictionaries give the dates of his various publications, of the offices he held and make mention of his best service: the collec- tion and preservation of an immense number of papers, maps, etc., relating to the French colonies in America, from their origin down to the French Revolution. Calendars of parts of them have been 170 ROSENGARTEN—MOREAU pe SAINT MERY. _ [April 20, printed by the Canadian Archive office and by the Wisconsin His- torical Society and in France. i His shop at Front and Walnut was the rendezvous of all the notable French exiles then in Philadelphia, and he entertained them - very modestly,—cooking his own simple meals in his rear office, and — sharing his good wine with them. . He figures in the “ Memoirs” of | Talleyrand and in the “ Travels ’’ of Rochefoucauld and in the books — on the United States by Brissot and Volney and other French writers. He translated and published two big quarto volumes on China by Van Braam, who had resided in that country as a member of a Dutch embassy. The book was dedicated to Washington. Van g | Braam became a merchant in Charleston in 1783 and was naturalized in 1784, then made his voyage to China, returned to Philadelphia in a 1796, bringing with him several Chinese servants, and a large collec- tion of paintings, drawings, maps and curios, which he exhibited in Philadelphia for several months, then kept in his house near Bristol, “China Hall” on the Delaware. In the appendix to the second volume of his book, there is a detailed account of his collection, — filling many pages. He too was elected a member of the Philo- sophical Society in 1797. Moreau de St. Mery printed a catalogue of the contents of his book store, of 72 pages, including many books in English, French, Italian, Spanish, Latin, German, Dutch, maps, music, and advertised - “‘a general business of stationers, booksellers and dealers in engray-— ings, a printing office and book bindery, to fill orders for books from Europe, deal in every kind of business on commission, and will not _ spare any care in studying to accomplish their enterprise intended to propagate and diffuse knowledge,” and at the end of the catalogue of books, etc., offered for sale, “ particular goods out of the book- sellers’ station, everything belonging to the Fleecy hosiery manu-— facture of New York, as foot and ankle socks, goutty mittens and — stockings, shirts with and without sleeves, drawers, muffs, ete. — elastic garters and gallices of different sizes.” Perhaps his field was too large, and the public not appreciative, for he failed for $5,000, and Philadelphia lost the advantage of such a bookseller, printer and publisher, as well as philosopher, author and translator. 911.] ROSENGARTEN—MOREAU pe SAINT MERY. 171 _ No doubt the same industry and energy led him to make the ve collection known by his name, now in Paris, of original docu- ts, copies, maps, etc., filling 287 volumes, bought by the French _ government, and now in its great archives for the use of students of colonial French history. _ That St. Mery was well thought of in Philadelphia, during his _ residence here is attested by the long list of subscribers to his book on Saint Domingo, including Vice-President John Adams, Adet, the French minister, Benjamin Franklin Bache, William Bingham, Thomas Bradford, Samuel Breck, Rev. Dr. Collin, Alexander James Dallas, P. S. Duponceau, Dupont, of Wilmington, Rufus King, Dr. Logan, Noailles, Timothy Pickering, Rochambeau, Talleyrand, John Vaughan, Volney, and many notable French exiles both in Phila- lelphia and elsewhere in the United States. Many of them were elected members of the American Philosophical Society, and its minutes show that its meetings were frequently attended by Talley- rand, Rochefoucauld, Volney, Van Braam, and its library has many books, the gifts of St. Mery and his fellow exiles. In a recent biography of Talleyrand we are told that when he landed here in 1794, it was the finest city in the United States, full of life, everywhere new buildings and work on them going on, the Streets full of elegant equipages, crowded with men of business, orkmen and sailors. _ Chateaubriand speaks of the beauty of the Quakeresses. Every Stranger from Europe was welcomed by the wealthy merchants,— ife was very expensive, board $8 to $12 a week, without fire, light or wine ; a negro servant cost $10 to $12 a month even with food and washing. Emigrés of all political creeds found a Noah’s ark of _ fefuge in Philadelphia. Talleyrand’s arrival was quite an event; 7 he found old friends, old soldiers of Lafayette, fellow members of ; the constituent assembly, among them Blacon, who had been deputy _ from Dauphine and one of the intermediaries between Mirabeau and the King. Hamilton gave him a warm welcome, but Fauchet, the French Minister, prevented Washington from receiving him, and Washington wrote to Lord Lansdowne, explaining why his letter of introduction did not enable him to meet Talleyrand. However, he 172 ROSENGARTEN—MOREAU pe SAINT MERY. [April did not busy himself with politics, but at once began speculating in land, then the great money-making business. The Scioto Company — was then all the vogue in Paris. The Holland Land Company w buying right and left. LaForest, the French Consul General, had bought an estate in Virginia in 1792. Noailles and Omer Talon in association with Robert Morris had bought large tracts of land on the Susquehanna for a colony of French royalist exiles, offering land, which had cost them 15 sous an acre for 6 francs, as a refuge from France. Talleyrand urged Mme. deStael and his friends in Europe to send money for investment, and he proposed buying land in Maine from General Knox. He told Moreau de St. Mery that he had a — plan for settling in Louisiana, and was a frequent visitor at St. Mery’s book store, meeting there his old friends and fellow exiles, Fayettists, Girondists, Constituents, Jacobins, Royalists, one of ther Count de Moré, says’ “ wandering like ghosts, full of regrets, lost hopes and disappointment over their shattered political careers.” — Moreau de St. Mery often spoke of the three days in 1789 w : as president of the Electoral College he was King of Paris. while the others were bewailing their hard fate, Moreau was bu with his shop and his books, and Talleyrand wrote to Paris schemes for revictualling Paris, starved by the Reign of Ter crying for bread, by ships loaded with rice, grain and fish, named t best merchants to deal with, and on the strength of his se secured the long-sought permission to return to France, and began that career of success which carried him safely through Republic, the Directory, the Empire, the Bourbon restoration at into the reign of Louis Philippe. oe Other Frenchmen had planned a great French colony,—twent four men, mostly young noblemen, had joined in forwarding Jo Barlow’s scheme of a great settlement on the Ohio,—the Scioto pany was organized, to buy 24,000 acres,—d’Epresmenil, their 1 lost his life on the guillotine ; Marnesia, after a tour through Americ returned to France, and with Lally Tollendal, Mounier and Malou 1¢ lost touch with their colony in the midst of the great events in the own country. — xgtt.] ROSENGARTEN—MOREAU pe SAINT MERY. 173 _ One of Moreau de St. Mery’s friends and visitors, Rochefoucauld Liancourt, wrote an account of the prisons of Philadelphia, which was printed by Moreau in French and in English in Philadelphia, (he was the translator), and later it was published in Paris, and in Dutch in Holland, and later still was made part of one of his six volumes describing his travels in the United States. Rochefoucauld spent four years in the United States and describes _ in great detail his experiences in the northwest and north, in Canada, _ in Maine, in the south, and in New York, New Jersey and Penn- 5 sylvania. Talleyrand from Philadelphia wrote to Mme. de Genlis, “ Roche- _ foucauld is here, making notes, asking information, writing, and _ more a questioner than Sterne’s curious traveller; he wants to see and know everything,” in his eager search for the truth. He met Knox, Sullivan, Jefferson, John Adams, Priestley, Livingston and Kosciusko. He appealed to Washington to intercede for the release of Lafayette from Olmutz. His inquiries included politics, consti- tutions, judicial organizations, army, agriculture, industries, statis- tics, charities, education. In Georgia he studied cotton and indigo plantations ; he condemned slavery and argued for the education of the negroes to prepare them for freedom; in Niagara and the great forests ‘he foresaw the sources of future industries. He established in France on his return societies like the Pennsylvania Prison So- ciety, and took home much that he had learned in the United States, which he introduced in France, useful reforms that made him a real philanthropist. . Another French settler in or near Philadelphia, Pierre Legaux, was elected a member of the American Philosophical Society in 1787. A Counsellor of Parliament, a member of the Academy of Arts and Sciences and of several foreign academies, employed in the French West Indies, he came to Philadelphia about 1786 and made his mark as a representative of French culture and scientific ability and by his charm of manner. He bought land on the Schuylkill near Conshohocken and planted vineyards. Washington and Mifflin and other notable men visited them and approved his enterprise. Jeffer- son, Genet, Brissot, Audubon, Wistar, were among those whose visits 174 ROSENGARTEN—MOREAU pe SAINT MERY. _ [April . and encouragement he recorded in his diary. He tried to get Jeffer- son to recommend to Congress protection for his infant industry. In 1791 he offered his country house to Washington as a home during the session of Congress and “hoped the country which owes its” liberty to your wisdom and military talent will owe her wine ee your generosity.” ee In 1793 the Legislature of Pcduapiicis chartered a company to promote the cultivation of vines, with a capital of $20,000 in $2 shares ; in 1800 the stock was fixed by a law passed by the Legisla ture at $1 a share down, and the balance of the $20 in easy instal- ments. Later he advertised that apprentices, black or white, would | be received, with terms of payment, and the promise of a gift o vines that they could take home and start the industry wherever they lived. In 1802 the company received its charter and organized. Among the stockholders were Thomas McKean, Robert Mor Genet, Duponceau, Stephen Girard, Alexander Hamilton, Aaron Burr, Jared Ingersoll, Muhlenberg, Bartram and other notable peopl of the 385 subscribers to the stock of the Pennsylvania Vine Co Legaux was elected superintendent at a salary of $600 a year with residence and living at the farm. Expenses soon outran receipts, the managers quarreled with Legaux, litigation brought ruin, and he, harassed, worried, disappointed, became a mere servant where had once been a genial host, finally succumbed and broken in spi died in 1827, and was buried at Barren Hill. Thus sadly closed another one of the frequent failures of French il oo in th United States.* Moreau de St. Mery kept a journal, cited by Pichot in his “ Souve- nirs intimes de Talleyrand,” in which he speaks of Talleyrand’ frequent visits to his book store, meeting there Noailles, Rochefou- cauld, Omer Talon, Volney and others less famous. While the host dined meagerly on rice and milk cooked in store, Talleyrand enjoyed drinking his own old Madeira, and was the life of the party. When Blacon called him monseigneur z the company burst into a hearty laugh. Talleyrand urged Na leon to erect a statue of Washington in Paris and to give France * Philadelphia Press, September 9, 1899, article by Samuel Gordon Smyth ROSENGARTEN—MOREAU pe SAINT MERY. 175 _ same perfect religious freedom that he saw practiced in the United _ States and he also advised the sale of Louisiana to the United States as a method of strengthening the ties between the two countries. One would like to see the journal kept by Moreau de St. Mery _ during his residence in Philadelphia. Did he in his palmy days as Ea member of the Council of State under the Empire or in the time a of his modest clerkship in the Marine Department, meet his old 3 visitors at his book store in Philadelphia, Louis Philippe, Talley- _ rand and Rochefoucauld and Volney and the other exiles now re- _ stored to their old prosperity, and did they recall the meetings of the _ American Philosophical Society and their attendance and share in 4 them? His large collection of historical papers, now rescued from oblivion by calendars by and for the American students of history, perpetuates his name and memory and services, more than do the volumes he wrote and printed and published at his book store at Front and Walnut Streets. : The latest historian of the French Revolution, Aulard, frequently mentions Moreau de St. Mery and his share in it, and refers to the collection of historical documents. His name does not figure in Dr. Mitchell’s capital novel, “ The Red City,” with its picturesque account of the French exiles living here in the closing years of the eighteenth century, nor in Kipling’s picturesque story of Philadelphia at that time. All the more reason therefore for an attempt to recall the memory of the French exiles who were members of the American Philosophical Society and especially of that one who figures most often and most usefully in its records of that time, Moreau de St. Mery. ___ Of the other French exiles during their residence in Philadelphia, there is occasional mention, as for instance in Talleyrand’s “ Me- moirs.” His two papers on the United States and the relations between France and this country, read before the French Institute, ere no doubt largely inspired by what he heard at the meetings of e American Philosophical Society, and his share in the sale of Louisiana to the United States helped to secure that vast territory for the future growth of the young republic and its ultimate great development. 176 ROSENGARTEN—MOREAU pe SAINT MERY. In the report of M. E. Richard on the Moreau de St. Mery coll tion, printed in the supplement to Dr. Brymner’s Report on Ca dian Archives for 1899 (Ottawa, 1901), he says it was stored in archives of the Marine at Versailles up to 1887, then removed to’ Ministere des Colonies, and stored in the attic of the Louvre. T were then fearful of the great risk of fire, but were se! removal to other quarters. In the reports of 1883-85 and 1887, mention is made of. volumes in the collection of Moreau de St. Mery, some forty of wh: relate to Canada, others to Louisiana and the French islands’ America. These belonged formerly to the Colonial Archives of Marine; of the collection headed Moreau de St. Mery sevente volumes contain description, etc., of the colonies, including a series memorials on Canada, 3 volumes are on the religious missions Canada, 12 volumes on Newfoundland, 12 volumes containing ro: instructions to governors, and decrees relating to Canada, 1 registers on Canada, Acadia, etc., 6 volumes on civil status Canada, 34 volumes on Louisbourg; an analysis was made of volumes of the Moreau de St. Mery collection for the Cane Archives. é It is open to the objection that “there is no strict order folloy in the compilation; it contains but a limited number of documer or even extracts from documents. It is difficult to understand 1 dominant idea of this collection.” ‘ This collection is, nevertheless, most valuable, for it contain considerable number of important papers, both transcrip.s and ori nals, not to be found elsewhere. On p. 5 of Richard’s Report, in a footnote, it is said Moreau d Mery, born in Martinique in 1750, studied law in Paris and pra in St. Domingo, where he became a member of the Superior Ce of the Island. Entrusted by Louis XVI. with the compiling of colonial code, he published in Paris “Les Lois et Constitutions Colonies Francaises de l’Amerique sous le Vent.’ Representi Martinique in the Constituent Assembly, he drafted the report of Committee on the Colonies. Forced by political events to leavt France, he fled to Philadelphia, where he remained from 1793 ROSENGARTEN—MOREAU pe SAINT MERY. 177 798, employing himself as a bookseller and publisher. He there blished his “ Description de la partie Espagnole de St. Domingue,” which he signed “‘ Moreau de St. Mery, member of the Philosophical ' Society of Phila.” He also translated or edited foreign works, and _ among them VanBraam’s “ Voyage to China.” Having returned to | France on the 18 Brumaire, he was, through his relationship with - Josephine de Beauharnais, appointed in 1800 to the position of Historiographer of the Marine. Napoleon appointed him to the _ Council of State, in view of his knowledge of eolonial affairs. In q 1802 he was administrator of Parma and Guastalla, but lost favor and was removed in 1806. He died poor and in receipt of a pension _ from Louis XVIII. _ While entrusted with a mission in St. Domingo, as publisher in liladelphia and historiographer in Paris, we find him everywhere observer and a worker, taking notes on everything. His collec- tion of manuscripts comprises 287 large volumes, and was purchased by the state after his death, that is to say the government had to ‘pay not only for the transcripts he had caused to be made, but even for the originals he had appropriated. _ Persons who take a special interest in the social and religious condition of the country, the disputes and conflicts between the uthorities will find in the Moreau de St. Mery collection far more than they could find in any other series. : That Moreau de St. Mery did a good work in preserving and making his collection is shown by the statement (in Richard’s Re- port, p. 8, etc.), that the Archives of the Ministry of Marine were utterly neglected that the precious papers were used during five eeks of the winter in 1793, as fuel to feed the stoves of the post of the Garde Nationale in the building where the archives were kept, and i in 1830 an employee gave up the archives to pillage and sold, weight for his own profit, whole piles of documents, bought by utograph collectors. __ Thanks to the suggestion of Prof. Cleveland Abbe, I found in the Monthly Weather Review for February, 1906 (Washington, Weather Bureau, 1907), at pp. 64, etc., ina notice by C. Fitzhugh Talman of the -U. S. Weather Bureau, the following: “ Foremost among the early 178 ROSENGARTEN—MOREAU pe SAINT MERY. [April writers upon the island of Santo Domingo, was Mederic Louis Moreau de St. Mery, who produced three voluminous works t the French possessions in the West Indies. Born at Fort Royal Martinique, in 1750, he passed his early manhood in Haiti, a settled at the then capitol of the colony, Cap Frangais (now Haitien). He held an important office in the administration of th colony, and also, under a commission from Louis 16th, travelled tensively through the French West Indies, collecting material for 4 work published in 1785, under the title ‘‘ Lois et Constitutions ¢ Colonies Francaises de l’Amerique sous le vent, de 1550 a 1785.” Returning to France he took an active part in the French Revolu until obliged to flee from his political enemies to the United Sta It was during a period of exile in the latter country that he publish two works descriptive of the island of Santo Domingo, one dev to the Spanish part of the island, the other to the French part. P lished by himself in Philadelphia in 1797, it was republished in Paris in 1875 by Morgand in 2 vols. 8vo. It is to this day regarded the Haitians as the highest authority upon the physical geogra of their country and is quoted at length in the latest Haitian gaze (Ronzier Dic. geog. et admin. d. Haiti, Paris, 1899). Mr. Tal reproduces St. Mery’s chart of the Island, and a full abate of h description of its meteorology. Moreau de St. Mery was active in the Philadelphia Society of Francais, and in the Library of the American Philosophical Soc there is the Ist vol. of its Proceedings,—no 2nd or later volume preserved,—it shows that Moreau de St. Mery was the leadin spirit in its activities. That his meteorological observations of Sa Domingo during his residence there in the eighteenth century, sh be found of value today, is but another proof of his useful activ His chief monument however is ‘his collection, bearing his name, ¢ original documents on the French in America, and by it now made known to students in the pages of Aulard, Be Thwaites and other historians. THE NEW HISTORY. By JAMES HARVEY ROBINSON. (Read April 22, 1911.) I propose to discuss in this paper the value of historical study. The question has long haunted me and certainly merits a more care- ful consideration than it has, so far as I can discover, hitherto re- ceived. It will be impossible to do more here than to analyze the problem and briefly state the general conclusions which that analysis suggests. _ The older traditional type of historical writing was narrative in character. Its chief aim was to tell a tale or story by setting forth a succession of events and introducing the prominent actors who par- ticipated in them. It was a branch of polite literature, competing with the drama and fiction, from which, indeed, it differed often only in the limitations which the writer was supposed to place upon his fancy. As Professor McMaster has recently said: “It was by no mere accident that Motley began his literary career with a novel called “ Merry-Mount,” and Parkman his with “ Vassall Morton.” These bespoke their type of mind. The things that would interest them in history would be, not the great masses of toiling men, not the silent revolutions by which nations pass from barbarism to civiliza- tion, from ignorance to knowledge, from poverty to wealth, from eebleness to power, but the striking figures of history, great kings and queens, the leaders of armies, men renowned for statescraft, and the dramatic incidents in the life of nations. Each must have his hero and his villain, his plots, conspiracies and bloody wars. Just as Froude had his Henry VIII.; just as Macaulay had his William III., Carlyle his Robespierre and Cromwell, and Thiers his Napoleon, so Motley had his William of Orange and Philip of Spain; Prescott his Cortez, Pizarro, Ferdinand and Isabella; and Parkman his Pon- _tiac, Frontenac and La Salle. History as viewed by writers of this _ PROC. AMER. PHIL. SOC., L, 199 L, PRINTED JUNE 27, IQII. 179 180 ROBINSON—THE NEW HISTORY. ‘school is a series of dramas i in each of which a few great men per- form the leading parts and use the rest of mankind as their instru- ments.” The commonly pevepied definition of history was long, “a record of past events” and these, naturally, the most startling and romantic and the best adapted for effective literary presentation. Doubtless there was some serious effort to describe conditions and institutions, since they formed the necessary setting for the events and anecdotes; sometimes they would even be assigned a place their own intrinsic merits; but what may be called the epic ideal o history prevailed until perhaps fifty or sixty years ago when, owing to the influence of the modern scientific spirit, a very fondanee revolution became apparent. Now let us review, by way of preliminary, what were deemed the advantages of the study of history of this older type. Lore Bolingbroke in his “ Letters on the Study of History,” written abil 1737, says: “An application to any study that tends neither to r us better men and better citizens, is at best but a specious and i genious sort of idleness; . . . and the knowledge we acquire by it a creditable kind of ignorance, nothing more. This creditable kind of ignorance is, in my opinion, the whole benefit which the gene of men, even the most learned, reap from the study of history: < yet the study of history seems to me of all others the most proper train us up to private and public virtue.” History, he quite pre erly says, is read by most people as a form of amusement, as th might play at cards. Some devote themselves to history in ord to adorn their conversation with historical allusions,—and the arg ment is still current that one should know enough of the past te understand literary references to, noteworthy events and persor The less imaginative scholar, Bolingbroke complains, satisfies him: self with making fair copies of foul manuscripts and explaining he r words for the benefit of others, or with constructing more or fantastic chronologies based upon very insecure data. Over these Bolingbroke places those who have perceived that history after all only “ philosophy teaching by example.” For “the e >“The Present State of Historical Writing in America,” reprinted f the Proceedings of the American Antiquarian Society for October, 191 Worcester, 1910, p. 18. 1911] ROBINSON—THE NEW HISTORY. 181 ples which we find in history, improved by the lively descriptions and the just explanations or censures of historians,” will, he believes, have a much better and more permanent effect than declamation, or the “dry ethics of mere philosophy.” Moreover, to summarize his argument, we can by the study of history enjoy in a short time a wide range of experience at the expense of other men and without risk to ourselves. History enables us “to live with the men who lived before us, and we inhabit countries that we never saw. Place is enlarged, and time prolonged in this manner: so that the man who applies himself early to the study of history may acquire in a few years, and before he sets foot in the world, not only a more _ extended knowledge of mankind but the experience of more cen- _ turies than any of the patriarchs saw.” Our own personal expe- rience is doubly defective; we are born too late to see the beginning, and we die too soon to see the end of many things. History sup-~ plies in a large measure these defects. There is of course little originality in Bolingbroke’s plea for his- tory’s usefulness in making wiser and. better men and citizens. Polybios had seen in history a guide for statesmen and military commanders; and the hope that the conspicuous moral victories and defeats of the past would serve to arouse virtue and discourage vice has been urged by innumerable chroniclers as the main justification of their enterprises. To-day, however, one would rarely find a historical student who would venture to recommend statesmen, warriors and moralists to place any confidence whatsoever in histor- ical analogies and warnings, for the supposed analogies usually prove illusive on inspection and the warnings, impertinent. Whether or no Napoleon was ever able to make any practical use in his own campaigns of the accounts he had read of those of Alexander and Cesar, it is quite certain that Admiral Togo would have derived no useful hints from Nelson’s tactics at Alexandria or Trafalgar. Our ‘situation is so novel that it would seem as if political and military precedents of even a century ago could have no possible value. As for our present “anxious morality,” as Maeterlinck calls it, it seems equally clear that the sinful extravagances of Sardanapalus and Nero, and the conspicuous public virtue of Aristides and the Horatii, are alike impotent to promote it. 182 ROBINSON—THE NEW HISTORY. [April we In addition to the supposed uses of history mentioned by Boli broke there was the possibility of tracing the ways of God to Augustine had furnished the first great example of this type of ! rative in his “ City of God” and thereafter history had very o monly been summoned to the support of Christian theology. Bos- suet, writing for the Dauphin in the latter part of the seventeenth century, says: “ Mais souvenez-vous, Monseigneur, que ce long chainement des causes particuliéres qui font et défont les empi dépend des ordres secrets de la Providence. Dieu tient du plus ha des cieux les rénes de tous les royaumes; il a tous les coeurs en s; main; tantot il retient les passions, tantot il leur lache la bride, par la i! remue tout le genre humain. Veut-il faire des conq rants; il fait marcher 1’épouvante. devant eux, et il inspire 4 eux a leurs soldats une hardiesse invincible. Veut-il faire des 1é; teurs; il leur envoie son esprit de sagesse et de prévoyance; il fait prévenir les maux qui menacent les états, et poser les fon ments de la tranquillité publique. Il connoit la sagesse humaine, toujours courte par quelque endroit; il l’éclaire, il étend ses vues, ¢ ) puis l’abandonne 4a ses ignorances; il l’aveugle, il la précipite, il 1 confond par elle-méme; elle s’enveloppe, elle s’embarrasse dans propres subtilités, et ses précautions lui sont un piége. Dieu ex par ce moyen ses redoutables jugements, selon les régles de sa j toujours infallible.”? It was assumed by such writers as Boss that in spite of the confessedly secret and mysterious character God’s dispensations it was nevertheless quite possible for the theologian to trace them with edifying confidence and interpret as divine sanctions and disapprovals, blessings and punishm trials and encouragements. For various reasons, which it is essary to review here, this particular method of dealing with the p and deriving useful lessons from it finds few educated defendeay the present day. r In the eighteenth century a considerable number of “ philows of history” appeared and enjoyed great popularity. They wer outcome of a desire to seize and explain the general trend of mi past. Of course this had been the purpose of Augustine and Bossu ” ?“Tjiscours sur l’histoire universelle,” concluding chapter. ROBINSON—THE NEW HISTORY. 183 but Voltaire devoted his “ Philosophie de l'histoire” (1765) mainly to discrediting religion as commonly accepted; and instead of offer- ing any particular theory of the past he satisfied himself with pick- ing out what he calls les vérités utiles. He addresses Madame du Chatelet in the opening of his “ Essai sur les Moeurs et l’esprit des nations” as follows: Vous ne cherchez dans cette immensité que ce ui mérite d’étre connu de vous; l’esprit, les moeurs, les usages des nations principales, appuyés des faits qu'il n’est pas permis d’ignorer. Le but de ce travail n’est pas de savoir en quelle année un prince indigne d’étre connu succéda a un prince barbare chez une nation ssiére. Si l’on pouvait avoir le malheur de mettre dans sa téte suite chronologique de toutes les dynasties, on ne saurait que des mots. Autant il faut connaitre les grandes actions des souverains qui ont rendu leurs peuples meilleurs et plus heureux, autant on peut ignorer le vulgaire des rois, qui ne pourrait que charger la _ mémoite. . . . Dans tous ces recueils immenses qu’on ne peut em- brasser, il faut se borner et choisir. C’est un vaste magazin ou vous prendrez ce qui est a votre usage.* Voltaire’s reactions on the past were naturally just what might have been expected from his attitude toward his own times. He drew from “le vaste magazin” those gs that he needed for his great campaign, and in this he did well, however uncritical his criticism may at times seem to a modern historical student. Herder in his little work, “ Auch eine Philosophie der Geschichte zur Bildung der Menschheit. Beitrag zur vielen Beitragen des Jahr- derts” (1774), condemns the general lightheartedness and super- ficiality of Voltaire and other contemporary writers who were, he thought, vainly attempting to squeeze the story of the universe and into their puny philosophic categories. Ten years later he wrote his larger work, “Ideen zur Geschichte der Menschheit,” in which he strove to give some ideal unity and order to the vast historic process, beginning with a consideration of the place of the earth among the other heavenly bodies, and of man’s relations to the egetable and animal kingdoms. “If,” he says, “there be a god in [April 22, vas 184 ROBINSON—THE NEW HISTORY. less excellent and beautiful than those by which all the celestiai bodies move. Now as I am persuaded that man is capable of know- ing, and destined to attain the knowledge of, everything that he ought — to know, I step freely and confidently from the tumultuous scene: through which we have been wandering to inspect the beautiful and sublime laws of nature by which they have been governed.” Hu-— manity is the end of human nature, he held, and the human race _ is destined to proceed through various degrees of civilization in var- ious mutations ; but the permanency of its welfare is founded solely and essentially on reason and justice. But it is a natural law that “if a being or system of beings be forced out of the permanent position of its truth, goodness and beauty it will again approach it by its internal powers, either in vibrations or in an asymptote, a : out of this state it finds no stability.”* Herder formulates from time to time a considerable number of other “laws” which he betiey emerge from the confusion of the past. Whatever we may think o these “laws” he constantly astonishes the modern reader not o by his penetrating criticism of the prevailing philosophy of his ti but by flashes of deep historical insight. He is clearly enough t forerunner of the “ Romantic” tendency that culminated in Hegel’ celebrated “ Philosophy of History” in which the successive migra tions and national incarnations of the Weltgeist are traced to its final and highest medium of expression, the German people. These genial speculations of the philosophers of history rested usually upon no very careful study of historical sources and thei: conclusions seem to us now very hazardous, even if we grant the cor- rectness of the data upon which they relied. It was inevitable that the historical students who, about the middle of the nineteenth cen- tury, commenced to feel the influence of the general scientific spirit of the period, should begin to look very sourly upon the earlier attempts to bring order and beauty out of a mass of historic asser- tions which were so commonly either erroneous or unproved, and t establish laws for events which one could not be sure had ever hap- pened. The reaction against the dreams of the philosophers of his- tory was, and is still, very clear. What may be called, for convéni- ence, the “scientific” modern school of historians believe that histo: ‘Opening sections of Book XV. 1gtt.] ROBINSON—THE NEW HISTORY. 185 _ like all other forms of scientific research, should be pursued first and - foremost for its own sake. The facts must be verified and classified by the expert, without regard to any possible bearing which his _ discoveries may have upon our attitude toward life and the proper way of conducting it. Attempts to draw lessons from the past have, it is plausibly maintained, produced so reckless a disregard of scien- tific accuracy and criticism, that the prudent historian will confine himself to determining “ how it really was ’—an absorbing and deli- _ cate task which will tax his best powers. Along with more exacting criticism and the repudiation of super- natural considerations and explanations came a revulsion against the older epic or dramatic interest in the past. The essential interest and importance of the normal and homely elements in human _ life became apparent. The scientific historian no longer dwells _ by preference on the heroic, spectacular, and romantic episodes, but strives to reconstruct past conditions. This last point is of such importance that we must stop over it a moment. History is not infrequently still defined as a record of past events and the public _ still expect from the historian a story of the past. But the conscien- tious historian has come to realize that he cannot aspire to be a good _ story teller for the simple reason that if he tells no more than he has good reasons for believing to be true his story is usually very fragmentary and uncertain. Fiction and drama are perfectly free to conceive and adjust detail so as to meet the demands of art, but the historian should always be conscious of the rigid limitations ‘placed upon him. If he confines himself to an honest and critical statement of a series of events as described in his sources it is usu- ally too deficient in vivid authentic detail to make a presentable story. The historian is coming to see that his task is essentially dif- ferent from that of the man of letters. His place is among the scien- tists. He is at liberty to use only his scientific imagination, which is surely different from a literary imagination. It is his business to make those contributions to our general understanding of mankind in the past which his training in the investigation of the records of past human events especially fit him to make. He esteems the events he finds recorded not for their dramatic interest but for the light that they cast on the normal and prevalent conditions which i a pal a ali igi a aa 186 ROBINSON—THE NEW HISTORY. gave rise to them. It makes no difference how dry a chronicle may be if the occurrences that it reports can be brought into some assign- — able relation with the more or less permanent habits and environment 4a of a particular people or person. If it be the chief function of his- tory to show how things come about—and something will be said of this matter later—then events become for the historian first and foremost evidence of general conditions and changes affecting con- — siderable numbers of people. In this respect history is only fol- — lowing the example set by the older natural sciences—zoology dwells _ on general principles not on exceptional and startling creatures or — on the lessons which their habits suggest for man. Mathematics no longer lingers over the mystic qualities of numbers, nor does the astronomer seek to read our personal fate in the positions of the planets. Scientific truth has shown itself able to compete with fiction, — and there appears to be endless fascination for the mind in the con- templation of what former ages would have regarded as the ra vulgar and tiresome commonplace. 7 In addition to the characteristics of modern history just enum- erated two great historical discoveries of the latter half of the nineteenth century have served still further to revolutionize our atti- tude towards the past of mankind. Curiously enough neither o these discoveries are due to historians. I refer to the well substa tiated fact that man is sprung from the lower animals, and secondly, that he has in all probability been sojourning on the globe for sev- eral hundreds of thousands of years. These discoveries have grave- ly influenced all speculations in regard to the earlier history of our race and have placed the so-called “historical period” in a new setting. The historian no longer believes that he knows anything about man from the very first but realizes that what is commonly called history comprises only a very recent and very brief period in man’s develop- ment. All history is modern history from the standpoint of pre historic anthropology. Lastly, a group of anthropological, psycho- logical and social sciences have made their appearance during t past fifty years which are furnishing the historian with many n notions about man and are disabusing his mind of many old misap- prehensions in regard to races, religion, social organization, and th psychology of progress. The older historians used such words as ROBINSON—THE NEW HISTORY. 187 race, human nature, culture, religion, church, people, Renaissance, _ Reformation, Revolution, almost as if they were the names of ani- mistic forces. These terms must be analyzed and reinterpreted in _ the light of the newer sciences of man. The kind of history, accordingly, the practical value of which we _ shall attempt roughly to estimate, and which for convenience sake "we may call the “new” history, is scientific in its methods, exact- ' ing in regard to the inferences it makes from its material; it rejects supernatural explanations and an anthropocentric conception of the universe; it studies by preference the normal and long enduring _ rather than the transient and exceptional; it accepts the descent of man from the lower animals, many of whose psychological traits he shares ; it recognizes that man has lived on the earth for not merely five thousand but perhaps for five hundred thousand years; it avails _ itself, when fully abreast of the time, of all the suggestions and criti- cisms that are constantly being contributed by the newly developed Sciences of anthropology, comparative, social and functional psychol- ogy, comparative religion, etc.» So much for the attitude of mind of the modern historian who realizes the changes which have over- taken his subject during the past fifty or sixty years. _ But if “history” be re-defined as no longer a record of past events but the attempt to describe with all possible scientific pre- cision what we know of the nature and conditions of human institu- tions, conduct and thought in the past, does not the term become hopelessly vague—as vague at least as the term natural science? Does not the historian sacrifice his only obvious clue to the past hen he gives up tracing a succession of conspicuous events, for only these lend themselves to an obvious and orderly selection and arrangement? Every human interest and achievement has its his- ry, every accomplished, and every vain dream. It would seem as every attempt to deal with the past must necessarily imply an rbitrary selection dictated by the investigator’s particular humor and tastes. This situation is still disguised by the continued pop- ularity of a standard variety of history, mainly political, dynastic and military, transmitted to us from the past and taught in our a * See “The Relation of History to the Newer Sciences of Man” in The Journal for Philosophy, Psychology and Scientific Methods, Vol. VIIL., No. 2 March, 1911, where I have elaborated this point. 188 ROBINSON—THE NEW HISTORY. \ schools and colleges and presented to the adult public in many » known older and newer treatises. In order to appreciate the arbitrary nature of the selection of — historic facts offered in these standard text books and treatises, 1 us suppose that a half dozen alert and well trained minds had ne happened to be biased by the study of anv outline of history and by some happy and incredible fortune never perused a “ standard historical work. Let us suppose that they had nevertheless learns a good deal about the past of mankind directly from the vast ra of sources that we now possess, both literary and archzeologica Lastly, let us assume that they were all called upon to prepare in pendently a so-called general history, suitable for use in the hig schools. They would speedily discover that there was no sing’ obvious rule for determining what should be included in their revi of the past. Having no tradition to guide them, each would what he deemed most important for the young to know of the p Writing in the twentieth century, they would all be deeply influer by the interests and problems of the day. Battles and sieges and courts of kings would scarcely appeal to them. Probably it wo occur to none of them to mention the battle of Issus, the Sa ni wars, the siege of Numantia by the Romans, the advent of Had the Italian enterprises of Otto I., the six wives of Henry VIII. invasion of Holland by Louis XIV. It is tolerably safe to assu that none of these events, which are recorded in practically all of ot manuals to-day, would be considered by any one of our writers ; he thought over all that man had done, and thought, and sufi and dreamed, through thousands of years. All of them would that what men had known of the world in which they lived, o thought to be their duty, or what they made with their hands, nature and style of their buildings, public and private, would < them be far more valuable to rehearse than the names of rulers and the conflicts in which they engaged. Each writer y accordingly go his own way. He would look back on the pa explanations of what he found most interesting in the present would endeavor to place his readers in a position to parti intelligently in the life of their own time. The six manuals completed would not only differ greatly from one another but wv 3911.) ROBINSON—THE NEW HISTORY. 189 have little resemblance to the fable convenue which is currently ac- cepted as embodying the elements of history. History in its broadest sense, is, in short, nothing less than the experiences of our race, so far as we can determine or surmise them. And what uses are we to make of the experiences of the race? The _ same kind of use that we make of our own individual history. We ‘may question it as we question our memory of our personal acts, situations and past ideals. But those things that we recall from the superabundant fund of our own experiences vary continually with our _ moods and preoccupations. We instinctively adjust our recollec- _ tions to our immediate needs and aspirations and ask from the past _ light on the particular problems that face us. Just as our individual history is thus not immutable but owes its value to its adap- tability, so with the history of mankind. As Maeterlinck has beauti- fully said, with increased insight, “historic facts which seem to _be graven forever on the stone and bronze of the past will assume an entirely different aspect, will return to life and leap into move- ment, bringing vaster and more courageous counsels.” History is ‘then not fixed and reducible to outlines and formulas but it is ever alive and ever changing, and it will, if we will but permit it, illuminate and explain our lives as nothing else can do. For our _ lives, are made up almost altogether of the past and each age should be free to select from the annals of the past those matters which have a bearing on the matters it has specially at heart. ___ If we test our personal knowledge of history by its usefulness to us, in giving us a better grasp on the present and a clearer notion of our place in the development of mankind, we shall perceive forth- _ with that a great part of what we have learned from historical works has entirely escaped our memory, for the simple reason that we have never had the least excuse for recollecting it. The career of Ethel- _ red the Unready, the battle of Poitiers, and the negotiations leading up to the treaty of Nimwegen are for most of us forgotten formule, no more helpful, except in a remote contingency, than the logarithm _ of the number 57.. The ideal history for each of us would be those facts of past human experience to which we should have recourse _ Oftenest to our endeavors to understand ourselves and our fellows. O one account would meet the needs of all, but all would agree 190 ROBINSON—THE NEW HISTORY. _ [April 22, that much of what now passes for the elements of history meet 1 needs of none. 7 It would take too long to attempt an analysis of the alae of 3 genetic treatment of the elements in our social life. It is perhaps the greatest single discovery of modern times that we understand situation best through its history, and this discovery has revoluti ized every branch of organic and social science. Indeed we or narily first get a fairly comprehensive notion of a given phenomen by tracing its origin and development, whether it be the human back: bone, the order of St. Benedict, the stock exchange, the Wagnerian opera, or the doctrine of stare decisis. In many cases the knowledge of the history of an institution not uncommonly gravely affects o attitude toward it. The United States Senate looks different who is not. The Puritan sabbath could never have sustained critical historical examination. One’s views of democracy, or of present laws of property, or of the prevailing economic organizati can readily be deeply affected by a study of the earlier conditio which lie back of present conditions. History has a disintegrati effect on current prejudices which is as yet scarcely appreciated. makes both for understanding and for intellectual emancipation nothing else can. Obviously history must be rewritten, or rather, innuae ale rent issues must be given their neglected historic background. present so-called histories do not ordinarily answer the questions 1 would naturally and insistently put to them. When we contemplat the strong demand that women are making for the right to vote, ask ourselves how did the men win the vote? The historians | consult have scarcely asked themselves that question and so do answer it. We ask how did our courts come to control legislation the exceptional and extraordinary manner they do? We look vain in most histories for a reply. No one questions the inalienz | right of the historian to interest himself in any phase of the px that he chooses. It is only to be wished that a greater number of ei historians had greater skill in hitting upon those phases of the t which serve us best in understanding the most vital problems of present. CoLUMBIA UNIVERSITY. _ THE ATOMIC WEIGHT OF VANADIUM DETERMINED FROM THE LABORATORY WORK OF EIGHTY YEARS. By DR. GUSTAVUS D. HINRICHS. (Read April 21, I9II.) Vanadium can no longer be considered a rare element. Ferro- vanadium is produced on a large scale for the manufacture of special vanadium steels. Strangely enough, it was in a kind of natural vanadium iron that Sefstrém detected this element eighty years ago. In 1830, while technical director of the famous iron works at Taberg in Smaland, Sweden, Sefstrém thought it might be interest- ‘ing to submit his high quality malleable iron to the Rieman Test or cold-short iron, notwithstanding the apparent absurdity of such an undertaking. Accordingly he took one of his bars. On a part of its bright metallic surface he drew the little circular ridge of tallow and poured dilute sulphuric acid into the shallow dish thus formed, expecting, of course, to see no change whatever of the bright metallic bottom of this improvised dish. But he was amazed to see that bright bottom instantly turn black while the shallow dish rapidly filled up with a black powder, omy as it does when the iron tested is badly cold-short. The distinguished disciple of the great Berzelius instantly realized that this striking contradiction between test and fact was a positive indication of the presence of a hitherto unknown chemical element. _ Accordingly he set about isolating this new element. Working up quite a number of pounds of his iron, Sefstrém obtained less than a decigramme of the substance from which the new element was to be separated. Hence he turned hopefully from the iron to its fresh slag and found it to yield a much larger per cent. of the black powder. He now soon succeeded in isolating the new element 191 192 HINRICHS—ATOMIC WEIGHT OF VANADIUM. | [April 21, which, as a good Scandinavian, he named vanadin after Vanadiis, a designation of Freya, the greatest Goddess in Valhalla. | Sefstrém had promptly informed his teacher of the discovery and soon after brought his entire stock of the new element to Berzelius, requesting him to continue the research for which his — own industrial work and the professorial duties at the Fahlun — Montan-School left him neither the leisure nor the facilities. For a short time Sefstrém worked with Berzelius on the new element in that famous ‘“ Kitchen Laboratory” where Berzelius alone com- _ pleted the splendid work of which he published a summary on pp. | g9-110 of the “ Annual Report” which he presented to the Swedish Academy of Sciences on March 31, 1831—exactly eighty years ago. For almost forty years the element vanadin of Sefstrom and Berzelius remained undecomposed, but the striking isomorphism | of the mineral vanadinite with the remarkable isomorphic group of apatite and pyromorphite presented the anomalous condition of the isomorphism of the element vanadin of Berzelius with the group PO of apatite and pyromorphite. This anomaly invited furth attempts of the reduction of vanadin in which Roscoe was success- ful, 1867, proving vanadin to be really the oxide VaO, in which V; is the symbol of the present element vanadium of the atomic weight 51. This fully explains the isomorphism of vanadinite os the oxide VaO, with pyromorphite, containing the conreseyas n oxide PO. In this first research of Berzelius on vanadium, the de master already determined the atomic weight of the new element; f his value 67 for what we now know to have been VaO gives Va 51 He devised and used five distinct chemical methods for this atomic weight determination to which not one new method has been add in the eighty years elapsed since that work was done by the g eat chemist in his kitchen laboratory. It is a well-authenticated histor fact, Berzelius not only made atomic weight determinations f vanadium, but they were as accurate as those made forty yea later by Roscoe, while some were as precise as correspondi determinations made eighty years later by Prandtl; besides, not only Roscoe and Prandtl, but all chemists have done this work by means 1911.] HINRICHS—ATOMIC WEIGHT OF VANADIUM. 193 of the methods devised by Berzelius which he practiced in his laboratory in 1831. It is therefore with great astonishment that I read in the first _ edition of the “ Recalculation ” of F. W. Clarke: “ Roscoe’s determi- _ nation of the atomic weight of vanadium was the first to have any 4 scientific value. The results obtained by Berzelius . . . were : _ unquestionably too high, the error being probably due to the presence of phosphoric acid in the vanadic acid employed.” q The same erroneous statement is repeated identically at the 4 oo of the chapter on vanadium in the succeeding two editions _ of the work as may be seen by comparing: p. 183, edition 1882; p. 7 _ 211, edition 1897, and p. 305, edition 1910. 7 The only new method, quite recently applied to the determination of the atomic weight of vanadium, is that of Edgar F. Smith.’ This admirable method strictly conforms to the Berzelian advice “to chose such chemical methods for atomic weight determinations that the final result shall depend as little as possible on the operator’s skill in manipulation.” In my summary of the work of one hundred years on the determination of the atomic weight of hydrogen? I have given this great rule of Berzelius, in his own handwriting, from his “ Sjelfbiografiska Anteckningar,” published by the Kgl. Svenska Vetenskapsakademien, 1901, p. 41. _ This rule requires to select such chemical reactions in which the physical and chemical characters of the substances weighed are so definitely fixed that the unavoidable errors of man and his instru- _ ments become negligible quantities. Such is the reaction no. 311 above referred to. Hence the work done by McAdam in the labora- tory and under the direction of Edgar F. Smith has furnished the highest direct chemical approximation obtainable to the absolute scientific truth that Va is 51 exactly. This will appear, we think, a from a careful examination of all the results actually obtained dur- ing the eighty years from 1831 to 1911 as plotted in our two dia- "grams no. 730 and no. 731 published with this paper. __ The above reference to the presence of phosphoric acid in the FEET IE GT AN _ *See Journal Amer. Chem. Society, 1910, p. 1603, in the December number. - *In the Révue Generale de Chimie, 1910, Nos. 22 and 24. 194 HINRICHS—ATOMIC WEIGHT OF VANADIUM. [April 2 vanadic acid used by Berzelius reminds us of the homely but sot scriptural advice habitually given by Berzelius to his disciples: “ not strain at a gnat while swallowing camels.” The phosphoric ac: in the vanadic acid used by Berzelius was detected by Roscoe in sample which Berzelius had presented to Faraday ; but the molybdic reagent necessary for the detection was not known to chemistry the year 1831 when Berzelius did his work on vanadium. a As a matter of fact, Berzelius did not see this gnat; but — work shows that he did avoid some of the camels that stalk about the laboratories and which were deglutinated unconsciously for and eighty years after Berzelius failed to strain that gnat. ‘ error-shares due to the oxygen are the fattest and most numerous of these camels, up to the present day. a Our METHOD OF REDUCTION. In order to solve the riddle of the conflicting experimental ¢ obtained in the chemical laboratories of the world during an en century of painstaking work, we have, especially in the last que century, carried on special researches on the proper mathema reduction of this kind of laboratory work. The final results of this extended research are briefly si marized in five tables of which two only have thus far been pub- lished. Our work itself has been published in the following boo and special papers: “The True Atomic Weights,” St. Louis, 1894, xvi 8vo, with 7 plates and many illustrations. Dedicated to Berth “The Absolute Atomic Weights,” St. Louis, 1901, xvi + 304 8vo, with portrait of Berzelius and three plates. “The Proximate Constituents of the Chemical Elemente Louis, 1904, with 7 portraits, many plates, 112 pp. text, 8vo. — is an inductive treatise of the subject. _ The “ Cinquantenaire,” 1910, gives some historical data, copi of older papers, letters in fac-simile and “ Fragments inédits ” fine diagrams; 66 pp., 4to, with plates and portraits. “Notes” published in the Comptes Rendus of the Academy | Sciences of Paris from 1873 to the present, almost sixty in numbée — Igtt.) HINRICHS—ATOMIC WEIGHT OF VANADIUM. 195 _ forming a volume of over 200 pp. 4to. The first, and in fact the _ greater number of the “ Notes,” were presented by Berthelot ; others were presented by Messrs. Gautier, Lemoine, Haller, Gernez, and _ other academicians. ; In the Moniteur Scientifique, from 1906 to 1909 more than a - dozen longer articles have appeared with many diagrams. The first two tables above referred to are found in the November num- _ ber for 1901, with discussions, pp. 731-744. _.._ The papers were originally written in four languages: Danish, : German, English and French. To these papers the reader may be referred by the Cinquantenaire and the list in the Prox. Constit. he results obtained, being in conflict with the dominant chemical hool, have not been widely circulated except as adopted children. For these reasons it is necessary here to give enough of the tails of the finally worked out practical method of reduction to ble the reader to repeat all the calculations required, so that he verify the results given. It will then be seen that the final method is quite simple; the ty was to get this method. Let a represent the absolute atomic weight of any chemical ele- ent, that is the whole or round number (4, or even }$) which the eriments indicate to be near the true atomic weight A, which tly to determine is the object of the reduction. The wumt dopted is exactly ~y of carbon-diamond which is practically iden- with +; of that of oxygen.* The departure of the true atomic weight from the absolute atomic weight we designate by the Greek letter epsilon (e) ; that is: A=a-+te. This departure, as a matter of fact, is found to be a | fraction of the unit; we invariably express it in thousandths f that unit. This departure—in units of the third decimal—is really our new able, the quantity to be determined. This apparently insignifi- t matter of form is really of the greatest importance. For this vy variable all products and powers become negligible quantities *Comptes Rendus, 117, p. 1075, 1893. PROC. AMER. PHIL. SOC., L, I99 M, PRINTED JUNE 26, IQII. 196 HINRICHS—ATOMIC WEIGHT OF VANADIUM. [April : in our necessary calculations, because the departures are small quantities ; hence, all calculations, even involving the most comple mathematical functions, are reduced to the simple rule of three an carried out by proportional parts. The importance will soon b recognized by the practice of the method. The actual Jaboratory work consists essentially in the deter mination of two weights which we denote by p and g and which re; resent chemically pure ‘compounds of the formula P and Q respe tively. The necessary condition is that the weight p has been com: pletely changed into g according to the exact formule P and Q by means of a suitable chemical reaction. Of such reactions we h tabulated and examined over three hundred that have been actu used for atomic weight determinations. We designate each su reaction by a number for ready reference. This number is simp marking their place in our table above referred to; it is arbits but a practical necessity. We have already above referred to remarkable chemical reaction, recently applied in the Harrison La ratory of the University of Pennsylvania as reaction no. 311. Substituting the absolute atomic weights q for the chemical bols in the formulz of the two compounds P and Q, we can rea calculate the value of the quotient P/Q which we call the ate ratio R and calculate the same to five decimals, the limit of pre today. On the following pages, giving the data for the ch reactions that have been used for the determination of the ; weight of vanadium there will be found examples of these a all other processes, to which we request the reader to turn as operations are defined. : On the other hand, the weights actually taken in the lab and designated by the letters p and qg will give the analytical ratic which we calculate also to five decimal places. The analytical determined by the different experiments with the same two -pounds P and Q will give hardly any identical values of r; we their extreme values, that is the maximum and the minimum given series of determinations made in the same manner with identical material. The differance between the greatest an least value of the analytical ratios of a series is the range 0 series. This characterizes the concordance of the different 1g11.] HINRICHS—ATOMIC WEIGHT OF VANADIUM. 197 -minations of any series without introducing any false theoretical notion, as is done by the calculation of the so-called probable error _ ofthe mean. The actual mean value we do calculate and use. While the individual analytical ratios vary for the different indi- vidual determinations in a series and even the means for the different series, it is found, as a matter of fact, that they bear a close relation to the atomic ratio. We call the excess of the analytical ratio over the atomic ratio, the analytical excess and designate it by the symbol e. That is: r—R-+e. The value of e is also expressed in units of the fifth decimal. - Tue EouaTION oF CONDITION AND THE SOLUTION Ex-A®guo. In the true atomic weights of 1894 (p. 139 to p. 169, esp. p. 158) the solution of the great problem is already shown to require an application of the method of the variation of constants. In the absolute atomic weights of 1901, the change or variation A of the atomic ratio for an increase of 0.1 in the atomic weight is determined for each reaction and applied for several important objects throughout the entire work. On pages 144-147 of that work _ the final solution is really given but implicit only, and lacking the equal distribution of the analytical excess among the elements present in the reaction. . The actual equation of condition was established in 1907 through _ long and difficult work, both analytical and geometrical. The gen- _ eral analytical deduction by means of Taylor’s formula was in the hands of eminent men abroad in the form shown in fac-simile (reduced to %) as printed p. 61 of my “ Cinquantenaire,” 1910. The most general construction, which permits the establishment of a _ criterion for the absolute atomic weight, is printed on p. 60 of the - same “ Cinquantenaire.” Here we will present the final practical solution of the resulting “insoluble” indeterminate or diophantic equation in the simplest and most direct manner, suitable for common, current, practical _ application. The true atomic weight, A, is the quantity sought, in the unit for which carbon-diamond is 12 exactly. ‘ - 198 HINRICHS—ATOMIC WEIGHT OF VANADIUM. _ [April 21, The absolute atomic weight a is indicated by the laboratory work; in case of doubt, the criterion just referred to has to be made use of. The departure ¢« is expressed in units of the third salical (thousandths) of the unit of atomic weights. Its exact determina- tion is the main object of this paper. i The atomic ratio R is a function of the absolute atomic weights, expressed by the quotient P/Q above given. If now the absolute atomic weight for any one given element in this ratio be increased by 0.1, that ratio will change or vary by an amount readily calculated from the formula of R as given; we use throughout seven place logarithms which give the precise value sought with the feast trouble. This change or variation we denote by the Greek capital delta A for the particular element of which the atomic weight was increased by 0.1 in the atomic ratio. In 1901 y made this calculation only for one element in the ratio; now it has to be made for every element in the ratio. In the tables here following this matter will become quite readily understood by simply repeating some of the calculations thus indi- cated. For the reaction no. 98 this work is quite simple, for only two elements are present, namely Va and O. In reaction 270, the work required is about double in amount, because four elements are in reaction, namely: Va, O, Cl and Ag. The analytical ratio r has to be calculated for each single dete mination made; it is considerably simplified if the weighings are given by the chemist to the hundredth of the milligram, are rounded off to the tenth of the milligram, which is as far as the weighings can be trusted; see, for example, my demonstration of this fact for the weighings of Richards made at Harvard-Berlin.* The analytical excess e is now obtained as it is r—R; it is ae expressed in units of the fifth place. Now we have in hand all the quantities required for obtaining the departure « sought, by solving the equation of condition. this indeterminate or diophantic equation is, of course, insoluble general, we have nevertheless obtained two practical solutions the same® of which the one properly named ex-@quo is the most * Moniteur Scientifique, Juin, 1909, especially pp. 384-385. *> Comptes Rendus, T. 149, p. 1074, 1909, with a most instructive figure. 1911.] HINRICHS—ATOMIC WEIGHT OF VANADIUM. 199 serviceable and by far the most readily understood and easiest applied. Our general deduction (really as indicated 1894 already: a method of the variation of the constants) leads to the simple form of the equation of condition 100 ¢== Ae, _ where the constant 100 presupposes that the analytical excess e and the variation A are expressed in units of the fifth place while the departure « is expressed in units of the third place or thousandths of the unit of atomic weights. It may not be amiss here to insist on the fact that since in every chemical reaction there are at least two elements present, the above equation contains at least two unknown departures « and is there- fore really an indeterminate or a diophantic equation. Our practical solution ex zquo of this equation is as follows: Let m be the number of elements involved in the chemical reactioz used, then the number of terms Ae in the above sum & is m. Ascribing to all elements an equal influence on the error or excess é, the part thereof due to each element will be e’=e/m. Hence the actual departure « for each element in the reaction will be determined by the simple relation 100 é €e=— A If the value of A be above a certain limit, this determination will be sharp; the corresponding reaction therefore may also be called sharp. But if the value of the variation A for any element is small, the reaction for that element will be dull and the determination of he atomic weight will be impossible with any high degree of pre- cision, as we have shown in Comptes Rendus, T. 148, p. 484, 1909, | the attempted determination of the atomic weight of Tellurium by reaction quite dull for that element. _ This one attempt strikingly shows the real condition of the work of the dominant school to be irrational. After having briefly explained the manner in which we have ed to solve the great problem of the deduction of the true atomic 200 HINRICHS—ATOMIC WEIGHT OF VANADIUM. [April 21, weights from the experimental work done in the laboratories, we may proceed to the full statement of the facts obtained for the el ment vanadium during the past eighty years and the final results” our discussion of the same. We shall present the facts in the most compact form of tables and: finally exhibit them to the eye in the form of accurately drawn graphics, from which we shall be able to read the final result the most readily and clearly. THE ACTUAL DETERMINATION OF THE TRUE ATOMIC WEIGHT OF VANADIUM. I.—AssoLute ATomMic DATA. Fundamental Constants, Calculated from the Absolute Atomic Weights. Only seven** chemical reactions have been used for the deters : nation of the atomic weight of vanadium, thus far; they are t following: No. 98: Pentoxide reduced by nydeicen No. 269: Oxychloride to silver. No. 270: Oxychloride to silver chloride. No. 311: Vanadate to chloride. (a)—Oxychloride to pentoxide. (b)—Sulphate to barium sulphate. (c)—Sulphate to pentoxide. The last three preliminary methods of Berzelius have been 1 by him, each once only, and by no other chemist, except that F made four determinations according to method (a). No. 3 but just been introduced by Edgar F. Smith, December, 1910 the chemical reactions used for the determination of the aton weight of vanadium, up to that date, were devised and firs by Berzelins in 1831, eighty years ago. It seems that his wo some scientific value, after all. In the following Table I. we have given the most important : damental constants required by our method of reduction. have all been calculated from the well-known absolute ~ weights: Va, 51; T, 16; Cl, 3514; Ag, 108; Na, 23; S, 32; 59 Tf we count 269 and 270 as distinct reactions. HINRICHS—ATOMIC WEIGHT OF VANADIUM. 201 1374; H, 1.008 instead of the exact value 1.00781 determined atomechanically by us (Révue gén. de Chimie, 1910, p. 386). TABLE I. FUNDAMENTAL CONSTANTS FOR VANADIUM. Atomic Ratio, R. Variation A for No. Formula. Fraction. Decimal.® Va. oO. Ci. Metal. 0. Ee 98 V2,0, 7ha 0.17 582 | —19 62 — Se vad), 173.5 269 3Ag 324.0 |.953 549) 31 31 93 |—49(Ag) va0Cl, 173.5 270 3AgCl 430.5 0.40 302; 23| 23) 42 |—28(Ag) NaCl 58.5 311 NaVa0, 1320 0.47 951 | 39 |—118 | 304 43(Na) Va,O, 182 fo). acl, 347 0.52 450} 27} 113 |— 9F ve Va,O, 182 4 2BaSO, 407 0.38 972} 43| 4° |— 17(S)|—17(Ba) Va,O; 182.00 . Va sulphate’ 398.06 0.45 722 27 | —35 — 23(S)|—140( H) _ JI—GeEneERAL SUMMARY OF THE EXPERIMENTAL Work DONE. In the common reviews of the eXperimental work done for atomic weight determinations, the amount of substance taken in each ‘experiment is not made the subject of special consideration. This neglect is due to the erroneous estimation in which the so-called “ probable error” of the mean is held. __ This probable error has caused the most serious errors ‘in all branches of physical science where it has been applied—in the un- _ fortunately common way, without proper understanding. We have ‘treated of this repeatedly, especially in our “absolute atomic eights, 1901, on the first hundred pages, to which we must refer. In the language of Berzelius already quoted we might say the above probable error is the gnat strained at which hides from sight the camel-like systematic and constant errors which are swallowed. We have, at last, seen one admission of the fact we have always * Between the second and third decimal of the five, we always leave a space to make the constancy of the first two conspicuous. *The crystallized sulphate is [VaO]:S:0s + 4H2O. 202 HINRICHS—ATOMIC WEIGHT OF VANADIUM. | [April 21, accentuated, that large constant and systematic errors may exist though the value found for the probable error of the mean is in- significant. In Clarke’s third edition (1910, pp. 93-98) the probable — error of the mean amounts to only one unit in the fifth decimal of the analytical ratio while the constant error of that ratio amounts - plotted the results of the individual experimental determinations — as ordinates to the weight taken as abscisse. In these diagrams — the scale selected for the atomic weights or the ratios must be very — great while that for the weight taken has to be small. In my diagram representing in this manner all the atomic weight determinations of hydrogen made in a century (Révue gén. de Chimie, 1910, p. 380) the unit of atomic weights is 30 meters (about 100 feet) while a decagram of water produced is represented by three centimeters (or a little over one inch). To permit this search for the really important constant and sys- tematic errors, we give the weight taken (to the decigram) in all our tables. For a series of determinations, we give the total weight taken for all the determinations of the series, and the mean weight taken for each determinatioh—which is obtained from the total by dividing this latter by the number of determinations made. e Table II. thus shows that 50 determinations have actually been — made for the determination of the atomic weight of vanadium on 4 grams each of the substance taken, not counting the seven pre- liminary determinations on I gram of matter each. It is also seen at a glance that by reaction 98 the work of Roseos ought to be the most reliable, while for 270 the work of Pran should be the best and that the work done by one chemist for 311 has been carried out under equally as favorable condition in regard to the weight operated upon. By means of the reference letter specified in the last column the corresponding line on the diagrams can be instantly identified. V may here already remark, that the length of this line, extending t the right or to the left from the vertical in the middle, marks the. magnitude of the departures for the elements as indicated by th chemical symbol added. 1911.1 =HINRICHS—ATOMIC WEIGHT OF VANADIUM. 203 Thus the least deviation or departure has resulted from the use of the Method 311 recently introduced by Edgar F. Smith. The preliminary work of Berzelius, done eighty years ago, according to methods a and 5b as represented by lines D and T on the diagram, showed departures for sulphur and barium extending beyond the limit of our diagram: 470 to the left (negative) in line D and 776 to the right (positive) in line T; the former is almost half a unit, the latter three quarters of a unit of atomic weight. TABLE II. SUMMARY OF THE EXPERIMENTAL DETERMINATION. Weight Taken, Grammes. - Number of : Letter on Reaction. Ch s . senpengs Dias. Total. oe Dixgwm. 98 3 1.6 4.7 Berzelius B I 0.6 0.6 Berzelius B* 5 6.0 30.0 Roscoe F 4 2.5 10.0 Prandtl Cc : 13 3.4 45.3 269 9 2.7 24.3 Roscoe N(M, R) 270 I 1.6 1.6 Berzelius G 6 1.2 7.4 Roscoe (A) H Mi 2 2.4 4.7 Roscoe (B) K 5 4.7 23.5 Prandtl, I. L 6 4.8 28.8 oe Il. P 4 8.1 32.4 ~ II. oO 24 4.1 98.4 gir 5 6.0 30.2 Smith-McAdam Q Total 51 3.9 198.2 or say: 50 4 200 PRELIMINARY REACTIONS: (a) Oxychloride I 1.6 1.6 Berzelius, 1831 A to Pentoxide 4 1.0 4.0 Roscoe, 1868 E _ (b) Va Sulphate : to Ba Sulphate 1 08 08 Berzelius, 1831 D (c) Va Sulphate to Pentoxide I 08 08 Berzelius, 1831 T Mean 7 1.0 72 204 HINRICHS—ATOMIC WEIGHT OF VANADIUM. April III.—Tue ANALYTICAL RESULTS. The chemical work, beginning with the determination of weight p and ending with the determination of the weight q (or the inverse) gives directly the atomic ratio r, by a simple division carried to five decimal places. A simple substraction now will give analytical excess e by using the atomic ratio calculated once for all for the reaction, The results obtained in this way are oie: Table III. We will here only call attention to the following peculiarly inter esting circumstances: “ Analyst B” working under Roscoe makes three determinations: 2 under reaction 269, and takes almost the same weight of the o> chloride for each one of these determinations; notwithstanding - fact that he really only repeats one and the same determination times, his results range over 501 units in the fifth place—an enor- mous range under so favorable conditions. In the language of Berzelius, this enormous range is aie a bi camel which was swallowed without an effort in the laboratory Roscoe, who strained laboriously at the tiny gnat of phosphoric a which a test of high delicacy, that was not yet known in the when Berzelius did his splendid pioneer work on vanadium a courteously presented to Faraday a sample of the vanadic acid had received from his disciple who discovered the element vanadi and which acid Berzelius had purified himself as far as the sci of his time permitted. These facts I gather from Becker, Smith Misc. Collections, 358, Washington, 1880, p. 132, quoting Liebig’s Annalen, 93, p. 6, 1868. TABLE IIl. THE ANALYTICAL Ratios DETERMINED. Reaction 98.—Atomic Ratio, R=0.17582. | Berzelius, 1831.—Meyer-Seubert, “ Atomgew.,” 1883, p. 28. Sums of Weights. r. é. Determ. Pentoxide. xygen. 3 4.6005 0.8120 0.17 278 — 304 I 0.6499 0.1124 204 — 288 - Diagram. Lines B and B*. x911.] HINRICHS—ATOMIC WEIGHT OF VANADIUM. 205 Roscoe, 1868 —Jour. Chem. Soc., 6, p. 330, 1868. No. ti 2. a, 4 5. Mean Weight Oxide 7.7 6.6 5.2 5.1 5.4 6.0 r=0.17 533 507 489 515 501 509 ‘= ae ee ee Oa OF BE eG Range 44.—Diagram, Line F. Prandtl, 1910.—Jour. Am. Chem. Soc., 1911, pp. 266-7, from Ztsch. anorg. _Chem., 67, 257. No. ts % a 4 Mean. Weight Oxide 9.1 9.9 87 12.3 10.0 r=0.17 261 376 395 394 356 r —32r —20 —187 —188 —225 Range 133.—Diagram, Line C. REAcTION 269.—Atomic Ratio, R=0.53549. Roscoe, 1868; Analyst A—Jour. Chem. Soc., Vol. 6. _ No, 5 2. 3- 4. Ss. ay Mean. ~Oxychlor. 2.4 47 4.2 4.0 0.9 1.4 2.9 r=0.53 425 528 533 510 530 532 510 e= —I24 — 2I —16 — 39 —I19 — 17 — 39 Range: 108.—Diagram: Line M. Roscoe, 1868; Analyst B. ss . Mean No. 7. 8. 9- Mean. of All 9. Oxychlor. 2.9 2.1 1.4 2.1 25 r=0.53 980 755 479 738 = 586 = 431 206 — 70 189 37 Range: 501.—Diagram: Line R. Range: 555. REAcTION 270.—Atomic Ratio, R 0.40302. -Berzelius, 1831.—1 determ.: 1.6385 oxychlor. gave 4.0515 Ag, hence r= 0.40442 and e 140. Meyer-Seubert, “ Atomgew.,” 1882, pp. 90-01. Diagram: Line G. Roscoe, 1868; Analyst A: ~ : No. 5 2. Si ys 5. 6. Mean. . Oxychlor. 1.9 0.7 0.8 1.4 1.0 1.6 1.2 r=040 323 531 537 337 399 174 383 e= 21 229 235 35 97 —128 81 _ Range: 363.—Diagram: Line H. ; ANALyst B: No. 7. 8. Mean. Mean of A and B. Oxychlor. 2.2 2.5 2.4 :5 r=0.40 391 333 362 378 e= 89 31 60 76 Range: 58.—Diagram Line K. Diagram: I. 206 HINRICHS—ATOMIC WEIGHT OF VANADIUM. [April Prandtl and Bleyer.’ Series 1., 1909. ie 3- cs 5: Mean, No. I. 2. Oxychlor. 5-5 5.9 3.2 5.3 3.6 or r=0.40 393 346 365 322 367 ae c= gl 44 63 20 65 ” Range: 71.—Diagram: Line L. “ * Clarke, “ Recalc.,” 1910, p. 307 for I. and II.; Journ. Am. Chem. 1911, p. 266, for ITI. Series II., 1909. No, oy 2. g, 4 5. 6. Oxychlor. 4.9, 39 5.0 6.5 4-3 4.1 r=0.40 331 286 318 315 308 3B e= 29 — 16 16 13 6 23 Range: 45.—Diagram: Line P. ene Series III., 1910. ame No. 1. 2. 3. *: Mean, Oxychlor. 7.8 8.4 10.7 5.5 Sa. r=0.40 301 311 321 333 aIFi= = — 1 9 19 31 Foe Range: 32.—Diagram: Line O. zi Reaction No, 311.—Atomic Ratio, R= 0.47951. Edgar F. Smith and McAdam, Jour. Am. Chem. Soc., nde p. 1614. No. Ee Mean, Vanadate 40 86 ae 38 6 : 6.04 r=0.47 931 927 941 937 921 * = 9BF e= — 20 — 24 — 10 — 14 — 30 ieee Range: 20.—Diagram: Line Q. Strictly this series consists of 3 determinations only, t which have been made twice, as follows: oe No. 1.3. 2.4. 5. Mean. Vanadate 4.6 5.7 9.5 6.0 r= 0.47 936 932 g21 931 e= — 15 —i19 — 30 — 20 Range: 15. This shows how even within narrow limits of weights (here from 4% to 914 grammes) the systematic error becomes evi dent as a function of the weight taken. Within the actual ran the analytical excess approaches zero with diminishing ar operated upon. : PRELIMINARY WorK. Analytical Reac- Rane Line. tion. : A. a I det. 1.6385 oxychl. gave 0.874 oxide 0.53342 E. 4 det. 4.0418 2.1258 0.52608 D. b 1 det. 0.351 pentoxide 0.913 Ba sulphate 0.38445 — 527 Berze T. c 1 det. 0.351 pentoxide from 0.775 Va sulphate 0.45200 + 432 Berzel 1.) HINRICHS—ATOMIC WEIGHT OF VANADIUM. 207 IV. THe ANALYTICAL EXCEss. _ The analytical excess e is only comparable in work carried out ccording to the same chemical reaction by different chemists. This ‘condition has determined the form of Table IV., in which the capital letter marks the line on our final diagram representing the work, hile the numbers below the letter represent: the first, the number f determinations made, the second giving the analytical excess obtained. This excess will naturally be found the greatest for all eliminary (or pioneer) work such as that done by Berzelius in his single trials of reactions (a), (b), (c). _ This table makes it very apparent that each succeeding chemist efited by the work and experience of his predecessor. This best under reaction 270 where Berzelius (1 determination) es the excess 140, Roscoe (8 determinations) only 76 and Prandtl 5 determinations) brings it down to an average of 28 only. For Reaction 98 this relation holds good for the work of Berze- ; and Roscoe only, while that of Prandtl, done as recently as 9, reaches almost the excess of Berzelius single first trial of hty years ago, although Prandtl used the mean weight of 2.5 ms while Berzelius had only half a gram for his work. TABLE IV. : Tue Resuttinc ANALYTICAL EXcEss. Reaction. 98 269 270 3tr a | é e Berzelius, B G rf I D 1831 3, —304 I 140 tr 892/11 —527/1 432 = te ¥. =28k8 Roscoe, F R H E 1868 5° —73\3 1189/6 8& 4 158 eed / e 3718. 76 ; M K I —39|2 60 LS L : 1909 4 —255 eat, re) ae P i 6: 2 th-McAdam, eae Q IgI0 5 —20 208 HINRICHS—ATOMIC WEIGHT OF VANADIUM. [April V. Tue Finat DEPARTURES. In this table (I1V.) we have. finally given all the departures f all the elements taking part in the different reactions used for + determination of the atomic weight of vanadium during the 1 eighty years—from Berzelius in Stockholm to Edgar F. Smith Philadelphia. We have fully explained the manner of calculati these departures; every reader can verify any of these values himself by the methods stated. Since the element most concern is vanadium, we have arranged the record of departures in the ord of the magnitude of the departures of vanadium, for which the e treme values are 1062 and minus 307, giving a total range of 13 say one and one third units of atomic weight; that is, from Va 52. of to 50.69. : If we omit the pioneer and preliminary work done by Ber in trying reactions (a) and (b), the extreme departures will be 80c and — 31, a total range of 831 only. The atomic weights will rur from 51.80 to 50.69, which is a remarkably fine showing for so lo a series of very difficult work on the rare element. To realize the generally excellent work of the early days of. zelius and of Roscoe—when the element was decidedly rare and cult to purify—we need only compare the limits of the dete tions by Prandtl of today with the range of the entire series cluding only, as we have done already, the reactions (@) and ( these extremes of Prandtl are 594 with reaction 98 and 13 in II. with reaction 270; a total range of 581 thousandths. The c sponding atomic weights of vanadium are 51.59 and 51.01 differ 58 hundredths. . We are greatly tempted to point out a number of interesting: i‘ tures on this table, but fear that the paper will asstme undue Ie ' and trust the reader will help himself. We only remind the reader that the letter in the first col this table permits the ready identification of the experimental re: expressed in numbers in this table with the graphical represen be t om on our two diagrams. 1911.] HINRICHS—ATOMIC WEIGHT OF VANADIUM. 209 TABLE V. Tue Departure. DETERMINED FROM EIcHTy YEARS OF WorK. : T Line in 2 Reac Ne Analyt, Departure, ¢, in Thousandths. Keoate Wik che — Spee Vg Va oO Cl | Metal.| Va. | 0. A |Berzelius a I 892| 1062! 263 | —326 52.06 | 16.26 B “ 98 3 \—304 800 | —245 51.80 | 15-75 BF sf 98 | 1 |—288| 784|—241 78 7 C_ Prandtl 98} 4 |—225| 594|—183 59 82 D_|Berzelius c I 432} 400] —310| —470%— 77” 40 E_ Roscoe a 4 158| 195 47|— 58 20 | 16.47 F 4s 98| 5 |} 73| 192|— 59 oe 19 | 15.41 _G_ (|Berzelius 270| 1 140} 153] 153 83 125 15 | 16.15 M Roscoe (B) 269} 3 189 150/} 150 50 |— 95 15 15 my (A) ajo} 6 | 81; 87) 87{| 47|—71| .09] — -09 I eo LA, BY) 13701 8 76 83 83 45 — 68 08 08 K > (3B) 270} 2 Go| 65; 65) 36;— 54| .07| .07 L_jPrandtl, I 270| 5 57 62 62 34 '— 51 06 06 N_ (Roscoe (M, R) | 269) 9 37 30 30 Io — 19 03 03 O |Prandtl, III 270| 4 14 15 15 7\— 12 or or P spate tina # F 270} 6 12 13 13 7; oI oI } Na: Q (|Smith-McAdams| 311; 5 |— 20 13 4|— 2l— 12 51.01 | 16.00 Ag: _R_ Roscoe (A) 269; 6 |— 39|/— 31|— 31|— Io 20 | 50.97 | 15.97 T us 6 ! 1 |—527!—307!—330! 776! 776| .69 | 15.67 . 3. sgt: * Ba. VI. Our GrapHIcs. The values of all departures given in Table V. are represented to the eye in our two graphics. Fig. 1 gives all the larger departures and as many of the smaller. ones as space would permit. The scale used is 200 thousandths of the unit of atomic weights to the inch ; or, what amounts to the same, the unit of the atomic weights is represented by five inches in length. Fig. 2 gives all the departures of the central region on a scale which is five times the one used in the construction of the first figure. Hence in Fig. 2 the unit is represented by a line of twenty-five inches ; or, in other words, it shows forty thousandths to the inch. The vertical of the ordinate represents the departures of vana- dium, while the departures of the elements combined with vanadium are set off on the horizontal as abscissz. The results for a complete analysis of any given compound are therefore set off on the horizontal line drawn through the point on the vertical determined by the departure for vanadium. a: @ A. e: e Perton. Departure looow & Va252. Yoo 5 goo } The Laboratory Work | of 4 \ of Eighty years S B. 3 £ bcd — @: astablishes +440 *2 > Va = 5) ecackly, . joo 4 . " _ i Gustavus D. Hinrichs fe! “Re peepee 4» 9 #73) Mare 1Git. 7609 Vv . 5 II. For att Departeres >200. soot * * For complete £ {209 | co ss see l; # 730. .] - s D © raok 1 VaS <_ minis 4,.--~. Pentox : Va hate. i sy7e @: a @: @ : sth Yoo + 300 q A / CL Q: wa Si. Va=51. Scale of —200 z shoes Departures E of the TOC O} Elements combined with. } 2 69. Variacklium. O Ct | Ag -/00 - Laboratory Work Vor J OQoene by -200 ° Berzelius,is3i; Js, Roscoe, lca: Ss ® Prandtl-B 190 “ ates 7 Ba. A - e : rt. e ~ Sutphates Va: B @ = 776 : lac sae “9 bei bu | ig EF Smmith-MAdan 1910. Departures § are eapressd in thousaadihs of he unit. ¥ a ie aren = Ree ae a) —j00 ~100 —/00 ° foo 200 Joo ¥0o0 Fic. 1. Departures. Scale 200 thousandths to the inch. ‘Your BY} O} syIpuRsNoYyj-ob :ajvog a ‘sainjiedaq ‘'@ ‘oy oss os! 'SYPZPUBVSNOYI Ut em —PDunysvdrgc Oo Os~ 2oc= OS: "€Z PN’ ‘gory § 66" La ‘90! 1G N= 87 Y OIA sory 777708 q0 - Ad Ne f - a mors tquny fo sYypuvsnoyy sis sranqauvdag j a UL Wy re YIU ee oer ee : bi yeon of L +E ‘ole{,@ Berend OlbI PEE OMS ad S| remvemoys oO7 eyy t17) 3 smemiendoq of 62 | He -60b) sPhague ypuvsg es se : reel eee) © - ©-£©@ rmO) Be et a "Zgg| ‘20250 © ha Peorrpna ff 69 € 1 “Jewry \ JERI SMM 2ez40G @ ‘ caweh os wm yuop . ‘ Ost WAaM hsogmsoqe Tre °4 burp soron ‘ Os + = \ ‘fg wucp yA hsoyosoqo] Oos~ z fygrm po 8) Vl 7 1 : ‘. ey Gace Ge re ae ee aie hat s “UN IPHWUW/A Poss “ is § Ri, ‘© isblargzeezz:s'9¢ =O ron rn iE is 7 , 4 Jo aybram > ymory amay a / ol z ' 401920944 “ roe om 9 =s0¢H:5¢1/=NYE NON ‘OL A ; bas #so= gerseliz= bye? 120% ‘697 a © r8 === 7% vesho= rgiite= 4O%A:'O ‘86 / : v? at + $U01490 2} oo ~ (ng oo! a 00/- AN ae oN rh £ © % oo ; Nv A. a> oO opp suo) ywuiuasayep do \ d ; #yA9g Yue APQvANL yar by G 45 DA V. / f . / @ —S9t ; © iW x os! 207 yr hx ai: TOR «0 $5~: 0 an a LAW wiiiaisiiiaaiacsieiasbiailinaiil iin 212 HINRICHS—ATOMIC WEIGHT OF VANADIUM. [April The chemist whose work is represented is indicated by the spec mark used to designate the point, as shown in the explanation, signs on the diagram. The small figure near the vanadium sign on the vertical indi the number of determinations made represented in that line. The line itself is marked by a letter, used in the tables ~ the purpose of ready identification. For each reaction, the geometrical place (or locus) is a st line passing through the center or the origin; the angle under whi ick it cuts the axes is determined by the ratio of the variation of | element concerned and that of vanadium. These lines can ‘herein ore -be drawn before any laboratory work is done, depending entirel. ‘the chemical formule of the compounds taken and obtained in reaction used. For further particulars, some of which are very interesting well as useful, we may refer to page 60 of our “ Cinquantenaire, where also a remarkable criterion is given, permitting to detect error in the assumed absolute atomic weight. The example t taken is copper. * Our two figures here inserted bring into clearest possible view 1 he fundamental fact that all these departures are co-related; that the experimental error is not thrown on the vanadium for which atomic weight is sought, but is distributed ex-zquo to all elem partaking in the reaction, as we have shown in formule, but w is here presented to the eye directly. We do not recognize or find the slightest pretext for the asst tion that any one element is immaculate and cannot be conceived partake in any error of whatever cause or origin; but we have fe that all elements in a chemical reaction are affected by the s cause of error according to the ties that bind them and which have read in the chemical formula and in the mathematical re atio first studied by Lagrange under the name of the Variation of art trary constants.** We know that it is absurd to suppose that oxygen is always fo to be 16, absolutely unaffected by any error, physical or chemical practice ; that next some other atomic weight of some other ele 2“ True Atomic Weights,” 1804, p. 158. 1911.] HINRICHS—ATOMIC WEIGHT OF VANADIUM. 213 can be determined and that this value also will remain unaffected in all reactions; and so forth. That errors rapidly accumulate in such an irrational process, we have shown as far back as 1893, almost twenty years ago. That paper was presented by Berthelot to the Academy of Sciences of Paris and published in its Comptes Rendus.** _ After having completed a thorough examination of all the atomic weight determinations made, we have, by a sort of crucial test, dem- onstrated that the present value Ag 107.88, implying a departure of 120 thousandths, is impossible; that O 16 requires Ag 108 exactly, according to all determinations made during an entire century in all the laboratories of the world.** ; It is this same principle that is demonstrated by the diagrams here printed and this demonstration is made visible to the eye: It is 30t vanadium alone that causes the error affecting the laboratory — work, but all elements in the reaction contribute to the error recog- nized in the final result of the analytical work. _ Instead of the common notion that the work of the different emists conflicts in the different values they have presented as the tesults of their determination of the atomic weight of vanadium our figures here inserted show to the eye that all determinations made ee in the common result of Va 51 exactly. While no experimental work of any kind, done by man, with instruments and by chemical reactions, all of which are but approximations to a mathematical ' perfection, can be expected to give perfectly exact results, we have proved that the final error cannot be ascribed to vanadium alone, as _ continues to be done by the dominant school, but that on the contrary, all the elements present in a reaction contribute, each one its share, to the excess or deficiency resulting. It was therefore necessary to find the laws regulating this participation of the different elements in the errors of the reactions and of the entire experimental work. ving discovered these laws, we have applied them here, to the atomic weight determinations made for vanadium and present in the vo graphics (Figs. 1 and 2) the final results thus obtained. These figures show plainly that all the departures from the abso- *T_ 116, p. 695. _ “See paper read December 2, 1910, before the American Philosophical Society; Proceedings, 1910, pp. 359-363. nec eae Binet 2 a iT ae 2k ee i =e ' 5 . 214 HINRICHS—ATOMIC WEIGHT: OF VANADIUM. [ lute values are converging to zero along each of the lines of wor pursued in the eighty years by the different chemists ; there is but insignificant exception, which we shall consider, when we take the recent work of Prandtl. i Our Figs. 1 and 2 proclaim that the atomic weight of vanad is exactly 51 in all these determinations, just as sure as oxygen the atomic weight 16 exactly and silver 108 exactly, chlorine exactly, sodium 23 exactly; in fact, all the elements have as a om weights exactly the absolute values given in our publications of last twenty years. i Even the very first determinations made by Berzelius, with a fraction of a gramme of material at service, and only in one si determination, by the reactions designated (a), (b) and (c), confirm the value Va 51; for the deviation noted for Va affects the other elements present as well, and therefore it would be abs to suppose that the atomic weight of vanadium could be obtai from a reaction which fails to give an exact determination for other elements present. Thus reaction (c) represented by line | Fig. 1, gives by the single determination made by Berzelius on decigrams of the rather complex hydrated vanadium sulphate, departure of 400 thousandths from 51 for the atomic weight vanadium; but the same determination gave the atomic weight oxygen 310 thousandths low as marked on the figure; it also the atomic weight of sulphur 470 thousandths low as indicated the edge of the diagram and by the arrowpoint: for the real cire mark falls far beyond the limit of the diagram. a Is it so hard to understand that a reaction that fails to giv cise determinations for all the elements it involves cannot ne be expected to furnish a value of precision for vanadium? Is about time for each individual chemist to begin to consider simple facts for himself, as was the practice in former days? — It would be interesting to trace the gradual approach to the where all departures are zero, as exemplified in the actual work the successive chemists. This will be found to hold good for with the single exception already mentioned. We will only p to a few special instances, expecting the reader to go over the em ground by himself. 1] HINRICHS—ATOMIC WEIGHT OF VANADIUM. 215 _ The preliminary analysis, represented in the line A at the top of Fig. 1, gave not only the greatest departure for Va, but also for in the only determination made by Berzelius. When Roscoe, forty years after, made a series of four determinations the result entered at E in our figure (both 1 and 2) he cut the departure from 1062 to 195 for Va and from 263 to 47 for oxygen. If this determi- nation were repeated with the benefit of all the progress made in laboratory work, the resulting mark would undoubtedly find its place much nearer the zero departure on the line (mainly dotted) -on our figure (1). _ Let us also look at the results obtained by the use of reaction 270, best seen on Fig. 2, lines G, H, I, K, L, O, P. Of these lines, is the most distant, representing the largest departures: it marks one determination made by the old master in 1831. Then we meet, in going toward the center of perfection or zero departure, the lines H, I and K bearing the mark of Roscoe and only about half as far from the zero as the line G of 1831. The total number of deter- tions made by Roscoe in 1868 was 8, represented by line I; 6 of these eight were “done by Analyst A” and are represented by line H ; the other two were “ done by Analyst B” and are represented by line K. Finally we have three series (I., II. and III., repre- sented by lines L, P and O, respectively) made by Prandtl quite recently ; of these, the last two series come quite close to the center, departure for Va being 14 only for the mean of the two series see Table V.). Since on our diagram the departures for O, Cl Ag are set off as abscisse to that of Va taken as ordinate, the wadual diminution of all the departures is strikingly shown in the lines for these elements converging to the point of zero departure. _At the same time we here have the positive evidence that Prandtl has produced two very concordant series (his IT. and III.) with a very small departure (mean 14 for Va and O) and one series (his first, I.) represented by line L for which the departure is 62, that is more than four times as large. : Now we may consider the results obtained by the reduction test, reaction No. 98, represented on our Fig. 1 by the lines B, C and F. lere we meet that one exception before referred to; for the greatest departure (line B) of Berzelius is but slightly diminished by Prandtl’s 216 HINRICHS—ATOMIC WEIGHT OF VANADIUM. [April 2 quite recent work (line C) and far surpassed by the small departures of the much older work of Roscoe (line F). It is to be hoped t Prandtl will also make one or two additional series under react 98, as he has done under 270; we dare say that, with due care, may repeat the experience he has reported for reaction 270 < greatly reduce the departure, at least to that of Roscoe in 1868. It will hardly be necessary to state that the numerical data have quoted in the discussion of our Figs. 1 and 2 are taken H ) tables II. and V. especially. _CONCLUSION. We think that the reader will have no trouble now in comple; his study of the facts placed before him in our tables and i figures which both comprise much more than their size oe to indicate. ; ) We therefore think that the reader will fully understand the fallacy of throwing all the errors of all kinds on the one el em: the atomic weight of which the modern chemist tries to dete: “jin the chemical laboratory and by experiment exclusively.” The reader will, we believe, now fully comprehend the si tion—both of the chemist and of his intended victim, the ele The victim—if it were conscious—would shiver in anticipation being made responsible for every error and mishap that may any of the elements present in the reactions, the apparatus usec even the chemist at work; for all these errors and shortcomings. modern chemical school de facto charges up to the element atomic weight of which it undertakes to determine. The only step—other than what the general progress of practical labora’ work may favor him with—will be the straining out a few m innocent gnats without in the least disturbing the ever attend herd of the old camels. | ‘It sometimes does seem strange that in twenty years this Be : lian picture from Saint Matthew (XXIII., 24) has remained true to Nature. It is not the fault of the individual chemists, cept in so far as they have surrendered a fundamental part of rightful domain to the International Atomic Weight Committee. _ THE ORIGIN AND SIGNIFICANCE OF THE PRIMITIVE 4 NERVOUS SYSTEM. By G. H. PARKER. (Read April 21, 1911.) Linnzus defined a plant as an organized, living, but non-sentient _ body and an animal as an organized, living, and sentient body. 4 Although no modern biologist would attempt to support the conten- tion that animals are sentient and plants are not, the distinction drawn by Linnzus is not without a certain foundation in truth, for _sentience in its full development and as Linneus probably under- stood it, is the exclusive and supreme possession of the higher ani- _ mals. That these animals possess intelligence as contrasted with 4 _all other natural bodies is a statement to which few naturalists will offer any serious objection. The seat of this intelligence is the ner- vous system and, though the integrity of the other systems of organs is essential in most cases to the well-being of the animal body, the fact that the totality of activities that makes up the mental life of human beings as well as that of other animals, is absolutely depen- dent upon the nervous system, is evidence sufficient of the paramount importance of these organs. It is, therefore, not without interest _ to inquire into the origin of this system of organs and to trace the early steps by which it passed from a position of initial obscurity to one of the highest significance. The nervous system of the higher animals, though enormously complex in its organization, is composed of relatively simple ele- ments, the neurones, arranged upon a comparatively uniform plan. This plan is well exemplified in the spinal cord of the vertebrates. g In this organ the sensory neurones, whose cell-bodies lie in the dorsal ganglia, extend from the integument through the dorsal roots _ to the gray matter of the cord. Motor neurones, whose cell-bodies _ are situated within the gray matter of the cord, reach from this _ region to the muscle-fibers which they control. These two classes 217 218 PARKER—ORIGIN AND SIGNIFICANCE OF [April of neurones would seem to be sufficient for all ordinary reflex o ations, but the cord contains within its limits many other neur which serve to connect one part of its structure with anoth These neurones, therefore, have beer! called association neurones, term which has unfortunately proved to be somewhat mislead because of its use in psychology for quite a different range phenomena. The so-called association neurones are interpolat : between the sensory and motor elements just described and m thereby lengthen and extend the courses of the reflex impuls Such neurones make up a large part of the substance of the cord al doubtless increase enormously its internal connections. In the t they not only add to the nervous interrelations, but they aff the region of the cerebral cortex the material basis for all intell tual operations. a The plan of neuronic arrangement as exemplified in the ve brates also obtains in animals as lowly organized as the earthworm In this form the sensory neurones, whose cell-bodies are situated the integument instead of being gathered into special ganglia, tend, as in the vertebrates, from the skin to the central ne organs, the brain or the ventral ganglionic chain. The motor 1 rones are essentially duplicates of those in the vertebrates in their cell-bodies lie within the central organs whence their fib extend to the appropriate musculature. Association neurones < also abundantly present in the earthworm though their function het in contrast with that in the higher vertebrates, is pure nervous tercommunication, for it is very unlikely that the earthworm sesses what in any strict sense of the word can be called in gence. Thus from a morphological standpoint, the nervous sy of the higher animals, even including such forms as the worm, have much in common, their three sets of interrelated rones, sensory, motor, and association, being arranged i an sentially uniform plan. Considered from a physiological standpoint, the nervous syste with its appended parts as just sketched falls in the higher anil ne into three well marked categories. On the exterior of these ani are to be found sense organs or receptors such as the free-ne terminations of the sensory neurones in the vertebrates or the sen- THE PRIMITIVE NERVOUS SYSTEM. 219 sory cells in the integument of the earthworm. These organs have for their function the reception of the external stimuli and the pro- duction of the sensory impulses. The receptors are connected by nerve-fibers with the central nervous organ or adjustor composed of e central ends of the sensory and the motor neurones and of the association neurones. Here the impulses arriving from the re- _ ceptors are directed toward the appropriate groups of muscles by _ which the animal may respond to the stimulus and, if the animal is highly organized, impressions are made upon the adjustor which, as memories, may become more or less permanent parts of the animal’s nervous equipment. Finally the adjustors are connected by nerve- fibers with the third set of elements, the effectors, which as muscles, electric organs, glands, etc., enable the animal to react on the en- vironment. Thus three physiological categories are to be distin- guished which in the order of their sequence in action are sense organs or receptors, central nervous organs or adjustors, and muscle, te., or effectors. It is to be noted in passing, that the physiological scheme just utlined includes a wider range of parts than is generally admitted under the head of the nervous system. The additional parts are the effectors, which, as will be shown later, form as truly a part of the whole system as do the sense organs or the central nervous organs. Since the term nervous system does not ordinarily include the effec- tors, it is perhaps best to designate the whole chain of related parts, receptors, adjustors, and effectors, as the neuromuscular mechanism and in dealing with the origin of the nervous system, it will be found important to keep this relation in mind, for in such an inquiry, the teal question that must be confronted is the origin of the neuromus- cular mechanism rather than that of the nervous system alone. _ The type of neuromuscular mechanism described in the preced- ing paragraphs in which a group of receptors is connected with a well centralised adjustor which in turn controls a complex system of effectors, is found only in the more differentiated metazoans. Cer- tainly in the simple metazoans, like the jellyfishes, corals, sea-ane- mones, etc., only the slightest evidence of this type of nervous orga- nization can be discovered. Nevertheless these animals possess a ‘neuromuscular mechanism but on so simple a plan that investigators 220 PARKER—ORIGIN AND SIGNIFICANCE OF _ [April have long been inclined to regard it as representing the first in the differentiation of the neuromuscular organs. This plan structure is well represented in the sea-anemones. Each of the 1 chief layers of cells that make up the living substance of the anemone’s body consists of three sublayers: a superficial or epi layer, a middle or nervous layer, and a deep or muscular layer. epithelial layer contains, besides many other kinds of cells, numbers of sensory cells which terminate peripherally in br like receptive ends and centrally in fine nervous branches. 7 fine branches constitute collectively the middle or nervous layer which occasionally large branching cells, the so-called gang cells, occur. Immediately under the nervous layer is the deep la of elongated muscle-cells. The condition thus briefly described present over the whole of the sea-anemone’s body and though » nervous layer is somewhat emphasized in the neighborhood of mouth, it cannot be said to be really centralised in any part. H this type of nervous system has been designated as diffuse in trast with the centralised type found in the higher metazoans. Not only is the structure of the nervous system of the sea-a one appropriately described as diffuse, but in its action this syst shows those peculiarities that would be expected from the pos of so diffuse an organization. Since each part of the animal tains its own nerve and muscle, it is not surprising that after tion many of these parts will respond to stimuli much as they when they were a constituent of the whole organism. Tentacles, instance, when freshly cut from the body of a sea-anemone will spond to pieces of food by encircling them, etc., in much th way as when these organs were parts of a normal animal. evidence of this kind has shown conclusively that the nervous tem of ccelenterates is no more centralized physiologically than anatomically, but is in all respects essentially diffuse. What is really present in the neuromuscular portion of the anemone’s body is a large number of peripheral sensory cells y deep branching ends connect more or less directly with the mu i. e., without the intervention of a true central organ. This romuscular system, if described in the terms already used, co said to be composed of receptors and effectors without an adju Pieer.3 THE PRIMITIVE NERVOUS SYSTEM. 221 or at least with this member present in only a most primitive state. In my opinion this is the condition in most ccelenterates. Judging _ from the more recent work on the nervous organs of these animals, a centralization can scarcely be said to be present at all in hydra; it is _ but little more pronounced in the sea-anemone; and, though most ' marked in the jellyfishes, it does not rise even here to a grade that entitles it to comparison with what is seen in such forms as the _ earthworm. The ccelenterates, then, are animals possessing recep- tors and effectors but without developed adjustors. Hence the : _ adjustor or central organ is in all probability an acquisition that 3 tepresents a later stage in the development of the neuromuscular mechanism than that seen in the ccelenterate. If the ceelenterates represent a stage in the evolution of the neuromuscular mechanism in which sensory cells and muscles are the only important parts present, it is natural to ask if there is not a _ still more primitive state from which the ccelenterate condition has ' arisen. On this question several hypotheses have already been ad- vanced. Claus and, subsequently, Chun maintained that originally the nervous system and the muscles were differentiated indepen- dently and that they became associated only secondarily. This view has deservedly received very little attention, for not only is it difficult to conceive that an animal would develop receptive ability without at the same time acquiring the power to react, but not a single exam- ple among the lower animals is known in which developed nerve and muscle are present and independent of each other. Much more worthy of consideration than the hypothesis of the independent origin of nerve and muscle is Kleinenberg’s theory of the neuromuscular cell. In 1872 Kleinenberg announced the dis- _ covery in the fresh-water hydra of what he designated as neuro- muscular cells. The peripheral ends of these cells were situated on

produced. In the light of the results here reported the temptation is strong to suggest that some phases of these pathologic states are E associated with the strenuous and exciting character of modern life a acting through the adrenal glands. Two of my students, Shohl and Wright, have recently shown that excitement of the cat results, in all cases thus far examined (more than a dozen), in glycosuria. Possibly i in the wild state emotions were useful in providing material 1 excessive muscular exertion that might follow, and that muscular activity would utilize the sugar so that it would not appear in the urine. This suggestion, however, must be put to further test. _ PROC, AMER. PHIL. SOC,, I, 199 0, PRINTED JUNE 28, IQII. THE CYCLIC CHANGES IN THE MAMMALIAN OVARY By LEO LOEB. (Read April 22, 1911.) The observations which I: wish to report to you are of interest from several points of view: 1. The process upon which the sexual cycle in mammals de has been analyzed, and a regulatory mechanism was found to one within the ovary. 2. A striking illustration is presented of the fact that the struc ture of organs is in many instances at least not a definite one, varies in correspondence with the functional condition of the org 3. The accurate description of the normal cyclic changes in » th mammalian ovary serves as a basis for the investigation of the path logical deviations which interfere with the natural course of th sexual functions and may lead to a temporary or lasting sterility. 4. We found in the ovary structures which must in all proba be interpreted as early stages of embryos developing spontaneo parthenogenetically within the ovary and it is probable that the d opment of these parthenogenetic embryos is related to certain p of the sexual cycle. In the ovary of the guinea-pig definite and very interesting cy changes exist which I made the object of my studies in the last fe years. : The mammalian ovary consists of two principal constituent pa namely: First, large and small bodies lined by granulosa cell filled with fluid, the so-called follicles; and, secondly, the corp lutea. Both follicles and corpora lutea have only a brief exist they develop to a certain point and then they degenerate and ually disappear. The follicles contain the ova. At certain pe of the sexual cycle a few follicles that have reached maturity rupt The ova reach the Fallopian tubes and uterus and after fertili 228 LOEB—CYCLIC CHANGES IN THE OVARY. 229 by a spermatozo6n form an embryo in the uterine wall. New fol- licles situated in the periphery of the ovary grow constantly to a cer- tain size and then degeneration sets in. The lining granulosa cells disintegrate and connective tissue grows into the cavity of the follicle. The ova in these degenerating follicles undergo frequently matura- _ tion and a few more or less regular divisions and then die. While _ thus the majority of follicles degenerate, become atretic, before they have reached maturity, a few follicles undergo certain progressive changes and become mature. They may rupture and discharge the _ ovum ; such a rupture is called an ovulation. The process connected with ovulation causes a degeneration of all with exception of the smallest follicles. These follicles grow and in the course of the next six days they have reached that size at which degeneration may set in. We find, therefore, degenerating follicles from the seventh day iter ovulation up to the time of the following ovulation. While ifter the seventh day medium-sized follicles constantly degenerate, follicles grow and take the place of the degenerating disappear- ing ones. It seems that it takes approximately ten days until some the follicles reach their full size. _ We may therefore distinguish two periods in the ovarian cycle: irst, the period of growth extending over the first seven days fol- ving ovulation, and, secondly, the period of equilibrium in which new follicles take the place of degenerating ones. A/fter the first large follicles have appeared, it takes a few days longer until large follicles become transformed into mature follicles that are ready to Tupture. We find, therefore, the first mature follicles to appear roximately eleven to thirteen days after ovulation and it would natural to expect that about fourteen days after the preceding a new ovulation should take place. The sexual period—that is the period between two ovulations— should therefore have a natural duration of approximately two weeks the guinea-pig. This is, however, not the case. The sexual period in this species actually lasts about twenty to twenty-five days. d this is due to the fact that a mechanism exists within the ovary t prolongs the sexual cycle. In order to understand this mech- ism, we must follow the fate of the ruptured follicle. A follicle RTA TEN MENTION NL EASES 230 LOEB—THE CYCLIC CHANGES IN that has ruptured at the time of ovulation does not degenerate in y similar manner as the other follicles do after they have reached full size, but they grow in a remarkable manner and form a new gland- like organ, the corpora lutea. Now these corpora lutea also degen- erate after a period of growth that lasts approximately seventeen to twenty days. In the corpora lutea resides the mechanism that pre- vents a new ovulation. It is necessary that they degenerate before a new rupture of follicles can take place. As long as they function they prevent ovulation. The fact that the corpora lutea degenerate 3 when seventeen to twenty days old, explains why a new ov takes place approximately every three weeks. If we excise the pora lutea at an early date after ovulation, a new ovulation oc very soon after mature follicles have made their appearance, appro? mately thirteen to fifteen days after the preceding ovulation. Und these conditions, the normal sexual cycle is reéstablished. During pregnancy the life of the corpus luteum is prolongec consequence of the changes occurring in the uterus or develoy embryo during the period of gestation and in consequence of the | longed life of the corpus luteum, a new ovulation is prevented. du: the whole course of pregnancy. Toward the latter part of p nancy, the corpora lutea again degenerate and directly after pleted labor a new ovulation can take place. Ovulation, therefore, depends upon three factors: First, upon maturation of ovarian follicles ; secondly, upon the time of dege tion of the corpora lutea; and, thirdly, upon less important, mo: less accidental conditions, as for instance, the process of copula The third class of conditions accelerates in many (not in all) ¢ ovulation, but it is not necessary for its occurrence. Even a preceding copulation, ovulation usually takes place, but in cases at a later date. Through what mechanism does the li the corpus luteum influence ovulation? It might be conceivable the corpus luteum delays the maturation of follicles thus preve a rupture. My observations have, however, shown that an inhi influence of the corpus luteum upon the maturation of follicles not exist. Mature follicles appear frequently during the life o corpus luteum, and especially during the period of pregna 3grt.] THE MAMMALIAN OVARY. 231 seems that pregnancy even favors the maturation of follicles. The ‘corpus luteum prevents, however, the rupture of the mature follicles. Pregnancy as such does not prevent the rupture provided the corpus luteum has been previously removed through excision. _ The structural changes in the ovaries are rhythmical and so reg- ular that a careful histological examination of these organs enables _us to decide within.a certain limit of accuracy at what périod of the _ sexual cycle the animal had been at the time of the removal of the ovaries. Having established the normal cycle I turned more recently my attention to its pathological deviations. It occurs in a certain num- ber of animals—and I have observed this to happen among females which showed no desire for copulation or in which notwithstanding an accomplished copulation an ovulation did not follow—that the ollicles do not grow to maturity, but that they undergo involution before they reach their full size, and that all, or almost all the folli- cles, become sclerosed, atretic, at a very early stage of their develop- ment. Under these conditions, an ovulation is impossible and the animal, in the ovaries of which such a deviation from the normal cyclic changes exists, are at least temporarily sterile; whether such a pathological condition may ever lead to a permanent sterility, future investigations must show. It will be readily understood that _here we have to deal with questions of the greatest importance to the physiology of the sexual functions. In order to appreciate thoroughly the conditions under which ‘such abnormalities in the sexual cycle occur, it is necessary to pro- duce the subnormal development of the follicles experimentally. Now it is of interest to know that such a premature involution of the varian follicles can be produced experimentally by burning a certain relatively small part of the ovaries with the thermocautery. The remaining larger part of the ovary remains apparently perfectly well, the cells functionate but the energy of growth of certain cells is diminished and subinvolution of the follicles with resulting temporary sterility follows. A comparable condition can be produced in tumors through heating, or through the influence of certain chemicals exerted in vitro as I found a number of years ago. Under such conditions 232 LOEB—THE CYCLIC CHANGES IN the tumors grow, but with markedly diminished energy. In both _ cases, in the case of the tumors as well as of the ovaries, we have to ‘deal with a state of living matter intermediate between its full ural vigor and latent life; we may regard it a state of partial shock of cells, in which the growth takes place but with a considerable decrease in energy. Besides the changes which I have just described, additional processes of the greatest interest take place in the ovaries of a certain number of animals and it is very probable that these processes usually commence at the period following ovulation are therefore, in a certain sense, a part of the cyclic ovarian changes. The process I refer to concerns an apparently spontaneous partial parthenogenetic development of ova in the mammalian ovary, occurrence of which I obtained convincing proof only within last few months. * Some years ago I described peculiaf structures that are found i the ovaries of guinea-pigs, and I expressed the opinion that they or inated in the ovarian follicles... Very soon after I had published observations, certain considerations suggested to me that these sti tures owe their origin to parthenogenetically developing ova. In as much as at that time I had not yet seen early stages of th structures referred to, | was unable to regard this hypothesis as su ciently founded to warrant publication. I continued, however investigations in this direction, and recently I succeeded in finding two animals the desired early stages. They must be interpreted embryos developing parthenogenetically within the ovary of 1 guinea-pig. We see in each case a chorionic vesicle with trophe blast, and plasmodia and syncytia penetrating into the neighbo tissues. “ There is present also a structure which is probably to interpreted as a neural tube. pe Aberrant blastomeres (remnants of dividing ova that failed participate in the embryonic development) cannot be seen in ovaries of guinea-pigs, and, inasmuch as the embryonic structu described in my former communication, are relatively frequent, curring in approximately ten per cent. of all guinea-pigs below th age of six months, and, furthermore, inasmuch as they are sitt * Archiv f. mikrosk. Anatomie, Bd. 65, 1905. THE MAMMALIAN OVARY. 233 _ in the cortex of the ovaries at a place where follicles lie normally and are found within follicle-like cavities, they can only be derived from ova developing parthenogenetically. Fertilization through sper- ‘matozoa can be excluded, inasmuch as the history of some of these animals is known to us and precludes such an interpretation. It is very probable that the parthenogenetic development sets in soon after ovulation, the altered conditions in the ovaries at that time (varia- _ tions in blood pressure, in intrafollicular pressure or changes in gas 4 exchange) supplying the necessary stimulus. This interpretation agrees well with my former observations concerning the parallelism "existing between the first segmentations taking place in non-fertilized va within the ovary and certain stages of atresia of follicles.? It is also of interest to note that frequently these changes are ‘multiple, several ova undergoing parthenogenetic development in the ‘same ovary. _ We may, therefore, conclude that in at least ten per cent. of all inea-pigs parthenogenetic development of the ova within the ovary ‘starts at some period of the life of the animal. The later stages of these developing embryos bear some resemblance to chorionepithe- iomata, certain tumor-like formations consisting of proliferating chorion tissue. During ovulation these structures are occasionally injured by hemorrhages and they are ultimately invaded and sup- planted by the neighboring connective tissues. These observations throw furthermore light on certain interesting tumors that are especially found in the ovaries and testicles, namely: the teratoid tumors and the chorionepitheliomata. My observations are a strong argument in favor of the view that teratoid tumors that _are found in the ovaries are not derived from misplaced blastomeres, as Bonnet and Marchand believed, but that the older view is correct according to which they are derived from parthenogenetically devel- oped ova, an opinion which I, also, expressed on previous occasions. € same statement can be made in the case of the chorionepithe- liomata that occur in the ovaries and in the testicles. I believe that e observations here recorded clear up the mechanism of the sexual ?“On Progressive Changes in the Ova in Mammalian Ovaries,” Journal Medical Research, Vol. 1, 1901. neous parsienneeneie development ss ova takes place during the life of the animal. — DEPARTMENT OF PATHOLOGY, _ ea BARNARD FREE SKIN AND CaNceR HosPiraL, — St. Louts, Missouri, THE SOLAR CONSTANT OF RADIATION.* By C. G. ABBOT. (Read April 21, 1911.) If we had no eyes we should still know of the sun by the feeling f warmth. The intensity of solar rays in any part of the spectrum can be measured by delicate thermometry. Vision and photography are both restricted within comparatively narrow limits of wave- length, and each differs in its sensitiveness from wave-length to ave-length. Ultra-violet, visible and invisible red rays, however, all produce their just and proportional influences on the bolometer, ‘or thermopile. This is not universally known, and there are still many who suppose we should distinguish between so-called actinic, visible and heat rays. Doubt has been expressed, for instance, ‘whether bolometric measurements give true indications of the ‘intensity of those rays which promote plant growth. Such doubts are not justified, and we may expect very valuable results in the future from the application of the spectro-bolometer to the interest-_ ing questions of radiation and plant physiology. _ We use heat units to express the intensity of solar radiation. The solar constant of radiation may be defined closely enough as the number of degrees by which one gram of water at 15° centi- grade would be raised, if there should be used to heat it all the solar radiation which would pass at right angles in one minute _ through an opening one centimeter square, located in free space, the earth’s mean solar distance. Experiments were begun about 1835 by Pouillet and by Sir John Herschel for the measurement of this great constant of nature. The investigation has been continued by Forbes, Crova, Violle, Radau, Langley, K. Angstrom, Chwolson, W. A. Michelson, Rizzo, Hansky, Scheiner and others. It is an ndication of the great difficulty of the research that entire uncer- * Published by permission of the Secretary of the Smithsonian Institution. 235 236 | ABBOT—SOLAR CONSTANT OF RADIATION. _ [April tainty as to the value of the solar constant of radiation between limits of Pouillet’s value, 1.76 calories, and Angstrém’s value, 4. calories per square centimeter per minute, prevailed at the beginn of the twentieth century. Professor Pringsheim has collected the following table* of ails constant values, as determined by different observers: Year, Observer. Calories. Year. Observer, Calories. 1837 Pouillet 1.8 1889 | Pernter 3.2 1860 Hagen 1.9 1896 | Vallot Fy fe 1872 Forbes 2.8 1897 | Crover and Hansky 3.4 aE Violle 2.6 1898 Rizzo 2.5 187 Crova 2.3 1908 | Scheiner io ee 1884 Langley 9.3 1908 = Abbot and Fowle ea* 1889 Sawelief 2.9 as He omits Angstrém’s 4.0, published in 1890 and withdra in 1900, but which is even yet sometimes quoted. He omits als Very’s 3.1, published in 1901 and independently obtained in 1 Recently published values of Kimball, Gorezynski and others, proximately 2.0, are based in part on work of Abbot and Fowle. — The determination of the solar constant involves: (1) corr measurements of the heat equivalent of the solar radiation at _earth’s surface; (2) a correct estimate of the losses which the rz have suffered in the atmosphere before they reached the measurinj apparatus. We shall now discuss these two branches of the wo r Pouillet invented, about 1835, his well-known instrument, pyrheliometer, for measuring the solar rays at the earth’s sur Many criticisms have been justly made in regard to the acc of this pioneer instrument, and attempts have been made by ma’ improve on it, or to substitute a better. In our practice at Smithsonian Astrophysical Observatory, we have substituted a s disk for Pouillet’s water chamber; inserted a cylindrical bulb mometer, radially instead of axially, in the disk; provided a lined wooden chamber to screen the instrument from the wind; added convenient adjuncts for shading and exposing the ins *“ Physik der Sonne,” p. 417. *This value was expressed in terms of a provisional scale of p rh ometry which has since been proved too high. 191] = ABBOT—SOLAR CONSTANT OF RADIATION. 237 ment.* Finally we have ceased to regard our instrument as giving more than relative measurements. It is only a secondary pyr- heliometer for convenient use. We standardize its readings by com- parison with an absolute pyrheliometer of another kind. ____No known substance absorbs radiation perfectly at a single en- a counter. Kirchhoff showed, fifty years ago, that a hollow chamber must absorb perfectly, because of the opportunity for an infinite number of absorption encounters within it. W. A. Michelson, in 1894, invented a standard pyrheliometer including a hollow chamber with a narrow opening for the admission of rays. The walls of the chamber were bathed by a mixture of ice and water, and the heating effect of the solar rays was measured by the amount of ice melted, which was determined by noting the expansion in volume of the mixture of ice and water. Nearly ten years later, being ignorant of Michelson’s pyr- heliometer (which was described in the Russian language), it occurred to me also to employ a hollow receiving chamber. I pro- posed to measure the solar heating produced in it by bathing its walls with flowfng water, and determining the rate of flow and rise of temperature of the water. After experiments lasting inter- mittently from 1904 to 1910, I am now satisfied that this device has proved successful, and that we have truly an absolute standard pyrheliometer. With the aid of my colleagues, Mr. Aldrich and Mr. Fowle, two of these water-flow pyrheliometers were carefully tested last year.* Not only did they agree in measurements of solar radiation, but test quantities of heat introduced electrically within _ the absorbing chambers were accurately recorded by the methods ordinarily used to measure solar heating. We believe of the abso- lute water-flow pyrheliometer, it gives the intensity of solar radia- tion at the earth’s surface in calories per square centimeter per ‘Minute within a probable error of 0.2 per cent. For convenience ‘we make our daily observations with secondary silver-disk pyr- heliometers, which have been standardized against the absolute water-flow pyrheliometer. *See Abbot, “The Silver Disk ‘Pyrheliometer,” Smithson. Misc. Coll. Vol. 56, No. 19, 1911. ) *See Abbot and Aldrich, Astrophys. Journal, Vol. XXXIII., 125, rgtt. 238 ABBOT—SOLAR CONSTANT OF RADIATION. [April Having perfected the standard and secondary pyrheliom to a satisfactory degree of accuracy and durability, the first brane of solar-constant work is accomplished by reading with the silve disk pyrheliometer at the earth’s surface, and reducing its ir cations to calories per square centimeter per minute. We now to the discussion of the second branch of the work, namely tl estimation of the transmission of the atmosphere for radiation. Lambert and Bouguer showed almost simultaneously, about 17 that the transmission of light through a homogeneous medium be expressed by an exponential formula, such as: sae E== Fa". Here E is the intensity transmitted, E, the original intensi the fraction transmitted by unit thickness, and m the actual thi ness of the transparent medium. _ 3 Pouillet applied Bouguer’s formula to the atmosphere. As atmosphere is not homogeneous, but decreases in turbidity density from the earth’s surface upward, this would seem at sight unjustified. But if we consider unit thickness to be the tl ness of the atmosphere traversed by a vertical beam, then as the departs from the vertical, it still shines through every layer v it did at first, and the path in every layer increases nearly secant of the zenith distance of the ray. Under these circums it can be shown that (subject to certain limitations to be ment the exponential formula given above should hold; if we consid to be the intensity at the earth’s surface, E, the intensity outsid atmosphere, a the transmission coefficient for a vertical — the secant of the sun’s zenith distance.® Owing to atmospheric refraction, the fractional increase i of the ray, as the zenith distance waxes, tends to be greater for 01 outer layers of the atmosphere than for its inner ones. other hand, the curvature of the earth’s surface produces an op site tendency. But for zenith distances less than 70° these may be neglected, and they are hardly worth considering *See Annals Smithson. Astrophys. Obser., Vol. I1., p. 14, 1908. ABBOT—SOLAR CONSTANT OF RADIATION. 239 zenith distance? Solar-constant determinations require no higher values of zenith distance than these to be considered. Radau and Langley proved the necessity of confining the atmo- spheric use of Bouguer’s formula to approximately monochromatic tays. In general, for a reason which Lord Rayleigh has shown, the ‘transmission of the atmosphere increases gradually with increasing wave-lengths. Thus in the violet the transmission for a vertical ray to sea level may be 50 per cent., and for a deep red ray 80 per cent. But besides this gradual change there are also spectral regions of almost complete absorption by atmospheric oxygen, and by water- ‘vapor, so that in these regions the transmission approaches zero. If we should disregard these differences, and determine the con- tants of the exponential formula above, by pyrheliometric measure- ents alone at different solar zenith distances, our result E, for the intensity outside the atmosphere must necessarily be too small.* _ Langley was the first to act upon this, and to devise apparatus and methods for measuring the energy and the atmospheric trans- mission at all parts of the spectrum. For this purpose he invented the bolometer about 1880, and automatic registration of its indica- tions about 1890. As we now use it the bolometer comprises two similar tapes of platinum, each about 1 cm. long, 0.01 cm. wide and .0OI cm. thick. These are coated with lamp-black by smoking over a camphor flame. They lie parallel to the spectrum lines, and about 0.8 cm. apart. One tape may be shined upon by the rays, ‘the other can not. Hence the heat absorbed from a narrow region of spectrum, usually about twice the extent comprised between the Dp lines, raises the temperature of the exposed tape with reference to the other. The two tapes and two resistance coils are combined to form a Wheatstone’s bridge, and the rise of temperature produced _as above stated deflects a sensitive galvanometer. The galvanometer ; needle ‘reflects a tiny spot of light on a photographic plate, which “moves vertically as driven by clock work. The same clock work “moves the spectrum slowly over the bolometer tape. In this way _may be produced in from eight to twelve minutes, according to the _ spectroscopic outfit employed, a bolograph, or spectrum energy 5 ce ali tae ta Pa eed ace ot ot * Loe. cit., p. 59. * Loc. cit., p. 16. 240 ABBOT—SOLAR CONSTANT OF RADIATION. _ [April 2 curve, extending from about wave-length 0.30 in the ultra-violet to about wave-length 3.0 in the infra-red. Its ordinates are deflec- tions of the galvanometer, proportional to energy in the spectrum, and its abscissee are proportional to differences of prismatic devia- tion. The Fraunhofer lines, and great oxygen and water-vapor bands, show as depressions in the curve. In order to eliminate dis- tortions which are due to differences, for differing wave-length: in the reflecting power and transmission of the mirrors and prism used in the optical train, special investigations of the transmission of the apparatus are made from time to time, and the curves ore accordingly. ‘ In our ordinary practice, from six to eight bolographs are tical in a single forenoon, between the times when the sun’s zenith dis- tance is 75° and (say) 30°. The curves are measured at ab thirty positions, uniformly spaced in the prismatic spectrum. group of six to eight measurements, at a single spectrum place, f nishes means of computing from Baga formula the aa made outside the atmosphere. The sum of the ordinates measu on any bolograph is approximately proportional to the total ener of all wave-lengths observed. Similarly the sum of the ordina computed for outside the atmosphere is proportional to the t energy there.® . In order to reduce the total energy, as determined bolometricz to calories per square centimeter per minute, the pyrheliometer read, while the spectro-bolometric work is in progress, on each ¢ of observation. Thus a factor is obtained for deducing from areas of the bolometric curves the true heat units corresponding. A complete determination of the solar constant of radiation requi *In the regions of great water-vapor and oxygen absorption the atmospheric curve is determined by interpolation between adjacent paratively unaffected wave-lengths on either side, for we know that there no oxygen or water-vapor absorption of these bands produced in the su that they ought not to show in the extra-atmospheric curve. Small allow are also made for the energy of lesser and greater wave-lengths than observed. * For further details consult Annals, Vol. II. ABBOT—SOLAR CONSTANT OF RADIATION. 241 about three hours of observation, under a cloudless and uniformly clear sky, and about three days of computing. We began to make solar-constant observations in Washington at the Smithsonian Astrophysical Observatory, in October, 1902, and continued them there whenever a favorable opportunity was " presented, until May, 1907. In all this time we made only 44 erably satisfactory determinations at Washington, for cloudless days were rare, and many days that promised fairly proved disap- pointing, by reason of the appearance of smoke, haze or clouds. Four important results came from the Washington observations. irst, no apparently good determinations yielded solar-constant va lues above 2.38 of our then provisionally adopted scale of calories, or 2.25 true calories. Second, the mean value in the true calories m 44 determinations was 1.960. Third, the transmission of the osphere was determined on many days, and for many wave- ths. Fourth, a strong probability was raised by the results of observations of 1903 that the sun is a variable star. This variation seemed to reach 10 per cent. in its extreme range, but no tendency ards a regular period was then found for it. A dependent ation in terrestrial temperatures seemed indicated. Primarily in order to make spectro-bolometric determinations of the solar constant, suitable to test the supposed variability of the un, an expedition under my charge went out to Mount Wilson in 190: , by invitation of Director Hale of the Mount Wilson Solar ne bservatory. The site proved excellent for the purpose, on account 3 f its considerable altitude, cloudless sky and freedom from wind. Much aid and comfort was furnished by Director Hale and his staff. The expedition was repeated in 1906, 1908, 1909 and 1910. 'e now occupy a cement observing shelter and living quarters there, n ground leased from the Solar Observatory. Our observations e generally occupied the six months, May 15 to November 15, in the last years we have made practically daily determinations the solar constant of radiation during this interval. * Astronomers have not yet very generally availed themselves of the accurate coefficients of atmospheric transmission obtained in our researches tor all parts of the spectrum, and from Washington, Mount Whitney and Wilson. *See S. P. Langley, Astrophysical Journal, Vol. 19, p. 305, 1904. 242 ABBOT—SOLAR CONSTANT OF RADIATION. _ [April It was thought doubtful by Langley, and others, if correct es mates of the atmospheric transmission can be made, even by the spectro-bolometric method of high and low sun observations. La ley, indeed, gave an argument tending to show that the values of solar constant thus obtained fall far below the true intensity of solar radiation outside the atmosphere. This argument, howev seems to be unsound.’* In order to test the accuracy of the me I made spectro-bolometric measurements on Mount Whitney (44 meters elevation) in 1909 and 1910 simultaneously with similar ol servations made by Messrs. Ingersoll and Fowle, respectively, ¢ Mount Wilson (1800 meters elevation). In 1905 and 1906 sola constant measurements were made nearly simultaneously at Me Wilson and at Washington (10 meters elevation). It does not. pear from these observations that there are any differences in solar-constant values depending on the altitude of the observer, not due to accidental errors of observations.’* In illustration of this conclusion I give the results obtained s taneously at Mount Wilson and Mount Whitney: Date. | 19 1909, Sages 3. rgro, Aug. 12. ! rg10, Aug. 13. | 1910, Aug. | 1.943 | 1.943 | 1.924 | 1.904 The very slight excess of the Mt. Whitney values is not I. enough to be significant. G We conclude that the solar-constant values computed from method of high and low sun observations do not depend on altitude of the observing station up to altitudes of 4,420 me provided the sky conditions are satisfactorily clear and unifc Reducing values published in Vol. II. of the Annals to stan calories at 15° centigrade, and including the mean values obi in later years,!® we have: : * See Annals, Vol. II., pp. 119-121. “As regards the Washington and Mount Wilson compariso Annals, Vol. IL, pp. 99 and 102. Note that the provisional scale of Annals values is 5 per cent. too high. at * Many of the values of 1910 are not yet reduced. ABBOT—SOLAR CONSTANT OF RADIATION. 243 SoLtar—ConsTtANT MEAN VALUES. Place, Washington. | Mount Wilson. | Mount Whitney. Date. 1902-1907 1995 1906 1908 : 1909 1910 | 1909 | Igto observed........ 44 59 | 62 | 113 | 9s | 28 1 3 PIO cscs, soc 1.960 1.925 1.921 1.929 1.896 1.914 | 1.959 | 1.956 - Our observations indicate as the mean value of the solar con- stant of radiation: 1.922 calories (15° C.) per square centimeter per minute. The observations having been obtained mainly near the time of sun-spot maximum we think it probable that their mean is hardly high enough to represent the average condition of the sun. We incline to think this because it has been shown by Koppen, Nord- mann, Newcomb, Abbot and Fowle, Bigelow, Arctowski and others . hat the earth’s temperature is a little lower at sun-spot maximum n at sun-spot minimum. This probable correction cannot exceed e or two per cent. There is another reason why our value of the solar constant may too low. We have not been able to observe, even on Mount Whitney, any radiation beyond the wave-length 0.29 in the ultra- “violet spectrum. Whether the rays of less wave-length are oblit- erated in the earth’s atmosphere or in that of the sun we cannot ow, but we do know that ozone, which is perhaps formed in the upper atmosphere, exercises powerful selective absorption beyond wave-length 0.29u. Hence it may be that we are forced to neglect some radiation not quite negligible. It is very improbable that the amount thus neglected exceeds I or 2 per cent. _ As for the supposed variability of the sun, our determinations ongly indicate that the so-called solar constant is not really a sonstant, but fluctuates over a range of about 8 per cent. This result is apparently the direct outcome of our observations, but the question may well be asked if the apparent fluctuation is not due _ either to the inaccuracy of the observations or to incorrect estimates of the transmission of the atmosphere. If it were due merely to accidental errors of observations, a gradual march, step by step, day y day; from a low value to a high one and return would be the ex- PROC. AMER, PHIL. SOC. L. 199 P, PRINTED JUNE 29, IQII. ae casas cunt cna mins alia tes isd id Aaa ea 244 ABBOT—SOLAR CONSTANT OF RADIATION, (April a1, ception. We find it to be the rule, hence we must exclude accidental errors as the main source of the apparent variability of the sun. As for the other explanation suggested, we find no material differ- .ence in the result derived for the solar constant on a good day whether we observe at sea-level, at one mile, or at nearly three miles elevation, though the pyrheliometer readings on the ‘ground differ by 25 per cent. between Washington and Mount Whitney Hence we may reasonably conclude that we do, in fact, correc estimate the loss which occurs in the atmosphere. The fluctuation in the solar-constant results therefore seems to indicate either a true variability of the sun, or else the interposition of meteoric dust, or other cosmic hindrance to the passage of radiation from the sun to the earth. These fluctuations, while not of regular periodici generally run their courses within five or ten days.*® 5 It is now proposed to test this conclusion by conducting solar-con- stant measurements simultaneously at Mount Wilson and in south Mexico. If the results of a long series of daily observations these remote stations should agree, it would seem quite unlikely that any apparently simultaneous fluctuations of the solar constant of radiation could be attributed to terrestrial influences. SUMMARY. Special apparatus, including the silver-disk secondary pyrh meter, the absolute water-flow pyrheliometer and the recor spectro-bolometer, has been employed by the writer and his leagues at Washington and Mount Wilson and Mount Whitne determine the mean value of the solar constant of radiation, its possible fluctuations. The observations, exceeding 400 in number, have been made all the years since 1902 to 1910, but most plentifully and accura in 1908, 1909 and 1910. The mean value of the intensity of : radiation outside the atmosphere, at mean solar distance, is fo to be 1.922 (15°C.) calories per square centimeter per minute, might prove I or 2 per cent. higher in years of less sun-spot acti The solar-constant values do not appear to depend on the altitude o! ** See Abbot and Fowle, Astrophysical Journal, April, 1911. " ABBOT—SOLAR CONSTANT OF RADIATION. 245 SELF-LUMINOUS NIGHT HAZE. By E. E. BARNARD. (Read April 21, 1911.) There is one phase of the night skies which does not seem to h ve received much or any attention. It is the occasional presence self-luminous haze. This matter does not seem to be similar» luminous night clouds, “ die leuchtenden Nachtwolken,” which observed by O. Jesse and others some twenty-five or thirty ago, and which were found to be clouds at such great altitudes the earth’s surface (upwards of fifty miles high) that they rece’ the sunlight long after or before the ordinary clouds. The obs tions of O. Jesse were printed in the Astronomische Nachrich Bd. 121, pp. 73, 111; Bd. 130, p. 425; Bd. 133, p. 131; Bd. 14¢ 161. In A. N., Bd. 140 (No. 3347), he gives a long list of altitt determined by photography, which range from 81 km. to 87 The mean value given by the observations from 1885 to 1891 82 km. (52 miles). These clouds were seen in the northern sphere only near the time of the summer solstice. In the sot hemisphere they were seen at the opposite season. From his p it is clear that these sunlit clouds were in no way related present subject, and I only mention them to forestall any su ges that they were similar to the ones seen by me. The objects described here were apparently at the altitude of the ordinary h clouds. They have been seen in all parts of the sky and at all of the night. In a paper on the aurora’ I have previousl attention to the frequent luminous condition of the sky at This feature long ago impressed itself upon me. Indeed any who has spent much time under the open sky hunting comets, will have been forcibly impressed with this peculiarity. I cases this illumination has been due evidently to a diffusion o 2“ Astrophysical Journal, 31, April, 1910. 246 BARNARD—SELF-LUMINOUS NIGHT HAZE. 247 meral star light, perhaps by moisture in the air. This latter con- _ dition is present as a whitening of the sky, which gives it a “ milky” pearance. At other times the sky is more or less feebly luminous, the luminosity is different from the other condition and is evi- dently not due to a diffusion of star light. In reality the sky seems be self-luminous. Sometimes the whole sky has this appearance, at other times a large portion only. At times the illumination so great that the face of an ordinary watch can be read with no - other light than that of the sky. It is indeed seldom that the sky rich and dark. In any determination of the total amount of the of the sky the results must be uncertain because of the great that so often take place in the amount of the illumination. 1e self-luminous condition frequently occurs when no ordinary ndications of an aurora are present. It is, nevertheless, doubtless an auroral nature, for Professor Campbell has shown that the Spectrum of the aurora is essentially always present on a clear dark - night. (Astrophysical Journal, 2, August, 1895, p. 162.) I have given an account” of the remarkable pulsating clouds of tht that are seen here occasionally and which usually, but not ways, have an easterly motion—generally southeast. They are ostly confined to the northern half of the heavens. There is another phenomenon that has been visible on a number of nights of year, and also in the present year, of which I have seen no cord. This consists usually of long strips of diffused luminous I believe that this is really ordinary haze, which for some ason becomes self-luminous. It is not confined to any particular on of the sky nor to any hour of the night. It always has a slow drifting motion among the stars. This motion is comparable ith that of the ordinary hazy streaky clouds that are often seen in e daytime. They are usually straight and diffused and as much as 50° or more in length and 3° or 4° or more in width. In some cases they are as bright, or nearly as bright, as the average portions of the Milky Way—that is, they are decidedly noticeable when one’s attention i is called to them. They apparently are about as transpar- ent as ordinary haze. Sometimes, when seen near the horizon, where _? Astrophysical Journal, 31, April, 1910, p. 210, etc. 248 BARNARD—SELF-LUMINOUS NIGHT HAZE. (April | they may be quite broad, they have strongly suggested the “dawn” or glow that precedes a bright moonrise. Their luminosity is uni- formly steady. The reason I speak of this matter as haze, and ia reason I think — it is only ordinary haze made self-luminous, is because on one occa- _ sion I watched a mass of it in the northwestern sky which was slow drifting northerly in the region of the great “dipper” of Ursa Major as daylight came on. These hazy luminous strips had been visible all the latter part of the night—new strips coming and g slowly, sometimes several being seen at once. As daylight killed them out I noticed, when the light had increased sufficiently, thi there were strips of ordinary haze exactly the same in form motion and occupying the same region of the sky. I am sure thi were the same masses that had appeared luminous on the night sk My impression, therefore, is that these hazy luminous strips v only the ordinary haze which had for some reason become s luminous. I am specially certain that these masses are not luminot as a result of any great altitude which might bring them within re: of the sun’s light, for they were frequently seen in such posit that the sun’s rays could never reach them. The sun or moon, the fore, had nothing to do with their illumination. It is also need to say that they are not related to the pulsating auroral clouds I have previously mentioned. I have not noticed this luminous haze in former years, though may have been present, and did it not seem unreasonable, one mi suspect some relation between this condition of the atmosphere the possible passage of the earth through a portion of the tail Halley’s comet on 1910, May 109. I will give here the observations which I have obtained of singular features. It seems to me that these objects should be served and a record made of the times of their visibility and motion, etc. It would be valuable to have records of them fre different stations to see if their luminosity is due to some get condition of the earth’s atmosphere at the time. It is not pro that this luminosity is in any way due to local conditions. In records here given, it is possible that on one or two occasions BARNARD—SELF-LUMINOUS NIGHT HAZE. 249 aurora was also present, but I have tried to confine the accounts to what I have called, and believe to be, self-luminous haze. They _ were not seen previous to June 7, Igo. THE OBSERVATIONS. 1910, June 7d 13 h om. These diffused luminous masses were seen in different parts of the sky. They were specially noticeable near the southern horizon where the appearance was that of a definite whitish light stretching along above the horizon for a considerable distance. Long bands of this matter were parallel with the southern horizon and above Antares. In the east a long strip 3° or 4° wide stretched from a Pegasi to a Andromedz and beyond. This moved wly eastward. At 13h 30m another was passing through the bow! the great “dipper” in the northwest with a slow easterly motion. very broad one was situated about 15°—20° from the zenith to the They were about as bright as the Milky Way in Cygnus. I waited until near sunrise, and could then see a long mass of ordinary haze, reddish with sunlight, occupying the position of one of the s that was seen near the bowl of the “ dipper,” which had been visible as a luminous mass until the dawn had killed it out. There were other strips and masses of haze at different points in the sky hen the sun rose. I think it was these streaks and patches of dif- used haze that were luminous during the night. They appeared as ordinary haze clouds in daylight. During the entire night there had been no ordinary trace of aurora. June 9. Though they were looked for several times none was seen until about 1oh 30m or 11h om. At Ith 25m a long broad hazy streak, as bright as the Milky Way in Cassiopeie was seen in the northwest. The lower end was in the “ sickle” of Leo near the horizon. Its upper end was 15° below the polar star. From a sketch at 11h 25m the following points were taken which were peeves in the hazy strip: a toh 5m 3+ 21°.5, GO" 3 + 49. t t extended beyond this latter point for quite a distance—roughly TRS Hs RE re a z7hom 8+ 67°. 250 BARNARD—SELF-LUMINOUS NIGHT HAZE. The stars were visible through it where it passed over them. — The motion was slowly to the northwest among the stars. Its. width was 5°. At 12h 0m a similar band passed over the “ dipper ” to the first one, evidently moving in the same direction. one at this time had either disappeared or was too near the horizon to be seen. At midnight I could read the time by my watch with only the illumination from the sky, which was milky and whitish or luminous. . June 10d 10h 45m. A long strip passed through Polaris and 5° below the bowl of the “dipper.” Its motion was towards the north by east horizon. 11h om a great number of luminous masses were scattered over the western sky (and extending to the south) to nearly as high as the zenith. These were mostly parallel strips with some irregular masses. They extended from the henge and seemed to diverge upwards. September 29. The sky was irregularly covered enereheel with a kind of luminous haze which occurred in great areas and in strips, with a few clear spaces between which were relatively dark. They were more or less conspicuous. At 8h 25m a diffused lumi- nous band stretched from Corona Borealis to the southwest horizon —nearly north and south. This continued northerly nearly to the pole and was diffused to the west. In the south and southeast for 20° above Fomalhaut to a Ceti was the upper edge of a luminous mass of haze covering the southeast sky to the horizon, O diffused areas of this matter were visible at different points over the sky. The whole sky was more or less luminous, but less notice able than the regions described above. By 8h 50m the broad 1 nous strip at Corona Borealis had drifted a little east among stars, but it seemed to go westward with them. At 11h lomaw could be read by the light of the sky. This was one of the bri est of the luminous nights that I have seen. The matter seemed be only ordinary haze but luminous for some reason. There no trace of aurora. The sky on which the luminous haze was was, at this time, brightened with a pale uniform illumination cove ing the entire heavens and nearly blotting out the Milky Way. These masses had very little motion. The sky was too luminous for long exposures with a portrait lens. a 11.] BARNARD—SELF-LUMINOUS NIGHT HAZE. 251 September 30. At 9h 15m for 10° above the east by north horizon a broad luminous band 50° long was seen just above and volving Aldebaran. It stetched to the south of the east point and in brightness resembled the appearance produced by the moon st before it rises. The light was soft, yet conspicuous. At oh 10m under Capella was a large soft diffused light—diffusing the east and beyond. This light was steady with no fluctuations. othing of a similar nature was visible in the north or elsewhere. The sky was dull and more or less luminous. At 1oh 55m the umination extended half way up to Aldebaran and the sky near e horizon was luminous like moonrise. This extended from 25° uth of east to nearly due north, rising much higher under Capella —a very soft and steady illumination. 12hom. The illumination was eeble and diffused. At 12h 30m it was very feeble and mostly in e northeast—scarcely noticeable. At this time dark smeary haze as visible all over the south. No evidence of an ordinary aurora as seen during the night. The sky was luminous all over, but not so much so as on the twenty-ninth. _ October 1. There was a bright aurora. _ October 2, 8h om. A pale illumination was seen in the low rth and also in the low east. The effect was probably auroral. October 6. The night was more or less luminous and misty. _ October 28. There was a luminous sky at night. October 30, 13h om. The night was very luminous with fully much light as would be caused by a one quarter full moon. The ilky Way was scarcely visible. Watch easily read by the glow at 14h om. At 15h and 15h 30m a luminous haze covered all the low northern sky as high as half way to the pole. This was not strong and did not look like an aurora. It seemed simply to be __ luminous haze. . November 1,12h 15m. The sky was remarkably luminous every- _ where. In the north from the horizon to halfway to the pole the . sky appeared more luminous than elsewhere. No trace of an arch. E _ The illumination did not look like that from an aurora, but at q 15h 30m a strong auroral arch had formed. November 10, 12h om. There was a great amount of luminous 252 BARNARD—SELF-LUMINOUS NIGHT HAZE. _ [April haze in the north and northwest. At 15h om a large mass 10° | was visible in the northwest. Later there was a long diffused s' 10° wide, which cut the Milky Way at right angles 20° a a Cygni. It was 40° or 50° in length and did not fluctuate. appearance was that of luminous haze. Below it was a region luminous haze that extended to the north. 1911, February 28, 15h 30m. For 20° to 25° siete all the east and northeast the sky was luminous with a soft aw light. There was no arch or intensification near the usual p for an aurora. This was not due to the presence of the Milky at that point. March 2, 8h tom. A long mass of luminous haze 6° 0 on broad was visible below 8 Leonis in the east. It diffused down the northeast horizon. It seemed to be brighter at times, but were no certain fluctuations of its light. It was not bright. 8h The region of luminous haze was passing over Arcturus and mo towards the east horizon. It was nearly horizontal and 30° with the north end the lowest. toh 45m. A long mass of lumii haze was visible one half way from Spica to the southeast ho j It extended south as far as Corvus and inclined to the so horizon. It was quite bright and steady in its light. All of southeastern sky strongly resembled the glow from an ex moonrise. Ith 35m. A strong glow from the southeast h extended up to 15° or 20° above Jupiter—like a strong moon all along from the east to the south and diffusing upward. I conspicuously strong. By this time the sky was increasing luminosity. In the meantime there had been no trace of ¢ during the night. These were the first of the luminous | haze that I had seen for a long time, except that of ro 28 when it appeared near the northeast horizon. Since the above observations I have not seen any of this lumi haze on the few clear nights that we have had in the absen the moon. YERKES OBSERVATORY, April 4, IQIt. BARNARD—SELF-LUMINOUS NIGHT HAZE. 253 Norte. Since this paper was in type Mr. C. F. Talman, Librarian of Weather Bureau at Washington, through Dr. W. J. Humphreys. called my attention to a paper, No. 22, of the Publications of the tronomical Laboratory at Groningen, “On the Brightness of the and the Total Amount of Starlight” by L. Yntema. Dr. Yntema calls attention to the frequent luminous condition of the and its effect on determinations of the amount of starlight. In ion 14 of his paper, which is devoted to earthlight, he gives erous records of this illumination. There does not appear to be y direct reference, however, to the main features of my. paper—the ninous hazy strips and masses. May 15, IQII. SPECTROSCOPIC PROOF OF THE REPULSION BY Tk SUN OF GASEOUS MOLECULES IN THE bei di OF HALLEY’S COMET. By PERCIVAL LOWELL. (Read April 21, 1911.) 1, The return of Halley’s comet has been noteworthy chiefly the possibility of employing upon it modern methods of instrum research. Since its last previous apparition have been devised t! two great engines of astronomic exploration, spectroscopy and ¢ tial photography. The former has afforded us our first direct kn edge of the substances composing comets, while the latter has gi us a means of easy and rapid registration of the visitant’s app ance. This is especially valuable in the case of a body as vast vague as a comet, free-hand drawing of which is peculiarly | lia to distortion. During the last return of Halley’s comet that body wa jected at Flagstaff to investigation by both instruments si neously. One result of this was the detection that gaseous cules—in contradistinction to minute solid particles merel directly repelled by a force emanating from the sun, prest the pressure of light. Previously this had been held impo Schwarzschild had, as he thought, demonstrated mathemati an able paper that molecules of gas were too small to be thus a by the forces concerned and Arrhenius had adopted his ded and published it as a fact in his “ Worlds in the Making.” — the bodies themselves would so soon refute this would not have | deemed probable and invests the detection with the more imm di interest. Incidentally we may remark that Schwarzschild had s given up his original opinion. 2. That the tail of a comet is due to repellant force exerted) by the sun is apparent from the direction the tail takes. For that dire 254 4911-] MOLECULES IN THE TAIL OF HALLEY’S COMET. 255 tion agrees with what would be shown by particles leaving the nucleus and travelling in hyperbolic orbits away from the sun, the sun being in the full or the empty focus according to the speed of _ Although the general fact is thus evident, to measure the reces- sion directly is to obtain both an observatioral proof of it and also something approaching an exact value of the velocity at a given time and place. Accordingly I determined to do this in the case of Halley’s comet at its recent apparition. At my disposal were the two hundred photographs taken of it at the Lowell Observatory between April 18 and June 6. To obtain trustworthy results the photographs to be compared must not be separated by too long an interval, since with time a general commingling of the various par- takes place which not only renders particular decipherment of different outbursts impossible but entirely alters the actual speeds. n the case of Halley’s comet this difficulty was erhanced by the uniformity of the tail. Irregularities, bunches or knots rare; the tail presenting as a rule, a remarkably orderly deport- t, dishearteningly same. Among the many plates, however, I s able to select a pair taken seriatim capable of recognition and measurement. Some of these handles to investigation were in the ture of bunches of matter, some of abrupt changes in its direction looking like promontories along the general line of the tail. I chose our of the more salient excrescences and selecting identical features them in the two regatives measured their respective distances from the nucleus on the two plates. The first plate was exposed gh 23m to gh 53m and the second from 1oh om to 1oh 53m, that the one followed directly on the other. _ When the angular amounts of the changes in place of the several ots were corrected for differential refraction and then reduced to speeds, account being taken of the distance of the comet from the earth and of the inclination to the line of sight of the respective positions along the tail, the results came out as follows: From these measurements the fact emerges unmistakably that a ulsive force directed away from the sun acted upon the particles on the tail. te 256 LOWELL—REPULSION OF GASEOUS MOLECULES [April 2 Tait oF HAtiey’s Comer. Angular Di from the Nucl Velocity of the Point of the Ta og ergy petoner nah or alas oe = Knot 1 1° 287 13.6 miles a second — Knot 2 rege 17.2 eee Knot 3 4° 367, 19.7 ae ae e 2 Knot 4 6° 15’ 29.7 (698 he 3. While the series of direct photographs was being taken. series of spectrograms were being carried on by Dr. Slipher, one an objective prism; the other set througha slit. The objective pri ones recorded simultaneously the spectrum of the nucleus and | together with that of the tail out to about 11° from the nu One of them was got on May 23 at the same time as the phe graphs measured; while others were obtained on dates before a after. Of the direct information afforded by these spectrograms the constitution of the comet an account is given in the exte: monograph on the comet published by the Lowell Observatory. 4. But a third result was obtained by the unwitting collaborat of the spectrograms and the photographs. While the photo; were giving their pictures of the tail, the objective prism s grams were doing the like, with this difference that they recorded a row pictures of it in the several colors of the spectrum, si : out into a band those made by each separate wave-length of They thus made it possible to tell to what wave-lengths the appearances were due. For it became evident at once from spectrograms that all wave-lengths were not equally concerned. the contrary, there were in the spectral image several distinc with spectral gaps between. By an analysis of the wave-l yielding pictures of the tail was thus offered a diagnosis of th stances composing it. In this way it appeared that CO,, « monoxide, was the chief constituent of the tail; CH,, mar. another ; CN, cyanogen, a third component; and minute solid cles, giving a more or less continuous spectrum, a fourth. That not one but a series of spectrograms was taken was tant. It not only gave us a constitutional history of the tail showed the necessity of simultaneity in photographic and spe graphic observations for comparative purposes. For the IN THE TAIL OF HALLEY’S COMET. 257 nonstrated that the constituents of the tail varied markedly from period to another. Thus from April 29, 1910, to May 7 the n of the tail was almost wholly emissive. On May I1 it 1 changed to one nearly continuous, while on May 23 it had be- ome largely emissive again and grew more so as time went on. representations of the tail a striking fact cametolight. Toappre- te this another point must be taken into account. In order to mpare properly a photograph and a spectrogram, both should be ide on the same brand of plate. No plate reproduces all parts of rays and depreciate others; the next will reverse the estima- Great error will then be introduced unless the plates be ‘Now the photographs measured were taken with a Brashear 5-in. ublet, an excellent lens, on a Lumiere = plate. The rays regis- ed by this plate extend from 3500 in the violet to 5160 in the where the sensitiveness ceases. Indeed the effect would have The light, therefore, of the photograph would be exactly ferentiated into its constituents by a spectrogram taken on a miere S plate. The only difference between the two would be due the absorption of the objective prism, an absorption relatively greater for the violet than for the blue or green. This would work much on one kind of light as on another of the same refrangibility as the two different kinds we are considering, the emission and = continuous spectrum, are about equally spaced in the region of e€ violet, the correction needed on this account is small. We may, en, directly gauge the character of the photograph’s light by that the objective prism spectrogram taken on the & plate. This we now proceed to do. On the exact date of the photo- aph no = plate was used with the two objective prisms though we ve objective prism spectrograms on Cramer Iso. Instantaneous on May 23, 25, 26, 28 and 29. The nearest plate to the date in stion was on May 29. Of this the best estimate gives for the nstituents of the light of the tail at a distance of 3° to 6° from he head : 258 LOWELL—REPULSION OF GASEOUS MOLECULES tas 80 per cent. of emission bands of carbon monoxide, 20 per cent. continuous spectrum, fas the hydrocarbon emission being, at this distance from the head feeble to show. Comparing now the spectrograms taken with a Voightiniiatats and a Cramer Iso. Inst. plate on May 23, 25, 26, 28 and 29 we f that the ratio of the two kinds of light varied in the direction of tively greater emission from the former to the latter date. On 23 itself the plates are affected by moorlight so that a direct parison of the relative ratios is too difficult to be made a basis direct comparison, but that of May 28 gives for the ratios in io 3° from the head: ‘ 70 per cent. emission of carbon monoxide, 5 per cent. emission of hydrocarbons, 25 per cent. continuous spectrum, Putting these facts together we shall not be far out of the in stating the ratio on May 23 of the emissive and continuous trum of the tail at a distance of from 3° to 6° from the he: the & plate as 70 per cent. emission spectrum, CO and CH,, 30 per cent. continuous spectrum. We have then this interesting conclusion: that the knots showed the action of a repulsive force exerted from the sun chiefly composed, not of solid particles, but of molecules of 5- Te clinch this deduction I next turned to comet More emphasized it. Before the time of measuring the velocities i tail of Halley’s comet I had done the like for comet Moreho knotted character of its tail offering promising inducement. not aware that Mm. Quenisset and Baldet, in France, and P J. A. Miller, of Swarthmore, Pa., had measured photographs « comet in this manner previously and detected the same accel motion away from the head which my own later measures sh My measures have also revealed why certain previous observer: as Barnard at Yerkes and Campbell at the Lick had failed 1 such evidence. IN THE TAIL OF HALLEY’S COMET. 259 Of comet Morehouse this observatory possesses about sixty nega- tives taken by Mr. E. C. Slipher. Among them are many pairs, the one plate following the other on the same evening. From the assort- ‘ment thus offered I have selected two sets for measurement, the one a pair taken on October 31, 1908, at 8h om + to 8h 42m + M.S.T. and from oh 14m + to 10h 8m = respectively ; and the other a triplet on November 16, 1908, No. 1 being taken at 6h 25m to 7h 13m; No. 2 at 7h 24m to 7h 50m; No. 3 at 8h om to 8h 32m, respectively. I chose four knots on one and five on the other with the following results: ; Tart oF Comet MoreHouse, OCTOBER 31. Plate I., Distance Knot | Plate ie a ig Knot! tyigerence 3 and ti: from Head. m Head. Knot 1 22’.8 24/.4 17.6 Knot 2 Y Eee | 76/.2 37.5 Knot 3 95’-5 99.4 3’.9 Knot 4 128/.4 134’.6 6/.2 Tait oF Comet Morenouse, NovEMBER 16. Plate I. Plate II. Plate III. Diff, 1-11. Diff. I1.-I11. Knot 1 48/.7 49’.9 51/.4 1/.2 17.5 Knot 2 63/.4 65/.4 667.9 2’.0 17.5 Knot 3 79’.7 817.6 83/.5 17.9 17.9 Knot 4 877.8 897.3 90’.9 17.5 17.6 Knot 5 211/.% 215’.0 217".7 3/.9 pat ‘ 6. It will at once be seen that both sets of plates show accelerated : velocity in the particles of the tail away from the head as the dis- - tance from the head increases. In the first set the acceleration is. fairly uniform, while in the latter the velocity does not increase until the distance out has become considerable. This affords the reason why some observers have failed to detect the motion. It is at times _ and in certain places masked. For this the following explanation a 5, may be offered: In the neighborhood of the head the several emis- __ sions are violently contorted as a mere inspection of the photographs show, and in consequence must be subject to collision with other por- _tions of the tail. Possibly they encountered here matter in space which speaks unspeakably of motion other than that due solely to repulsive force. If now an observer chanced to make his measures PROC. AMER. PHIL. SOC. L. 199 Q, PRINTED JUNE 29, 1911. 260 LOWELL—REPULSION OF GASEOUS MOLECULES, ; at this inopportune moment he would naturally conclude ‘that 4 repulsion existed while in truth another motion was tem oe it. due practically wholly to emission; in other words to glowit Here, then, we have not only corroboration of the fact, forward from study of Halley’s comet, to wit: that molecules of are repelled by the sun, but, from the light of the tail being posed solely of gaseous molecules, any supposition that they not the cause of the visible effect, is entirely excluded. We reach then this interesting conclusion: that molecules 2) not only may be but demonstratedly are repelled by the < the sun and that though we have reason to suppose that min particles may be similarly inipressed it is of the former not the | that we have direct proof at present. LoweLL OBSERVATORY, April 10, IgI1t. THE NEW COSMOGONY. BY 23. 4. SEE. (Read April 21, 1911.) The results established in the writer’s “ Researches on the Evo- ution of the Stellar Systems,” Vol. IT., 1910, have given a new basis . our conceptions of the cosmogony. Instead of the traditional octrine of throwing off, we now have that of capture, which means entially that the nuclei originated in the distance and have since own by accretion as they approached the centers about which ey now revolve in greatly reduced orbits of small eccentricity. - Not only have we witnessed a radical change in the point of view, but also in the method of research employed. And along with these changes has come the introduction of rigorous mathematical and namical criteria by which the mechanical principles involved may e extended over an almost unlimited period of time. Not the least important of the improvements recently introduced is that resulting from a careful examination of the premises under- ying our reasoning. Nothing is adopted from tradition, nor taken granted, nor from any authority however high; but every ques- tion is examined on its merits and from the very ground up. As the subject is new it naturally follows that much still remains to be done; the general trend of nature’s laws seems to be well established, cosmogony begins to assume the form of a true science. Ac- cordingly it may not be without interest to the general reader to summarize in one connected view the leading principles of the new Science of cosmogony, with brief analysis of the criteria by which they are confirmed. ' 1. Babinet’s criterion based on the mechanical principle of the servation of areas, by which we are enabled to calculate the nes of rotation of the sun and planets when expanded to fill the bits of their attendant bodies, as imagined by Laplace. This enables us to say at once that the attendant bodies could never have 261 PEAR A RR Re ER re epee me 262 SEE—THE NEW COSMOGONY. been detached by acceleration of rotation, as handed down by tion from Laplace’s original nebular hypothesis of 1796. : 2. As the planets and satellites could not have been thrown off, they must have been captured and added on from without, or have been formed from the agglomeration of fine dust right wh they now revolve. This latter alternative, however, is easily shoy to be impossible, owing to the feeble mutual gravitational attr of small masses of matter under the stronger tendencies to disp by tidal action, which always exist near large centres of attra There remains therefore no possible mode of origin for the p and satellites save that of capture, or addition to the — without. 3. When first captured the satellites must therefore tei already of such considerable size that they were able to ga ; and consolidate with their globes, numerous smaller masses r ing in the vortices about the planets. The collisions arising. in process of the gathering in of smaller bodies by larger ones strikingly illustrated by the craters noticed in the face of the } which were formed by impact, the embedded satellites being in s cases at least twenty miles in diameter. Hig 4. Thus while the satellites were all captured, and were inally further from their planets than they are at present, they grown larger in the course of ages as they revolved in the re medium about the planets, just as the earth and primary p still growing larger by the impact of meteorites against t faces, as they slowly approach the sun. The earth sweeps u 1,200,000,000 meteors, and the amount of this dust is ca to form a layer a millimeter thick in a century. 5. We know the satellites must have grown in mass since were captured, because they have been drawn nearer and near several planets, by increase of the central attraction, as in the brated problem of Gyldén.* But if the mass of the sun increased, by the downfall of cosmical dust, so also must the *Since this was written the capture of Satellites has been ind pen confirmed by Professor E. W. Brown, in an important paper in the M Notices of the Royal Astronomical Society for March, 1911, p. 453. %* A. N., 2593. SEE—THE NEW COSMOGONY. 263 of the planet or satellite have been correspondingly augmented by the me cause. 6. For whilst the decrease of the major axis of the orbit of a ite might result wholly from the growth of the mass of the and satellite, yet the decrease of the eccentricity of a satellite orbit can be explained only by collisions in the nebular resisting medium. This cause and no other whatsoever will explain the roundness of the orbits so characteristic of the solar system. 7. Accordingly as most of the satellites suffered collisions suffi- ‘cient to reduce and well nigh destroy the eccentricities of their bits,? it necessarily follows that all these bodies should have their rfiaces indented by impacts with smaller masses, just as is shown the craters on the moon. 8. For whilst Oppolzer, Gyldén and others have proved that the growth of the masses by the downfall of cosmical dust would crease the central attraction and bring the bodies close together, is proved by the mathematical researches of Airy, Herschel, Leh- ‘mann-Filhés, and Str6mgren, which I have carefully verified, that _ this decrease in the major axis does not decrease the eccentricity. _ Hence the decrease of the eccentricity is traceable to no cause what- soever but the action of a nebular resisting medium, as held in my “Researches, * Vol. II., p. 146. 9. The craters on the moon can therefore be due to no cause whatsoever other than the collisions which our satellite has suffered from other small bodies in space, and all divisions of opinion on the subject are henceforth swept away forever. For as the other satel- lites have had their orbits rounded up in nearing their several planets, it is necessary to suppose the same cause to have acted also on our ‘moon, even if the eccentricity of the orbit in this case has not been ndered excessively small. 10. This solution of the problem of the roundness of the orbits— leading problem in the cosmogony of our solar system—is what athematicians call a unique solution. It reveals not only a possible, but also the only possible cause of the extremely circular move- --? In section 548 of his “General Astronomy,” edition of 1904, the late Professor C. A. Young remarks that the “almost perfect circularity of the ‘satellite orbits is not yet explained.” 264 SEE—THE NEW COSMOGONY. ment characteristic of the planets and satellites. The solution thus possesses all the rigor of a theorem in geometry, and meets the requirements of the most rigorous of the mathematical sciences. 11. The existence of planets beyond Neptune is indicated by tl extreme roundness of Neptune’s orbit; for this shows that nebulosity was much too dense at that point for the system to” minate at the present known boundary. Moreover, as I have shown that the planets were originally connected with the comets, and the comets recede to their home in a spherical shell thousands of times the earth’s distance from the sun, it necessarily follows that | planetary system extends on almost indefinitely. Several planets considerable size must be assumed to revolve beyond Neptune, they may yet be discovered by observation or photography, thot at that great distance the practical difficulties will increase, owing the feebleness of the sun’s light and the slow orbital motion, wh will require exposures of the photographic plate exten : many hours, and perhaps on successive days: 12. The planets have been built up out of cosmical dust, cc and satellites; so that all the matter now in the planets come or nally from the heavenly spaces. This follows from the fact : the nebular development is from the outside toward the center, formation always beginning in the distance and proceeding by aceré tion as the bodies gravitate towards the sun, and revolve in smaller and rounder orbits. This order of development is di verified by the phenomena of the spiral and ring nebule; for the movement is proved to be towards the center, where sei develops for the domination of the system. 7 13. And just as our planets have been added onto the sun f without, not thrown off, as was erroneously taught for more tha century by Laplace and his successors, so also will similar p have been formed by the same process about the other fixed s Thus there are undoubtedly systems of planets about the fixed : and they are habitable and inhabited like those revolving | the sun. Moreover, the other suns have their systems of co and their planets have captured systems of satellites as in our pla tary system. This grand conclusion rests on an incontestable be and is of transcendent philosophic interest. by: ‘tgrt.] SEE—THE NEW COSMOGONY. 265 14. The causes which have operated in the development of our _ solar system are thus general throughout the sidereal universe. __ Everywhere repulsive forces are dispersing fine dust from the stars to form the nebulz, and the nebule in turn are settling down and _ whirling around to form stars with planetary systems about them. 15. Professor Barnard’s magnificent photographs of the Milky _ Way show that cosmical dust everywhere pervades the heavenly — spaces. And it is proved that variable stars are due chiefly to attendant bodies revolving in resisting media. When considerable bodies come into collision, as a large planet with a sun, the result _ is a temporary star or Nova. 16. The new cosmogony thus embraces within its scope the chief _ problems of the universe, and the dynamical causes assigned are - deduced from simple phenomena operating according to known laws which are actually verified in the solar system. The arrangement of the nebulz on either side of the Milky Way is the natural out- come of the operation of repulsive forces, the canopy of nebulz con- _ gregating as far from the stratum of stars as possible. This assigns _ a known cause for the great order of nature first brought to light by the telescopic explorations of the elder Herschel in 1785. x Like astronomy itself it is obvious that cosmogony is at once the oldest and newest of the physical sciences. Having renewed its youth by the introduction of definite principles and exact methods, _ it has recently taken on such vigor that it promises to become the © most majestic of the sciences. Nothing is more worthy of the at- _ tention of philosophers than the study of the great laws of the phy- sical universe, and the marvelous processes of development by which the beauty and order of the cosmos came about. This was the _ great problem which gave rise to the development of the physical _ sciences among the Greeks, and it will always occupy a position of transcendent importance in the domain of natural philosophy. U. S. Navat OBsERVATORY, Mare Is_tanp, CALIFORNIA, April 3, Igtt. THE EXTENSION OF THE SOLAR SYSTEM BEYOND NEPTUNE, AND THE CONNECTION EXISTING _ BETWEEN PLANETS AND COMETS. By TJ. J. SEE. (Read April 21, 1911.) One of the most remarkable results of the writer’s recent researches on the origin of the solar system has consisted in the development of a satisfactory proof that the primordial nuclei of the planets were formed at great distances from the sun, and that their primitive orbits were highly eccentric like those now described by the comets; so that in the last analysis it is shown that the tw classes of bodies are merged together, or rather that the planets have been built up by the agglomeration of cosmical dete form of comets, and other fragments of matter, from our ancient nebula. The following is a brief outline of the thread of argument if leading to this conclusion: 1. It is shown by the exact data supplied by Babinet’s criterion that not one of our planets could have been thrown off from sun, by acceleration of rotation, as imagined by Laplace in 1796, that the nuclei must have started in the distance and since neared sun, by insensible degrees, as the masses were gradually augmen by precipitations from the surrounding nebular medium. 2. When it was thus demonstrated by exact calculation that premise handed down by Laplace is erroneous, our theory planetary genesis was placed on a new basis by the proof that roundness of planetary orbits is due to the secular action of resisting medium, which has reduced the size of the plate or! and rendered them almost exactly circular. 3. In order to be so exactly circular, as they are now found to these orbits must originally have been very large, and also hig eccentric, like the orbits of comets; the orbits accordingly have b reduced in size by encounters with the other minor bodies, th 266 rg1t.] SOLAR SYSTEM BEYOND NEPTUNE. 267 absorption of which also increased the masses of the planets enormously. ; _ 4. If one asks for ocular evidence that the planetary bodies have been in collision with smaller masses, this evidence is found in the phenomena shown in the face of the moon, which was formerly an independent planet, and is so small a globe as never to have devel- oped water or atmosphere; so that it is a kind of hermetically sealed celestial museum, so near us in space that it serves for the illustra- tion of the process of absorption and capture in cosmogony. The type of collisions visibly illustrated by the dents in the moon’s face necessarily have occurred with all the planets; but the moon as our ‘nearest planetary neighbor alone enables us to study the process of accretion by collisions with bodies of all sizes, from particles to ‘satellites as large as twenty miles in diameter. 5. The obvious deposits of dust over the older lunar craters give them an aspect of great age, and in many cases the outlines of the craters are practically obliterated. In other cases newer craters are formed over the older ones; so that we can certainly infer by direct servation that the moon has been built up by accretion, dust being gathered in to be deposited over dust, and crater over crater. This is the same process which we see at work on the earth, except that the meteorites now swept up by our planet are generally small and consumed i in the air before reaching the earth. 6. Since the planets were begun as independent nuclei in our nebula, and since augmented by the gathering together of an infinite number of small bodies, such as comets, the matter of planets and comets must necessarily be the same, for they are common products _of our ancient nebula. The planets have been built up by the gather- _ing in of satellites, comets and smaller particles of cosmical dust. 7. Now we have pointed out that Neptune’s orbit is too round or it to be the outermost of the planets of the solar system. If the _ resisting medium was dense enough at that great distance to produce such extreme circularity in the motion of Neptune, there was enough of the nebulosity beyond that planet to make several more planets of comparatively large size. Thus it is certain that our _ System does not terminate at Neptune, but extends on almost inde- finitely. It is probable that in time we may be able to discover 268 SEE—SOLAR SYSTEM BEYOND NEPTUNE. several trans-Neptunian planets ; but the recognition of these remote bodies will be difficult, owing to their slow motion and the faintness of the sun’s light at that great distance. 8. The notable expansion of our ideas of what constitutes a : nebula will thus be of great practical use in the progress of astron- omy. The overthrow of the theory of Laplace is only a small part of the service to science brought about by the discovery of the true laws of the development of.our system. As the comets recede to distances amounting to thousands of times the earth’s distance from the sun, so also must embryo planets be imagined to bridge over ; gap heretofore separating the planets and comets. And we ma imagine planets to extend to at least 100, perhaps 1,000 times - earth’s distance from the sun. Some of the comets may go Ic times further yet, but at such great distances we can never knot much about their motions in these remote regions of space. S 9g. When we contemplate the vast extent of our primo ‘nebula implied in the distances to which the comets recede, a remember the large apparent areas covered by many other neb in the sky, we see that our solar nebula evidently was of the ordi type, and that it certainly was not a gaseous mass in equilib under hydrostatic pressure and extending only to the orbi Neptune. Of course all these old doctrines of Laplace are quite abandoned, but they long deceived us, and kept cosma in a stationary condition for over a century. . 10. The origin of the primordial nuclei in the distance is a n sary consequence of the working of planetary bodies towards dominant center of attraction—the sun. Hence the formation system of planets is necessarily from without inward, just ‘th reverse of the traditions handed down by Laplace. This harmoni: perfectly with the new theory of the spiral nebule, which makes ring nebule particular cases of the more general spiral te The formation in all cases is from the outside towards the ce! Planets form in all nebulz, and since small bodies approach the ce more rapidly than large ones, under the action of a resisting mec it follows that the planets thus capture systems of satellites su we observe attending the planets of the solar system. U. S. NavAL OBSERVATORY, March 7, IgITt. THE SECULAR EFFECTS OF THE INCREASE OF THE SUN’S MASS UPON THE MEAN MOTIONS, MAJOR AXES AND ECCENTRICITIES OF THE ORBITS OF THE PLANETS. a). he SEE. (Read April 21, 1911.) In the days of Newton, Lagrange and Laplace, it was assumed _ that the formation of the planetary system was essentially complete, and the sun’s attraction rigorously constant from age to age; and it was scarcely deemed necessary to consider the secular effects of slight modifying causes such as the downfall of cosmical dust upon _ the bodies composing the solar system. But the progress of the past century has shown that the Newtonian hypothesis of a constant mass and a central attraction depending wholly on the distance, but not on the time, is at best a very rough approximation to the truth; for in addition to the downfall of cosmical dust upon all the bodies of our system, it has been shown by the researches of Arrhenius, Schwartzchild and others, that the sun especially is losing finely divided matter under the action of repulsive forces such as we see illustrated in the streamers of the corona and the tails of comets. In our modern studies of the orbital motions of the heavenly bodies, therefore, we have to take the central mass as variable with the time, and consider the small secular changes which will follow from a variation of the central attraction incident to a gradual change of These questions have been treated in some form by many of the successors of Newton; and even this great philosopher himself in _ one case supposed that the central mass might be varied by a comet - falling into the sun.t_ Laplace devotes considerable attention to the secular equations for determining the effects of the decrease of the sun’s mass due to loss of light, then supposed to be of corpuscular *“ Principia,” Lib. III, last proposition. 269 270 SEE—SECULAR EFFECTS OF THE character.*. The modern discussions based on the analytical methods | of Gyldén are, however, much more satisfactory than those of the age of Laplace; and I propose to give a brief account of them, chiefly with a view of summarizing the state of our knowledge, and of removing some inconsistencies which may mislead those who are unfamiliar with the literature of the subject. For example, in the late Professor Benjamin Peirce’s “ ideality: in the Physical Sciences,” Boston, 1881, p. 131, the following curious statement occurs: ; The constant increase of the solar mass would have an influence on the planetary orbits. It would diminish their eccentricities, according to a law of easy computation. Hence it is possible that the orbits of the planets may have been originally very eccentric, almost like those of the comets; and their present freedom from eccentricity may have resulted from the growing mass of the sun. What modification of the nebular theory may be involved in this supposition cannot easily be imagined, without the guidance of some indication from nature. 2 This statement is misleading and erroneous, and the only way 4; a can explain its appearance in the writings of Peirce is by the fact that his last lectures were prepared when he was at an advanced age and in ill health; and thus it is probable that some confusion occurred. Quite recently an analogous confusion has appeared the Astronomische Nachrichten, No. 4454, in a short article by | R. Bryant, on the secular acceleration of the moon’s mean motion. In order to place before the reader a summary of the ch investigations bearing on the problems now under discussion we cite the Beret papers: “The Problem of the Newtonian Attraction of two Bodi ae masses Varying with the Time,” H. Gyldén (A. N., 2593). i “Ein Specialfall des Gyldén’schen Problems,” J. Mestscher (A. a , 3153 and 3807). 3. ““ Ueber Central Bewegungen,” R. Lehmann-Filhés ae 3479-80). 4. “ Note on Gyldén’s Equations of the Problem of Two Bodi with Masses Varying with the Time,” E. O. Lovett (4. N., 3790). 5. “Ueber die Bedeutung Kleiner Massenanderungen fir di Newtonsche Central Bewegung,” Dr. E. Stromgren (A. N., 3897). * Mécanique Céleste, Liv. X., § 20. tort.) INCREASE OF THE SUN’S MASS. 271 __ The last of these papers is the most important, since it supple- _ ments and extends the results of the earlier investigators. Professor Stromgren’s method is one of great generality and appears to be the most satisfactory yet devised ; and we shall base our brief discussion chiefly on this paper. If o be a very small quantity, and ¢(t) some function of the _ time, the original unit of mass becomes 1 + o¢(t), and the differ- ential equations of motion become _ + BU t 060] 5=0, d*y BZ wtFli + o$(t)] = O; where k? is the gravitation constant, and the mass is unity at the initial epoch to. The new constant of areas becomes ay ax | ae eg de" "hows k/T1 + of(t)]p = const. (2) _ Other formule of interest are: — I v= e[1 + o9(0)] (2-4), (3) 8a = a,od(t) — 20° 0 f ¢ (2) “dt ; (4) I I ——e— r a [1 + 22/, (ie) cosi(e + nt — 7)] (5) And finally after a careful investigation of all effects due to errors of the first order of the disturbing force, o¢(t), Stromgren finds: | ba = —ao[/+ 2% (sin E—sin Z)], | aaa — “ g(sin E—sin £,), : (6) or = aes pos £— cos £,). 272 SEE—SECULAR EFFECTS OF THE Here n is the mean motion and’ E the eccentric anomaly. It will be seen from the first of equations (6) that the semi-axis major is diminished by a secular term depending on t, and by a periodic term depending on the difference of the sines of the angles E and E,, or the position in the orbit. Thus the mean distance is — t both periodic and secular variation. : In the case of the eccentricity, however, the second of the equa- tions (6) shows that there is no secular term, and only periodic changes occur. A similar remark applies to the longitude of the perihelion as shown by the third equation of (6). - We conclude, therefore, from Strémgren’s careful anata that there is no secular decrease in the eccentricity due to a steady growth of the central mass; and that the views expressed by Peirce and Bryant are due to confusion, or to some error in the chain of ; reasoning. “5 This conclusion accords with the result reached by Professo# Lehmann-Filhés, in paper No. 3,° cited above. For Lehmann- Filhés shows that 7 € COS r= @, COS x, + periodic terms, “f e sin r= é, Sin 7) + periodic terms ; and remarks that when the attracting mass slowly increases orbit slowly narrows up, but yet always remains a similar cor section. He adds that.this is true for any eccentricity whateve The results of Lehmann-Filhés and Strémgren, each worked « independently ofthe other, and with much detail, are therefore entire accord; and as Stromgren’s development is given in full, every step in his analysis is quite clear, we must reject the cone sions of Peirce and Bryant as not well. founded. This concllusion that the steady increase of the central mass V not diminish the eccentricity also confirms the results reached Airy* and by Sir John Herschel. For these eminent authorities show that a central attractive disturbance decreases the eccentricit; as the planet moves from the perihelion to the aphelion, but increa * Cf. A. N., 3479-3480. *“ Gravitation,” pp. 50-51. *“ Outlines of Astronomy,” tenth edition, 1869, p. 463. 19116] INCREASE OF THE SUN’S MASS. 273 correspondingly in going from the aphelion to the perihelion; so that only periodic changes of the elements e and = occur. _ Accordingly it follows that the only possible cause which could have diminished and practically obliterated the eccentricities of the orbits of the planets and satellites is the secular action of a resisting medium, as fully set forth in Volume II. of my “ Researches on the Evolution of the Stellar Systems,’ 1910. Increasing the central mass accelerates the mean motions, and thus becomes very sensible in the theory of the motions of the planets; but it has no effect on the shape of their orbits. The almost circular form of the planetary orbits, therefore, may be referred to the secular action of a resisting medium and to no other cause whatsoever. _ This result is of no ordinary interest, since it refers the round- ess of the planetary orbits to but a single physical cause, and gives us what mathematicians call a unique solution of the leading problem f cosmogony. For Babinet’s criterion shows beyond doubt that the planets never were detached from the central bodies which now _ govern their motions; and the argument given in Volume II. of my “Researches” proves that all these bodies were formed in the stance and afterwards neared the central masses about which they now revolve. The demonstration of the true mode of formation of our solar system is therefore supported by the necessary and suffi- ent conditions usually required in mathematical reasoning ; and we may say that the laws of the formation of the solar system have been confirmed by mathematical criteria having all the rigor required in the science of geometry. This generalization will, I think, add not a little to our interest in the geometry of the heavens; and it is equally worthy of the attention of the astronomer, the geometer and the natural philosopher, who so long struggled to unfold the wonder- ful process involved in the formation of the planetary system. U. S. Nava OBSERVATORY, _ Mare Istanp, CALIFORNIA, March 20, I9gII. ON THE SOLUTION OF LINEAR DIFFERENTIAL TIONS OF SUCCESSIVE APPROXIMATIONS. By PRESTON A. LAMBERT. (Read April 20, 1911.) The object of this paper is to apply to the solution of linear ential equations, both ordinary and partial, the method of « into series used in the solution of algebraic equations in read by the author before the Philosophical Society in . and in April, 1908. Let the given differential equation be dy d*y a"y (1) ; f(x 4, I; > ax’ ax?’ x oa Sey qa)m° The method of solution consists of the following steps: > (a) Break up the left-hand member of the differential into two parts, ois a*y a"y i(nn% a dy d’y ad"y f(« “Is de’ Gan ga) such that the first part equated to zero can be integrated known method, and multiply the second part by a parameter . pendent of + and y. Replace the given equation by and = dy d"y dy ay | (2) f(x Pie 2 ego ia) + Sl ay ¥, ae’ ae (b) Assume that (3) Y= Yo + WS + Yo? + Ye? + Y4S* + + makes equation (2) an identity. 274 rgtt.] LINEAR DIFFERENTIAL EQUATIONS. 275 : (c) In this identity arranged according to the ascending powers _ of S equate to zero the coefficients of the different powers of S. (d) Solve the differential equations thus obtained in regular order FOF Yo, Vas Vos Var Vas °° *- : (e) Substitute these values in (3) andmake S unity. The result- ing value of y, if it contains a finite number of terms or if it is a uniformly convergent infinite series, is a solution of the given dif- ferential equation. _ The method of solution of linear differential equations as here outlined does not seem to occur in mathematical literature except as - developed by the author. The method will be exemplified by applying it to two differential equations, important in mathematical physics—Bessel’s equation, a _ second order ordinary differential equation, and Fourier’s equation _ for the flow of heat, a second order partial differential equation. _ Bessel’s equation is a"y ay Replace Bessel’s equation by (#35473 —1'v) + Sixty =0 and assume that ¥=Jo t+ IS + oS? +455* + y,S* + --- _ makes the latter equation an identity. _ When arranged in ascending powers of S this identity is ay, a*y,|S* 4+ +--+ =0. + xy, + 277, *This method gives a formal solution of non-linear differential equations, but up to the present time the author has been unable to test the resulting series for convergency. PROC. AMER. PHIL, SOC. L. 199 R, PRINTED JUNE 30, IQII. 276 =~ — Wy, + ry, = 0, ay : ea te o2 — ny, + 2, =O. The equation in y, is a homogeneous linear differential e and its solution is ae Ax + Br, becomes d’y, A Fe taal Ps — wy, = — Art? — By This equation becomes exact when multiplied by et, resulting equation integrated gives a linear equation of the first the solution of which is —Agt Beer n+) * n= 1) Substituting this value of y, in the equation for dete m1: nin and proceeding in the same manner Ag Ber $5 2*.2!(z + 1)(” + 2) = 24.21 (n — 1)(a — 2) In like manner I= _ act Bre 31+ 1H + 2H + 3) 7 231 — Ne — 2) — and so on. Iq 26 1911.) LINEAR DIFFERENTIAL EQUATIONS. 277 Substituting these values of yp, ¥,, V2, Vs, °** in Y=YVo + YS + YS? + 35° + 90° ioe and making S unity, . ee Cc. I a yn ae|i— $12 * Gs 1)(z + 2) 24-2! I s ~ (+ 1 + 2)(H + 3) 23! | ei I e n—1 2+ (w— 1m — 2) 2-2! +Be| 14 I = @— 1m — 2Xn — 3) 2317 |. _ When 1 is not an integer the terms of both series in this value of y continue indefinitely according to the law of formation which inspection makes evident, both series are uniformly convergent ex- cept when ro, and both series are solutions-of the given differ- ential equation. _ When 2 is a negative integer the law of formation of the terms oi the first series changes after the (m)th term and when m is a positive integer the law of formation of the terms of the second series changes after the (#)th term. The second case will be con- sidered. When 1 is a positive integer the (#)th term of the second series is PES aphate Ja 29°) n mos I)! (% _ 1)! s Substituting this value of y,_, in the differential equation for : : In determining yn,,,Yn.2s Yass, °*-, the second term in the bracket 278 LAMBERT—ON THE SOLUTION OF gives the terms of the first series in the value of y constant. This new series is combined with the first series in value of y. The first term in the bracket gives on oe ee - ati = 2-1 (nn — 1!Ih an we i)* 2*(" + 5(: + t+ eS g a"** log x | Inta = 2t—Ty 1 (n — 1)! | 212m + 1)” + 2) ant I ~ 2!12n+ 1)(" + 2) a+iI . . . . . . . . . (: + 2 +s The solution of Bessel’s differential equation when n is a p integer is therefore ve oe I # ae Te 28 (aw + 1m + 2) 22! : tae ee ~ (a + 1)(% + 2)(m + 3) 2°23) ee I Pa. +e E Be ge 2 (x — 1)(% — 2) 2*-2! Bx" log x I x I ~ Shelia api I x ("+ 1)(% + 2)(% + 3) 2°- Bx I ( Eee ~ 2“Ini(n—1)!La4+1 heat 2° I I I I )z Ee ~ (" + 1)(” + 2) (het aan eee 2 This is also the solution of the differential equation whe n negative integer. git.) LINEAR DIFFERENTIAL EQUATIONS. 279 _Fourier’s partial differential equation for the linear flow of heat is : oV #V vor ~* an Replace Fourier’s equation by OV CV : a —** oe and assume that V=V.4VS4+V 2S? 40,5 +++ -makes the latter equation an identity. __ When arranged in ascending powers of S this identity is © LALLA av, eV, |S+---=0 mem oe ta eV, eV, eV, foe) ae | = at _ Equating to zero the coefficient of the powers of S in this identity, there result the following partial differential equations for the deter- ‘mination of V,,V,,V.,V3,---, av, aV. eV, Siegen yo oa SO aV, eV. av. eV, an? a3 * aa => These partial differential equations solved in regular order give = Hx), Vix eK), Von =o. Substituting these values of V,, V,, V., V;, . . . in the assumed value of V and finally making S unity, there results A) V= oz) + 9"(ane) + 92) DY 5 gre) oils . 280 LAMBERT—ON THE SOLUTION OF [April 20, which is a solution of Fourier’s equation for all values of ¢(4#) for — which V either contains a finite number of terms or is an infinite series uniformly convergent both in x and in ¢. | 3 The following table shows several values of ¢(+) and the cor- responding solutions of Fourier’s equation, : I (1) $4) = 4, Vn A; (2) $(%) = Az, V = Az, (3) $(4) = Ax’, V = A(z? + 2K), (4) $(2) = A sin (nx), V = Ac" sin (nx), (5) $(4) = A cos (nx), V = Ae~™* cos (nx), (6) $(+) = Ae”, Vm Ac tnKt, (7) $(4) = Ae™, V x Act Bh reg rae (8) $(2) = Ae™ sin (x2), V = Ae™ sin (nx + 2 Kt). It will be noticed that in these solutions ¢(1) is the value of when t==0, that is V—=¢(-) is the initial heat distribution. It will also be noticed that in all these results + may be replaced by «+a. This statement is true of the results in the several lowing tables. If Fourier’s differential equation is replaced by and the assumption made that “V=V,+V,S + VS? + VS* + -:: makes this equation an identity, this oe. arranged in ascend powers of S is OV,|S' +++. =o. eV, HV, OV | os Kaa|t+Kga\St+k ga | "+k oe _ oY, POR So ey Ot fara Or Equating to zero the coefficients of the powers of S in identity, 1911.] LINEAR DIFFERENTIAL EQUATIONS. 281 fV, OY, 1 OF, FV 1 OV eee we aS Oe a Oe ___ Solving these partial differential equations in regular order for ke V,, V,, V2, V;, ---, substituting these values in the assumed expres- sion for V, and Getty Bee S unity, the ss ape Va drt EHO G+ WOR i) I I Cay : + Ot) + a=, +t is a solution of Fourier’s differential equation for all values of b(t) and @(t) for which V either contains a finite number of terms or is an infinite series uniformly convergent both for x and for t. Solutions of the differential equation when ¢(t) =o correspond- ing to several values of @(t) are as follows— Ot) =0, (1) (2) — A; V= A, (3) 04) =A4#, Vn A(*t tsa), a x* (4) (2) = AA, y= Asli + SI(aKA ~ aia 2%" : + cia |, (5) A(t) = Ar, Vs ae = (6) A(z) = A#, oites V= aal I+ ee re a te ee St ee Par (21? 3-5-7 = ae 6! (2k2F ; (7) O¢@)=Ar?, V= ar-} o. 282 LAMBERT—ON THE SOLUTION OF [April 20, (8) @(¢) =Asin(n/), V= A sin (nt) + +e cos ut —- | (9) (4) = A log ¢, Y= A| tog e+ ee Ki 61 Kio ae AF ee i ee? +e ate at x It will be noticed that i in these solutions = 059 is the heat dis- 7 tribution when +o. i: Solutions of the differential equation when 6(t) =o corespont= ing to several values of $(t) are as follows: ee (10) O(t) = Ae”, V= Ae! 1 III 0(t)=0, (1) $(2) = A, V = Ax, x (2) $(¢) = Ay, VaA Al a4 73 zh 2x x wy anmar rabies tide (4) $(7) = AA radilse da a ae + 33K, ween . (5) o(¢)= Ar}, Ve Ara E = (6) jase aoe E (7) $4) = Asin(n#), V=A [sin (ut)x + ¥ cos (nt) "4 ig MEP a Sows ) ye LINEAR DIFFERENTIAL EQUATIONS. 283 3 2 ee aoe 7 ORS og SEY Gt RA ot It will be noticed that in this set of solutions V =o when ro. Let u,—f,(*,t), u.—=f.(y,t), %s—=f,(z,t) represent solutions f the three one-dimensional Fourier’s equations, ively. It is readily proved that Vy = a,u, aid: are solutions respectively of the two-dimensional Fourier’s equation OV PV vida a ~*(a3+ 3) d the three-dimensional Fourier’s equation OV K sae Aa as a tot This shows how solutions of the two- and three-dimensional ourier’s equations can be obtained from the solutions of the one- dimensional equation. For example, from the one-dimensional solutions ae yaa —n?Kt ot V= oe and V= Ae sin (nx) he three-dimensional solutions Ean! Ae ee (2) = Ve Aews*+8*+ AF sin (ax) sin (By) sin (72), respectively, are obtained. ee 284 LAMBERT—ON THE SOLUTION OF If the solution of the three-dimensional Fourier’s equa ior oV K ae halle A ES ay + is a function of r and t¢ only, so ae V =f (r,t), where r= (2? + y? + 2*)}, the transformation of the given equation from rectangular to 0 coordinates shows that the solution is u : 3 Vass : ‘ : where wu is a solution of the Fourier’s equation Ou Ou a ae ; It follows that solutions of the three-dineusionsdl: cual of tl form V f(r, t) are obtained by replacing x by r in any solut ic of the one-dimensional equation OV ee oe — * oe and dividing the result by r. In this manner are obtained the solutions A V (1) i ees. —r Ae*kt * ee (3) V = oe (ur + 2n°K?). : -It is interesting to compare the solutions of Fourier’s partial ferential equation obtained in this paper with the solutions tabu by Sir William Thomson in thé mathematcal appendix of the ar 1 “Heat” in the “ Encyclopaedia Britannica,” ninth edition. Sir William Thomson obtains his results by summation, xgtt.] LINEAR DIFFERENTIAL EQUATIONS. 285 _by integration, from the solution IV (1) above. All his results occur directly in the above tables or are combinations of two of these solutions. It is evident that there are several misprints in the results as printed in the “ Britannica.” _ Of course there are many solutions of Fourier’s equation which must be built up from elementary solutions, however found, by means of Fourier series, or which must be obtained by the methods of harmonic analysis. The solution III (5) above is the series used by Sir William Thomson in his solution of the problem of the secular cooling of the earth.? An interesting result in pure mathematics is obtained as follows: Sir William Thomson shows that for a continued point source of , if the rate is an arbitrary function of the time, f(t), the solu- of Fourier’s equation when K=1 is given by the definite sant et Vm dsfe— )eaw _ The second part of the general solution (B) above shows that r? ves |i porronsro a +-| is also the solution of Fourier’s equation for the same conditions. dee It ows that [en Ga- pak SOAPS ODF APO + « | sa general formula for computing the definite integral. _ LenicH UNIversity, BETHLEHEM, Pa. _ #“Mathematical and Physical Papers,” Vol. III. PROBLEMS IN PETROLOGY. By JOSEPH P. IDDINGS. (Read April 21, 1911.) parts of the earth, and to their composition and structure. Not that there is less need than formerly for accurate observati tion, texture and occurrence, but the introduction of greater defi ness into conceptions of their modes of formation, and the widening — and synthetic investigations of the geophysicist, have advanced t study of rocks from the accumulation of data and statistics, to the — formation of laws and relationships, both as regards minutest de of composition and texture, and with respect to petrographical p inces, and their connection with the dynamical history of the regi of the earth in which they occur. As a consequence of this advance new problems present th selves, and invite the codperation of workers in several bran of inorganic science. Leaving out of consideration for the preser the great problems of metamorphism, some of which are being cessfully treated by Adams, or have been under investigation Van Hise and Leith, I wish to call your attention to certain ph of the study of igneous rocks that may be grouped under heads for present purposes, as follows: (1) The actual min composition of igneous rocks, (2) the mathematics of the pee 08) of igneous rocks, and (3) petrographical provinces. 286 7 IDDINGS—PROBLEMS IN PETROLOGY. 287 1. ActuAL MINERAL COMPOSITION OF IGNEOUS Rocks. Although the minerals constituting various rocks are their most obvious features, aside from their general color and texture, and _ have been the chief object of study by petrographers since the intro- duction of microscopical methods of investigation, they still remain among the most important problems before the petrologist. The exact composition, crystal characters and optical properties of many of the minerals are well known. But some of the common- est, such as the micas, amphiboles and aluminous pyroxenes, are not _ perfectly understood chemically, and the relation between their com- position and optical constants is not so definite that one may be employed to determine the other, as is the case with the lime-soda- _ feldspars. Moreover, the exact amounts of the component minerals in various kinds of igneous rocks have not been determined, except in a very few instances; nor has the precise composition of those minerals that occur in mixed crystals, that is, the principal ferro- _Magnesian minerals, been determined in the vast majority of the _ rocks described. There is, therefore, a great field of research, imperfectly culti- _ vated, capable of yielding immediate returns of the first importance ‘ for the solution of other problems connected with the mineral com- _ position of these rocks. Similarly, more definite and specific study and description are needed of the crystal forms and arrangements of the mineral con- _ stituents of igneous rocks than have heretofore appeared in petrog- faphy, in order that the texture of various rocks may be clearly understood, since texture is a very definite exponent of physical _ conditions that attended the crystallization of each igneous magma. F ‘Up to the present time petrographers have been content with very _ vague and incomplete descriptions of rock textures, as well as of _ kinds and amounts of minerals composing various igneous rocks. _ The determination of the kinds and amounts of the minerals in _ every rock leads to the problem of the formation of the minerals in each instance, and a comparison of the mineral composition of _ a rock with the chemical composition of the magma from which it 288 IDDINGS—PROBLEMS IN PETROLOGY. solidified. This involves the chemistry of solutions of ino compounds, chiefly silicates; the mutual interaction of the va: chemical elements that appear in an analysis of the whole r together with the possible catalytic action of constituents, no’ water gas, that may not become parts of the fixed compounds, may escape in greater part upon the solidification of the magma. Some of thé minor problems, or factors, within this large ¢ may be alluded to briefly as follows: The first and most of result of a strict correlation of the mineral composition of with the chemical composition of the whole mass, repres the fixed components of the formerly liquid magma, is the nition of the nonappearance in certain kinds of rocks of ; minerals whose presence is necessary to satisfy the chemical req ments of the magma solutions. This is the case with comp! crystallized but exceedingly fine-grained lavas of particular co positions, notably andesites. ‘ Minerals that should be present to the extent of as ode as per cent. in some instances are not visible, are occult, and TABLE I. I. & 2. SiO, 59.87 quartz 13.02 AlOs 15.02 orthoclase 17.24 Fe.0Os 2.58 albite 28.56 FeO 3.40 anorthite 17.28 MgO 4.06 diopside 3.99 CaO 4.79 hypersthene II,II Na.O 3.39 hornblende —— K.O 2.93 biotite — H:0 1.86 hematite a TiO, 0.72 magnetite 3.71 P.O; 0.26 © ilmenite 1.37.4 F 0.02 apatite 0.50 Etc. 1.10 water 1.51 1.00.00 Ete. 1.26 99.51 be hidden within the substance of those minerals that are v that is, they must be held in solid solution within other kind crystals. An example will illustrate the case. A magma whose chemical composition is shown by analysis IDDINGS—PROBLEMS IN PETROLOGY. 289 ‘Table I., under favorable conditions should form the mineral com- pounds in the proportions shown in column 2. In this there are 13 per cent. of quartz, and 17 per cent. of orthoclase, together making 30 per cent. of the whole. And this amount of quartz is the least amount of free silica capable of separating from a solution of such a chemical composition, assuming that the minerals formed are those known to occur in igneous rocks. A magma of this chemical com- position commonly crystallizes as a pyroxene-andesite composed, so far as the microscope can determine, of lime-soda-feldspar, pyrox- _ ene, and magnetite, with no visible quartz or orthoclase. And, yet, _ from the chemical analysis of the rock there should be 30 per cent. these compounds. __ The orthoclase molecules may be readily imagined in solid solu- tions within the lime-soda-feldspars, although in coarsely crystal- lized forms of such a magma, diorite, orthoclase crystals appear as independent individuals. It has been shown in the Geophysical Laboratory of the Carnegie Institution. that orthoclase and anor- thite molecules form homogeneous mixed crystals when melted together and cooled in an open crucible. The disappearance of 17 ‘per cent. of orthoclase in this particular andesite is, therefore, due to the conditions of solidification of the rock. The non-appearance of the quartz may be explained in part by its existence in solid solu- tion in other minerals of which, however, we have not sufficient evidence at present; or it may occur in minute crystals mistaken for _andesine feldspar, since the optical properties of the two that may be recognized in minute crystals, are almost identical. In coarser grained forms of chemically similar magmas the quartz appears, but _the conditions attending crystallization in the contrasted cases may favor its disappearance through solid solution in one instance, and its separation as quartz crystals in the other. __ In this connection it is to be pointed out that the apparent actual mineral composition of certain igneous rocks may not be the real eral composition by as much as 30 per cent. of the whole. For the occult minerals in solid solution are as much a part of the rock though visible. “Moréover; the percentages assigned to the min- ls that aré seen must be in error by the amounts of the occult 290 IDDINGS—PROBLEMS IN PETROLOGY. [April | minerals in solution. The problem of the determination of mineral composition of rocks is for this reason more complex than at first appears, and is further complicated by the difficulty of determining the amounts of colored and colorless crystals, bs: they appreciably overlap one another in thin section. om Another obvious result of a comparison of the actual 1 : composition of igneous rocks with the chemical composition o . their magmas is the notable variability in the combination of miner that may in some instances result from the crystallization of ong mas of like chemical composition. This is true both as to kinds and amounts of the resulting minerals. A striking illustration of | variability is found in the mineral composition of three rocks fi the same region, the parish of Gran, Norway, which have described by Brégger. Analyses of the three are shown in colut I, 2 and 3, Table I. The first rock is an essexite, the second camptonite, the third a hornblendite, and while the compositi differ slightly in percentages of silica, and to a less extent in. ot constituents, the chemical resemblances are striking, and the tl analyses lie within the range of many well-known series of a of particular rocks. TABLE II. I. . SiO, 43.65 40.60 37.90 4295 orthoclase 72 Al:Os 11.48 12.55 13.17 12.29 albite ad Fe:Os 6.32 5.47 8.83 3.80 anorthite 189 — FeO 8.00 9.52 8.37 7.05 leucite 3-9 MgO 7.92 8.06 9.50 13.09 nephelite ‘II. CaO 14.00 10.80 10.75 12.49 diopside 26.9 — NasO 2.28 2.54 2.35 2.74 olivine B68: K.0 1.51 1.19 2.12 1.04 magnetite 11.6 H.0 1.00 2.28 1.40 1.82 ilmenite 10.2 CO, tr 268 —etc. 0.62 hematite 08 TiO, 4.00 4.20 5.30 1.82 apatite —_— P:Os tr — tr .99 H.0 1.4 100.16 100.79 99.69 100.19 Etc. pacnittig The first rock consists of lime-soda-feldspar and augite, some olivine and mica, and rarely a little hornblende. The s IDDINGS—PROBLEMS IN PETROLOGY. 291 while the third is almost wholly hornblende, only 2 per cent. being pyroxene and nephelite. The same magma might have crystallized as nephelite-basanite, as appears from the calculated mineral com- position shown in column 5, and from comparison with the analysis and mineral composition of a nephelite-basanite from Colfax County, N. M., shown in columns 4 and 6. __ This is only an extreme case of variations well known to exist in most groups of rocks that may be referred to chemically similar magmas. And the magma already cited as capable of furnishing a pyroxene-andesite may also yield a quartz-mica-diorite, whose com- sition is shown in column 3 of the first table. It is evident from these examples that the minerals called horn- blende, or more properly amphibole, in the descriptions of these tocks differ widely in chemical composition, and often represent totally different mixed salts. Thus in the hornblendite of Gran, the hornblende contains all the components that might, under other con- ditions, have crystallized as pyroxene, olivine, feldspar, leucite, nephelite and magnetite. _ Any attempt to correlate igneous rocks on the basis of the actual mineral composition, without taking into account the actual chem- al composition of the minerals involved in each case must lead to confusion. . _ One of the most important problems in petrology is the elucida- tion of the laws controlling the production of mineral compounds from molten magmas. A consideration of the simpler chemical reac- tions that may be expected to take place in silicate solutions like Tock magmas, and wht do take place in crucibles in the laboratory, iopside, hypersthene, olivine, thnghtibe mere some cheat feel minerals. _ Minerals like mica clearly involve the chemical action of water, its components, hydrogen and hydroxyl, since hydrogen enters into its constitution. According to Penfield hydroxyl, and sometimes fluorine, enters into the composition of hornblendes, forming bivalent radicles with aluminium, and ferric iron. In pyrogenetic analcite, and in other possibly primary zeolites in igneous rocks, H,O enters PROC. AMER. PHIL. SOC. L. 199 S, PRINTED JUNE 30, IQII. 292 IDDINGS—PROBLEMS IN PETROLOGY, into the silicate compound. The physical conditions which control the chemical equilibrium within magma solutions that yield th mineral compounds are problems for the geophysicist, though their nature may be inferred in a general way from the mode of occur- rence of the rocks containing the minerals in question. Indications of a catalytic action of H,O within rock magmas a furnished by the association of free silica with orthosilicates con- taining magnesium and iron, such as the common occurrence of quartz and biotite in granitic rocks; the frequent association quartz and tridymite with olivine in lavas; and of quartz, tridymi and fayalite in lithophysae in certain highly siliceous lavas. The instability of these systems under changed conditions equilibrium is shown by the inversion of hornblende to an ede tion of pyroxene, magnetite and feldspar, in some lavas; and by the solution of quartz phenocrysts in some basalts, accompanied by formation of shells of metasilicates surrounding them. Already laboratory research has established the range of stab of some of the rock minerals under laboratory conditions: the in sion temperatures under atmospheric pressures of the various forn of SiO,, quartz, tridymite, crystobalite; of the simpler compound: crystallizing as orthorhombic and monoclinic pyroxene, and corresponding amphiboles ; of a simple system involving alumini magnesium calcium silicates; and of other series of compow The value of these definite contributions to the problems of mineral composition of igneous rocks is great. Much mor needed. And the necessity for eventually approaching nearer the physical conditions obtaining in rock magmas is apparent, the probable efficiency, chemical and physical, of highly he gases under strong pressures is taken into consideration. Rese under such conditions is attended with great difficulties, and s risks. Enough has been mentioned to show a wide range for fv at study by the geophysicist, the chemist, and the petrographer. oe 2. THe MATHEMATICS OF THE PETROLOGY OF IGNEOUS ROCKS. The study of igneous rocks involves the consideration of gre of intricate relationships, the exact expression of which is at p IDDINGS—PROBLEMS IN PETROLOGY. 293 ent beyond our competence. Abstract conceptions of some of the simpler relationships, based on partial knowledge of the factors involved, serve to point the way along which quantitative investiga- tion may be profitably pursued. The stoichiometric character of the chemical compounds that con- stitute rock minerals relates them as definite functions to the chemical constituents of the liquid magma from which they crystallized. The existence of mixed crystals, and of solid solutions, introduces the _ treatment of series into the problem of the expression of the relation- ship between the mineral composition of a rock and the chemical mposition of its magma. In such an expression the fixed compo- nents alone are invdlved. But there are definite quantitative rela- ips to be expressed regarding those chemical components of a ‘magma which may act only catalytically in producing the actual eral combination constituting the rock. Such actions may be mical, in the sense that compounds form that subsequently dis- ear, as should H,O combine with SiO, to form hydrogen ortho- silicate, H,SiO,, and subsequently resolve itself into water and quartz tridymite. Or they may be physical, in the sense that increased ecular mobility in the magma liquid may affect the character of crystallization by changing the freezing point and the nature of compounds stable under the conditions obtaining at the time. In broadest sense, then, the mineral composition of an igneous rock is a function of the chemical composition of the magma. _ Since the physical conditions attending the solidification of rock gmas affect the chemical equilibrium of the constituents, as well s the physical character of the liquid, its temperature and viscosity, also influence the chemical composition with respect to the gas- $ components capable of being held in solution under pressure, he mineral composition of an igneous rock is also a function of the physical conditions attending its solidification. _ Toa notable extent is this also true of the texture of such rocks, their degree of crystallization, size of grain, and the shape and angement of the individual minerals. In the expression of these Telationships the treatment of serial functions must be a pronounced feature. The gradual variations of temperature and pressure are as 294 IDDINGS—PROBLEMS IN PETROLOGY. [April essential factors in the consideration of the physical conditions rock magmas, as the variations in texture and in mineral composi are universally characteristic features of igneous rocks. The existence of definite quantitative relations between the min- eral composition and the texture of igneous rocks on the one hand, ~ and the chemical composition of the magma and the physical condi- tions attending its eruption and solidification on the other, rests the obedience of the component elements to the laws of physi chemistry. These laws are not fully established, or known, at this time, and the relationships involved may be too intricate to be com- pletely expressed in customary mathematical terms, nevertheless, definiteness of the quantitative relationships can not be doubted, < approximate expressions of them become problems for petrole of the future. ae In the consideration and correlation of all known igneous rock: variability in composition and texture and the existence of continuc series are the most conspicuous general characteristics. The va bility in the composition of igneous rocks indicates heterogenei magma solutions. This may be inherent in them, and represent a condition of existence before the initiation of eruption; or, as is more probably the case, it may result from differentiation of hor geneous magmas during periods of eruptive activity, within more less extended regions. Differentiation results from diffusion of co pounds in solid molecules, or less complex ones, either at the tim separation as crystals, or earlier, through convection currents, diff ences in density, or differences in solution pressure. The resu magma solutions differ only in the quantities of various che compounds ; the amount of some in extreme instances reaching Subsequently formed compounds are not inherently different those in other magmas except by reason of the amounts of « chemical components, which may be concentrated in some entiated parts; as in the concentration of the rare elements in pegmatites ; or by different combination of chemical elements th catalytic agents. There are no inherent, or inherited, chara of form, organism, or immaterial traits, as in living beings. magmas are simply differently mixed solutions of inorganic pounds. IDDINGS—PROBLEMS IN PETROLOGY. 295 Magma solutions possess different degrees of heterogeneity as shown by the composition of various bodies of igneous rocks. In some there are slight differences in different parts, extending through large masses. In others marked differences occur within short dis- tances in small masses. Variability in the composition of igneous rocks from place to place is a universal characteristic, resulting in series of varieties of composition within single bodies, and among different masses. The aggregate of all such series of variations in one region may form a continuous series of wide extent; or there _ may be gaps in the series in one region, which may be filled by the _ phases of composition exhibited by rocks in another region. ____Inone region the composition of a nearly homogeneous rock mass of considerable magnitude may assume a certain local petrographic _ importance, while in another region it may appear only as a facies of another rock body. There appears to be no chemicophysical rea- son for the production of a magma solution of one mixed composition rather than of another very nearly the same. But it is known that ‘magmas of intermediate, or more mixed, compositions, are more abundant than those of extreme, or simpler, compositions. The accumulated evidence of chemical analysis, microscopical study of rock sections, and observation in the field, shows the exist- ence of wide serial variations of composition, continuous along numerous lines, owing to the number of variable components. This evidence also shows that there is no one definitely composed magma solution more abundant throughout large areas of the earth than others; none that deserves special consideration, or may be recog- nized as a universal type. It is true, as already remarked, that in certain regions there are large bodies of rock having nearly uniform composition that assume local importance, and serve as types for reference in particular regions. But it must be admitted that the idea of type is subjective, inherent in the petrographer, not the rock. And when all known series of igneous rocks are treated as products of chemicophysical reactions universal in their application, the for- tuitous character of the chemical composition of particular bodies of ‘erupted magma becomes apparent, and the significance of such local types disappears in a systematic treatment of the whole body of 296 IDDINGS—PROBLEMS IN PETROLOGY. [April 21, petrographical facts involved in a comprehensive description of igneous rocks, fee ; Recognizing the existence of continuous series of petrographical factors, chemical, mineral and textural, necessary to the complete description and definition of igneous rocks, the problem presents itself of dividing the complex series of rocks so characterized into parts that may be described in a comprehensive and systematic manner. — A familiar example of a physical series divided in a regular man- — ner for purposes of exact use is that of temperature, partitioned in degrees of definite proportions of a continuous scale. It is undoubt- edly an arbitrary method and differs distinctly in three commonly employed usages. It might be a more “natural” method to express temperature with reference to the melting points of a series of sub- ~ stances; and the value of certain of these definite points as datum points is well known. But the merits of the arbitrarily, but vei naturally, divided scale are attested by its universal employment. The proposal to partition the petrographical series into quantita- tively definite parts, as has been done in the Quantitative System of Classification of Igneous Rocks, the size of the divisions being arbi- trarily chosen, has excited criticism by some petrographers, who consider it arbitrary, artificial and not “natural.” But the objection, that measured precision condemns a classification of igneous rocks, because it makes evident “its aloofness from the scheme of nature based not on arithmetical but on physical and chemical principles,”? _ suggests a lack of appreciation of the mathematical precision of stoichiometric chemistry, and a failure to grasp the definiteness of ‘quantitative physics, whose natural expression is found in high er mathematics. Both of these sciences are fundamental to that petrology; and as mathematics is the language, or expression, ¢ quantitative relationships, the more definite the knowledge of the quantitative factors and relationships obtaining in igneous rocks, the more natural will become their expression in mathematical terms. ‘ Acknowledging the usefulness of such terms as “ consangui and “ parent ” magmas, in emphasizing the fact that there is relati ship between rocks in certain instances, it must be admitted that fl too frequent use of these and other biological terms, as “ families” * Harker, A., “The Natural History of Igneous Rocks,” 1909, p. 366. — > ott.) IDDINGS—PROBLEMS IN PETROLOGY. 297 of rocks, minerals of “first and second generation,” and the like, tends to convey by implication the idea that there exists among _ igneous rocks genetic relationships analogous to those sustained by living organisms. In fact, this idea has been clearly formulated _ by Harker? in stating that the mutual relationships ef igneous rocks will furnish a “ fundamental principle analogous with that of descent, _which lies at the root of natural classification in the organic world.” The significance of the term “natural” when applied by some _ petrographers to petrographic classification appears to be pregnant _ with biological conceptions. But what is proper and natural in the treatment of assemblages of organisms is not for that reason, neces- sarily, proper, or natural, in the treatment of a series of chemical _ solutions and their solidified phases, however much the various solu- _ tions may be related to one another by reason of differential diffusion _or fractional crystallization. 3. PETROGRAPHICAL PROVINCES. Although the fact has been recognized for twenty-five years that there are regions within which the rocks erupted during any par- ticular period exhibit certain peculiarities of mineral composition and texture that distinguish them from rocks belonging to the same gen- eral group, erupted simultaneously in other regions,* little or no attempt has been made to define more precisely what constitutes the characteristics of any so-called petrographical province. _ It has been pointed out that in some regions many of the igneous _ tocks are especially rich in alkalies; in some sodium being promi- nent; in others potassium. But nothing approaching completeness of definition, either as to composition of the rocks, or extent and limit of the region of occurrence, has ever been attempted. _ And yet some very general and far-reaching speculations have _ been indulged in on the basis of hastily formed impressions, both as _ to the character of such groups of rocks and their relationship to _ assumed structural features of the earth. Asa result certain petrog- _ raphers have grouped all igneous rocks into two contrasted cate- *Tbid., p. 362. * Judd, J. W., Quar. Jour. Geol. Society, London, 1886, Vol. 42, p. 54. 298 — IDDINGS—PROBLEMS IN PETROLOGY. gories, without considering the probability of their being many phases of combination of the variable factors of igneous rocks that must characterize all petrographical provinces of the earth. The assumption that rocks must either belong to what have been called the “Atlantic” or the “ Pacific” provinces, without serious — definition of either of these rather comprehensive terms, has led to — the humorous conclusion that the igneous rocks of Great Britain — belonged in some periods of geological history to the “Atlantic,” in others to the “ Pacific” provinces ; indicating the flexible, one might say caoutchouc-like, nature of these provinces. . The igneous rocks of the Andes and of the western Cordillera of North America have been referred to as representing the “Pacific” province, while the more alkalic rocks of Scandinavia and of some other parts of Europe are considered to represent the “Atlantic.” The igneous rocks of Great Britain belong to neither of these dis- tinctive groups as a whole. And the rocks erupted at different geological periods in Great Britain, while they exhibit some va tions in extremes of composition, which might result from different degrees of differentiation of chemically similar magmas, bear some of those resemblances to one another that are supposed to charac terize rocks of one petrographical province. The misconception underlying the generalization responsible for the terms “Atlantic” and “ Pacific,” as applied to petrographi provinces, appears from the facts brought out by Cross regarding the alkalic character of some of the lavas of Hawaii, and by Lacroix regarding alkalic rocks in Tahiti; to say nothing of similar rocks in New Zealand and elsewhere in the southern Pacific. Moreover, in the midst of Europe, in Hungary, there are groups of rocks iden in all respects with those of the Great Basin in western America. From this it is evident that one of the most important and inter- esting problems before petrologists is the investigation and e definition of the districts and regions of igneous rocks in all parts the world, with the purpose of obtaining the data with which to f definite conceptions of what have been termed petrographical p: inces. Enough is known already to make it evident that there ai many kinds of such groups of igneous eruptions and not two strongly contrasted series; that they blend into one another in composition t9tt.] IDDINGS—PROBLEMS IN PETROLOGY. 299 that the delimitation of the regions, or provinces, may be pronounced in some instances, and ill-defined in, others. The character of the rocks in different provinces, and the distri- bution of provinces throughout the earth, together with their rela- tions to the geological structure and dynamical history of the region in which they occur, furnish problems of the first magnitude in petrology. One of the questions to be answered is: the relation of the com- _ position of igneous rocks of different parts of the earth to its isostasy. : The configuration of the earth’s surface demands the presence of - material of different densities beneath the surface. Does this show : itself in the character of the material erupted in different regions. An answer to this can not be given offhand. The requirements in density are relatively so slight when great volumes are concerned, as pointed out recently by Hayford ;* the series of igneous magmas of any region is so diversified in composition and density; and the estimation of their several volumes is so hazardous an undertaking that a reasonable solution of the problem can only be expected after the accumulation of a great amount of exact data. Whether there is any relation between the kinds of magma erupted in a particular region and the dynamical events within the region is another problem yet to be solved. Assertions to the effect that there is a definite relationship have been made, but they are in the nature of broad generalizations upon questionable premises, producing the results already discussed in connection with the terms “Atlantic” and “ Pacific.” It is possible that differences in the sequence of dynamic events ‘in various regions, or in one region at various periods of its history, ‘may be accompanied by differences in the processes and results of differentiation of chemically similar magmas; that is, in series of erupted rocks, but the existence of such relationships has yet to be clearly established. For it is also possible that the material of the earth may be heterogeneous in composition, differing somewhat from place to place, and yielding different kinds of magmas in different *Hayford, J. F., “ The Relations of Isostasy to Geodesy, Geophysics and _ Geology,” Science, N. S., Vol. 33, No. 841, 1911, pp. 199-208. PEM Tae es CYS co solution of which involves the cooperation of petrographers chemist, the geophysicist and ihe geologist. A STUDY OF THE TERTIARY FLORAS OF THE ATLANTIC AND GULF COASTAL PLAIN. By EDWARD W. BERRY. (Read April 21, 1911.) INTRODUCTORY. The observations recorded in the following pages may be said to represent a preliminary sketch of a small chapter in the study of the South Atlantic and Gulf Coastal plain undertaken by the United States Geological Survey in codperation with the various state sur- veys under the direction of Dr. T. W. Vaughan. Neither geologist nor biologist fully appreciates the magnitude, complexity or uniqueness of the coastal plain of the southeastern United States. The present coast line, a boundary first recognized by the aborigines and early explorers and so emphasized by ‘geog- taphers, is from the standpoint of the student of geologic history a continually shifting demarcation which does not, nor perhaps never, marked the seaward limit of the physiographic unit known as the Coastal Plain Province, for the gently sloping land surface continues seaward beneath the waves of the present Atlantic and Gulf waters varying distances up to 100 miles or more and then precipitately de- scends several thousand feet in a few miles, forming the majestic escarpment which is regarded as the continental boundary. In the past the coast line has advanced inland over the present emerged ‘portion of the coastal plain and receded seaward over the present submerged margin, many times. At one time the waves of the Gulf of Mexico broke in southern Illinois, at another they were confined 100 miles south of the present sites of Mobile and New Orleans, 600 miles to the southward. On the whole, the history of events in Tertiary times has been a Progressive adding to the land area of the continent, the most im- _ *Published with the permission of the director of the United States Geological Survey. * 301 302 BERRY—TERTIARY FLORAS OF THE portant elevation being that of the early Miocene which was followed — by a subsidence, which was, however, less in extent than those which | had preceded it. 3 No part of the coastal plain is so favorably situated for the study of the floras which preceded the present, extending backward to a time which marks the first recorded appearance of angiosperms, as that of the Gulf states. No single part of North America contains so continuous a series of Tertiary deposits carrying fossil plants. — Here we find abundant floras in the lower and middle stages of the Eocene, considerable floras in the Oligocene, some in the later Mio-— cene, and rather abundant fossil plants in the Pliocene. The Rocky Mountain region is rich in Eocene fossil plants and there are some Miocene floras, but no Oligocene or Pliocene floras are known. The Pacific coast region likewise furnishes Eocene and Miocene fossil plants but none of Oligocene age. The fossil floras of the coastal plain are found in an area where it is possible to attain to some measure of accuracy in predicating the general character and cour of ocean currents and winds and other physical features of the en- vironment. On the other hand the western floras just mentioned grew in areas where vulcanism was great at times; in areas of ' a orogenic activity, where changes in topography were numerous and elevations of several thousands of feet are recorded; areas in which r climatic conditions not only varied from place to place, but pa through a large cycle of secular changes. All these factors gre complicate the floral history. The floras of the southern coastal plain are moreover check for the most part by very abundant marine faunas in intercal beds, or the plant-bearing beds which represent the coastal swai and the shallow water deposition of the old embayment m er laterally with the contemporaneous limestones or marls which forming in more open waters along the coasts to the southward, that there is a considerable body of facts bearing on depth, cha of the bottom, and marine temperatures, with which to compare temperatures. These criteria have been admirably worked out the Florida area by Doctors Dall and Vaughan for the post-Ec and their results furnished a reliable datum plane for the deducti to be derived from the study of the fossil floras of these times. agtt.] ATLANTIC AND GULF COASTAL PLAIN. 303 So far as I know I was the first paleobotanist to explore the south Atlantic and Gulf coastal plain and that exploration has only just begun. Professors Fontaine and Ward visited the region and collected a few Cretaceous plants a score of years ago. Professor Lesquereux a generation and a half ago described a few Eocene plants collected by Professor Hilgard in Mississippi and by Professor Safford in Tennessee, and Doctors Knowlton and Hollick have iden- tified various small collections made by others in different parts of this vast area. With the exception of fragments of the petrified stems of con- ifers, palms and dicotyledons the plant-remains are in the form of ‘impressions, mostly of foliage, but with a goodly sprinkling of fruits and seeds, and in some few cases even flowers are preserved. While the oscillations of the Gulf area have been numerous they have been, as I have just mentioned, inconsiderable in amount, only a few hundred feet at most, and the coastal region has uniformly been one of slight relief. The various floras show a complete absence of upland types. This is in striking contrast to the Euro- pean older Tertiary floras. The only large area of the globe which has been thoroughly studied, Europe, was far less stable than this region in Tertiary times and lying much farther toward the pole was subjected to the rigors of Pleistocene conditions whose influence never reached our southern states. The object of the writer’s work may be classed under three _ heads: (1) To determine the correlation of the various Tertiary for- _ mations particularly in the upper portion of the Mississippi embay- ‘ment where marine fossils are largely absent, (2) To obtain data regarding the physical conditions under which the various floras _ flourished, (3) To accumulate biological data regarding the geo- graphical distribution, specific differentiation and evolution of the Tertiary floras. ‘Thus one of the principal phases of the study for the geologist might be embraced under the term paleoecology. The methods include a study of the old shore lines of the different epochs, of the character of the sediments and their genesis, of the contained animals and plants, and the alternative climatic and edaphic factors which their grouping may indicate. 304 _ BERRY—TERTIARY FLORAS OF THE [April 21, It is the chronologic and ecologic senuens upon which I wish to dwell in the present connection. The paleobotanical record of the Atlantic and Gulf coastal cas a furnishes a history which extends back as I have just mentioned — beyond the oldest known angiosperm to a time (Lower Cretaceous) when the flora was made up almost entirely of tree-ferns, conifers. and those interesting cycadophytes (Cycadeoidea) whose trunks are sometimes preserved with such marvelous perfection that the out- lines of the embryos in the ovules can often be made out in detail. — Coming a step nearer my present theme, a step of some millions. — of years from the Lower into the Upper Cretaceous we find the first _ great modernization of the floras of the world due to the seemingly sudden evolution of the main types of angiosperms. These upper Cretaceous floras are well represented in the coastal plain from — Marthas Vineyard to Texas. They extended northward to Green-_ land and southward to Argentina in South America, and are found to indicate very different physical conditions from those which — prevail at the present time. I do not intend, however, to dwell upon the Upper Cretaceous floras in this connection but pass to a con- sideration of the succeeding Eocene stage of plant evolution. In — this as in subsequent times the chief emphasis will be laid upon that — section known as the embayment or old Mississippi Gulf, although where the record is more complete in other parts of the coastal plain I will not hesitate to use it. BasAL EOceENEe. The Eocene as defined by Lyell was marked by the dawn of recent species of marine mollusca. It is equally well marked by sudden expansion and evolution of modern types of mammals | plants after a long antecedent Cretaceous development. The floras become thoroughly modernized as compared with those which pre- ceded them, although they are still very different in their general facies and distribution from those of the present. In the earliest stage of the Eocene known as the Midway, tl relations of sea and land in the Gulf area differed in only minor par- ticulars from that of the late Cretaceous. The waters of the Missis- grt.) , ATLANTIC AND GULF COASTAL PLAIN. 305 -sippi Gulf were, however, deeper. This factor combined with a much less influx of fresh water from the tributary streams, due in ‘some measure to the low relief of the land, enabled marine faunas to reach well toward the head of the gulf. These faunas indicate ‘subtropical bottom temperatures northward as far as Paducah, Ky. ‘The known floras are very scanty and unsatisfactory and in the present state of our knowledge do not merit an extended discussion. Lower Eocene. _ The Midway Eocene was succeeded by a long interval during which a great thickness of deposits was laid down which are col- _ lectively known as the Wilcox Group. The character of these sedi- ments and their faunas show that the gulf was somewhat restricted and much shallower than in the preceding stage, with true marine conditions prevalent only in its lower portion. The shores were 3 low and relatively flat. They were flanked by current- or wave-built bars and separated from the mainland by shallow inlets or lagoons. The lower courses of the streams were transformed into shallow ; estuaries or broad swamps through which the smaller streams meandered. The accompanying sketch map (Fig. 1) shows the rela- tion of land to water at this time. The shore line along which the strand flora migrated is approximately indicated, and some of the localities where fossil plants have been discovered in the littoral _ deposits of this age are indicated by stars, while the general move- ment of the warm ocean currents is indicated by arrows. A mag- _nificent flora is preserved at a large number of localities in the clay lenses which were formed in these estuaries and marginal lagoons. _ This flora shows a mingling of tropical and subtropical types as far a northward as where the Ohio River now flows into the Mississippi. It is of unparalleled richness and preservation and will bear a more : extended analysis. __ Among the ferns it contains representatives of the genera Acro- _ stichum, Pteris and Lygodium, none of which appear to be common. _ Both feather and fan palms are not uncommon. Conifers are rep- _ resented by a single occurrence of a species of Arthrotaxris—a genus _ which in the living flora is confined to the coastal swamps of Tas- 306 BERRY—TERTIARY FLORAS OF THE [April 21, mania but which is widespread in European Eocene floras. A large variety of dicotyledonous forms are preserved, representatives of about two hundred different species of which about one third have thus far been satisfactorily identified. These include seven or eight species of leguminous shrubs and trees represented by pods as well as leaflets—evidently strand plants, as are numerous modern species ie ~ Dy PS Cal Ca — = - * 7 A Ca - 7 cial Fic. 1. Sketch map showing the approximate relation of land to water in the Lower Eocene. Stars indicate fossil plant localities, diagonal lining indicates submerged areas. of Acacia, Cesalpinia and Dalbergia. Evergreen lauraceous forms — are also abundant, the genera Cinnamomum, Laurus, Malapoenna, Persea, Oreodaphne (Ocotea), etc., being represented by several species. Figs are abundant and of several species, embracing both 1911.] ATLANTIC AND GULF COASTAL PLAIN. 307 the pinnately veined and the palmately veined types. There are three or four species of Sapindus—another strand type of the mod- ern equatorial and subequatorial zones. Other members of the strand flora include representatives of the genera Conocarpus, Guetteria, Mimusops, Persoonia, Terminalia, etc. Leaves of several species of live oaks (Quercus) are abundant. The collections also include fruits of the families Anacardiacee and Umbellifere, and of the genus Aristolochia. Curious elements common to Europe are several species of Banksia, an antipodean genus in the existing flora. There is a fine species of Cercis, a very common Euonymus and at least two species of Engelhardtia based upon the characteristic fruits as well as leaves. The latter genus has a single existing species in Central America and several in Asia, where they range from India to the East Indies. It is common in the European Tertiary, but has not previously been known with certainty from orth America. An interesting member of this flora is a large digitate species of Oreopanax, a modern tropical type, abundant in Central America. _ The flora as a whole contains no strictly temperate elements, although many of the genera contain modern forms which range for more: or less considerable distances in the temperate zone. Such a flora could scarcely flourish under existing conditions north of latitude 29°. In its general facies it is subtropical and a number of e forms indicate a high percentage of humidity, and well dis- tributed and abundant seasonal rains, although this latter feature tends to be obscured by the large number of the inhabitants of the ‘sandy shores which are preserved while the inland and river bank dwellers are less fully represented. A majority of the elements in this Wilcox flora could be duplicated today on the Florida Keys and the southern peninsular mainland of Florida. Additional members of this flora not enumerated in the preceding ‘paragraphs include representatives of the genera Apocynophyllum, Calamopsis, Ceanothus, Celastrus, Celtis, Cordia, Diospyros, Dryo- : phyllum, Magnolia, Malpighiastrum, N. erium, Rhamnus, Rhus, Saba, Sapotacites, etc., nearly all of which are new to science. PROC. AMER. PHIL, SOC. L. 199 T, PRINTED JUNE 30, IQII. 308 BERRY—TERTIARY FLORAS OF THE [April ar, MippLe Eocene. Middle Eocene floras are less abundant than those of the Lower Eocene since this period is marked by a considerable subsidence and deeper waters in the Mississippi Gulf, which, however, eventually \ A 03, | = aS 7 oo c rake ~ cs , / 1 et GCap ~ F > 4 ra “? © es fat - ye Cn s ~~ c , 7 : * aS ) pak, P 4 om WS \ Ne ‘\ A < 8 if e : : BIAS y fs ig > ™ x : ” Cal = - - or - - - kia ia aac Fic. 2. Sketch map showing the approximate relation of land to water in the Middle Eocene. Stars indicate fossil plant localities, diagonal lining | indicates submerged areas. or became shallower again and duplicated in a measure the Lower Eocene conditions. a At a number of localities in Georgia and at two or three in nora ern Mississippi and in Arkansas representatives of the Middle Eocene flora have been collected. In Georgia where the plants are associated with shallow water and estuarine invertebrates I found 4 ATLANTIC AND GULF COASTAL PLAIN. 309 the remains of a typical mangrove flora associated with types which today characterize the tropical and subtropical beach jungle. This ora includes an Acrostichum closely allied to the modern Acro- stichum aureum Linné which is such an abundant fern in the man- ve and nipa tidal swamps. Other genera represented by fossil ns are Conocarpus, Dodonea, Ficus, Malapenna, Pisonia, Momisia, Rhizophora, Sapindus, Terminalia, and palms of the genus Thrinax. Botanists familiar with the flora of the torrid zone will ecognize at once that this is a typical strand flora of the tropics ich might almost have been taken bodily from Schimper’s classic ‘Indomalayan Strand Flora, or which can be seen today along the rida Keys and in the West Indies. The plants of this age from Mississippi and Arkansas do not indicate such a well marked ecological group nor quite such high mperatures as those from Georgia, nevertheless they also are a ely subtropical coastal types and embrace species of Sabal, Nt mnus, Panax, Ficus, Dryandroides, Persea, Sapindus, etc. One 9f the most interesting forms abundantly represented in north- Arkansas is a citraceous form with alate petioles which I ‘named Citrophyllum. Additional genera which are present are ‘ectandra and the coniferous genus Arthrotaxis. In Fig. 2 is shown the approximate position of the shore line g which the mangrove and the tropical beach flora migrated hward in the path of northerly flowing tropical ocean currents. Upper Eocene. No upper Eocene floras are known from the coastal plain but is believed that future discovery will reveal their presence when area where they are likely to occur shall have been examined detail. _ Lower OLIGOCENE. The Lower Oligocene has yielded no plants except petrified agments of the wood of palms and dicotyledons. The sediments € more or less impure marine limestones, and if marginal deposits with plants were laid down they were subsequently destroyed by rosion, or have not yet been discovered. 310 | BERRY TERTIARY FLORAS OF THE Extensive marine faunas indicate even more torrid conditions than in the preceding epoch, uniformly distributed over this whole area. MIDDLE OLIGOCENE. The Middle Oligocene deposits are those of shallow tropicz waters with a bottom temperature of at least 39° C. (70° F.), mz r toward the east with true reef corals in Georgia, but becom: brackish or fresh toward the west, by reason of their shallown and the increased volume of fresh water from the Oligocene as sippi and Tennessee rivers and other streams. The flora is but includes tropical swamp types, the fern genus Acrostic by pel the most abundant form collected. = The accompanying sketch map (Fig. 3) shows in a tec way the relation of land and water in the Middle and Upper O cene. It is to be noted that the great Mississippi Gulf sas reduced to a very wide and shallow reentrant. UPPER OLIGOCENE. Toward the close of the Oligocene a widespread emergence the land was inaugurated accompanied by a slight lowering of peratures. The floras are not abundant but are represen : western Florida and central Mississippi. They. contain very abut ant remains of several species of Sabal-like palms; the large le of a species of Artocarpus or breadfruit; leaves of figs; of th ‘Cin- namomum or camphor tree; representatives of the genera A Bumelia, Diospyros, Pisonia, Gyminda, Gleditsia, Nect Sapotacites, Rhamnus, Ulmus, etc.—the latter being the only which is a strictly temperate type in the modern flora, although of the genera enumerated have representatives in the warmer of the temperate zone at the present time. MIOCENE. A long interval followed the close of the Oligocene, during the coast line of southeastern North America was considerabl ward from its present position, in consequence of which 4 ro1t.] ATLANTIC AND GULF COASTAL PLAIN. 311 Fic. 3. Sketch map showing the relation of land to water in the Middle ' and Upper Oligocene. Stars indicate fossil plant localities, diagonal lining _ indicates submerged areas. _ of this age are unknown. This interval comprises the first half of _ the Miocene age and when renewed submergence furnishes us _ with a record we find very different conditions from those pre- viously enumerated. Either because of the diversion of the gulf _ stream to the eastward due to the emergence of peninsular Florida ; _ or as a result of changes in depth off the Hatteras anticline, a cool _ inshore current seems to have swept southward along the coast and _ through the Suwannee Strait across northern peninsular Florida, i ‘carrying with it a northern marine fauna which replaced the tropical _ fauna that had previously occupied this region. [April 21, 2 BERRY—TERTIARY FLORAS OF THE a 31 The fossil plants of this age are unfortunately rare and are as yet unknown south of the Maryland-Virginia area. The accom- panying sketch map (Fig. 4) shows in a generalized way the upper _ Miocene conditions after the resubmergence of the area, the maxi- mum emergence during the lower Miocene being unknown. The ue Fic. 4. Sketch map showing the approximate relation of land to wale 2. ae in the Upper Miocene. Stars indicate fossil plant localities, diagonal indicates submerged areas. the supposed directions of the ocean currents. x The known fossil plants from the Atlantic coast Mioe ne, exclusive of diatoms, include the following species from the M ry- land area near Washington described by Hollick:* Be: * Hollick, Md. Geol. Surv., Miocene, 1904, pp. 483-486, tf. a—b. rort.] ATLANTIC AND GULF COASTAL PLAIN. 313 Quercus Lehmanni Holl. Ulmus basicordata Holl. Cesalpinia ovalifolia Holl. Rhus Milleri Holl. Pieris scrobiculata Holl. Phyllites sp., Holl. In addition to the above the following have been described from € same horizon at Richmond, Va., by Berry :* Salvinia formosa Heer? Taxodium distichum miocenum Heer. Salix Raeana Heer. Carpinus grandis Unger. Quercus calvertonensis Berry. Rhus Milleri Holl. Planera Ungeri Ettings. Ficus richmondensis Berry. Platanus aceroides Goeppert. Podogonium? virginianum Berry. Dalbergia calvertonensis Berry. Celastrus Bruckmanni Al. Br. Nyssa gracilis Berry. Fraxinus richmondensis Berry. These plants indicate that the coast was low, which explains the absence of any but the finest terrigenous materials in the shallow _ water deposits which constitute the Calvert formation. The flora from Virginia indicates the presence of extensive cypress swamps, e latter type of plant being the most abundant fossil collected and other plants identified being for the most part similar in their physiological demands upon their environment. The flora from ryland is the natural counterpart of that from Virginia in con- taining several typical elements of just the sort of a plant associa- tion found on sands (inner beaches and more or less stationary dunes) along the present coasts in the temperate zone. __ _ Regarding age the plants are clearly Middle Miocene according European standards. They indicate less conclusively the climatic * Berry, Journ. Geol., Vol. 17, 1909, pp. 19-30, tf. 1-11. 314 BERRY—TERTIARY FLORAS OF THE conditions which prevailed along the Miocene coast in this latitude. There is considerable evidence of a scant rainfall, that is to say of less than 30 inches anrfually but this may well have been merely coastal condition. Indirectly the lack of land derived sediments in the deposits points to the conclusion that relatively dry conditions extended over wider areas. The mean annual temperature is di cult to determine. Several of the closely allied modern plants such as the existing bald cypress do not extend north of Maryland in the existing flora, while Ficus does not fruit north of Virginia, whi also marks the northern limit of Planera. However, the Miocene forms enumerated are all different specifically from the existi1 members of their respective genera and the conclusion is reached that the Calvert flora would grow under the climatic conditions pre- — vailing at the present time between Sandy Hook, N. J., and Cape — Henry, Va., and that the mean annual temperature which they indi- cate is between 50° and 55° F. | PLIOCENE. Pliocene floras have been unknown from North America until last year when deposits of this age with abundant fossil plants we discovered in southern Alabama. The most remarkable form this flora is the fruit of Trapa, the water nut, which Raimann Engler and Prantl segregates from the family Onagracee to f the family Hydrocaryacee. In the existing flora this genus has three species of southern Europe and southeastern Asia but it well known in the older Tertiary of North America and Ew and in the later fossil floras of Europe. Another interesting s; in this Alabama Pliocene flora is a species of Glyptostrob coniferous genus allied to our bald cypress which is now co: to eastern Asia, but which appears to have been cosmopoli Tertiary times. Other elements of this flora are abundant live (Quercus) ; several species of elm (Ulmus) ; abundant twigs, se and cone scales of a species of cypress which is very close existing bald cypress (Tarodium). Additional elements are §; of Nyssa, Hicoria, Planera, Betula, Dioscorea, Prunus, Pinus, This flora is quite modern in its facies and is a mixture of swar —igtt.] ATLANTIC AND GULF COASTAL PLAIN. 315 _ types and those of live-oak barrens. Among existing localities which I have visited which impress me as duplicating the climatic and other physical conditions indicated by this late Pliocene flora are the estu- aries along the gulf coast of Alabama and western Florida, among which Apalachicola, Mobile, Perdido and Pensacola bays are the larger. The Santa Rosa peninsula which separates the latter from the Gulf of Mexico supports a flora that is very similar to this Pliocene flora and one or two of the species represented in both are closely allied and may even be identical. PLEISTOCENE. Pleistocene plants are also common throughout most of the _ coastal plain and when they shall have been thoroughly studied they _will yield a large body of exact facts which will throw much light _ upon the immediate ancestry and migrations of our existing flora. Already more than one hundred species have been recorded, most of which are still existing and these indicate a very different geo- _ graphical distribution from that of the present coastal plain flora. CONCLUSION. _ I have only had time in the foregoing remarks for a very frag- mentary and incomplete sketch of the present study which has really only just commenced. With the complete exploration of the area _ and the additional collections which it is hoped to make it is believed that the combined results of the speakers studies of the fossil floras and those of his associates on the fossil faunas and the areal geol- ogy will furnish a basis for reconstructing the physical, faunal and floral history of the southern states, during the several millions of years from the Cretaceous to the present, which will constitute a _ lasting contribution to the history of the earth. AN OPTICAL PHENOMENON. By FRANCIS E. NIPHER. (Read April 21, 1911.) In 1871 in a letter to Tyndall, Joseph LeConte gave an interest- ing discussion of an ocular illusion which had been previously — described. Tyndall communicated it to the Philosophical Magazine — (XLI., p. 266). The phenomenon was observed in the manner here | described : eos Pierce a card with a pin. Hold it before the eye at a distance — of four to six inches, looking. through the hole at a bright back- ground. Place the pin in front of the eye with the head central ce in the pupil and in close proximity. The pin head will be “ seen in the hole,” and in an inverted position. As was pointed out by LeConte, this is not an ee image but a shadow. As proof of this he cites the fact that if a series ot holes are made in the card, a similar appearance of the pin head 1s seen in each hole. He adopted the idea that objects are seen erect, because the nerve fibers at the lowest point on the image see the tof of the object in the direction along which those rays have com He also argued that the inverted appearance of this shadow, whi was erect on the retina, was in harmony with this explanation. 4 The well-known fact that this point in the image is the vertex of a cone of rays, whose base is the pupil of the eye, and that diverging bundle of rays, when traced outward, does not define | position of any external point, is sufficient explanation of the that this line of reasoning has not been generally adopted. — dently the fact that there are no rays has also been taken into sideration. It does not seem quite evident that nerve fibers at lower point of the image on which ether waves collapse and de their impulses could “see” that these waves had their origin definite point, at the top of an object, at a definite distance from refracting media, in which the radii of curvature of these w were reversed in direction. And these waves from this point on tl object are involved in a summation of waves 5 from other and adj ing points. 316 ort. j NIPHER—AN OPTICAL PHENOMENON. 317 Many observers have doubtless had experiences like those which the writer had years ago while doing survey work. Two transit instruments were available, one of which showed the object viewed in erect, and the other in inverted position. A few days of use of either instrument enables the observer to give proper signals to the rod-man in a perfectly automatic way. After having thus become alternately educated, an attempt to use these instruments at random for brief intervals, relying wholly on what he sees through the instru- ment for the information which is to guide him in making his sig- nals, leads to the most helpless confusion. The observer even seeks to find his way out of his difficulties by comparing what he sees _ through the instrument with the impression received by a direct view. Such experience as this appears to justify the conclusion that we see external objects as we have learned how to see them, by help of our other senses. Even then it is a matter of never-ending wonder that we have in our possession certain nerve-fibers that can be trained to see. There are many interesting features of the phenomenon which LeConte discussed which appear to have escaped his attention. His claim that the sharp outline of the pin head seen in the hole could not be an optical image, since such an image would be so much out of focus as to be invisible, is justified to this extent. The object is in fact also visible in its real position in shadowy outline. It appears transparent, and the inverted shadow of the pin head is mentally _ projected outward and appears to be visible through the object itself. Every detail of the letters on a printed page is visible through this enlarged and transparent appearance which the object itself presents, due to an out-of-focus image on the retina. _ The sharpness of outline of the shadow decreases as the hole is made greater in area. This is due to penumbral effects. A black card gives more sharply defined results than a white one. A tube having the pierced card at one end and the pin head at the other may be applied to the eye, in such a way as to cut off all side light. The _ head may be covered with a black cloth, which is also wrapped around the tube. The shadows are then as sharply defined as an optical image could be. If the black sateen cloth be thrown over the head, and the eyes be directed towards a bright sky, a multitude of cir- 318 NIPHER—AN OPTICAL PHENOMENON. [April 21, cular images like pin hole images will be seen between the crossed fibers. Some of these are due to the right eye and some to the left. A pin head in front of either eye will show multitudes of inverted pin head shadows. . A circular disk of white paper having a diameter of I mm. or | slightly less, mounted upon a black card will also have upon ita sharply defined black shadow of the pin head, if the side facing the __ observer is illuminated. The paper disk must be near enough to the eye so that its image on the retina is out of focus, as in other _ cases where the pin hole is used. At various points on the glowing _ end of a cigar, when observation is made in a darkened room, similar — 3 shadows may be observed. A small blot of ink on a sheet of white paper will yield a white shadow of the pin head. The same result is given by a hole in a white card, if the card is illuminated and ~ observation is made through the hole at a dark background. fr If the reflected image of the full moon or of a bright star front — the convex surface of a lens be used instead of the pin hole in a card, the inverted shadow will be observed. . If the reflecting surface is concave, the shadow will appear erect if the eye is placed between the reflector and its principal focus. If the eye and pin are in the divergent beam beyond the principal focus, the shadow of the ae head will appear inverted. It is evident that when the shadow on the retina is erect, it oneen inverted, and vice versa. ; The eye lens and retina may be replaced by a convex lens anda __ paper screen upon which an image of the moon may be cast. A pin closely in front of the lens will show no shadow. If another conve le front of or behind the screen, according to the position of the second lens. The shadow of the pin will then appear. The capacity for accommodation of this artificial eye is unis ited, and the second lens may be dispensed with. The screen being placed between the lens and the image, the shadow of the pin wi appear erect on the screen. When placed beyond the image, it will appear inverted. ; a If an opera glass be focused on a street lamp 50 meters away a torr] NIPHER—AN OPTICAL PHENOMENON. 319 pin head between the eye and the eye-lens will produce no shadow on the retina. If the glass be focused for a nearer object, an erect shadow will appear. If focused for a more distant object, the shadow will appear inverted. A hole through a card and with a bright background may be viewed by means of the opera glass. The hole may have any diameter from 0.05 to 1.5 cm. The distance of the card must be adapted to the diameter of the hole, and may vary from close contact with the object lens to three or four meters, the glass being focused for a more distant object. The results are as indicated above. The setting sun surrounded by bright clouds may be used as an object, if viewed through the foliage of trees thirty or forty meters distant, the glass being focused for an object more distant than the trees. The mass of foliage will be dotted with pin head shadows. Each opening through the leaves acts in a manner similar to the pin hole. In all of the cases described, the shadow upon the retina is by some mental act projected outward in space. An interesting ques- tion arises concerning its apparent position. LeConte says that in his experiments it appears in the hole in the card. Perhaps it would be proper to say that it is seen through the hole. The hole itself may have a diameter of about one third that of the pin head, and the pin head then appears smaller than the hole. Its apparent size depends somewhat on the diameter of the hole. If a pin is placed back of the card and in erect position so that it is visible through the hole, it may be so placed that it has the same apparent size as the shadow. If the pin is at a distance of 30 cm. from the eye, and the card is at a distance of 15 cm., the shadow and the pin will have the same apparent size. The appearance of the inverted shadow and the erect pin is as shown in Fig. 1. PG 1. This suggests an interesting device whereby the line of sight of the two eyes and the capacity for muscular adjustment may be exam- 320 NIPHER—AN OPTICAL PHENOMENON. [April 21, ined. Pierce a card with two pin holes, at such a distance from each other that when placed at half the distance of distinct vision from the eyes, they may be seen as one. This can be done by drawing lines across the ruled lines of a page of white paper, and crossing the ruled lines symmetrically so that at the top of the page the lines are farther apart and at the bottom they are nearer together than the two eyes. Pierce pin holes at each intersection of the ruled lines with the cross lines. If held in front of the eyes so that the cross lines are seen double, the two inner images of the lines will appear to cross. At this distance apart thus determined two holes will appear as one. Place a card having holes thus placed in front of the eyes. Mount two pins in front of the pupils so that the two shadows appear superposed in the superposed images of the holes. Two pins may now be placed back of the card so that when viewed through the holes they will also appear superposed. The two holes and the four pins will then present the appearance shown in Fig. 1. This ar- rangement locates two points along the line of sight of each eye. The holes may be in separate cards which close the ends of two tubes, through which the observations are made. These tubes, together with the pins, should be capable of screw adjustments. When the pin hole is viewed through a tube which is lined with dark paper, the card serving to close the outer end of the tube, it may be used for an examination of certain imperfections in the eye. For example, in my own case one eye shows a minute hole with a bright background to be of uniform appearance. Viewed by the other eye a rather sharply defined shadow is shown in the center of the hole. This is due to a slight irregularity in the curvature of the outer sur- a face of the cornea. This is due to a grain of gunpowder which was _ blown into the eye from a horse-pistol which was discharged from a _ distance of about 35 cm., into the lower part of the face, about fifty years ago. The grain of powder was visible for many years, * but has been gradually absorbed. A slight distortion of closely ruled — parallel lines indicates that an irregularity of the surface still per- sists. The shadow seen in the pin hole shows that light is not uni- : 2 formly spread over the retina when a slightly divergent beam of light enters the pupil. Any opacity in the crystalline lens would also produce a shadow upon the retina. Lye PROCEEDINGS AMERICAN PHILOSOPHICAL SOCIETY HELD AT PHILADELPHIA FOR PROMOTING USEFUL KNOWLEDGE Vou. L Juty—AucustT, 1911 No. 200 SYMPOSIUM. : THE MODERN THEORY OF ELECTRICITY AND MATTER. I. THE GENERAL PRINCIPLES. By DANIEL F. COMSTOCK. (Read April 22, 1911.) The field of the present discussion is so large and the time for . so limited that I feel sure I can serve you best by foregoing the luxury of an historical introduction and by entering somewhat ptly into the heart of the subject before us. I wish, then, to ss before you the general ideas and beliefs respecting the ulti- nature and relations of matter and electricity which are in the oreground at the present time. In dealing with progress of scientific explanation it is necessary 9 remember, what we too often forget, that the verb “to explain,” when applied to a new complex phenomenon, means merety the ex- pressing of it in terms of something else either more familiar or more fundamental. An exaggerated example of the first type is nished by all the old anthropomorphic explanations of natural PROC. AMER. PHIL. SOC., L. 200 U, PRINTED JULY 31, IQITI. 321 322 COMSTOCK—THE MODERN THEORY _ [April 22, phenomena in which the less familiar actions of the outside world were expressed in terms of the intimate and much more familiar workings of the human mind. “ Nature abhors a vacuum,” and like expressions, show the type. a The progress of science exhibits countless examples of ‘the second type of explanation, for, wherever two or more concepts are merged into a profounder synthesis, there we have an expression in terms of something more fundamental. When, for instance, it is said that the phenomena of tidal action are caused by the gravita- tional attraction of the moon, it is stated merely that this action is really one with countless other phenomena which, although dis- similar on the surface, merge with it into the profounder synthesis known as the law of gravitation. It is important also to remember in this connection, that in this process of explanation we always have left the one concept into which the many have merged, so that as time goes on the alterna- tives of explanation grow fewer and fewer, and in the end—if we can imagine an end—there is no explanation, because there is no more fundamental fact. eis You will pardon me, I am sure, if I say one word more with reference to this question of ultimate explanation. : We have all heard people say, “Isn’t it wonderful that so rome is known about electricity, and yet no one knows what electrici is!” Now, doubtless the observation has some meaning, but cer-— tainly not as much as it seems to have; for, after all, what do th mean by “ what electricity is”? Do they expect the announce: ~ some day that electricity is a liquid similar to water, or a gas sim to air? : It is becoming more and more probable that electricity is | chief constituent of the atoms themselves, and an electron, whi a particle of electricity, if anything is, is certainly less than a ter thousandth the size of one of the atoms in a water molecule. The fore after the “is,” following the word “electricity,” there is not ing to put which is already familiar, and when the profounder cept does come, it will be extremely fundamental, but it will ca the layman no thrill of long-anticipated disclosure. | A special type of critical attitude is necessary in dealing witl 1911.) OF ELECTRICITY AND MATTER, 323 fundamental physical concepts and it is an attitude which we seldom assume, so that these remarks have been necessary to introduce ‘properly the three fundamental realities which modern physical theory now contemplates, namely: the atom, the electron, and that ‘mysterious but perhaps even more fundamental entity known as “energy. A few years ago I would have mentioned also the ether, but I am a little reluctant to do so now. Not that there has been a sudden revolution in the realm of thought, resulting in the complete over- throw of the old regime, but rather that development has been such as to render the concept of an ether less and less impressive—one ‘might say—and less and less important. Changes of opinion in such matters are, it seems to me, partly questions of emphasis, and ‘radiant energy in all its nakedness is now usurping much that the ether has long stood for. Of course, the loss, or rather the dimness, of the ether concept implies a certain loss of concreteness; but, as has been said, con- ereteness in new concepts, founded as it is on familiarity, is a secondary virtue and is of far less importance than the value of a concept in furthering the great process of induction which leads us _to more and more general truths. Although the concept of the ether is slowly dimming, the concept of the atom is becoming more and more definite and vivid. The study of the scintillations caused by radium rays and the work of Rutherford in counting the alpha particles give us, for the first time in the history of physics, definite observable results which apparently can only be due to the action of single atoms, and which therefore furnish proof beyond reasonable doubt that the atom and molecule are names of actual realities, and are not merely two prominent words in the statement of a useful hypothesis. ‘The kinetic theory of matter, carrying with it the concept of temperature as violence of atomic vibration, has also been strength- ened enormously by the work of Perrin and Einstein on the Brown- jan movement. They find that microscopic particles in solutions have a perpetual motion in close agreement with the kinetic theory. Indeed, they act in all ways like big molecules, obeying the kinetic laws deducible from mechanics. Their observable agitation is part 324 COMSTOCK—THE MODERN THEORY [April 22, of the general vibratory motion which distinguishes a hot body from a cold one. a The electron, too, is now well within what we might call “the exclusive circle of the truly real.” This minute charged body has by many researches, among them the recent one of Millikan, been — shown to be a definite reality, present in all matter and entirely or largely responsible for all the phenomena we know as electrical. __ The other fundamental entity, energy, is also, if I may be allowed the phrase, “mysteriously real.” Radiant energy leaves the sun _ eight minutes before it reaches the earth, and must, therefore, exist during that time in the space between, dissociated from ordinary matter. When it finally strikes some object and is absorbed, it gives” the object a thrust—that is, communicates momentum to it, as a_ bullet would do—and at the same time, of course, it increases the — total energy of the body. In the same way, a body radiating energy recoils during the emission in a way similar to a gun. All this is remarkably like the action of ordinary matter. We can, however, say even more. There is very good reason for be-- lieving that, were it possible to shut up a large amount of radiant energy in a hollow box, the inner surfaces of which were perfectly reflecting, so that the rays would be reflected back and forth indefi- nitely, we would find this confined energy to possess both mass and weight. Not many decades ago such an idea would have seemed — absurd, but it is hard now to avoid the conclusion that such would be the case. a8 I can do no better now than to describe in a few words the pic- ture which we have today of the ultimate structure of matter. A piece of matter is composed of particles called atoms, which, by uniting in groups in various ways, form the characteristic ag; gates which we know as molecules. There are about one hund e different kinds of atoms, varying in relative weights from one to 24 and in relative volumes from one to about 16. The approxin diameter of an atom is one one-hundred-millionth of a centi . These atoms are in ceaseless motion to and fro, the energy 0 this motion determining what we call the temperature of the bod Within the atoms and in the spaces between them are large num of very much smaller particles known as electrons. They "4 : E c f | t E. gt.) OF ELECTRICITY AND MATTER. 325 carry a negative charge, which is relatively enormous, considering _ the fact that the approximate electronic diameter can scarcely be more than one one-hundred-thousandth that of the atom. When any cause sets up a general movement of the electrons within a body, we have a current of electricity, while the random _ vibratory heat motion of atom and electron is the cause of continual _ emission of radiant energy to other objects or to outside space. A piece of matter is thus a complex system composed of an in- conceivably large number of ultimate units, atoms and electrons, in ceaseless motion to and fro, and permeating all is the mysterious, _matter-like entity which we have called radiant energy, and which ever seeks to escape with an enormous, though perfectly definite, velocity into the space outside. Some of it succeeds in escaping, but there appears to be a vast amount which in some way is im- -prisoned in the atom-electron aggregate and thus never becomes “radiant” in the strict sense of the word, though it resembles fadiant energy so closely as almost to warrant the same name. Having thus obtained an impressionistic view of the structure of a piece of matter, I would like to call attention to the properties of the space surrounding an electron, or, what is the same thing for our present purpose, the space surrounding any body, say a small _ sphere, possessing an electric charge. This space is the seat of what we call electric force, and is known as an “electric field.” Now, it is a well-known conclusion and one which cannot at present be in any way avoided, that the energy which the body possesses, by virtue _of its charge, that is, the energy originally required to charge it, _ resides in the electric field around the body, and not on or in the body itself. The existence, without apparent motion, of this energy in what seems to be empty space is very remarkable, but the con- clusion that it is there is unavoidable, and after all there is no great _ difference between the discarnate state of this energy and the state of the radiant energy of the sun on its way to the earth. From _ the older point of view this energy in the electric field was “ strain energy” in the ether, the so-called strain being similar to what we would get in an immense block of rubber if a pin-head embedded at . its center were to swell into the size of an egg. The rubber would _ be pushed back in all directions and the energy of this compression 326 COMSTOCK—THE MODERN THEORY [April 22, would be stored up throughout the whole mass, a great deal of it near the center, but an appreciable amount even out at the very limits of the block. From the present point of view, we think more _ about the energy and less about the ether, but the general isi ee the same. a) Let us call this energy located in the space surrounding an electric charge “bound energy,” to distinguish it from the closely similar type of self-propagating energy which can also exist in space and to which we apply the term “ radiant,” and let us then ask what properties, if any, this bound energy gives to the body. which it surrounds. I stated before that radiant energy, when it struck a body, com municated momentum to it in the same general way as a material | 4 body, say a flying bullet, would do. The radiant energy thus acts as if it had mass, and the question now is, “ May the ‘bound energy’ surrounding our charged sphere also be considered to _ possess mass?” We may answer this question in the affirmative, for this bound energy is electric energy, and, thanks to ‘the great founders of electrical science, we know the laws of electric action a pretty completely. Applying these well-known laws we find that when our Jindal sphere is moved, it acts like a current of elecricity and sets up a magnetic field about it, and the formation of this field acts, by the — well-known laws of induction, to retard the motion of the charg Thus the setting in motion of the sphere is made more difficult by reason of its charge, an effect which is equivalent to “yr tha the sphere has added mass. If this were all, we might say that the added charge of tricity had mass, so that the mass increment is on the surface | the sphere where the charge is known to reside. This is mot a however, for by letting the sphere expand we can decrease the energy in the space about it without changing the magnitude of its: charge; and we find, by further simple application of well-known electrical laws, that the new mass will change exactly as the energy changes. If the new mass had been proportional to the charge, it would remain constant with it instead of changing with the energy Thus what is known as the electromagnetic mass of the sphere is tgtt.] OF ELECTRICITY AND MATTER. 327 proportional to its energy. I was able to show, several years ago, that, with certain limitations, this same result is true for any electric system. : It may be said, then, to follow from what might be called ele- _ mentary electromagnetic principles, that the electromagnetic mass of an electric system is proportional to its electric energy. Now _Hasenhorl, at Plancks’s suggestion, I believe, had already shown that a similar result applied to radiant energy properly speaking; for he found that from known laws of radiation it followed that a hollow box, like that mentioned earlier, with perfectly reflecting _ walls, would, when filled with radiant energy, act as if it had added mass. That is, the pressure of radiation would be so changed by the increasing velocity of the box as to oppose the force causing this increase, and this inertia, this added mass, he found would be proportional to the amount of energy present. My proportionality constant agrees with his. Lewis, from a totally different point of view, has reached a similar conclusion. : It is, as you know, one of the profoundest generalizations in modern physics that light and other forms of radiant energy are in reality all forms of electromagnetic energy. Hasenhorl’s results, therefore, that confined radiant energy possesses mass, combined _ with the result obtained in connection with the bound energy sur- _ rounding electric charges, gives us the general result that all elec- : tromagnetic energy, whether bound or radiant, possesses mass, and this mass is proportional to the quantity of energy present. You see that the concept of energy, although in some ways very illusive, is getting singularly definite and persistent. Since we see that electric mass is proportional to electric energy, the question naturally arises: How much of the mass of the electron is due to the electric energy surrounding its relatively enormous charge, and _ how much is the “ordinary mechanical mass” of its body proper? We have a means of distinguishing between the two masses, for _ the electric mass does not remain constant when the velocity of the charge becomes great. Electrical laws tell us that it increases, very slowly at first, then more rapidly, and that as the velocity of light S approached it becomes very great. Of course, in ordinary me- 328 COMSTOCK—THE MODERN THEORY [April 22, chanics, on the other hand, the mass of a body is considered to be a constant and to have nothing whatever to do with the velocity. a By studying experimentally the deflection of the beta-rays of radium, which consist of streams of electrons travelling at veloci- ties very near that of light, Kaufmann has shown that the experi- mental change in mass fits the mathematically deduced change when, and only when, the “ ordinary mechanical mass” is negligible. g In other words, as near as measurement can yet go, the mass of the © q electron is entirely the electromagnetic mass of the surrounding — energy, and it has no appreciable mass of what I might call “the _ old-fashioned mechanical kind.” . This is a result of extraordinary importance in physical theory, — for it immediately suggests the general question, “Is all mass of this — origin?” Since an ordinary piece of matter is permeated with electrons and also with the radiant energy which all parts of it are — constantly absorbing and emitting, it is an absolutely unavoidable — conclusion that at least part of the total mass is of this electro- magnetic type. But the question is, “Is all mass electromagnetic?” q or, more properly speaking, “Js all mass of the same type, and does — it all depend upon the velocity in this same way?” An affirma- tive answer would imply a profounder unity in physical phenomena than has hitherto been recognized and would thus correspond to’ passage to a deeper synthesis. Of course, the deeper concept which unites two or more others should, in strictness, be made independen of these others ; but, although definitely foreshadowed in the pres case, the detailed statement of this deeper law is at present i possible except as regards changes due to motion; so that th taken with the fact that electromagnetic phenomena are so fami that we may be said to know their modus operandi in terms Ke magnetic fields, electric forces, and the like, renders it provisio allowable to state the question in the form: “ Are the laws of tricity and optics the laws of matter in general?” , We have, during the last few years, been attaining with gr and greater surety a definite answer to this question. It has through the gradual adoption of a remarkable concept, profoun its meaning and very far-reaching in its consequences. I refer the so-called “ Principle of Relativity.” tgtt.) OF ELECTRICITY AND MATTER. 329 There are times in the history of science when various contra- dictions require that the process of building stop until the most fundamental concepts are reéxamined. This was true in the his- tory of astronomy at the time of Copernicus. The prevailing con- ception of the earth as a fixed center about which all the other bodies revolve had been practically sufficient for a long time, but gradually difficulty after difficulty arose until it was no longer possible to patch up the old theory to meet the accumulation of stubborn facts. Only by a change in the most fundamental con- ception, namely, that of the earth as a fixed center, could harmony be brought out of chaos and a new period of development com- menced. We seem to be passing through a somewhat similar period in physics, and the “Principle of Relativity” contains the modified concepts. _ By way of transition, let me make one or two statements at this point about electric and magnetic systems in general. It can readily be shown to follow from known electromagnetic laws that two electric charges of the same sign moving side by side with the same velocity repel each other Jess than when the two are stationary. This is due largely to the fact that, when moving, each charge is surrounded by a magnetic field and this magnetic field introduces new forces. _ This simple statement introduces a far more general one, for it may be shown that a steady motion of any electromagnetic system so changes the force between the various parts of the system that it tends to take up a new position of equilibrium. The forces are such that the whole system tends to contract along the line of motion. Ti it be allowed so to contract until it reaches this new position of equilibrium, then everything will be as before the system was set in ‘motion, with two important exceptions, if the system has any in- ternal motions caused by electromagnetic force. First, two mo- tions, say the oscillation of two charges, one in the front of the System and the other behind, which in a stationary system take place simultaneously, will, in the moving system, take place not ‘quite simultaneously ; for the forward one will be somewhat behind in time ; and, second, all such motions will take place a little slower than they did in the stationary system. The term electromagnetic 330 COMSTOCK—THE MODERN THEORY —_ [April 22, system, used in this sense, includes, of course, all radiant energy, — for it will be remembered that this type of energy is also: electro- — magnetic. : . Thus, on grounds of well-established electromagnetic theory int . without any new fundamental conceptions whatever, we can make — a general statement that any electromagnetic system, when set in motion, tends to assume a new state of equilibrium, and if this — change be allowed to take place, then all effects, electric and optical, take place in the changed system in a manner exactly corresponding © to the way they did take place before the system was set in motion. A moment’s thought will make it evident that we are here ap- parently in the presence of a fundamental lack of harmony between electromagnetic phenomena and the phenomena connected wi matter in general; for, from the ordinary point of view, the parts of a “rigid body” maintain their mutual relations unaltered whether or not it is set in motion, while, as we have seen, this is not true electromagnetic systems. Now, we have no choice but to consider a real body as a com- bination made up of the two kinds of systems if there are two, the electromagnetic type and the “rigid body” type; hence it is clear that, when such a mixed system is set in motion, there will be a considerable amount of what we might term internal discord, owing to the conflicting tendency of the two types. It would be very easy to distinguish such a mixed moving system from the same syste at rest, because its parts would bear quite different relations to ez other than they did before. Setting a real system in motion w be like heating an object made up of two substances having diffe t coefficients of expansion. The parts of such a system would bear totally different relations to each other than when the wh thing was at rest, and the internal dissension would increase the velocity and would depend on its direction. . Now as a matter of fact, we all live on such a moving system, a system which is going around the sun with a velocity ¢ nearly twenty miles a second; and yet, although many caref planned and executed experiments have been carried out to de differences in the actions of various electrical and optical syst according as they are made to face with the earth’s velocity sort] OF ELECTRICITY AND MATTER. 331 across it, every one of them has given negative results, and has thus shown that the relative parts of the system bear the same rela- tion to each other, no matter what the direction of motion is. It is like finding that an object made up of metal and glass had no strains set up in it when heated, a result which could only be _ attained if the metal and glass had the same coefficient of expansion. ; e What are we to conclude in the case before us? There seems to be no alternative. We already have seen that electrons are among the fundamental constituents of all atoms; we have seen that radiant energy is electromagnetic and that such energy permeates all matter. We have seen that energy resembles matter in possessing ‘ s, and that, therefore, to the same degree, matter resembles gy. The necessary conclusion seems to be that all physical phe- _ nomena obey the same general laws of which the known electro- “magnetic laws are as yet the completest expression. And now to what have we committed ourselves by this con- clusion as regards changes due to motion? Merely this: that all systems being ultimately “electromagnetic” in the above sense, : dergo certain changés when set in motion, but these changes are such as to leave all parts bearing the same relation to each other. _ Thus since the knowledge of an observer travelling with the system only relative, he is not able to detect such absolute changes, just we are not able to detect the motion of the earth. The changes in his system would be noticeable to an observer whose instruments did not move, but cannot be detected by moving instruments. __ The kind of change which we have said is produced in a system by setting it in motion has one property which is important and pro- foundly significant. It is that the moving observer sees precisely same change in stationary systems which he is passing as the Ste tionary observer sees in the moving system; so that not only can the moving observer not detect his motion by means of his instru- ments, but the two observers together, if their memory fail, can by © means tell which is moving. There is, in other words, a very complete symmetry with regard to what the two observers can actually find out about their systems, although we called one of them stationary i in the beginning. Because of this complete symmetry the most conservative among 332 COMSTOCK—ELECTRICITY AND MATTER. [April us will admit that the concept of absolute motion need not be very often, at least, in the science of the future; if the foregoing views are true ones. The modern group of conceptions known as the “Principle of Relativity” teaches that the idea of absolut motion is entirely superfluous, and that the time honored concepts | space and time, as independent of all motions, do not perio the real universe and should be modified. Modified in what way? Real time is measured by as cle and real distance is measured by real rigid bodies, and we find unexpected discord between moving clocks and stationary clocks and between moving rods and stationary ones. We have, therefore no possible use for what might be called “universal time.” — might form a vague concept of a “cosmic second” pervading universe, but we could do nothing with it, and: it would theref be entirely artificial. Whenever we wished to think about an act moving object and wished to measure some vibration frequency it, let us say, we would have to use some actual clock-beat or periodic phenomenon as a unit. So that the actual universe — hopelessly in its grasp, and our concepts of space and time to valuable must be in harmony with the habits of real things. The principle of relativity, therefore, makes changes in fundamental concepts of space and time for moving systems second in a moving system is longer, the meter, in the dire motion, shorter, than in stationary systems. The units then « harmony with real happenings in such systems, and this m possible to introduce the last great synthesis of modern theory, deeper unity of physical law under the dominance of what we known as electromagnetic principles; and this brings us o nearer to the last, ultimate generalization which is the une ideal of science. MASSACHUSETTS INSTITUTE OF TECHNOLOGY, April 20, 1911. zt: RADIOACTIVITY. By BERTRAM B. BOLTWOOD. (Read April 22, 1911.) _The study of the discharge of electricity through gases and the _ properties of radioactive substances has done much to broaden our owledge of the relations of electricity and matter. It has served throw a new light on the ultimate constitution of matter itself, while confirming the older theory of a discontinuous or atomic icture, has led to the presumption that the chemical atom is not - divisible into still smaller entities, but that in some cases it can ergo a spontaneous disruption accompanied by the ejectment of ertain of its constituent parts at high velocities. All this has opened oad and attractive field for more or less legitimate speculation id conjecture. - Since the first recognition by Becquerel in 1897 of the radio- active phenomena exhibited by the element uranium, the extension of knowledge of the radioactive substances has steadily and pro- gressively advanced. This development has been due in great part 0 the early formulation of the theory of atomic disintegration, pro- | in 1902 by Rutherford and Soddy, which has served as a ematic foundation and has afforded an orderly basis for the terpretation of the otherwise somewhat complicated relations. According to this hypothesis the atoms of the primary radioactive ements are considered to undergo spontaneous disintegration and n this manner to give rise to a series of successive radioactive prod- _ ucts, differing from the parent substances as well as from one another in physical and chemical properties and in the relative stability of their atomic systems. Simultaneously with the disruption of the atoms certain characteristic radiations are emitted by the systems, 333 334 BOLTWOOD—RADIOACTIVITY., and it is these radiations which led to the discovery of the radio- elements and which particularly distinguishes them from all ot types of matter. Ce The characteristic radiations emitted by radioactive substz nees are three in number and are nis as the alpha, the beta and gamma rays. Rett For our knowledge of the alpha radiation we are indebted chi to the work of Rutherford and his associates, which has shown co clusively that these rays consist of discreet particles of atomic din sions, projected with high velocities and bearing a positive cha The earlier investigations were conducted with a view to deter ing the mass of the particles from the deflections suffered by rays in electric and magnetic fields of known strengths. A value the ratio of the charge to the mass (e/m) was obtained in this r ner and this led to the conclusion that, if the charge on a pa was the same as that carried by the hydrogen ion in electrolysis, mass of the particles was approximately twice that of the hydrog atom. Strong evidence was also obtained that the alpha parti emitted by the different radio-elements are identical in nature r spective of the character of the particular radio-element from» they originate. A very ingenious experiment was then devised by Ruther and Geiger, in which a known fraction of the alpha particles er by a radioactive source was allowed to enter a small ionization ber containing a gas at low pressure. Under the influence of a s electric field the ions formed by the entering particles acquit high a velocity that additional ions were produced by their ca with neutral gas molecules and the electrical effect was increa a point where the action of a single particle could be readily d It was shown, moreover, that each of the scintillations appea a screen of Sidot’s blende placed in the path of the rays corres to the impact of a single alpha particle. In this manner the actual number of alpha particles emits ( radioactive substance was accurately counted, and it was fou nd 1 one gram of radium itself emitted 3.4 X 10’ alpha particles second. By measuring the electric charge imparted to an insult BOLTWOOD—RADIOACTIVITY. 335 by the impact of a given number of alpha particles, the charge car- _ tied by each particle could be readily determined. This was found _ to be equal to 9.3 X 107° electrostatic units. It was then shown in ' am ingenious manner that this charge was twice that carried by an electron or by a hydrogen ion, although preceding determinations of _ the latter magnitude had indicated that its value was somewhat less than one half of 9.3 X 107°. The recent determinations made by Mil- likam of the charge on an ion have shown, however, that 4.65 X 107° is not far from correct and have confirmed the conclusion reached by Rutherford and Geiger that the charge on an alpha particle is equal to twice the unit charge of the hydrogen atom in electrolysis. _ With this modification, the ratio of the charge to the mass of an 5 alpha particle indicates that the mass is equivalent to an atomic weight of four. This corresponds to the mass of an atom of the gaseous element helium. _ The final proof of the intimate connection of the alpha particle with the helium atom was supplied by Rutherford and Royds who roved by spectroscopic methods that readily detectible amounts of helium could always be obtained when large numbers of alpha par- ticles were allowed to penetrate through a thin glass wall into a highly evacuated receptacle or into a screen of lead, from which the helium was ultimately released by fusion of the metal. It is a significant fact that although the alpha particles from the different types of radioactive matter appear to be all of a similar nature and to consist of. atoms of helium bearing a double, positive charge, the velocities at which they are ejected are different for the different radioactive substances. The velocity of the particles emitted by the atoms of any one type of matter undergoing transformation is, however, always the same and is a characteristic constant for that particular radio-element. Attention was first called to this impor- tant relation by Bragg and Kleeman and it is undoubtedly significant ‘its bearing on the constitution of the radio-atoms. The observed velocities of the particles appear to lie between the limits of 1.5 X 10° and 2.25 X 10° centimeters per second. Owing to their high velocities the alpha particles are capable of Passing through thin layers of ordinary matter, and can penetrate into air at atmospheric pressure for distances of from somewhat less 336 BOLTWOOD—RADIOACTIVITY. than three to about eight centimeters. They produce phosphor cence and chemical action in substances on which they impiage? 2 strongly ionize gases through which they pass. ote Through measurements of their deflection in electric anda nagneti fields the beta rays emitted by radioactive substances have been teal to consist of negatively electrified particles or electrons with apparent mass about one eighteen-hundredth that of the ee atom. The velocity with which they are projected is consider higher than that of the alpha particles and in some cases cana ir tenths the velocity of light. They are capable of penetrating thr moderate thicknesses of ordinary matter and for considerable tances in air. They cause phosphorescence and chemical action substances on which they fall and produce ions in gases a which they pass. : Although the beta particles emitted by the different epee | radioactive matter are in all respects identical in nature, they chib the same peculiarities with respect to the velocities with which are initially projected that has been Observed in the case of the < rays. The velocity of the beta rays from any given radio-ele is the same within certain limits for every disintegrating atom of t element, but is different from the velocity of the rays emitted other elements. The velocity of the beta rays is therefore chat teristic for each of the substances which give rise to this 1 radiation and has undoubtedly a significant bearing on the co tion of the radio-atoms. It appears probable from some experiments performed by Hahn and von Beyer, in which a mag spectrum of the beta rays emitted by certain radioactive subs was obtained, that the transformation of some of the radio-a accompanied by the expulsion of a series of beta particles of ent velocities. These results are very suggestive of the model devised by Sir J. J. Thomson in which the atom was asst ‘ed built up of concentric shells of electrons revolving wit . velocities in circular orbits. It would seem quite possible such a system a rearrangement of the parts might result in the sion of electrons from several layers simultaneously. The third type of radiation associated with radioactive t mations, known as the gamma rays, is similar to the X-rays — igtt.] BOLTWOOD—RADIOACTIVITY. 337 supposed to consist of electromagnetic pulses in the ether. A rather _ ingenious corpuscular theory as to the nature of these rays has been proposed by Bragg, but has not met with general acceptance owing _ to the fact that it appears to be not altogether in accord with the __ experimental evidence. _.___ The origin of gamma rays seems to be very intimately connected with the emission of beta particles, for the two types of radiation _ have been observed to appear simultaneously and bear a certain defi- c- nite relation to one another. When the expelled beta particle has a *% : “high velocity the gamma ray is of a very penetrating character, while if the velocity of the beta particle is low the gamma ray emitted has but little power of penetrating ordinary matter. The disintegration of the radio-elements takes place according to a very simple law and the number of atoms of any radio-element which undergo transformation in the unit time is a definite propor- tion of the total number of atoms initially present. The number of atoms N remaining unchanged after any time ¢ is given by N=WNoe-', where No is the initial number of atoms and AJ is the fraction undergoing transformation in the unit of time. This frac- tion has a fixed and invariable value for each separate radio-element -and for this reason is known as the constant of change for the ele- _ ment in question. It is a relatively simple matter to determine the period of time required under thése conditions for exactly half of any given quantity of a radio-element to be converted into other substances and this time is known as the half value period. The rate of change and the corresponding half value period is a definite characteristic for each of the radio-elements but is very different for the different radioactive substances. The half-value period of ura- ‘nium, for example, is over five billion years while the half-value period of certain other radio-elements is only a few seconds. Although the disintegration of some of the radio-elements has been examined over wide extremes of temperature and pressure, and __ under various other special conditions which would greatly influence ___ the course of ordinary chemical reactions, it has not been found pos- sible to definitely alter or effect the rate at which transformation _takes place to the slightest measurable degree. It is therefore evident PROC, AMER, PHIL, SOC., L. 200 V, PRINTED JULY 31, IQI1. 338 BOLTWOOD—RADIOACTIVITY, [April 22, that the disintegration of the radio-active substances is of a wholly different character from the ordinary chemical changes. This is exactly what would be expected if the radioactive processes occur within the atoms themselves, for, in accordance with our general theories, chemical forces appear to be restricted in their action to the exterior of the atomic systems only. at We have thus far considered only the laws which govern the transformation of radioactive matter and the radiations which | c= company the disintegration of the atoms; let us now turn our ai tion to the substances themselves. Investigation has brought to li; three main groups of radioactive elements—the uranium series, th thorium series and the alkali metals. Of the last mentioned knowledge is not very extensive. A type of beta radiation appeat to be emitted by the salts of potassium and rubidium but their to be considered as true radio-elements is not as yet entirely clear The uranium series, in addition to the parent substance, con ten products which may be properly considered as in the main of descent. These are uranium X, ionium, radium, radium en tion, radium A, radium B, radium C, radium D, radium E and r: F. Each of these products exhibits a characteristic chemical vior which is different from that of the parent element uranium. | half value period of uranium has already been mentioned as e | ing five billions of years and the disintegration of its atoms is a c . panied by the expulsion of alpha particles. Uranium X, ionium radium are solids, the two first having chemical properties simi 1 thorium, while radium has those of an alkali earth and particu! a resembles barium. Uranium X has a half value period of al twenty-four days, and it emits only beta and gamma rays. The’ of disintegration of ionium is not as yet known with any degr accuracy but it is certainly a relatively stable product and is t formed but slowly. Its half value period is probably of the ord ten thousand years. It emits alpha rays only. The half value p of radium is approximately two thousand years. Rutherford Geiger have deduced a somewhat lower value, namely 1,760 as a result of their experiments, but this value is probably an t estimate, as will be explained later in this paper. Radium alpha rays and probably very low velocity beta rays also. tort.) BOLTWOOD—RADIOACTIVITY. 339° _change following radium is a striking one for the product in this case is gaseous. It is known as the radium emanation and has the inert chemical character of the rarer gases of the atmosphere, helium, neon, argon, krypton and xenon. When the atoms of radium emana- tion undergo transformation, the succeeding product known as ra- dium A is formed. This is a solid and is deposited in the form of a thin coating of active matter on the walls of a vessel containing the _ emanation. This acquirement of activity by the surface of objects in contact with the emanation was observed some time before a satisfactory explanation of the phenomenon was suggested. It was. therefore known as “imparted” or “induced” activity. It is now called the active deposit. Radium A, which has a half value period of three minutes and emits alpha rays, is followed by radium B, which emits beta rays and is half transformed in about twenty-six minutes, and this in turn is succeeded by radium C with a half value. ‘period of about nineteen minutes. The transformation of radium C is accompanied by the expulsion of alpha, beta and gamma rays.: An interesting product known as radium D then ensues, its transfor- mation being characterized by the absence of any detectible radiation. whatever. A product of this sort is known as a rayless change and other examples to such a change occur in both the thorium and actinium series. On account of the similarity of its chemical prop- erties to those of ordinary lead, radium D is known as radio-lead. It undergoes transformation more slowly than the immediately pre- ceding products and has a half value period of about sixteen years. It is followed by radium E, a beta ray change, with a period of five days, and this is succeeded by radium F, otherwise known as polo- nium. Polonium emits alpha rays only and is half transformed in one hundred and forty-three days. __ In addition to the ionium-radium series, uranium is also the pro- -_genitor of another group of radio-elements of which actinium is the first and most stable member. The rate of change of actinium has ‘not yet been determined, but is comparatively slow and is probably of the same order as that of the radium. Actinium offers another example of the rayless changes which have already been referred to, and no radiations have been observed to accompany its trans- formation. 340 BOLTWOOD—RADIOACTIVITY. [April 22, Six products subsequent to actinium have thus far been identi- fied in this series. The first is radioactinium, an element having a half-value period of 194 days and emitting both alpha and beta par- ticles. The subsequent product is known as actinium X. The half — value period of actinium X is about ten days and its atoms disin- tegrate with the expulsion of alpha particles. The next step in the series of transformation is the gaseous product known as the actin- — ium emanation. This, like the radium emanation, is chemically inert — and incapable of entering into combination with other elements. — Actinium emanation is a very short lived substance and has a half — value period of only 3.9 seconds. It is transformed successively — into three other products, which are solids, known respectively as _ actinium A, B and C, and together constitute the so-called active — deposit from the actinium emanation. Actinium A has a half value — period of 36 minutes, actinium B of 3.1 minutes and actinium C of — 5.1 minutes. The first emits beta rays, the second alpha — and © the third beta and gamma rays. . As already stated, actinium and its products are genetically con- nected with uranium and are, in some manner as yet obscure, derived from it. The evidence in support of this conclusion is quite con- vincing. All uranium minerals contain definite quantities of actin- ium and in the older minerals the relative proportions of uranium % and actinium present are so constant as to permit of no other expla nation. But the actual genealogical history of actinium is st obscure and we are not yet in a position to clearly trace the line of descent. Whether actinium is formed directly from uranium by special kind of transformation which involves only a small propor- tion of the total number of the atoms changing, or whether its pro- duction occurs at a later stage in the uranium-radium series, is at present an open question and the discovery ,of the true relations is one of the most interesting problems now awaiting solution. The thorium series presents another group of radio-elements comprising ten successive members. I shall not stop to enumera these in detail, but their principal physical and chemical character- istics have already been determined. Thorium itself, the parent sub- stance, has a very slow rate of change, which is probably not more than a fifth that of uranium. 1911.] BOLTWOOD—RADIOACTIVITY. 341 Its first product, mesothorium I, is another example of a rayless change like radium D and actinium. Owing to the fact that large quantities of monazite are commercially treated for the extraction of thorium used in the manufacture of incandescent gas mantles and that the technical separation and isolation of mesothorium appears to be an economic possibility, there is some prospect that mesothorium may become a competitor of radium for scientific and therapeutic uses. Its life compared with radium is relatively short, however, its half value period being 54 years, but this in itself is not neces- sarily a serious disadvantage. The chemical properties of meso- thorium are similar to those of radium and barium. The fifth product in the thorium series is known as the thorium emanation and is a chemically inert gas like the radium and actinium emanations. The remaining four products constitute the thortunt active deposit. The combined uranium and thorium series includes 28 radio- elements, of which only the two parent elements were known before the development of radioactive methods. Radioactivity‘ has there- fore added a considerable quota to the known types of matter. An interesting relation which is met in the study of radioactive change is the so-called radioactive equilibrium. Ilia relatively long- lived radio-element A is the parent of a less stable product B, and if A is initially entirely freed from B, then a certain definite fraction of the atoms of A will undergo transformation each second to form _ atoms of the product B. The number of atoms of B produced from _ A in this manner each second will be essentially constant and the amount of B will increase. But the atoms of B also undergo trans- formation at a constant rate and, as the quantity of B increases, a continually increasing number of its atoms will be transformed in the unit of time. A point will finally be reached where the number _ of atoms of B which disintegrate in any given time will be exactly equal to the number of atoms of B formed from A in the same _ interval. The relative amounts of A and B will then remain con- _ stant and the conditions can be expressed by the equation 1,P=A,Q, where P is the number of atoms of 4 and A, its constant of change, 342 BOLTWOOD—RADIOACTIVITY. [April 22, and Q is the number of atoms of B and A, the constant of change of | the product B. Under these conditions the substances A and B are said to be in equilibrium with one another. The general mathemat: ical theory of successive changes of this kind has been i or Rutherford. : In the discussion of the shasartetiatio’ of the sipba rays it has been pointed out that the evidence supplied by the determination of the ratio of the charge to the mass of these particles indicates that their nature is the same in all cases. Let us consider satiny. the additional facts which are in support of this conclusion. The presence of considerable proportions of helium in cigiha minerals containing uranium and thorium has been very frequentl) noticed. It was found by Ramsay and Soddy in 1903 that helium could be detected in the residual gas set free when a specimen of crystalline radium bromide was dissolved in water, and shortly aft this they showed that the spectrum of helium appeared with time a tube which initialiy contained only radium emanation. — Debierne found that helium was produced by a strong preparation of actinium, and conclusive proof has also been obtained by Strutt and by Soddy — that helium results from the disintegration of both thorium and uranium. During the past year I was ‘able to éxperineriale demonstra the production of helium by ionium, and some earlier carried out by Professor Rutherford and myself showed that helit appeared during the disintegration of polonium also. The lat conclusion has since been confirmed by the work of Mme. Curie ¢ Debierne. ; The data supplied by the counting cepetinn ak of Ruthie and Geiger afford a basis for the calculation of a number of impo tant physical quantities, such as the mass of the hydrogen atone a number of atoms in one gram of hydrogen and the number of me cules per cubic centimeter of any gas at standard pressure and tem perature. In a similar matter Rutherford and Geiger have ca lated the amount of helium produced per year by one gram of radi containing equilibrium amounts of its three alpha ray products, th emanation, radium A and radium C. The number obtained in this way was 158 cubic millimeters of helium per year per gram of tort.) BOLTWOOD—RADIOACTIVITY. 343 radium. Measurements of the rate of production of helium by a radium salt have been carried out by Sir James Dewar and have given results somewhat in excess of this, namely 182 and 169 cubic millimeters. As a confirmation of the accuracy of Rutherford and _Geiger’s values, however, it may be stated that an investigation of the production of helium by radium made last year by Professor Rutherford and myself gave results in excellent agreement with the calculated value. An account of these experiments will be published Shortly. a3 In connection with these results there is, however, one rather important point which should be mentioned. This is the fact that the rate of production of helium and the half value period of radium _as calculated by Rutherford and Geiger from the results of their counting experiments, are directly dependent of the purity of the salt of radium used as a standard in their measurements. They assumed that the material of their radium standard was pure anhy- _drous radium bromide containing 58.5 per cent. of radium. If this was not the case and the material used as their standard contains less than the theoretical amount of radium, their calculation of the number of alpha particles emitted per gram of radium and the rate of production of helium is too low, and their estimate of the half value period of radium is too high. If, on the other hand, the mate- tial of their standard consists in part of some other compound of _tradium containing a higher proportion of this element than is con- _ tained in the bromide, their value for the number of alpha particles emitted and the rate of production of helium is too high and their _¢alculated rate of disintegration of radium is too low. _ There are certain reasons which lead me to believe that the radium standard used by Rutherford and Geiger actually contains a higher proportion of the element radium than they have assumed in _ their calculation, and that the true half value period of radium is greater than 1,760 .years as they have deduced it. In 1908 I pub- lished an account of some experiments on the growth of radium in _ionium preparations, which pointed to two thousand years as the _half-value period of the former. This estimate was quite indepen- _ dent of any radium standard and I am of the opinion that it is nearer the true value than is the estimate made by Rutherford and Geiger. 344 BOLTWOOD—RADIOACTIVITY. [April 22, The point can be definitely settled, however, by a comparison of Rutherford’s standard with a standard of indisputable purity. Such a standard is in prospect in the not distant future and its prepara- tion has been undertaken by Mme. Curie on behalf of the Interna- tional Radium Standards Committee appointed at the recent Radio- logical Congress in Brussels. A very interesting action which has been observed to accompany — radioactive transformations is known as the recoil phenomenon. a When a plate bearing a thin layer of very active material is placed — in close proximity of another plate which is inactive, a portion of the active matter becomes detached from the film and is —— on the surface of the second plate. The effect is increased considerably if the receiving lies is charged negatively with respect to the plate bearing the active coat- ing. This action is apparently due to the fact that, when the alpha or beta rays are expelled at a high velocity from a radio-atom under: going transformation, the reaction on the residual atom causes this to move in the opposite direction with sufficient force to detach it from the plate. The action is analogous to the recoil of a rifle attending the expulsion of a high velocity bullet. When, for exam- ple, the active coating on the first plate consists of radium A th the active matter received on the second plate is composed almost exclusively of radium B; and when the film consists of radium B th material thrown off is for the most part radium C. This and oth similar effects which have been noted are all of such a nature as to suggest that the explanation proposed for this interesting phenom- enon is the correct one. The effect of the electric field indicates | in some way these “rest atoms” acquire positive electric charges. From the standpoint of the disintegration theory, it is evid when we consider the three principal groups of radioactive : stances, the uranium-radium group, the actinium group, and thorium group, that the radioactive phenomena exhibited by atoms abruptly disappear after they have passed through a ce series of transformations, which terminates with radium F in instance, with actinium C in another and with thorium D in third. The apparent explanation of this circumstance would sé to be, that, following the last active change, the residual atomic 1911.] BOLTWOOD—RADIOACTIVITY. 345 nucleus finally attains a permanently stable form which undergoes no further alterations. If such is indeed the case, then we might expect that these ultimate end products of radioactive decay would accumulate in old radioactive minerals where the process of trans- _ formation has been proceeding for long geological periods. This _ line of reasoning has enabled us to identify at least one of these q products and that is, in all probability, the one following radium F. _ The residual atom in this case appears to be no other than the atom of ordinary lead. There are, moreover, certain theoretical argu- ments which point to the same conclusion. The accepted ‘atomic weight of uranium is 238.5. It has been found that two alpha par- ticles are emitted during its transformation and one by the succeed- _ ing product, ionium. This would correspond to the loss of three alpha particles or helium atoms with an atomic weight of four or a total of twelve units. Two hundred and thirty-eight and one half _ less twelve give two hundred and twenty-six and one half for the _ atomic weight of radium, which corresponds to the value obtained in the actual determination of the atomic weight of this element : by Mme. Curie. The transformations of the atoms of radium, the emanation, radium A, radium C and radium F are each accompanied __ by the expulsion of another alpha particle, making five in all. Five _ times four is twenty and two hundred and twenty-six and one half _ less twenty is two hundred and six and one half. The latter number is sufficiently near to two hundred and seven and one tenth, the most recently determined atomic weight of lead, to support the conclusion _ that lead is the ultimate disintegration product of radium. It has not yet been possible to determine the end products of the actinium : or the thorium series but they will undoubtedly be identified among . _ the various elements occurring in small proportions in the older ura- nium and thorium minerals. = Before completing this necessarily brief résumé of the present ___ Status of the study of radioactive phenomena it is necessary to make some reference to the series of papers published by Sir William __ Ramsay associated with A. T. Cameron and F. L. Usher. These _ papers, which deal with the action of radium emanation on various other substances, suggest the occurrence of certain changes, which if they really took place would be of fundamental importance to the 346 BOLTWOOD—RADIOACTIVITY, theory of the constitution of matter. Unfortunately, so little we can be attached to these results and the conclusions reached authors, that they have received no serious oe fr most competent to judge their value. | formations. In those cases where it has been possible its influence, the loss of an alpha particle is always accc a corresponding decrease in the mass of the atom from Separated. a the disruiption of a Lainie is é times that of the beta particles. So far as the results of ‘ments have enlightened us we have not yet been able | resolving of the by far the greater proportion of the effé an atom into anything other than a further subdivision: matter. ies YALE Linicenve. April 22, 191t. III. HE DYNAMICAL EFFECTS OF AGGREGATES OF . ELECTRONS. By OWEN W. RICHARDSON. , (Read April 22, 1911.) I. ELECTRONS AND MATTER. he enormous difference in the behavior of different materials ds electric force is a matter with which everyone is familiar; is one of the triumphs of the electron theory that it has given very satisfactory picture of the difference between insulators conductors of electricity. We are to regard all matter as made primarily out of electrons. They are the stones with which the ‘ial structure is built up, the electrodynamic forces are the t which holds the stones together. There are, however, two rent ways in which the electrons may exist in a given portion matter. They may be located in position of stable equilibrium, which case a very small force will displace them to a small extent ‘a perfectly enormous force would be required to dislodge them ughly and give rise to instability; or they may be so loosely that they are able to move about in the interstices of the material, much after the fashion in which we believe the molecules move ‘ina gas. In the former case, when the electrons are practically d, we say the substance is an insulator; in the latter case, where are wandering about, the substance is a conductor. _moment’s reflection will show that this difference is sufficient explain the difference between insulators and conductors. Con- what happens when a slab of the first kind is placed in an ic field. There will be a displacement of the electrons, it is but the displacement will be small and they will all return to original equilibrium positions as soon as the external field is se 347 348 RICHARDSON—DYNAMICAL EFFECTS OF [April ; removed. There will be no transportation of electrons and that what, on the electron theory, constitutes an electric current. electric field across the slab is, nevertheless, different from w it would be if the material were not present. The difference | tween different insulating materials in this respect depends sol on the comparative ease of displacement of the electrons they co tain. The specific inductive capacity of dielectrics, which, you remember, was discovered by Cavendish and Faraday, is, in fact, measure of the product of the number of electrons in unit volw of the material by the average displacement which they undergo unit field. : The behavior of the second kind of material is quite diffe Even in the absence of the electric field, the so-called free elect are moving about in it in an irregular manner in all directions. effect of an external field is to superpose on the irregular moti definite drift, on the average, in the direction of the current. drift of the electrons involves transportation of electricity and the fore implies the existence of an electric current. ? _ All the laws which regulate the transference of electricity acr conductors, such as, for example, Ohm’s Law, which states that current is proportional to the applied electromotive force, and J Law, which states that the rate of production of heat by a is equal to the product of the resistance of the circuit by the of the current, follow at once from this simple hypothesis. It necessary to suppose that all the electrons in the material are p in the free condition; some of them may be, and in all prob the majority are, ina state of equilibrium similar to that which in insulators. All that is necessary is that some of the should be able to move without restraint. When the other con are the same the magnitude of the current which a given m will transport is proportional to the number of carriers ava that is, to the number of free electrons per unit volume. It is in the explanation of the relation between the cond of substances for electricity and for heat that the electron has scored one of its most notable triumphs. Everybody know: the best conductors for electricity are also the best condui heat. It is not so generally known how very close the relati AGGREGATES OF ELECTRONS. 349 etween the two phenomena is. In the accompanying table two columns of figures are shown. Cocthcient of Soeficient of Ratio: Thermal Conductivity, this Ratio, Material. Electrical Conductivity. Per Cent. _ Copper, commercial .......... 6.76 X 10” at 18° C. Seer CE) POPE Sc. oi... een, 6.65 X Io” at 18° C. me Copper (2) pure. .........-..- 6.71 X 10” at 18°C. BUMS OEPO sooo uses cones = 6.86 X 10” at 18° C. Og Bore Sy one 7.27 X 10” at 18° C. oO Be Seer eeeres 7.09 X 10” at 18° C. Se peer ree 6.99 X 10” at 18° C. MME OED ea twee eg a bes ene s 7.05 X 10° at 18° C. Se eee 6.72 X 10” at 18° C. Poe A SOREN, Pure ......:......- 7.06 X 10” at 18° C. INN eg wie pin soc © 7.35 X 10” at 18° C. SIR os on sg ods son a's 6.36 X 10” at 18° C. 6 68 Gare enor 7.76 X 10° at 18° C. | NSE eae 7.54 X 10" at 18°C. ME bce Prat a cu ekes eebc eas 9.03 X 10” at 18° C. ED is coh sinavacssiasss ee 9.64 X Io” at 18° C. _ Constantan (60Cu, 4ONi) ....11.06 X 10” at 18° C. 23 _ Manganin (84Cu, 4Ni, 12Mn).. 9.14 X 10” at 18° C. 27 They represent the results of measurements by Jaeger and Dies- orst of the electric and thermal conductivities of a large number metals and alloys. The first column of figures gives the ratio of e thermal to the electrical conductivity for each of these substances nd the second gives the percentage change of this ratio when the temperature is increased one degree. It will at once be noticed that numbers in each column are almost equal, particularly if we p to the pure metals. Thus for every pure metal the electrical ductivity bears to the thermal conductivity a proportion which is nost independent of the metal: and the ratio of the thermal con- ‘ivity to the electrical conductivity increases by almost the same punt for one degree rise of temperature for each metal. The efficient of increase of this ratio with increase of temperature is ) very nearly equal to the coefficient of increase of the volume of all es with temperature, when the pressure is maintained constant. GRESREIELELSSSSSSSS | 350 RICHARDSON—DYNAMICAL EFFECTS OF [April These interesting relations were shown to be a consequence of electron theory of conductors by Drude. He proved that they low inevitably from the assumptions (1) that a metal con electrons which move about freely like the molecules of a gas, (2 that they possess a certain average mean length of free’ path during the traversing of which they are only acted on by external applied electric force, (3) that this path is termir ed a collision and that the new motion which then ensues is, o1 average, independent of the previous motion; and lastly (4 their average kinetic energy is the same as the average kin energy of translation of a molecule of any gas at the same temps tur as the metal. electron during its free path is Xe and its acceleration Kem the velocity of the particle at the beginning of the Path is Xe velocity at the end will be # + =e where ¢ is the average Gnd tween two collisions. ae ae velocity in the direction electric field is therefore > x= ¢ since the average value of u over a large number a Ss is zero. Now the free path ; A equal to vt where v is the mean speed. Thus the average velocity of the any: in si direction of the electric field written in the form — X= pont If m is the numises of ‘ele unit volume, the ee “e oe which, in unit time, drift. unit area drawn perpendicular to the direction of the elect Xr 4 X will bes nX <=. Each of these carries a charge e so quantity of dedigitiey transported across unit area, or in oth : the electric current density will be Now it is a necessary consequence of the principles which lie the kinetic theory of matter that mz should be equal to a) 6 is the absolute temperature and a is a universal constant which AGGREGATES OF ELECTRONS. 351 be calculated from the properties of gases. This assertion is the a a statement of the relation (4) enumerated above. ; ; er vity of the material is o = —- = gi In this formula e? and a have the same value for all substances, m and A are constants charac- teristic of each substance, v is independent of the nature of the ‘material but is proportional to the square root of the absolute temperature. It is a well-known result of experiment that the specific conduc- tivity of all substances is versely proportional to the absolute tem- perature. We therefore conclude that the product mA for all metals ‘must be inversely proportional to the square root of the absolute 3 _It is a well-known result of the kinetic theory of gases that the thermal conductivity of a gas is equal to 4nAva. Hence = -1% é. ‘Thus this ratio should have the same value for all metals at the same temperature and the temperature variation should be the same s that of the volume of a gas at constant pressure. These are the relations which are exhibited by the experimental results of Jaeger and Diesselherst. : The electron theory of metallic conduction has ‘enabled us to understand a number of curious effects which occur when a con- ductor is placed in a magnetic field. One of these, the Hall effect, onsists in a deflection of the line of flow of a current which is caused by the magnetic field. Another effect, which is especially marked in the case of Bismuth, is an alteration of the specific resist- of the material caused by a magnetic field. These effects are intimately connected together and have a simple explanation on the lectron theory. It is well known that any electrified particle moving n a magnetic field is acted on by a force which is perpendicular to the plane containing the magnetic force and the direction of motion. The superposition of this force upon the other forces acting on the electrons in a metal carrying a current will cause all the electrons ‘to curve round in the same general direction, giving rise to the Hall effect. It will also increase the average curvature of the paths of the 352 RICHARDSON—DYNAMICAL EFFECTS OF [April 2a, electrons. In this way the time which is required for electricity to be transferred will be greater so that the specific electrical conductivity will be diminished. This is the explanation of the second effect. Both these effects are complicated by the action of the electrons the atoms so that the foregoing description is only to be regardeg as a rough outline of what really occurs. i So far we have only considered the way in which the aiectifal theory of conduction explains a number of phenomena which w familiar before it was enunciated. The power to do this is a neces- sary attribute of every scientific theory. A scientific theory, how- — ever, is often much more useful than this in that it leads to the pre- diction of phenomena which would hardly have been foreseen withou its aid. The present theory has been able to prove its usefulness in this way, as the principles underlying it have enabled us to develo a new chapter in physical science, a chapter to which I have ven to give the name of Thermionics. Thermionics relates to the emi sion of electrified particles by hot bodies and the phenomena to whict they give rise. a It is found that all bodies when heated to a sufficiently high ren perature give rise to an emission of both negatively and positi charged particles. In many ways the negative emission is the mo interesting as the particles emitted are negative’électrons having p erties identical with those of the carriers of the cathode rays. — connection between this emission of negative electrons and the tr: portation of electricity in a metallic conductor is very intimate. — have seen that, in order to explain the phenomena exhibited by tallic conduction, it is necessary to suppose that such conducte contain large numbers of “free” electrons. If these electrons maving about freely inside the conductor, as we have suppose question at once arises as to why they do not escape into the st rounding atmosphere. It is clear that they do not do so, othery there would be a leakage of electricity from the surface of all ductors at ordinary temperatures. The answer must be that t are forces at the surface of the metal which are sufficiently great prevent them from escaping. Now consider what we should pect to happen as the temperature of such a body israised. Wehay supposed that the average kinetic energy of the contained electrot 1911.] AGGREGATES OF ELECTRONS. 353 __ is higher the higher the temperature. Clearly, at a sufficiently high ___ temperature some of the particles will have enough energy from their heat motion to be able to break through the surface. Moreover, the number which are able to escape will be greater the higher the temperature. ‘ A theory following these lines has succeeded in predicting the ___ way in which the emission of the electrons depends upon the tem- __ perature as well as a number of other interesting relations between | the thermal and electrical behavior of substances. It will be re- marked that the view which has been outlined is very similar to the view of the phenomenon of evaporation which is afforded by the kinetic theory of matter. According to that theory the particles of _ the liquid escape into the vapor when their kinetic energy (to be accurate we ought to say that part of it which depends on the com- 4 ponent of velocity normal to the surface) exceeds the work they : a _ have to do in order to pass through the surface. Thermionic emis- a sion may be looked upon then simply as the evaporation of electrons a which may be regarded as dissolved in the metal. Just as water is E cooled when it evaporates and heated when steam condenses into it; so we should expect a conductor to be cooled when it emits electrons __ and heated when it absorbs them. Both these effects have recently ce been discovered, the former by Wehnelt and Jentzsch and the latter _ by Richardson and Cook. There is one point in this connection which is worthy of further _ consideration. We have seen that it is necessary to suppose that the electrons in a metal behave like the molecules of a gas. The same will be at least as true of them after they have been emitted. Thus when a metal at a high temperature lies in an air-tight en- = closure there will be two atmospheres of electrons, one at a high = _ pressure inside the metal and the other at a low pressure in the Eg enclosure outside of the metal. If the principles of the kinetic _ theory of matter are well grounded it can be shown that in both of _ these atmospheres the electrons are moving about with all possible : speeds but that the proportion of them which have a given speed is _ the same for each atmosphere. Moreover, the proportion is the same known function of the temperature in each case and in each case also PROC. AMER, PHIL. SOC., L. 200 W, PRINTED AUG. 4, IgIiI. 354 RICHARDSON—DYNAMICAL EFFECTS OF [April 22, the average kinetic energy should be the same as the average kinetic energy of a molecule of any gas at the temperature of the enclosure. In fact the laws of the kinetic theory of gases can be applied without change to the atmospheres of electrons ; and the above asser- tions are simply statements of a theorem in the kinetic theory of © gases called, after its discoverers, the Maxwell-Boltzmann Law. — According to this law if a large number of molecules are selected — at random out of any gas the proportion of them which have speeds — lying between certain assigned values let us say u and w’ is a certain — definite function of u and uw’. The value of this function, which in addition to u and wu’ depends only upon the temperature of the gas — and the mass of its molecules, was first deduced by Maxwell. Max- well’s deduction of the value of this function, though sufficiently con- vincing to those who are familiar with the methods of mathematical _ physics, was, neverthless, a highly abstract piece of reasoning; and — nature of a direct test of it by experiment on gases. With the — atmospheres of electrons we are, however, able to do a great deal — more than we could with a gas made up of uncharged molecules. By placing them in a suitable electric field we can bring forces to bear on each individual electron which are enormous compared with the forces exerted on a molecule by the earth’s gravitational field. For example if the electrons are being emitted from a heated flat plate we can place another flat surface a little in front and charge it up, so that the electric field tends to drive the ions back into the surface at which they originated. Under these circumstances only those electrons will be able to cross from one plate to the other if their kinetic energy is greater than a certain value depending on the ele tric field between the plates; thus the current that gets across will a measure of the number of electrons emitted whose kinetic ene exceeds a known value. By experiments of this kind, and oth based on similar principles, we have succeeded in determining law of distribution of speed among the individual electrons wl are emitted. It is found to agree in every particular with that p dicted by Maxwell for the case of a gas whose temperature is th same as that of the metal emitting the electrons and whose molec weight is equal to the mass of an electron. In particular the average 1911.] AGGREGATES OF ELECTRONS. 355d kinetic energy of the electrons is the same as that of the molecules of a gas at the temperature of the metal which emits them; and we can calculate the value of the well-known constant R in the gas _ equation pv== 6, where p is the pressure, v the volume and @ the absolute temperature of the gas, from purely electrical experiments of the kind indicated. It follows from the results of these experiments together with a simple application of the principles of the dynamical _ theory of gases that the free electrons inside a metal must have the distribution of velocity which is required by Maxwell’s law and in particular must have the same average translatiorial kinetic energy 4 as the molecules of a gas at the temperature of the metal which __ contains them. 2. MATERIAL MEDIA AND ELECTROMAGNETIC RADIATIONS. The action of light on insulating media is a rather complicated, but extremely important, phenomenon on which the electron theory has thrown a great deal of light. Maxwell showed, many years ago, that light is an electromagnetic phenomenon. A beam of light is in fact a wave of oscillating electric and magnetic force, the electric -and magnetic forces being at right angles to one another and to the direction of propagation. When such a wave falls on an insulating medium the oscillating electric force will set into vibration the com- paratively stable electrons which, as we have seen, are embedded in the medium. The electrons will execute what are appropriately called forced oscillations, about their original equilibrium positions, nd these oscillations will have the same periodic time as the light. Thus when it traverses a material insulating medium the light has not only to keep itself going; it has to keep the electrons which make up the medium going as well. Roughly speaking one may say that the electrons in such a medium behave like a load on the luminif- erous ether. We should therefore expect them to diminish the speed of propagation of light through it and this is found to be the case. The exact expression for the velocity cannot be obtained without going more deeply than we have time to into the electro- “Magnetic theory of light. It was first given by Maxwell, who _ showed that the refractive index, to which the velocity of propaga- _tion is inversely proportional, was equal to the square root of the 356 RICHARDSON—DYNAMICAL EFFECTS OF [April 22, effective specific inductive capacity of the medium. Now the spe- cific inductive capacity of an insulating medium is equal to unity plus the product of the number of electrons per unit volume by their average displacement in unit electric field. When the material is subjected to constant electric forces the displacements of the elec- trons are always proportional to the forces and the specific inductive capacity is therefore a constant quantity. When the force is an oscillating one the matter is complicated by the fact that the elec- trons try, as it were, to strike a balance between their own natural period of oscillation and that of the force acting on them. They end by oscillating with the same frequency as the force which — excites them but the distance they travel from their equilibrium position depends a good deal on their natural periods as well. Thus the specific inductive capacity for oscillating forces will not be a constant quantity but will depend to some extent on the frequency of oscillation of the force. By the effective specific inductive ca- pacity we mean the specific inductive capacity for electric forces which oscillate with the frequency of the light under consideration. It is evident from what has been said that the refractive index of an insulating substance depends upon the frequency or, in other words, upon the color of the light. We see at once why a beam of white light is split up by a prism into the constituent spectral colors. For each ray is deviated by the prism according to the value of its refractive index. : Perhaps the most interesting question in this part of our subject — is that of the behavior of a substance towards light whose frequency — is close to that of the natural periods of the substance. In that case. the electrons are set into violent motion owing to the occurrence of what are sometimes called sympathetic vibrations. The nature of this phenomenon may best be illustrated by considering a simple mechanical analogy. Imagine a spiral wire with a weight at one end to be hung from a shaking support. If the weight is pulled down and let go it will oscillate backwards and forwards with ; definite natural frequency which depends on the stiffness of th spring and the heaviness of the weight. If the shakiness of the support arises from tremors in the building, to the walls of which we will suppose it bolted, as a rule the frequency of its vibrations wi 1911.] AGGREGATES OF ELECTRONS. 357 be very great compared with the natural frequency of the spring. In that case the shakiness of the support will have very little effect on the spring. If however the frequency of the tremors happens to be equal or nearly equal to the natural frequency of the spring the latter is set into very violent agitation, for the reason that the natural swings of the spring are continually being helped by the oscillations of the support. A precisely analogous effect takes place when the period of the light is close to the natural period of the electrons. In fact it can be shown that, if there is nothing analogous to a frictional force to damp down the vibrations of the electrons, they will execute oscilla- tions of infinite amplitude when there is exact coincidence between the periods. Since the displacement of the electrons in unit electric field is the important factor in determining the refractive index we Should expect its value to change very considerably in this region. As a matter of fact, in the extreme case where there is no damping, the value of p? falls rapidly from a small positive quantity on the short wave-length side of the position of coincidence to the value — © at exact coincidence (A—A,). As the period of exact coinci- dence is passed pw? changes suddenly to + o and on the long wave- length side falls rapidly to a rather larger positive value than the one that it had at a great distance from the natural period on the short wave-length side. __ Several very important deductions can be drawn from the results _ which have just been described. In the first place we notice that provided we always keep to the same side of the natural period, no matter which side we choose, then the refractive index p» always diminishes as the wave-length A increases. Hence, since the devia- tion of light by a prism is greater when the refractive index is greater it will be smaller the greater the wave-length. The blue light will therefore be deviated more than the red light in the spec- trum. This is the well known kind of dispersion which is exhibited by prisms of glass and similar colorless transparent substances. When part of the spectrum lies on one side of the natural period and part on the other there is a sudden increase in the value of the tefractive index when the natural period is crossed. The spec- trum will then consist of two groups of colors, that on the long 358 RICHARDSON—DYNAMICAL EFFECTS OF [April 22, wave-length side being more deviated than that on the short wave- length side, although in each group the colors are in the normal order. This is the so-called anomalous dispersion which was dis- covered by Kundt and which is exhibited by all transparent colored bodies, like the aniline dyes, which possess a metallic shimmer. Immediately on the short wave-length side of A,, we have seen that »? has a negative value in the case we are contemplating. yp in this region has therefore what mathematicians call an imaginary value. It can be shown that this imaginary value means that the waves are incapable of entering the medium. When a train of waves of this wave-length falls on the medium they are not absorbed, properly — speaking, but are completely reflected. The substance would appear to be opaque to light of this wave-length not because it absorbs the — light which falls on it but because it reflects it completely. If mixed light which contained some of this particular wave-length — were made to undergo a sufficient number of successive reflections — from plates of the substance, only light of this particular region of © frequency would ultimately be left over, since a certain percentage — of the other wave-lengths always gets through. This principle has — been utilised by Rubens to isolate radiations of definite wave-length in the infra-red part of the spectrum. These radiations are called, very appropriately, residual rays. The foregoing discussion does not touch the very interesting questions of the absorption of light by insulating media. There w be no absorption, properly speaking, unless there are forces acti on the moving electrons which tend to dissipate the energy of light. Such forces must in general exist and it is usually assum that there is a retarding force proportional to the velocity of the — moving electrons, chiefly because this is the simplest assumpti which can be made which is not in contradiction with fact. existence of forces of this kind modifies the foregoing conclus to a considerable extent in detail but it does not affect their gen character. Planck has pointed out that it follows from the principles of electromagnetic theory of light that the radiation from the movi electrons gives rise to a retarding force which may be taken to” proportional to their velocity. Such a force must unquestionably 19tt.J AGGREGATES OF ELECTRONS. 359 a exist but its magnitude is quite small. It is of interest to see if it _is sufficiently large to account for the known cases in which the _ dissipation of energy is smallest. These are unquestionly the cases in which residual rays are obtained. I have developed a formula which expresses the percentage of incident energy which goes into the residual rays, which includes the case where dissipation is taken _ account of. This formula leads to two separate methods of esti- a mating the order of magnitude of the dissipation. Both these methods show that the dissipation must be of the order of 10% in cer- tain units where Planck’s theory leads to a dissipation of the order _ 10* in the same units. Thus the source of dissipation pointed out by Planck is about 10* times too small to account for the smallest case of dissipation known to us. I am inclined to think that the most general type of absorption of light by bodies of this class is of the following character. We have seen that the electrons execute forced vibrations under the influence of the incident light. When the period of the light ap- _proximates to the free period of the electron the electrons absorb a great deal of energy from the light. In general this absorption of _ energy will go on until the vibrations carry the electron out of its ' region of stability. A rearrangement of the system will then take _ place and during this rearrangement a great deal of the kinetic _ energy which the electron has accumulated will be transferred to other parts of the substance and will make itself felt as heat. As far as that particular electron is concerned the sympathetic vibrations will have to be established all over again. It is not necessary to suppose that during this process the electron is actually carried out of the atom when it breaks loose from the region of stability. The whole occurrence may take place in the one atom. On the other hand we know a great many cases of bodies which emit electrons under the _ influence of light and in these cases the electrons must get carried _ out of the atom. It seems to the writer to be an advantage _ Of this view that it connects the absorption of light with the so- called photo-electric effect. As a first approximation this view of the absorption of light leads to the same relation between absorption and frequency as does the assumption of a retarding force propor- tional to the velocity. 360 RICHARDSON—DYNAMICAL EFFECTS OF [April 22, There is another point in this connection that is not without in- terest. On any theory of absorption the natural periods of a sub- stance are characterised by conferring on it either intense absorp- tion or intense opacity. It is therefore evident that they can be detected very readily by experiment. From an analysis of the nat- ural periods of a large number of substances which has been carried out by Drude it appears that there are two types of vibrations which occur. In the one the electron forms the vibrating system and in the other one of the constituent atoms or a group of atoms vibrate as a whole. Owing chiefly to the enormous difference between the mass of an electron and that of an atom there is an enormous difference between the frequency of the two types. The electronic type always gives rise to frequencies in the ultra-violet part of the spec- trum and the atomic type to natural frequencies in the infra-red. It is therefore not an accidental circumstance that almost all chem- __ ically pure substances which are not conductors of electricity are transparent in the visible spectrum. o The action of Roentgen rays on matter is a subject of great inter- est. According to the ether pulse theory of these rays elaborated by Sir J. J. Thomson, the relation between the Roentgen rays and sodium light is similar to that between a series of sharp cracks anda * ~ musical tone. And on the modern view of the nature of white light the difference between white light and the Roentgen rays is one of degree rather than kind. The cracks corresponding to Roentgen rays are much sharper than those which correspond to white light. According to the principles of harmonic analysis which we owe to Fourier it should be possible to resolve either of these kinds of — radiation into simple harmonic elements. I have estimated that the average frequency of these elements for the Roentgen rays would © be 10,000 times greater than that for those which form white light. — This estimate is based on the view that the kinetic energy of the electrons emitted by bodies under the action of ultra-violet light and Roentgen rays is a function of the frequency of the equivalent vibra- — tions. The experimental results indicate that the functionality isa linear one and there is considerable theoretical support for this view ° Some investigators have maintained that the square root of the — energy is proportional to the frequency; but even if this extreme — 19tt.] AGGREGATES OF ELECTRONS. 361 view is taken the estimated frequency is not changed enough to affect the general argument. If the Roentgen rays are so much like white light you will at once ask why they are not deviated by a prism. The answer is very simple. It follows from the principles of the electron theory that the refractive index » of a substance for electromagnetic vibrations Vv z mp,’ — p*) _ charge on an electron, m its mass, ps its natural frequency and vs the number of electrons of type s in unit volume of the material. In general this formula will not be exact on account of the interactions of the electrons on one another but it will give results of the right order of magnitude. Now e and m have the same value for all electrons and in the part of the spectrum near the visible v,e?/mp,? is of the order unity for such frequencies p, as fall within that region. Now vr, will always be of the order of the number of mole- cules per cubic centimeter whichever of the s classes of electrons we consider, so that we may draw the following conclusions: (1) On account of the very great absolute value of p, »* will be equal to unity except for a very narrow range in the immediate neighbor- hood of p=/;. (2) Only such substances will be capable of re- fracting the Roentgen rays as have natural frequencies p, which lie within the range of values of » embraced by the Roentgen rays ex- perimented on. In any event it is clear that with a mixed group of rays such as is emitted by an ordinary X-ray bulb, practically the whole of them will pass through a prism without deviation. _ Barkla’s experiments on secondary rays show that the Roentgen tays exhibit phenomena very much akin to fluorescence in optics. One interpretation of Barkla’s results would be that there are in material atoms natural frequencies comparable with the frequencies in the Roentgen rays. In that case, although almost the whole of a beam of Roentgen rays would be undeviated by a prism there should be a small amount which would be deviated. At present I am making experiments to detect this effect. OF course if one adopts the corpuscular view of the Roentgen rays recently developed by Bragg, effects of the kind described are _ of frequency p is given by p?=1+% where e is the 362 RICHARDSON—DYNAMICAL EFFECTS OF [April 22, not to be expected. At present the balance of evidence seems to be decidedly against the corpuscular view. I am inclined to think that the primary Roentgen rays originate largely as the result of secondary actions due to the stirring up of the electrons in the atoms of the anti-cathode by the rapidly mov- ing cathode rays which impinge on them. On this view the con- stituent frequencies of the rays would be, to a considerable extent, a matter of the atoms in which they originate ; and it may be that the gap in electro-magnetic radiations between ultra-violet light and the Roentgen rays, which exists at present, may never be filled up; as there may be no atoms which have natural periods in the neighbor- hood of these frequencies. i Recent years have seen the accumulation of a very large quantity of material relating to optical effects which are produced by a mag- netic field. It is impossible, within the limits of this discussion, to attempt to show the enormous usefulness of the electron theory in_ the development of the science of magneto-optics; but there is one phenomenon which we cannot afford to pass by entirely, if only on account of its historical importance. I refer to the Zeeman effect. This effect was called after its discoverer, who showed that the spectral lines which are emitted by all gaseous substances under suitable conditions of excitation, were slightly displaced by a v strong magnetic field. The true explanation of this phenomenon was at once given by H. A. Lorentz. He pointed out that if monochromatic light was emitted by vibrating electrons the fre- — quency of the vibrations would be altered if the atom which tained the electrons found itself in a magnetic field. This change in the frequency, of course, corresponds with a change in the wa’ lengths of the emitted light. He also predicted that the emi light would be polarized in a certain way and this was confirmed experiment. Lorentz showed, in addition, that the value of | electric charge of an electron, divided by its mass, could be cal from the displacement of the spectral lines in the magnetic The results of these calculations showed that the value of this ratio was the same as that found by Sir J. J. Thomson and Wiechert for the cathode rays in a discharged tube. Thus the Zeeman effect an the cathode rays were the first two phenomena which afforded ex tort.) AGGREGATES OF ELECTRONS. 363 plicit evidence of the existence of these minute charged particles, whose mass is nearly 2,000 times smaller than that of the lightest known chemical atom. 3 The emission of ordinary heat radiation such as is given out __ by all substances, and in increasing amount the higher the tempera- - ture, is very intimately connected with the theory of electrons. As : : _ is well known this radiation is electromagnetic in character and in- | eludes visible light as a particular case. We also know that when ' an electron is accelerated it emits electromagnetic radiation. It is natural therefore to attribute the origin of this radiation to the motions of the electrons of which material bodies are made up. By making use of the principles of thermodynamics we can prove that the nature of the radiation of this character which is to be found in any enclosure maintained at a given temperature is inde- pendent of the nature of the walls of the enclosure. The amount and character of this radiation is thus independent of the material from which it originates. If therefore we can calculate the amount of this radiation of each wave-length for any particular substance, for a series of temperatures, we shall know what it is for all sub- ‘stances at the same temperatures. Unfortunately when such a calculation is carried out in the most logical and natural way it leads to results which are not in agreement with those given by experimental measurements. Another mode of calculation given by Planck leads to a formula which agrees with the experimental values. It has been shown however that Planck’s calculation involves the implicit assumption that energy is an atomic or discontinuous quan- tity. This idea is distasteful to many physicists and it is so revolu- tionary that it is not desirable to adopt it without very convincing evidence. One of the authorities on this very intricate subject, eans, maintains that the reason for the discrepancy between the less revolutionary theory and experiment is due to the fact, as he asserts, that the experiments do not measure the true equilibrium radiation. However this may be, the difficulties which lie in the explanation of the connection between radiation and temperature do not belong to the electron theory proper but are outside of it. No matter how it may be decided the outcome of this question is : not likely to shake the foundations of the electron theory of matter. 364 RICHARDSON—DYNAMICAL EFFECTS OF | [April 22, 3. THE NUMBER OF ELECTRONS IN AN ATOM. The behavior of very rapidly moving electrons in their passage through matter is a very interesting subject of investigation. Thanks to the discovery of the radio-active substances we are able to experi- ment, if we wish, with electrons whose speed is almost equal to that of light (3 X 107° centimeters per second). These rapidly moving electrons are able to shoot right through the atoms of bodies, but when in their flight they pass very close to one of the constituent — electrons theif paths are deviated. When a group of them passes — through a considerable thickness of matter these deviations tend to — accumulate ; so that a group of electrons, all of which were moving a parallel to one another to start with, becomes divergent. Sir J. J. Thomson showed how the average deviation arising from a single impact could be calculated and also how the average divergence of the beam, which was caused by its passage through a given amount of matter, would depend on the number of electrons present in the matter traversed. Experiments made by Barkla showed that when matter, which — was made up of elements of low atomic weight, was traversed by Reentgen rays it was caused to emit so-called secondary Roentgen rays, which were precisely similar in character to the primary Reentgen rays which excited them. A careful study of these secon ary rays showed that they were primary rays which had been s tered. The phenomenon is in fact very analogous to that whi gives rise to the blue color of the sky, which was shown by Lo: Rayleigh to arise from light scattered by innumerable small partic present in the atmosphere. The amount of such scattering depe on the number of particles which are engaged in doing it. In the case of the Roentgen rays these particles are the electrons pre: in the matter, each one of which is set into violent motion by 1 Roentgen ray pulse. The exact way in which the amount of scattering should depend on the number of electrons engaged in operation was figured out by Thomson who showed from Bar experimental results that the number of electrons reckoned per ato of the material was comparable with the atomic weight. ‘ We have seen already that information of a like nature may AGGREGATES OF ELECTRONS. 365 as was then available Thomson found his conclusion, that lumber per atom was comparable with the atomic weight, was ngthened. Quite recently J. A. Crowther has made a very care- il experimental Meeeent covering both these lines of inquiry. ly by his experiments and that the number of electrons which to make up each atom is three times the atomic weight of the April, 1911. A IV. THE CONSTITUTION OF THE ATOM. By HAROLD A. WILSON, F.R.S. (Read April 22, rort.) According to Sir J. J. Thomson’s theory? atoms may be rega as rigid spheres of positive electricity containing negative elec which can move about freely through the positive charge. — total negative charge on the electrons in an atom is equal to positive charge on the sphere. This theory has many advant: over the theory of Sir J. Larmor, who regards atoms as systems positive and negative electrons in rapid motion. In the first p the sphere of positive electricity provides a rigid and stable founc¢ tion which is lacking in the other theory and which seems very ne sary to explain the extraordinary stability of atoms. It is difficult see how Sir J. Larmor’s atoms could possibly survive the shocks continual violent collisions with other atoms. Sir J. J. Thomson’s theory has also the great advantage th explains the fact that only negative electrons can be isolated saa positive electricity is always associated with atoms or molecules matter. ee It also explains the fact deduced from the Zeemann effect spectral lines are emitted by vibrating negative electrons and not positive electrons. It is consistent with the fact that atoms lose a few negative electrons without their identity being destr which does not seem possible on Sir J. Larmor’s view. The k theory of gases agrees best with the facts when the atoms J. J. Thomson’s theory. This theory therefore may be used as a working hypothesis enables a mental picture of the atom to be formed. It leaves nature of electricity and of the ether an open question and is « *“The Corpuscular Theory of Matter,” 1907. 366 WILSON—CONSTITUTION OF THE ATOM. 367 quently much less fundamental than, for example, Lord Kelvin’s _ vortex ring theory. The negative electrons and the positive sphere '* may or may not turn out to be modes of motion of the ether; at _ present we cannot say. 3 One of the first questions which naturally arises in connection - with this theory is, how many negative electrons are there in each atom ? This question has been answered approximately by examin- _ ing the effect of matter on light and Rontgen rays. When electric _ waves pass over electrons the electrons are acted on by the electric _ forces in the waves and so emit radiation. This means that the electrons scatter the incident radiation. The amount of radiation _ scattered by one electron can be calculated on the electromagnetic theory and hence from the amount observed to be scattered by a wn amount of matter the number of electrons in the matter can be estimated, the number of atoms in a given amount of matter can be exactly calculated because we know the charge carried by one atom in electrolysis and the total charge carried by the matter. Hence we can get an estimate of the number of electrons per atom. The total energy scattered by a mass containing N electrons is N—E, where ¢ is the charge on one electron, m its mass and E the incident energy. This formula is due to Sir J. J. Thomson. _ The most recent determination of the energy scattered when Rontgen rays pass through matter is that by Crowther.? He finds that the number of electrons per atom of aluminium is 85, which is about three times the atomic weight. Previous experiments of a similar character have given nearly the same result for other elements. It seems very probable therefore that all atoms contain a number of electrons proportional to their atomic weights and not very much greater. The mass of a negative electron is only one seventeen-hundredth part of that of an atom of hydrogen, so that the negative electrons ‘only account for about one six-hundredth of the mass of any atom. ‘The rest of the mass therefore must be the mass of the positive sphere. According to this theory therefore the mass of matter is not electromagnetic in its origin, for the electromagnetic mass of the ? Proc. Roy. Soc., A, Vol. &5, p. 20, IQII. 368 WILSON—CONSTITUTION OF THE ATOM. [April positive sphere is negligible. This theory therefore does not support the view which is the basis of the “ principle of relativity,” that all phenomena are electromagnetic in character. The mass and rigidity of the positive spheres are assumed to exist and cannot be explained by electromagnetic forces. There is no reason why the motion of these spheres through the ether should not produce effects capal of being detected and which would enable us to determine the vel ity of the earth relatively to the ether. The fact that this has | yet been done does not prove that it is impossible. According to Sir J. J. Thomson’s theory the properties of ¢ di f ferent atoms are due to the number and arrangement of the electron in the positive sphere. The problem of the distribution of electrons in a positive sphere has not been solved and is ver complicated, so Sir J. J. Thomson investigated the much simp problem of the distribution of n electrons in a plane when they ar all acted on by forces of attraction proportional to their dista from a fixed point in the plane. This problem can also be solved experimentally by mean Professor Mayer’s floating magnets. The electrons arrange th selves in concentric rings. Thus six give a ring of five and one the middle. Seventeen give a ring eleven, a ring of five and ¢ in the middle. Thirty-two give rings of fifteen, eleven, five and c in the middle. Forty-nine give rings of seventeen, fifteen, el five and one in the middle. With two in the middle we get a series of rings containir 12, 16, 19 and 22 electrons respectively and a similar series three in the middle and so on. . This leads to a very interesting suggestion with regard series of elements which have similar properties for examp: Helium, neon, argon, krypton, xenon ; hydrogen, lithium, sodiu potassium, rubidium, cesium; fluorine, chlorine, bromine, iodine. _ Sir J. J. Thomson suggests that each element in such a : may be derived from the one before it by the addition of < ring of electrons the arrangement of the inner rings remaini changed. This explains the similarity of the properties elements in such series. ; On this view an atom of bromine is an atom of chlorine wi 4911.) WILSON—CONSTITUTION OF THE ATOM. 369 2 _ addition of oné more ring of electrons together with the additional __ amount of positive electricity required to keep the atom neutral. _ Sir J. J. Thomson has shown that many of the facts connected with Mendeléeff’s periodic law can be explained on this theory. _ In the atoms of course the electrons are not really confined to one plane but are distributed throughout the volume of the positive _ sphere, so that instead of concentric rings of electrons there are "concentric spherical layers. An atom of bromine is therefore de- tived from an atom of chlorine by the addition of one more layer, the inner layers remaining unchanged. Although the exact solution of the problem of the distribution of ‘n electrons inside a positive sphere is too complicated to be worked out I find that an approximate solution can be obtained without much difficulty, which enables the results of the theory to be com- pared with the atomic weights of the elements. Consider an electron having a negative charge e inside a sphere of positive electricity of uniform density of charge p per c.c. Close to the electron the electric field is of strength e/r?, where r is the distance from the electron, so that 4xe tubes of electric force come out of the electron, if the number of tubes per sq. cm. is taken to be equal to the field strength. Consider one of these tubes of force and let ds be an element of its length and q its cross section at ds. __ The charge in the length ds is pads, so that Fa— (% aa 5, (Fa)ds) = 4mpads, where F is the electric force along ds. Hence $ —d(Fa) = 4mpads, ___ which gives a F,a, — Fa= 4xpfads, e 3 where F,q, denotes the value of Fa at the surface of the electron. This shows that as we go along the tube Fa diminishes and when F,a,=4rnpfads it will be zero and the tube will end. Now F,= t _e/a®, where a is the radius of the electron and a,—=a?/e, so that _ F,a,==1, hence 4xpfads from the surface of the electron to the end of the tube is equal to unity. Thus the positive charge in each * PROC. AMER. PHIL. SOC., L. 200 X, PRINTED AUG. 4, IQII. 370 WILSON—CONSTITUTION OF THE ATOM. [April 22, — tube is 1/47 so that its volume is 1/4mp. The total volume of all the 4me tubes is therefore e/p. Thus the tubes of force starting — from the electron occupy a volume e/p and this is true in any case whether other electrons are near or not. Also since every tube of — force must end on positive electricity it is clear that the volume e/p @ can only contain the one electron from which the tubes start. Thus . when any number of electrons are present each one will be sur-_ rounded by its own field which will occupy the volume e/p. The positive charge in the volume e/p is equal to e, so that if the sphere — has a positive charge equal to the total negative charge on the elec- — trons in it, it will be divided up into n equal volumes, each contain d ing one electron. The energy in an element of a tube of force is equal to F*qds/8x, and if the tube is slightly distorted this element will still have the same volume and also (Fa) will remain unchanged so that the change in the energy in the element will be due to the change in F. The energy will be a minimum when the tube is in equilibrium so that F will be as small as possible and therefore a as large as possible. This means that the tubes tend to become as short as pos- sible, their volumes remaining constant. The effect of this will evidently be to make the field round each electron tend to become as nearly spherical as possible with the electron in the middle. Consequently to determine approximately the distribution of the » electrons in the positive sphere it is sufficient to find how the sphere can be divided up into N equal volumes, all as nearly sph ical as possible and put an electron at the center of each of th volumes. . When » is large it is easy to see that this requires the electr to be arranged like the centers of the shot in a pile of shot. with thirteen electrons we should have one in the middle and tw arranged around it, all at the same distance from it. ‘Suppose the volume of the field of one electron is v, sal N,N,N, etc., denote the number of electrons in the atoms 0 series of similar elements. Each element is formed by the addi of a spherical layer to the one before it and it is clear that all the layers must be of nearly the same thickness if the fields of all electrons are to be nearly spherical. Consequently if 1,,72, 73, . grt.) WILSON—CONSTITUTION OF THE ATOM. 371 denote the radii of the atoms in the series we should expect to have e f,—1,—=1,—!,—1, — 1, — ete. A,, A,, A;, etc., denote the atomic weights and suppose B4,=n,, BA,—n,, etc., where 8 is a constant. Then we have Sain =H = PUA, $27 mis? = Mm.V = BUA m,- 47 (45) ¢ fatio —r,)= A on —A m=, iS haee C is a constant which should be the same for all series of similar elements. Also (Tm,,;—1Tm)*—v approximately, so that According to the theory therefore we ought to be able to find the mber of electrons per atom from the atomic weights. In the figure the values of 4? for series of similar elements are fed against the order of the elements in the series. For some es a constant has been added to the values of A3 to prevent the _ different lines falling too close together. It will be seen that values of A for each series fall nearly on straight lines id that the different lines are nearly parallel. This shows that mat — Am 3==C is nearly constant, as was to be expected from the heory. The mean value of C is 0.81. Hence we get B=8 so that he number of electrons per atom comes out 8 times the atomic ght in all cases. : This estimate agrees as well as could be expected with the num- , deduced from the optical properties of the elements which ht be expected to be too low. _ We have assumed that the electrical density of the positive spheres is uniform so that the approximate agreement of the atomic ghts with the theory confirms this assumption. It is easy to that the arrangement of the electrons in the positive sphere is affected by a change in the size of the sphere provided its density ains uniform and its total charge the same as before. It is 372 WILSON—CONSTITUTION OF THE ATOM. [April 22, tT tf :: es shit : : att $3 YY 33 H 3 seek } H Ott } i =: = :: Hy oe HH *-, +- - v = E - E i if vo E _ rt < H : ““ Hy ° E LE pes i H Stitt ° c i } I +42 % % vie - ee EEE Ve > Ss i Sessesess 3 - 7 ose seseeris PA xe E HH } “a e3 E ti = — r +t t q sesse te ae Sitttts ie TI e A . i E Pettit i sbeanecs: t t x 44 : ae” gees - sete rate? es t $ cone po tr ree he ps +444 3. oe +4 HH aH : oS ts $33 : iP 68 4) fi Hi |: iH HHT Ret + ( : . Hf Ae ¢ bess He ttthtit tis bid i= TS oe 4 7 — — ‘i if HT TH HHH HH i ati sy i} 3 / H H St E Sits? of CERELSSEES teste sreds cattesesss rr 7; 7 : : E 1 a i Hatt H aa nee’ of t E os 4 fs = = “sia? : + +4 + cy + sees t HH . Het HST Eethesey tees Hesty Tt Hs : ares “ee iy sseise Rutt sossees ttt Tay boas stoosed (bec 2ctie | sit tt est Eskssheats unosesy Tiytt Oo +H tee wis 9 She 4 tit Order in Series. possible therefore that the addition of new layers increases the density of the positive spheres instead of increasing their size. If. WILSON—CONSTITUTION OF THE ATOM. 373 is were so the calculation of 8 given above would not be affected can be easily seen. The equation (42/38)3=Ami —Am_,3 gives tam (ar+ee—0(S5)) equation enables the atomic weights of a series of similar nts to be approximately calculated if that of the first in the is known. For example, if we take A,=1 we obtain the m F. BE I I H==3 2 6 Licey 3 18 Na=.23 4 40 K= 39 5 77 Rb=85 6 129 Ce 2542 ve take 4, =9 we obtain the following numbers: m As I 9 Be==9 2 24 Mg=24 3 51 Ca=40 ' 92 Sr== 67 5 150 Ba==137 6 230 Ra= 226 will be seen that the numbers given by the approximate a deduced from Sir J. J. Thomson’s theory agree approxi- y with the atomic weights. I think this must be regarded as x evidence that there is a considerable element of truth in THE HIGH VOLTAGE CORONA IN AIR. By J. B. WHITEHEAD. (Read April 21, 1911.) The term “corona” as employed by electrical engineers refers the luminous envelope which surrounds a bare electrical condu when its potential is raised above a certain value. As the voltage long-distance transmission lines has been raised to higher and higher values in order to reduce the size and cost of the conductors and s increase the distance of economical transmission, a limiting conditi has been found in the insulating properties of the atmosphere. Fo each definite space separation and size of conductors, above a cer value of voltage the regions immediately surrounding the conduct become luminous, and a power loss sets in which increases rapidly with further increase in voltage. These facts were first noted by electrical engineers in this coun tr in 1896. It was promptly recognized that the region in the im me diate neighborhood of the conductors is subject to the greatest elec intensity and that the phenomena are due to local though restri break-down of the air. This was corroborated not only by the ence of the luminous envelope immediately around the condu for voltages above that at which the loss begins, but by study o effect of changing the size and separation of conductors; decre separation and size both increase the surface electric intensi therefore lower the voltage at which loss begins. The electric sity at the surface of the conductor may be readily calculated i cases that occur from the voltage and from the separation an of the conductors. It is directly proportional to the voltage in cases. The term “corona” was first used by Steinmetz in 1898 describe the luntinous envelope and has been generally adopte engineers, Many measurements have been made on electric power transn 374 1gtt.] WHITEHEAD—HIGH VOLTAGE CORONA IN AIR. 375 sion lines in efforts to determine the law connecting the voltage at _which loss begins with the physical constants of the line. These measurements have shown marked inconsistencies among themselves, the results on the same lines on different days being often at variance. A number of laboratory investigations, in which the widely varying conditions of a transmission line are under control, have naturally followed. They have indicated with a rather wide variation in nu- merical values that the critical voltage or voltage at which corona begins on round wires varies inversely as the temperature and directly as the pressure; also that the electric intensity under which the air near the surface of the conductor breaks down has not a constant value but increases markedly for conductors of small diameter; and _ further that the value of the intensity at which break-down begins is _ that corresponding to the maximum value of the alternating wave, and is independent of the material of the conductor. The general nature of the influence of temperature and pressure could probably have been predicted from numerous investigations of the discharge of electricity through gases; the quantitative relations for pressures near that of the atmosphere do not, however, appear to have attracted the physicist, nor indeed have they as yet been satisfactorily determined for the voltage of corona formation by _ experimental engineers. The accumulated results of physical inves- _ tigation and theory, however, offer no obvious explanation of the ‘rise of the critical surface intensity for smaller wires, nor of the influences of the form and frequency of alternating voltage. The fact that the corona voltage is that corresponding to the maximum value of the alternating wave has been proven by stroboscopic methods and by the use of distorted wave shapes. It indicates that the time element involved in the process of break-down of the air is short compared with the periods of the common alternating cur- rent circuits. The apparent sharpness of the connection removes many objections to the use of the alternating electromotive force as a means of investigation, and renders available its many advantages. It is only necessary to know the shape of the alternating wave and this may be obtained readily by several well known methods. Effec- tive values as read on direct reading instruments may thus be used 376 WHITEHEAD—HIGH VOLTAGE CORONA IN AIR. [April 21, and the corrective factor for the maximum value is obtained from the shape of the wave. i Many of the inconsistencies among the measurements on existing _ transmission lines and those made in laboratories arise in the difficul- — ties of measuring the power in high voltage circuits ; the instruments — must be placed in the low voltage side of transforming apparatus, the — losses in which, being generally greater than those to be measured, — introduct a troublesome source of error. The appearance of the — visible corona has been used by laboratory workers as an indication — of the beginning of loss through the air. With proper precautions — this method may be very reliable but its use is generally attended by — danger of subjective and other error. As a result of the discrepan- _ cies among these approximate determinations of various investiga- tors, there has appeared much speculative suggestion of the presence _ of other unrecognized influences, as for example the moisture con- tent of the air, the presence of “free” or natural ionization, an ab- E normal property of air when near a small wire, etc. : The present problem therefore resolves itself into two parte: first, a satisfactory method for the determination of the law under which the air in the neighborhood of a long straight and usually — cylindrical conductor breaks down under electric strain; and second, the law governing the amount of loss when the voltage is carried above the critical value. A year ago the writer’ described a method by which it is possible to determine the voltage at which the breaks down near a round wire to’a maximum inaccuracy of a few tenths of one per cent. The original paper may be consulted for the details, but the principle is simple and may be described briefly. The wire is stretched along the axis of a metal cylinder and t voltage is applied between them. Air may be passed through the cylinder by means of two lateral tubes near the ends, the walls ¢ the cylinder at these points being drilled with a number of r | holes. Close to one set of these holes and outside the cylinder a wire — mesh electrode connected to a sensitive electroscope is placed. soon as the air around the wire breaks down under increasing volt- age, copious ionization sets in which causes a rapid leak from the *J. B. Whitehead, Proc. A.J.E.E., p. 1059, July, 1910. 1911.) WHITEHEAD—HIGH VOLTAGE CORONA IN AIR. 377 charged electroscope. The initial discharge of the electroscope is very sharply marked. Observations may be repeated at will and * Fic. 1. Arrangement of apparatus. and of temperature on the electric intensity at which atmos- air breaks down. The experiments on temperature, moisture n ent and diameter of conductor are given in the paper mentioned ibove. The results of the remaining investigations are first given No attempt is made to describe the details of the experiments. For these the reader may refer to the earlier paper and also to one hortly to be presented to the American Institute of Electrical Engi- in which the practical bearing of the results will be discussed. nfluence of Diameter of Conductor—For convenience of refer- a-condensed table of the results on this portion of the work 378 WHITEHEAD—HIGH VOLTAGE CORONA IN AIR. [ is given in Table I. and the results are plotted graphically in F For comparison, points observed by other investigators are shown. The values are all corrected for temperature, pressure wave form and give the maximum values of the electric ir at which the air breaks down under a pressure of 760 mm. TABLE I. Sete RELATION BETWEEN DIAMETER AND CRITICAL SURFACE INTENSITY Diameter, Material of | Diameter of| Material of | Critical Pri Ratio cm. Wire. Tube, cm, Tube. Volts Co: u : 0.089 Copper 4:9 Brass 74.5 125. 0.122 ie is J 44.0 250.1 be s< “ “c 87.8 : 125.09 oO. 156 ““c “e “cc 97.0 “ce “ec “ec “6 “ec 97.0 ““ 0-205 6“ “c 6“ 109.5 “ec “cc sé “ee ““c 55.0 250,18 “ “cc 6.35 “cc 60.5 “ “sc “cc “é “ec 60.2 “ec “c “ “e ““ (8) 59-9 “cc 0.254 Aluminum “f " 65.9 se 0.276 Copper at ee 68-9 t¢ ““ se sé “< 68.8 “cc “é “6 “é “e 68.5 “e se “é “ec “c 69.0 “ “é “cc “ “<“ (8) 68.9 “ “ . 9.52 Steel 77-3 “ 0-325 Aluminum 6.35 Brass 72.8 $f 0.347 Copper 6.35 ne 75-13 : © 3405 a 9.52 Steel 85.7 ae 0.399 Steel ec “ec 92.5 “e 0.475 “cc “cc a3 100.6 “cc mercury, and at temperature 21° C. I am indebted to P. Alexander Russell for pointing out that these results obey simple law. If E be the critical electric intensity in kiloy centimeter and d the diameter of the conductor in centin curve of Fig. 2 obeys closely the equation: E=32-+ 13.4 1/Vd. The observed values of Table I. are compared with the v culated from the above formula in Table II. The percenta; is also given, and it is seen that with one exception the diffe is well within one per cent. The exception refers to an 2 wire which could not be polished to a clean surface; a rou igtt.] WHITEHEAD—HIGH VOLTAGE CORONA IN AIR. 379 TABLE Il. Kilovolts per Centimeter. Diameter, cm. Difference, Per Cent. Calculated. Observed. 089 76,950 77,100 + 0.19 122 70,400 70,875 + .67 -156 65,950 65,880 ae .205 61,600 61,680 + 13 254 58,600 58,750 + 25 .276 57,500 58,000 + 87 325 55,600 55,000 — 1,08 -340 54,980 55,100 + 21 -347 54,780 54,500 — sf -399 53,230 53,050 — .23 _ 475 51,460 51,400 — Wl 80 ~ 70} a ~ 5 aN = ol PLA = Sal |, & Ps on : ty. Bigg AC. # 50} a Yn & o T=-~—1 D.C 2 = 8 49 ¥ “ > . a ee) = SA 3 20} — Fees: a E wu 2 a o-RYAN ec @-MERSHON 2 eee Bey 10 1 2 3 4 5 6 DIAMETER MILLIMETERS Fic. 2. Relation of critical intensity and diameter. 380 WHITEHEAD—HIGH VOLTAGE CORONA IN AIR, [April 21, physical investigation of the nature of the process involved in the electrical break down of the air. The formula indicates that the value of electric intensity of a uniform field or near the surface of a plane conductor, at which air would break down is 32 kilovolts- per centimeter. The work of von Scweidler, Townsend and others - indicates that at about 30 kilovolts per centimeter, secondary ioniza- tion or ionization by collision sets in between parallel plates sub- jected to a difference of potential. In a later paragraph several other experiments are described which indicate that the start of the corona in air is due to secondary ionization. So far as the writer is aware however no theory has been advanced to explain the influence of the curvature of the conductor. That the nature of the — molecular structure of the air is concerned there can be no dou but the variation of values of critical intensity occurs within a range of diameter many orders of magnitude greater than molecular dimensions, and is related to the diameter in a way which offers no suggestion of explanation. Effect of Stranding the Conductor.—It is quite obvious that the surface intensity is the determining factor in the voltage which corona occurs, then a stranded conductor should have — critical voltage lower than that of a solid conductor of a diameter equal to that of a circle tangent to the strands. On the other it is not obvious that the critical voltage of a stranded conduc would be less than that of a solid conductor of equivalent cros: section, for the diameter of the latter will always be less than that of the enclosing circle of the former. Evidently also the rela tions will vary with the number of strands. The question is importance since all of the larger transmission lines consist cables or stranded conductors. : A series of observations was made on a number of cables stranding ranging from three to nine conductors uniformly fil the outer layer. The interior space was filled with a single wire several wires of suitable size, but in each conductor the wires of outer layer were all of the same size, .162 cm. diameter. The ca were clean and smooth and drawn tight along the central axis of outer cylinder of the apparatus. The results are condensed in T III, in which comparison is made between the diameter of a solid — 191t.] WHITEHEAD—HIGH VOLTAGE CORONA IN AIR. 381 TABLE III. Strands, (a) (2) () | (e) | (ea) _| Spiral Pitch, 3 -349 .272 .247 1.10 .708 | Tog 4 404 332 32 1.04 792 | 8.65 5 45 381 -37 1.03 822 | 9.89 6 49 .430 42 1.026 857 12.3 7 541 48 465 1.032 .868 12.3 8 589 53 516 1.027 877 10.78 9 .64 581 .567 1.025 886 10.9 3 -336 -27 -207 1.305 616 none 4 378 312 25 1.248 665 none circular conductor having the same critical voltage as that observed for the cable (column c) and the diameters of the circle just enclos- ing the cable (a) and that given its equivalent section (bd). The ratios b/c and c/a are given in the last two columns, and are plotted in their relation to the number of strands in Figs. 3 and 4 respec- tively. Fig. 3 indicates that if instead of a solid conductor a slo 4 as, | N are ~~ c & a | | 3 aS a) é 7 s Ss Numeer of Srranos. Fic. 3. voltage will be lowered, but by less than three per cent. if the num- ber of strands is greater than 5. For a three-strand cable the lower- ‘ing is ten per cent. The ratio c/a is more important however. This ratio compares the diameter of a solid conductor having the same critical voltage _as the cable, with the actual overall diameter of the cable. It there- fore refers the behavior of a given cable, with regards critical voltage, to a solid round wire whose diameter is expressed as a fraction of the overall diameter of the cable. This is a more logical 382 WHITEHEAD—HIGH VOLTAGE CORONA IN AIR, [April 21, | a basis of comparison than the other since the interior of a multi- — strand cable may be made up in such manner as to cause a con- — siderable variation in its cross section. In fact many transmission cables have centers of hemp, or other material, the entire conducting - ; section residing in the single outer layer of strands. Thus Fig. 4 _ shows that a three-strand cable has a critical voltage which is that of a single wire of seven tenths the overall diameter of the cable. At nine strands the equivalent sage diameter is still less than 9 that of the cable. In a stranded conductor the strands are always spiralled. The pitch of the spiral for the cables described above is given in Table III. The spiral arrangement of the strands tends to lessen the value of the electric intensity on the outer surfaces of the strands since the equipotential surfaces are rendered more nearly cylindrical about the axis of the cable. The values of maximum surface electric intensity for cables of various numbers of strands and in which there is no spiral may be computed from an expression given by Jona? and due to Levi-Civita. This expression involves a hyper geometrical series whose evaluation requires some labor. As makes no allowance for the spiralling of the strands no deduction may be drawn from the present observations as to the actual inten- sity at which corona occurs on the stranded conductor. Values deduced from the expression should, however, be of great value the study of the nature of the breakdown of the air when taken conjunction with measurements on cables without spirals. For i these cases the maximum electric intensity at the outer edge o strand would obtain over a narrow circumferential distance, wh the same intensity reached at the surface of a single wire obte over a whole circumference. A comparison of corona voltages the two cases should throw light on the distances involved in process of secondary ionization and kindred phenomena. | At the bottom of Table III. there are given the results of obs vations on a three- and a four-strand conductor in which there v no spiral. The size of the strands was the same as that of the foregoing cables. The strands were carefully straightened, polish ? Jona, Trans. Int. Elect. Congress, St. Louis, 1904, Vol. IL., p. 550. 2971-1 WHITEHEAD—HIGH VOLTAGE CORONA IN AIR. 383 a and built up by soldering with a fine blow flame so that the strands a were uniformly tangent to each other throughout. The results indi- 3 ; cate the further lowering of the critical voltage when spiralling is ' absent. The ratio c/a falls from .71 for the spiralled three-strand | E. to .61, and the difference for the four-strand is somewhat greater. _ The pitch of the spirals of the cables investigated does not appear _ to follow any regular rule. This irregularity however does not appear to have any corresponding effect on the points of the curve oe wo & RATIO <= N 34 SN 3 4 7 9 NumsBer of STRANDS. Fic. 4. _ of Fig. 4. From this it may be concluded that for a pitch of spiral _ less than twelve diameters there is no gain on the ground of lessened ‘Surface intensity due to the more uniform distribution of the elec- tric field. At this writing the author has been unable to obtain solutions of __ Levi-Civita’s expression as applied to three and four strands. These would permit by the foregoing results a knowledge as to how the Maximum corona intensity for a round wire compares with that 384 | WHITEHEAD—HIGH VOLTAGE CORONA IN AIR. [April ar, , at the surface of the same wire when made up into a three- or four-strand cable without spiral. Influence of Frequency and Wave Form.—By the use of cathode ray oscillograph in the high voltage circuit Ryan in 190. showed that the appearance of corona was accompanied by a h or peak on the charging current wave in the neighborhood of maximum of voltage. The writer by stroboscopic methods shown that the corona is periodic, appearing every half cycle that its first appearance with rising voltage coincides accurately w the maximum of the voltage wave. Also the duration of the coro: with steady circuit conditions, may be reduced with lessening volte to a very small fraction of the period of the alternating electromoti force. Thus a corona which was found to exist for only oni twentieth of a period at the crest of the voltage wave of a 60 circuit was plainly visible in a darkened room. It is evident, the fore, that the interval of time involved in corona formation < cessation is extremely short. For these reasons it has been supp Os that the appearance of corona depends only on the maximum of voltage occurring in the cycle, and is therefore independent the frequency. Experience with existing lines indicates that if is an influence of frequency it is small for the range between and 60 cycles. The closeness with which the critical voltage m be read by the method described gave promise of discovering ¢ comparatively small differences due to variation of frequency. ‘Sev eral series of tests were therefore made with different sizes of | The observations are not recorded here as the points on the c of Fig. 5 are a sufficient indication of their accuracy. The r from 15 to 90 cycles was obtained from two generators, and voltage from a 10-KW. 25-cycle 100,000-volt transformer. T transformer had also a low voltage secondary coil. On the ct the values of voltage are those measured at the terminals of coil; these values are therefore proportional to the voltage in high tension winding and therefore to the electric intensity at surface of the wire. These observations were made with rods cm. and .635 cm. in diameter placed at the center of a pipe 120 long and 30 cm. in diameter. The observations were taken | rgtt.] WHITEHEAD—HIGH VOLTAGE CORONA IN AIR. 385 continuous set, interruption being necessary for only a few seconds to change generators. There were consequently no appreciable _ Variations in temperature or pressure. ____ The results as taken are plotted in the lower curves of Fig. 5 _in which observations for ascending and descending values of fre- _ quency are plotted as crosses and circles respectively. The irregular shape of these curves repeated itself accurately in experiments over o fx nw = —j—| Be: 2 Fin te 2 ey et Pay, 28 Z m™L a 4 Lo rR WES : EFF VE youTs -3f RAL - -7iecm, OAM. g s°B)- Ss |so xs 2a - es ‘ inte so oO Soni Be Sees ae ~~ by oe Ee See, ae a —~_! ; _—s Be 2 oa 40 60 i ae GO Creies PER Second Fic. 5. the same range of frequency with other wires. Since the trans- former was operating over a wide range of frequency at approxi- mately the same value of voltage, and its magnetizing current was therefore variable, a variation of wave form due to the armature Teaction of the generator appeared probable. Oscillograms were therefore taken of the voltage at the terminals of the low voltage secondary coil at frequencies 20, 35, 55, 60, 65 and g1 cycles, and at transformer excitation corresponding to 50 volts on the same coil. The ratios of maximum to effective values of these waves were __ PROC. AMER. PHIL. SOC., L, 200 Y, PRINTED AUG. 5, IQII. 386 WHITEHEAD—HIGH VOLTAGE CORONA IN AIR. [April ; then determined by micrometer measurements of ordinates take every 7.5 degrees over two half waves. The several values of “a : ratio so obtained revealed a minimum at 55 cycles thus expla the rise in the lower curves of Fig. 5 at that frequency. ie upper curves the points indicated are the voltage of the lower ct multiplied by the ratio of maximum to effective value as calcula from measurements of the oscillograms for the correspond frequencies. The upper corrected curves of Fig. 5 show a lowering of critical voltage with increasing frequency. The result leaves some- thing to be desired in the accuracy of location of the points upon the curve. It should be noted however that owing to the me cation of the scale, the error of the points off the upper curves the 25-cycle portion of the lower curves is only about 1 per c Several other sets of observations for different sizes of wire rey curves of the same general characteristics. The measurement of possible error of this nature. The curves therefore show wi fair accuracy the nature of the variation of the critical voltage mercial frequencies 25 to 60 cycles per second, is only ab per cent. Influnce of Pressure——The influence of pressure on the va forms of spark discharge has been closely studied. Paschen’ states that the sparking potential for a given spark length is proportional to the pressure; his investigations covered the of pressure between 10 and 75 cm. of mercury. " Carr* has s that this linear relation extends down to pressures of a iat n meters if the spark lengths are not greater than I cm. but does. obtain for lower pressures. Townsend® has shown that the pot * Paschen, Wied. Ann., XXXVII., 79, 1880. “Carr, Proc. Roy. Soc., LXXI., 374, 1903. *Townsend, Phil. Mag., V1., 1, 198, 1901. WHITEHEAD—HIGH VOLTAGE CORONA IN AIR. 387 gradient at which secondary ionization sets in when electricity is passing through a gas is directly proportional to the pressure. Wat- ‘son® investigated the spark length between spheres up to fifteen ‘atmospheres and found that the spark potential increases with the ‘pressure in an approximately linear relation. From the general similarity between the corona and the brush form of spark dis- charge, therefore, a linear relation between pressure and critical ‘surface intensity, or the potential gradient at which corona begins ; to be expected. Apparently the only study of the influence of ressure on the formation of the alternating corona is a single set f observations by Ryan’ on a wire .32 cm. in diameter placed at the enter of a cylinder 22.2 cm. in diameter. He observed the alter- ting voltage at which the visible corona appeared for the range of pressure between 45 and 90 cm. of mercury; the alternating frequency was 130. The resulting linear relation is given as between _ the kilovolts K actually applied and the pressure in inches of mer- iry, K=2.93 + .g02b. | _ In Table IV. are given the results of a typical series of observa- tions on the influence of pressure on corona voltage; the values are those for a wire .152 cm. in diameter. The wires were clean and ight and centered accurately on the axis of the outer cylinder the apparatus which has been briefly described. This cylinder a diameter of 9.52 cm. The ends were closed with ebonite ps of the same diameter and 5 cm. deep. The side tubes were 0 closed by caps, and the leading-in wire to the discharge electrode ssed through a column of sulphur supported in hard rubber; no roubles with either insulation or air leak were encountered with this “arrangement. All joints were sealed with a mixture of bees wax d resin and pressures between 30 and 100 cm. of mercury were iched without trouble. The discharge electrode was placed inside _the upper side tube and within one or two millimeters of the grating lormed by the holes drilled in the outer cylinder ; in the earlier work t was found that a flow of air from the cylinder over the electrode contributed little to the sharpness with which the condition of _ * Watson, Electrician, 62, 851, 1909. "Ryan, Proc. A.J.E.E., XXIIL., 101, 1904. 388 WHITEHEAD—HIGH VOLTAGE CORONA IN AIR. L breakdown was indicated, the initial discharge of the elects occurring at the same value for both moving and stationar The results of Table IV. are plotted in the lower line of F: TABLE IV. Manometer, Crit, Prin. Volts. Ratio, 1: 125. Right Left, Diff. 102.2 102.2 102.2 487.5 587.5 | —100 97-5 97-5. 97-2 459-5 605.5 146 91.3 91.3 91.2 427 628.5 201.5 87.2 87.5 87.8 407.5 642 234-5 83 83.2 83.4 386.5 656 269.5 79-9 80 80 367.5 | 669.5 302 80.5 80.7 80.6 371 666.5 295-5 74 74 74 340 688.5 . 68.1 68.1 68.1 313.5 707 393-5 94-2 94-2 94.2 439 617 178 106.5 106 106.2 499 576.5 77-5 114.5 114.9 114.8 545-5 545-5 Ratio 1: 250. 57-5 57-4 57-5 545-5 545-5 ° 59-3 = 59-9 59.8 570.5 | 5305 | + 40 61.8 61.6 61.7 592 516.5 75-5 64 64 63.8 618.5 499-5 66 641 486 67-7 67.7 67.7 661 473-5 69.8 70 69.9 687.5 | 457 71.6 71.6 71.7 710 444 between the values of voltage at the primary terminals of former and the pressure in millimeters of mercury. Thi is directly proportional to the corresponding value of pe gradient at the surface of the wire. The ratios of transfor were I to 125 and I to 250, the frequency 60, and the rai oO maximum to the effective value of the alternating wave of motive force, as measured from an oscillogram as already was 1.46. The temperature was 24° C. The results for wire are also plotted in Fig. 6. The equations of the lines in Fig. 6 have no significance since they apply to a partic bination of wire and outer cylinder. The values of surface po gradient have therefore been calculated from the expressic dV £E © r log— gtr.] WHITEHEAD—HIGH VOLTAGE CORONA IN AIR. 389 _ in which £ is the maximum value of the potential difference between | wire of radius r and outer cylinder of radius R, and which in this "case is the effective voltage multiplied by 1.46. Expressed in terms of electric intensity at which corona begins, in kilovolts per centi- ae ees ? 4 Fé iL mt ¥ LZ fg oe a 77 SF; Sg P id e “a ve i ae a Fa od pz Ls Pf At. Z , 321 >< = 4 Cx pe s PRESSURE (mm. Hq.) Fic. 6. meter, and pressure in centimeters of mercury, the equation for the 52 cm. wire is: = d(KV Ae = aS 15.2 + .673p, (3) and for the .276 wire: SAY) = 11.6 +595 f, (4) 390 WHITEHEAD—HIGH VOLTAGE CORONA IN AIR. [A While both equations are linear it is seen that the slope of for the smaller wire is the steeper, that is that the variation | critical surface intensity with the pressure is greater the smaller wire. It is interesting to note that the values at 76 cm. pressut 66.2 and 57 correspond extremely closely with the values 66.4 57-7 observed a year before and so calculated from the bas ic of Fig. 2. If Ryan’s results for a .317-cm. wire be expressed in oe terms used in the above formulae, the resulting equation o line is: HRP) 6354 aap The slope of this line is greater than that of either the .152-c1 the .276-cm. wire as expressed in equations (3) and (4), alt the larger size of wire should cause the slope to be less; also tl initial constant term is considerably less; further the va ue ) critical surface intensity at 76 cm. pressure indicated ge f (5) is 62.6, while that calculated from formula (1) and t frequently observed by the writer is 55.7. Ryan used iny the visible corona for indication of initial breakdown; some Q results on wires of different size are plotted as circles in where they are seen to be very irregularly located. Aside fr uncertainty of the method of observation, the wave form an quency may have introduced considerable error in the results | reported, although that due to frequency would have tende lower rather than a higher value than for 60 cycles. a Further experiments on the variation of the pressure equ with the size of wire are in progress. : Influence of Temperature and Moisture—No satisfachaae i tigation has been made of the influence of temperature on voltage. Ryan reports a series of observations on the visible ca for temperatures between 70° and 200° Fahrenheit. The size wire is not stated. The results are admittedly wanting in acct but indicate a linear relation between corona voltage and tem ture; in fact, Ryan states that the maximum value of corona varies inversely as the absolute temperature. t911.] WHITEHEAD—HIGH VOLTAGE CORONA IN AIR. 391 The writer has conducted a short series of tests between 6° and 41° C. on a .27-cm. wire for the purpose of obtaining a correction factor for his various observations as taken at different tempera- _ tures. The result as stated in the paper already referred to is that - the relation is linear and that for each degree rise or fall from 21° _C. there is a lowering or raising in the value of the critical voltage of 0.22 per cent.; Ryan’s results indicate 0.27 per cent. for this value. Expressed in terms of surface intensity in kilovolts per centi- meter and temperature in degrees Centigrade the writer’s results may be expressed by the formula: | KV./Cm. = 61 —.132¢. (6) n view of the observations of the effect of variation of pressure on different sizes of wire, it is not improbable that the constants of equation (6) will also vary with the size of wire. Further investi- gation in this direction is therefore desirable. ve Moisture content up to amounts quite close to saturation have no effect on the values of voltage at which corona begins. While there is still some dissent from this opinion among electrical engi- “neers, the author’s results on this question, described in the earlier paper, appear very conclusive, and have been widely accepted. An influence of moisture on the amount of power loss above the critical voltage appears quite probable, in the light of the ionization theory in which the mass of the ionic carriers, which make up the current / are an important factor in its value. DISCUSSION. ’ So far as the question of the value of voltage at which corona will start on a given transmission line is concerned, it is probable that a solution will be reached sooner or later by means of experi- “ments of the general character as those described above, supple- mented by observations on existing lines. Also, there is good reason to suppose that a comparatively simple law will be found. For the irface intensity for any arrangement and size of cylindrical con- luctors, corresponding to a given voltage, may be expressed in terms of these constants; and the critical or corona intensity, under stand- 392 WHITEHEAD—HIGH VOLTAGE CORONA IN AIR. [April a1, ard conditions of temperature and pressure, is a simple function of the diameter of the conductor. The relation between pressure | within the range of the atmosphere, and critical voltage, for a given size wire, is linear; and although the slope of the linear relation changes with the size of wire there is good reason to suppose that — a simple law connecting them can be found. Much the same may be said of the influence of temperature; preliminary experiments show ing that the linear relation exists over a fairly wide range. The effect of stranding the conductor has been studied for only one siz of strand as yet, but it seems .a simple matter, with some further investigation, to express the effect of each of these influences in terms of the diameter of the conductor. . The influence of the frequency does not offer promise of expres- sion as a simple relation; this influence is small however within the limits of frequency met in practice. The state of the atmosphere — appears to be of small importance, for moisture does not influence the critical voltage, nor does its state as regards ionization, as is indicated by several considerations given in a later paragraph. Dir and impurities which on settling cause irregularities on the surf of the wire, may lead to localized brush discharges; and if th are sufficient in number they may cause a noticeable loss ae: normal critical voltage. It is of great interest, however, to consider the results in th relation to present theories of the nature of the electric conductiv and breakdown of a gas. It is assumed that the reader is fa with the general features of the theory of ionization. Under theory the neutral atoms and molecules of matter may be separa into smaller charged particles, and the motion of these particles un electric force constitutes an electric current. In a gas there always a small number of these free ions present; this number maj be greatly augmented by Rontgen rays, ultra-violet light and ot! well known ionizing agents. When so ionized currents of mag tudes within easy measuring range are obtained between term subject to a difference of potential. If this difference of potential i increased, a point is reached where the current increases sharf showing the presence of some new saurce of ionization. The the 1911.) WHITEHEAD—HIGH VOLTAGE CORONA IN AIR. 393 states that these new ions are formed by the impact of those already existing, and moving with higher velocity in the increased electric field, with the neutral molecules of the gas. This phenomenon has been called ionization by collision or secondary ionization. _ The results of the experiments which have been described above are for the most part consistent with the ionization theory. The various circumstances surrounding the appearance of corona all indi- cate that it is an instance of secondary ionization. Formula (1) ‘indicates that near a conductor of large radius or near a plane, the corona intensity approaches a value 32 kilovolts per centimeter; secondary ionization Between plane electrodes in closed vessels at _atmospheric pressure has been noticed by several physicists to begin in the neighborhood of 30,000 volts per centimeter. The mass of elementary negative ion or electron is approximately 5.9 X I0-* gms. and the charge it carries is 4.6 X 10° electrostatic units. In an electric field the mechanical force acting on the electron is the product of its charge and the strength of field. Hence by the laws of simple mechanics it is possible to calculate the acceleration, the velocity and the kinetic energy attained by an electron in moving a given distance under a given electric intensity. If the mean free path of the electron, about 6 X 10° cm. at atmospheric pressure, be the distance between collisions, it is thus easy to calculate the kinetic energy of the electron due to the electric field, when it collides with amolecule. This energy is readily seen to be equal to pVe, where p is the mean free path, V the electric intensity in electrostatic units, and e the charge of the electron. If now the voltage between plane parallel electrodes be raised until secondary ionization begins, the value of the voltage makes it possible to calculate the energy re- quired to ionize a molecule of a gas. In fact the values of the energy required to ionize a molecule which are now generally ac- cepted are largely based on determinations of the value of electric intensity at which secondary ionization begins. It has been pointed out above that the values of this intensity as determined by Town- d and others are in close agreement with the value 32,000 volts per centimeter indicated by equation (1) as the lowest value at which corona appears. To one skeptical as to the correctness of the theory _ of ionization therefore (and there are many such) all that may be 394 WHITEHEAD—HIGH VOLTAGE CORONA IN AIR. [April ba said so far is that the phenomena of sudden increase of current above | a certain value of electric intensity as observed by Townsend, and — that of corona formation, are probably due to the same causes. EB there are several other independent methods of determining the energy required to ionize a gas. The values are commonly ex- pressed in terms of the potential difference in volts through which — the electron must pass in order to acquire energy sufficient to pro: duce an ion by collision. .The value pertaining to the method d scribed above is from 10 to 12 volts. Rutherford, from the relati between the heating effect of radium and the number of ions it p1 duces, gives the value 24 volts. Stark and Langevin by independe methods conclude that the values are 45 and 60 volts respective While the extreme values differ by the factor 5 or 6 it must be membered that the actual amount of energy required to produce ar ion is about 5 X 107" ergs, so that all of these values indicate t same order of magnitude; therefore when taken together they cc stitute a very strong reason for supposing the value 5 X 10" erg: is close to the correct one. If this be true it is good evidence that formation of the corona is actually due to the liberation of ions the neutral molecules of the gas, when the latter suffer collision a free electron moving under the force of the electric field. 1 the electron and not a gaseous ion or aggregate is the active is shown by the shorter free paths of these latter which © relation already given results in a lower value of kinetic ene the time of collision than those given above. The writer has shown by stroboscopic methods that abou : critical voltage the corona begins and ends at a point on the nating current wave which corresponds very closely in every with this critical value. It is well known that since secondary zation depends only on the velocity of the ions and thus electric intensity, it should within wide limits be independ the number of ions already existing in the gas. The corona sharply on the descending side of the voltage wave showing the copious ionization present during the existence of corona not aid it in persisting to a lower voltage than that at wh starts. The presence of a greater or less amount of free or s taneous ionization in the atmosphere has been advanced by s - -agtt.] WHITEHEAD—HIGH VOLTAGE CORONA IN AIR. 395 writers to explain the discrepancies, among different observers, in the voltage at which the corona starts. The foregoing facts seem _ fairly conclusive that this supposition is not correct. In order, however, to further remove doubt on this point a simple experiment _ was performed in which the air surrounding the conductor was ionized from an independent source. A clean polished wire 15 cm. _ in diameter was stretched vertically along the axis of a cylinder 17.5 _ €m. in diameter and about 120 cm. long, made of woven wire with a Icm. mesh. The high voltage was applied between them, the wire cylinder being also connected to ground. A large Rontgen ray tube was enclosed in a light-tight box and placed close ‘to the cylinder. When this tube was excited a crude electroscope placed 20 or 30 em. on the other side of the cylinder was immediately discharged showing that the air of the neighborhood was strongly ionized. In the darkened room the starting of the visible corona on the wire could be located readily and the corresponding voltage determined _ by successive trials within an error of two or three tenths of one per cent. By the use of independent observers it was established without doubt that the presence of the R6ntgen ray tube caused _ no variation in the value of voltage at which the corona starts. _ The general influence of a decrease in pressure or an increase in temperature toward a lower critical voltage is quite consistent with the ionization theory. For under the kinetic theory of gases the free paths of the vibrating molecules and ions are lengthened in these two conditions. During the free path or interval between collisions the ions are acted on by the electric force, and the longer the interval the greater the velocity acquired and the more kinetic energy and ionizing power. Hence a given amount of energy will be acquired at a lower voltage if the free path is lengthened. __ The lowering of the critical voltage by an increase in frequency is not to be explained so simply. However if within the molecule or atom there are a number of electrons in motion or free to move, and there is some indirect evidence to this effect, it is evident that the forced vibrations set up by an external alternating field will, with the increasing frequency of these vibrations, cause the mutual attractions within the structure of the atom to become less and less strong, and therefore more liable to be broken when in collision 396 WHITEHEAD—HIGH VOLTAGE CORONA IN AIR, [April 21, with an extraneous ion. It is surprising however that this effect should be noticeable at frequencies so low as 60 to 90 cycles, for they are incomparably slower than those suggested by theory for the vibrations within the atom. The close relation between the first appearance of corona and the peak or maximum of the voltage wave is natural in the light of theory, for at atmospheric pressure the mean free path of an electron is about 6 & Io cm. long, and under a field sufficiently strong to ionize this path is traversed in about 2 X 107? seconds. Perhaps the most interesting problem in connection with the phenomenon of corona formation is the explanation of the greater — values of electric intensity required to start corona around smaller oa wires, 7. e., the upward trend of the curve of Fig. 2. Why should the a properties of the air change with a slight alteration in the size of a a conductor whose diameter is fifty thousand times as great as the © mean free path of a molecule? No tenable explanation has been ~ offered. The attraction to the conductor of oppositely charged — ions which pile up as it were and reduce the actual gradient below that calculated, and at the same time increase the gas pressure, has been suggested. Both suppositions immediately include an in- fluence on corona voltage of the amount of ionization already present, and this as already noticed is contrary to observation. Simple calculation also will show that the charge sufficient to mate- rially reduce the gradient at the surface of a conductor at corona potential would require a number of ions far in excess of the num- bers commonly present in the atmosphere. The writer by a sensi tive optical method could find no indication of an increase of pres sure at the surface of the conductor. It appears probable that the explanation will be found in the decreasing surface of the smaller — conductors. Secondary ionization probably begins with the col- lisions of a few electrons which have free paths longer than # average. With decreasing area of conductor, the number of neigh- i, boring electrons whose free paths exceed a certain length, and the same time are subject to the maximum electric intensity, will be decreased, and consequently the corona forming electric intensi must be higher. : JoHns Hopxins UNIVvERsITy, April 20, I9grt. : DISRUPTIVE DISCHARGES OF ELECTRICITY THROUGH FLAMES. By FRANCIS E. NIPHER. (Read April 21, 1911.) : In a paper published by the Academy of Science of St. Louis’ _ the author pointed out the essential difference in character between the effects of X-rays in the ionization of air and that produced in a column of air exposed to the positive terminal of an influence ‘machine. _ The action of X-rays is to dislodge negative corpuscles from ‘some of the air molecules and load them upon others. Such a mass of air is said to have the property of conduction. Some of the mole- cules in it will accept negative corpuscles from those to whom they have delivered them or from the terminal of a negatively charged electrometer. Other molecules will deliver their overload of nega- tive corpuscles to an electrometer terminal from which negative cor- puscles have been drained, or to the molecules which they have robbed. Ti left to itself such a mass of air soon loses its property of conduc- tion. The average corpuscular charge of a molecule in such a mass of air is the normal amount. _ In a mass of air which forms the positive column due to the action of an influence machine the negative corpuscles have been drained, or are being drained into the positive or exhaust terminal. In air of ordinary pressure it is found that in air thus drained of “negative corpuscles, a disruptive discharge diffuses into the drained region. The disruptive channel widens and apparently ceases to have a disruptive character within the region thus drained. In a few cases the disruptive channel has re-formed on the other side of such a cloud-like mass which had apparently drifted over the * photographic plate and away from the positive terminal. *Trans., Nos. 1 and 4, Vol. XIX., and No. 1, Vol. XX. 397 398 NIPHER—DISRUPTIVE DISCHARGES [April 2 An illustration of this action is shown in Fig. 1. A photographic plate had the heads of two pins resting upon the film. They formed the terminals in a gap in a discharge line from the negative terminal of an eight-plate influence machine to ground. Between this gap and the machine was another gap of about 1 mm., which was at the large knob of the machine. : In order to produce the effect shown in the figure, the mach was turned very slowly for several minutes. Small discharges occurred at the small gap. When there was danger of a spark be- tween the pin-heads, the machine was stopped for twenty or thirty seconds and then continued. This resulted in draining the neg corpuscles from the air around the grounded pin-head. A pro plates in which this operation was continued for an increasing interval, the plates being then developed. at In Fig. 1 after continuing the slow driving of the mail about three minutes, its speed was then suddenly increased disruptive discharge passed over the photographic film beennes pin-heads. This plate is one of many hundreds that have shown this nomenon of a diffused conduction in the region around the posi end of the disruptive channel. This channel began at the nega pin-head, in the midst of the negative glow. That region wa in a condition of conduction for the negative discharge, and ha been in any case observed. Fig. 1 is one of a few cases where discharge wandered considerably from the line joining the heads. In some cases the plate was in the positive line. In cases the two pin-head terminals were directly connected to tt positive terminals of the machine with minute gaps at the machi In all cases the diffusion area was formed at the positive pin-h terminal. In all cases the appearance shown in Fig. 1 was observ The appearance is that which might be caused by a volley of neg tive corpuscles discharged from the end of the disruptive channe and aimed at the pin-head forming the positive, in this case grounded, terminal. The pin-héad shielded that portion of the f which was behind it and in line with this discharge from the fog- 1911.J OF ELECTRICITY THROUGH FLAMES. 399 ging effect observable around it. The air-film which carried the dis- charge was in close contact with the film, as is shown by the char- acter of the shadow. The lowest part of the rounded pin-head only was effective in this shielding of the film, as is shown in Fig. 1. The interior of the disruptive channel is also a drainage or con- duction channel. It is in a highly rarified condition, approaching a that of a vacuum tube. The discharge which passes through it is in the nature of a cathode discharge. The air molecules which form the stepping stones for this conduction discharge are urged in the opposite direction from that in which the corpuscular discharge is Fic. 1. passing. This is incidental to the fact that the conductor is in gaseous ls form. These air molecules have in some cases produced effects at = the negative terminal, similar to those shown in Fig. 1. They are, however, less marked in character. They are in the nature of _ “canal rays,” as observed in a vacuum tube. A photographic plate showing such effects was reproduced in a former paper.? In a copper wire the transfer from atom to atom likewise occurs. There the atoms cannot yield, they are nearer together, and the phenomena ‘of conduction are much more simple. * Trans. Acad. of Sci. of St. Louis, Vol. XIX., No. 4, plate XXII, Fig. A. 400 NIPHER—DISRUPTIVE DISCHARGES [April 21, An attempt was made to compare the conduction-properties of a drainage column of air like that shown in Fig. 1, with those of the flame of a blast lamp. Fig. 2 shows a camera photograph of dis- ruptive discharges between a red-hot ball of iron hung on a wire suspension by means of which it was grounded, and the negative terminal of the influence machine. The ball was heated by a blast lamp, the air being fed from a tank at about two atmospheres pres- sure. A similar flame was placed between the hot ball and the nega- Fic. 2. tive terminal, so that the discharges passed through it. On account of the long exposure, the contrast between the flame and the indi- vidual sparks is not very distinct. Some of the sparks show a par- tial photographic reversal. The discharge lines are, however, all more or less clearly visible within the flame. Fig. 3 shows a single spark, made under the same conditions, although the flame was exposed for nearly half a minute before the spark passed. Fig. 4 shows a similar photograph in which the exposure to the flame was not over half of a second. There are two discharge lines visible, r9tt.J OF ELECTRICITY THROUGH FLAMES. 401 although only one discharge could be distinguished by the sound. The fainter discharge came from the red-hot ball, and crossed the track of the brighter spark, which came from a hook serving for suspension of the ball on a grounded wire. The track of the fainter spark is as sharply defined within the flame as that of the brighter one. In Figs. 3 and 4 the discharge was in the positive line. The hot ball was grounded. Fic. 3. It is evident from these results that the conduction of the gases within the flame of the lamp is very much less than is shown in the positive column near the anode terminal in Fig. 1. In that figure, the air within the disruptive channel is highly rarefied. This channel is a hole bored through the air. The discharge through this channel issued from the end and continued as “sheet lightning” across the drainage area surrounding the grounded anode. This drainage area PROC. AMER. PHIL. SOC., L. 200 Z, PRINTED AUG. 5, IQII. 402 NIPHER—DISRUPTIVE DISCHARGES {April 21, is not in the rarefied condition which exists within the disruptive spark channel. This part of the discharge must be practically noise- less. The sound produced by the spark is caused by the collapse of the spark channel in a manner similar to that caused by the crack produced by the end of a whip-lash, which also cuts a hole in the air. When an electrical discharge occurs between clouds or between Fic. 4. a region containing an excess and one having a deficiency of elec- trical corpuscles, the latter region must be in a condition like that surrounding the grounded anode in Fig. 1. The disruptive channel will diffuse into it. This region is one which is properly called a region of conduction. The other end of the discharge channel must penetrate regions where the air is super-charged with corpuscles. It is not in the a Pe an Tee ee ee a ee _ OF ELECTRICITY THROUGH FLAMES. 493 me sense a region of conduction. Here tributary discharge chan- $s will form. These discharge channels branch out from the main ular stream is flowing. This end of the discharge is called eh Peels in most cases the ends of the discharge 404 TRELEASE—THE DESERT GROUP NOLINEZ. [April aseee io wee Pe mews cas® . . . w ° Zz * ett ewe wees . ’ * eet Mae tees ~ III} cALIBANUS LY NOLINA = DASYLIRION SX BEAUCARNEA DISTRIBUTION OF NOLINEAE THE DESERT GROUP NOLINE. (Prates I-XVII.) By WILLIAM TRELEASE. (Read April 21, 1911.) History. The four genera Nolina (Michaux, 1803), Dasylirion (Zucca- ini, 1838), Beaucarnea (Lemaire, 1861) and Calibanus (Rose, 1906) form so natural a group that many botanists have considered a single generic name, Dasylirion, sufficient for all, though they differ enough in fruit to have caused the founder of this genus to question the propriety of including in it all of the species that were known even in his day; and they show marked differences in habit. Except that Dasylirion was based in part on a Hechtia, which led its author—who later recognized the error—to place it among _ the Bromeliacee, and that on his suggestion it has been connected transiently with the Juncacez, this genus and its immediate rela- tives have been accorded place generally among what are now con- sidered as Liliaceze,—though not always under that family when its rather heterogeneous components have suffered temporary segrega- _ tion. No better arrangement has been found than that of Engler _and Prantl* who locate the Nolinee between Yuccee and Dra- cenez as part of the Draceanoid Liliacee. From the Yuccee they are sharply differentiated, among other characters, by their small _ polygamo-dioecious flowers (never 10 mm. in diameter), few- ovuled pistil, and small usually indehiscent fruit rarely more than _ one-seeded: and the Draczenez differ from them in a usually some- what gamophyllous perianth, perfect flowers, and prevailingly fleshy fruit,—but in all of these respects the group of Dracenee _ Offers a good deal of latitude. 405 406 TRELEASE—THE DESERT GROUP NOLINEZ:. {April 21, : DISTRIBUTION AND ORIGIN. Like the Yuccez, the Nolinez are all North American, and they are comparably distributed except that none are known from the — West Indies. They are among the characteristic plants of the d temperate backbone of the continent. None extend north of southern Colorado, and no species is known to have a very extended — range. Their focal center is evidently the temperate Mexican tableland, on which the genera are all represented and to which majority of their species are confined, Beaucarnea alone, in its most typical form, being characteristic of the hot country and ranging into Central America. Of the two genera that reach the Uni ¢ States, Nolina only enters into the Californian flora, and that only in the southern desert. Though unrepresented in the intermediate region, from which it may be assumed to have disappeared, genus also appears in the South Atlantic states, apparently as offset from the grass-leaved Texan stock, rather than indicating primal home (map). The ontogeny of the group is scarcely more than a matter speculation. No reason is apparent for considering it to be ancient. Though evidently related to the typically septi Yuccez, it seems rather more likely to have had a closer evolutio connection with the typically loculicidal Draceneze. More s factory hypotheses may be held concerning the affinities of component genera. Nolina may be taken as most closely appre ing the prototype of the group because of its extensive range, number of species composing differentiated groups, and confo to the liliaceous plan in its 3-celled pistil and cotyledonary Calibanus appears to be an offset of Nolina. Beaucarnea Dasylirion, with a single-celled pistil, may represent parall shoots from Nolina or a no-longer recognized derivative o genus; and the question may be raised whether Beaucarnea is than a well-marked subgenus of Dasylirion which, strictly li itself consists of two quite dissimilar groups. These affiniti be indicated as follows: hives Je alibanus. \ 7 Beaucarnea. NDasylirion. gt.) TRELEASE—THE DESERT GROUP NOLINEZ. 407 BIo.Locy. All Nolinez are perennial, and, as would be expected from their habitats, they are pronounced xerophytes with a rather succulent ‘caudex,’ either small and insignificant or moderately developed, and ‘then either prostrate or erect, or even of tree size (pl. 1-4), and tather hard usually rough-edged or even prickly leaves** covered by a well-cuticularized epidermis, the stomata usually arranged in lines overlying the parenchyma between strong fibrous bundles and either furnished with an outer vestibule as in Agave, etc. (Dasyli- rion), or located between prominent ribs that, especially in Nolina, are often covered with more or less interlocking papillae.*** They occur most strikingly in such desert associations? as count Agave, Yucca and Hechtia among their characteristic components (fl. 2, 4). In many species the tip of the leaf shreds into a sometimes brush- ‘like bunch of fibers, and in one (Nolina Bigelovii) the margin breaks away sparingly—in kind, rather than quantity, recalling the fibrous exfoliation characteristic of many yuccas and of one large group of spicate agaves. From a study of the leaf-tip of Dasylirion -acrotriche, Zuccarini®* was led to believe that what passes for the leaf is really a petiole with ventral ligule, the blade, considered as peltate, being represented by the more dorsal shreds only. The prevalent dorsal insertion of the haustorium on the cotyledonary sheath in seedlings of this group is worthy of note in connection with this opinion (//. 15). _ Though sometimes weakened or even destroyed by flowering “under cultivation, all of the Nolinee appear to be normally poly- carpic. The terminal inflorescence+?* is essentially of one type though varied from a thin lax raceme-like wand into a stout com- pound spike with short and broad divisions or an open simple, com- pound or even decompound panicle (p/. 5). Whatever its form, _ the flowers are clustered, usually two or three together, in the axils of small prevailingly denticulate bractlets, either on cushions so short that they appear to come from the main axis, or, more com- monly, on evident secondary or tertiary branches (p/. 6, 7). The primary branches appear to be 8-ranked** and the bracts are often large and conspicuous, those which support the ultimate flower _ Clusters being scarcely larger than the bractlets. Tn aan ei a A cea ee 408 TRELEASE—THE DESERT GROUP NOLINEZ, [April 21, The sometimes slightly fragrant’? polygamo-dicecious flowers are borne on slender pedicels never greatly exceeding their own length, which are always distinctly jointed, usually about the — middle. Though the flowers are small, their at first petaloid, then scarious-persistent distinct entire or toothed segments are usually whitish, though more or less tinged with green, violet, rose or cream—a coloration supported by the usual whiteness of the scari- ous bractlets and, often, by similarly colored large bracts. The small elliptical anthers are introrsely versatile, their filaments slightly adnate to the base of the perianth segments. Three connate carpels, with typically two anatropous basal ovules each, constitute the pistil which is 1- or 3-celled in different genera. The stigmas are essentially apical, on more or less free and divergent style tips in Nolina, crowning the rather narrowed ovary in Beaucarnea, along the rim of a distinct funnel-like though sometimes cleft style in Dasylirion, or as sessile points in Calibanus (pl. 8). Essentially unisexual and often dicecious, the flowers are perfect in plan; and abortive stamens are found in the fertile flowers, and more or less recognizable rudimentary pistils in those that are functionally staminate. In fertile flowers nectar is secreted by small septal nectar-slits in the base of the pistil,—often very evident aft this has enlarged into a fruit (pl. 9); and in staminate flowers the rudiments of the carpels that perform the same function.** Though prevailingly 3-merous, the flowers may show deviatir from this pattern. Preda*’ noted that about one-fifth of the flowe of a pistillate plant of Dasylirion glaucum were 4-merous; and | examining large numbers of the fruits of this genus I have obse: 2-, 4- or 5-winged fruits of several species and one 4-carpellary of Calibanus (pl. 11). Several observers have found that p developed fruits may occur now and then on staminate plants*; own observation shows that well developed stamens may be foun some pistillate flowers; and Bouché’ records the transformation - staminate into pistillate individuals,—suggesting an interesting ? of study for those who may observe and experiment with the plants as they grow under natural conditions. Observations on pollination do not appear to have been corded, but the flowers are clearly entomophilous and their pollina- tort.) TRELEASE—THE DESERT GROUP NOLINE. 409 tors are to be sought probably among the Hymenoptera and Dip- "tera, as has been suggested to me for Dasylirion by Sr. Patoni, of _ Durango. _____ Normally fertilized, the ovules develop into 3-sided or 3-grooved _ seeds with micropyle by the side of the hilum, a slender often a scarcely discernible raphe, and thin and smooth or somewhat thick- ened and wrinkled envelopes composed of thin-walled cells and representing essentially the seed-coats though often with a terminal umbo or apiculus representing the base of the nucellar tissue. The ‘bulk of the seed consists of rather firm endosperm through which _ the finger-like embryo passes upward from near the micropyle oward the morphological base of the nucellus. The endosperm consists of moderate-sized polygonal cells with glistening white rather thick pitted walls and coarsely granular contents destitute of starch. The walls of these cells are of the “ reserve-cellulose ” type, but they are colored blue by neither iodine nor chlor-iodide of zinc, though they swell so greatly in the latter reagent that in a thick section the contents, in which large and abundant oil drops separate out, promptly extrude, sausage-like, from any chance eak (pl. 10). Went and Blaauw” have reported partial embryo ‘ormation in some ovules and much more complete endosperm de- velopment in others, in a pistillate Dasylirion—apparently without concurrence of male nuclei. Usually only one of the six ovules pro- duced by a normal pistil matures in the 1-celled fruit of Dasylirion and Beaucarnea or the 3-celled ovary of Calibanus; but with the 3-celled fruit in Nolina, though a single seed is the rule, two or three are not infrequently seen,—usually only one to a cell, though ex- ceptionally both ovules of a carpel develop. The ripened fruit is dry-walled: subglobose with three low ribs _ Calibanus, triangular with strongly developed dorsal wings on the carpels in Dasylirion and Beaucarnea, and deeply 3-lobed be- “tween the wingless carpels in Nolina. In the first three genera it s not dehisce, but in Nolina, though the delicate walls are often egularly torn—sometimes even before maturity of the rather mly attached seed, or the fruit may remain long unopened—loculi- dal dehiscence is more or less prevalent (pl. rr, 12). If observations on dissemination have been published, they have 410 TRELEASE—THE DESERT GROUP NOLINEA, [April 2t escaped my search, but the process may be inferred with sor t probability from the character of the fruit. In all, the ripened fruit with its enclosed or attached seed or seeds and the persistent bu unenlarged perianth, falls by disarticulation of the pedicel,— cle to the fruit in Dasylirion and Calibanus, somewhat further from i in Beaucarnea, and usually at a still greater distance in Nolina. provision for dissemination other than through rolling or being blown over the ground appears in the round fruit of Caliba s The winged fruits of Dasylirion and Beaucarnea are as evid wind-scattered as the similarly disarticulating and equally s fruits of Rumex,—though in the latter the wings are not carpe but consist of the enlarged persistent sepals. The very diff fruits of Nolina are likewise evidently wind-disseminated, — more or less inflated carpels giving them a character interme between winged and balloon fruits. Toe of Klebs’ Asphodelus-Tradescantia type,?°*® the seed—freed the remnants of the fruit in Nolina or still contained in them i other genera—remaining in the ground with the arched haus! elongating with the cotyledonary sheath so as to reach a len even 10 mm. In Nolina longifolia and in specimens of N. par: preserved by Dr. Rose, the haustorium is apical, though elbowing is sometimes seen near the the top of the arch s sometimes straightens and lifts the seed from the ground. lings of Beaucarnea and Calibanus preserved by Dr. Rose show in these genera the sheath is produced above the arch in form pointed ventral ligule, as is true in such species of Dasyli have observed. In these cases the haustorium appears to tinctly dorsal on the sheath, along which it is often she fracted (pl. 13-15). Initial growth is evidently at the p ‘ expense of the granular protoplasm, oil and “ reserve-cellu the endosperm. In Calibanus and Beaucarnea, as is shown in lent specimens in the National Herbarium prepared by Dr. Ro: e formation of the thick trunk follows germination quickly. 1911] TRELEASE—THE DESERT GROUP NOLINEZ. 411 UsSEs. Though none of the Nolinez can be considered as of great _ present economic importance, many of them are utilized in one way or another and it is probable that more use can be made of _ some species than is now the case. In the great bend of the Rio Grande I have seen the trunks of Dasylirion split open to give stock access to the rather watery pith; and they are sometimes cut for _ feeding.2** In Mexico the trunks of Dasylirion are roasted and eaten similarly to those of the mezcal agaves; and Dr. Gregg notes similar use of a Nolina on the label accompanying a specimen of _ From such roasted trunks of Dasylirion, after fermentation, an alcoholic beverage very similar to mezcal spirits is distilled, and under the name of sotol*?"**21 it is very commonly used through _ the extensive Mexican territory over which this genus occurs. As in Yucca, Agave, and some other plants, the sap of those now under consideration contains, as a water conservation provision, a saponi- fying substance, and the roots of Nolina Palmeri are said to serve as an amole.* The leaves of Dasylirion and Nolina—and presum- of Beaucarnea—are used for thatching,** basket work, coarse _ and similar plaited-ware, either entire or shredded.®?-19 | less employed than that of yuccas and agaves, their fiber is also somewhat used locally, and the narrow leaves of the eastern bear-grass have long been used in their entirety for hanging meat id similar domestic purposes for which strength rather than finished cordage is needed. Some thought seems to have been given also to the preparation of paper pulp from the fiber of Dasylirion.* SYSTEMATIC REVISION. In revising the forms known to me I have had the privilege of ng an unusual amount of typical material, for which I am greatly ebted to Professor Radlkofer of Munich (Zuccarini types), Dr. inson of Cambridge (Watson types), Dr. Rose of the National keley, whose collection contains numerous critical forms. Ow- to Engelmann’s early interest in the vegetation of the Texano- exican region, his herbarium, now at the Missouri Botanical 412 TRELEASE—THE DESERT GROUP NOLINEZ:, [April 21, Garden, is rich in representatives of this, as of other groups char- acteristic of that arid region,—as herbarium representation of such plants goes: and in it, as well as in the herbarium of the New York Botanical Garden and in the National Herbarium, have been found types or cotypes of the species of Scheele and Torrey. I do not venture to think that anything like the last word on the group is here said,—the sparse occurrence of the representa tives of admitted species through a vast and greatly diversified area, as shown by the distribution map, would speak against such a view 5 but the following rather tersely cast synopsis is published in the hope that it may render the work of filling gaps in both range and forms easier than it has proved in the past. Space is not taken for a full bibliography,—though this would not have been very exten- sive; but the principal revisions of each genus are noted, as well as the various names under which a species has appeared; and ref ences are given to all illustrations that have been found. SyNopsis oF GENERA, Ovary 3-celled. Fruit wingless. Fruit deeply 3-lobed, often inflated: seed nearly globose, rather fleshy- walled. Pedicels articulated rather far below the flowers. Peri segments entire, papillate-pointed. Leaves strongly ribbed with papillate grooves, at most serrulately roughened on the margin. florescence a panicle (or racemosely reduced). Now Fruit globose-triangular, not lobed or inflated; seed melon-shaped, thin walled, occluding the sterile cells. Pedicels articulated close to flowers. Perianth segments nearly entire, rounded. Leaves as in Nol Inflorescence a panicle. CALIBAN Ovary 1-celled. Fruit 3-sided and 3-winged, not lobed or inflated. : Pedicels articulated somewhat below the flowers. Perianth seg entire, acute. Seeds 3-grooved or 3-lobed. Leaves somewhat the grooves not usually papillate, at most serrulately roughened on margin. Inflorescence a panicle. BEAUC. Pedicels articulated close to the flowers. Perianth segments dentic rather obtuse. Seeds 3-grooved or 3-sided. Leaves not ribbed, margin (in all except one square-leaved species) armed with st prickles and usually also serrulate-roughened. Inflorescence a compound spike. DasyLt NOLINA. Michaux, Fl. Bor.-Amer. 1: 208. 1803.—Watson, Proc. Acad. 14: 246-8. 1879.—Rose, Contr. U. S. Nat. Herb. 10: agtt.] TRELEASE—THE DESERT GROUP NOLINE. 413 g06.—Sometimes merged in Dasylirion or Beaucarnea, and made to include the latter genus by Hemsley, Biol. Centr-Amer. 3: 371, which is conformed to the views of Bentham and Hooker, Gen. Plant. 3: 780—At first monotypic, based on N. georgiana. Leaves thin and grass-like (but hard-fibrous), linear, rarely over 5 mm. wide, rather flat, usually not brush-like at tip. Bracts not very showy. Acaulescent (pp. 413-416). GRAMINIFOLIZ. _ Inflorescence commonly as long as the minutely serrulate-scabrous essentially green spreading leaves, peduncled, unbranched or with slender usually simple branches 15-25 cm. long. Floriferous bracts _ small, not imbricated. Pedicels remaining filiform, increasing to 8 or Io mm. and equaling or exceeding the usually rather large and inflated fruit. Seed not prominently exposed. Leaves smooth and rather open between the ribs. Panicle not com- pound. Lower bracts much shorter than the subtended branches. Bractlets barely serrulate. Notrwa GrorciaANa Michaux, Fl. Bor-Amer. 1: 208. 1803.— M(asters), Gard. Chron. n. s. 15: 688, 697. f. 126. Phalangium virgatum Poiret in Lamarck, Encycl. Méth. 5: 246. 1804. ee Leaves 3-5 mm. wide. Inflorescence simply pani- j ‘Oi. cled with rather spreading branches. Flowers rather — large. Fruit subelliptical, rather pointed, 7-9 X 8- tomm. Seed 2X 4 mm.—PI. 5, 11. Central South Carolina and across central Georgia. _ Specimens examined: Georcra. Milledgeville (Boykin, 1836). Augusta (Cuthbert, 1877). Belair (Eggert, 1899). Big Lott’s _ Creek (Harper,965,1901). Columbia County (Chapman). Thom- son (Bartlett, 1174, 1907). N. atopocarpa Bartlett, Rhodora. 11: 81. 1909. _ Leaves 2-4 mm. wide. Inflorescence unbranched or mply panicled. Fruit more or less unsymmetrically yg obovate, shallowly notched, or pointed, scarcely inflated, a ee, 414 TRELEASE—THE DESERT GROUP NOLINEZ. [April a1, Leaves (as in all except the two preceding) with the sides ok ‘the ribs microscopically papillate. Lower bracts sometimes ab: equaling the subtended branches. Bractlets toothed. Fruit (as | . all except the two preceding) conspicuously notched. ina N. Brirronrana Nash, Bull. Torr. Bot. Cl. 22: 158. 1895. ue Leaves 5-10 mm. wide. Inflorescence simply pani- cled with rather erect branches. Fruit depressed- — orbicular, 8 10 mm. Seed 3 X 4 mm. North-central Florida. Specimens examined: Fiorma. Eustis (Nash, 450, ae - type; Webber, 406, 1896). Clermont (?MacElwee, 1895; Williar son, with the close-ribbed leaves of this species, but fruit —— georgiana). N. LinDHEIMERIANA Watson, Proc. Amer. Acad. 14: 247. 1871 Dasylirion Lindheimerianum Scheele, Linnea. 25: 262. 1852. D. tenuifolium Torrey, Bot. Mex. Bound. 215. 1859. Beaucarnea Lindheimeriana Baker, Journ. Bot. 10: 328. 187 Leaves 2-5 (exceptionally 9) mm. wide. Infl o PAR cence simply panicled with spreading branches — less than 10 cm. long, or the lower of these slender branchlets less than half as long. Fruit s SXLr_ what depressed-orbicular, 7-8 & 8-10 mm. — mm.—PIl. 12. os Central Texas.—In the region of N. texana and Das) texanum. Specimens examined: Texas. Vicinity of New. Bra (Lindheimer, 213, 1846,—the type of D. Lindheimerianum; 552, 1846; 1214-1217, 1849). Sabinal River (Wright, roro, 185 —the type of D. tenuifolium). Austin (Hall, 634,1872). Bande Pass (Reverchon, 1606, 1884). Cherry Spring (Jermy, 8 Edwards County (Hill, 39, 1895). North of San Antonio (Hi ings, 81, 1910).—Gillespie County (Jermy,—with leaves +74 wide). Western Texas (Wright, 673, 1849). Inflorescence rather dwarf, panicled. Bractlets rather conspicuous, or less lacerate. Leaves glaucescent, raggedly serrulate-sea Agr.) TRELEASE—THE DESERT GROUP NOLINE#. 415 Be Pedicels rather slender, at length equaling or exceeding the fruit. Floriferous bracts not imbricated. Lower bracts linear, leaf-like. Panicle simple. __N. pumiza Rose, Contr. U. S. Nat. Herb. 10: 92. 1906. Leaves 2-4 mm. wide. Inflorescence 30 cm. long, 3 the upper two-thirds narrowly and simply panicled a 9 with short weak branches scarcely 2 cm. long. _ Fruit suborbicular, 6-7 mm. in diameter, the pedicels pa somewhat thickened upwards. Seed (immature) ae 2X 3 mm. West-central Mexico. Specimens examined: Tepic. Sierra Madre Mountains near Santa Teresa (Rose, 2165, 1897,—the type). Lower bracts dilated and scarious. Panicle compound. N. Hartweciana Hemsley, Biol. Centr.-Amer. 3: 371. 1884. _ Cordyline longifolia Bentham, Plant. Hartweg. 53. 1840. _ Roulinia longifolia Brongniart, Ann. Sc. Nat., Bot. ii. 14: 320. 1840. __ Dasylirion junceum Zuccarini, Abhandl. Akad. Miinchen. C1. II. : 4 (=Denkschr. 19): 19. 1845. _D. Hartwegianum Zuccarini, 1. c. 21. 1845 —Bentham, /. c. 348. 1857. _ Beaucarnea Hartwegiana Baker, Journ. Bot. 10: 327. 1872. Shortly caulescent? Leaves 3-4 mm. wide, somewhat fibrous- shredding at tip. Inflorescence 25-50 cm. long, short-stalked, ovoidly compound-panicled with pyramidal divisions 8-15 cm. long and short stiffish branchlets ——PI. 16. Central Mexico. Collected about Zacatecas by Hartweg in 1837. _ The characters are extracted from the descriptions of Zuccarini nd Baker and from a photograph of a Hartweg co-type (406) in the Delessert herbarium which I owe to the obliging kindness of Ml. de Candolle and reproduce here with his permission. Pedicels thickened, about half as long as the rather large fruit. Floriferous bracts imbricated. Panicle simple, scarcely half as long as the leaves. HUMILIS Watson, Proc. Amer. Acad. 14: 248. 1879.—Hemsley, _ Biol. Centr.-Amer. 5. pi. 93. 416 TRELEASE—THE DESERT GROUP NOLINEZE. [April a Beaucarnea humilis Baker, Journ. Linn. Soc., Bot. 18: 237. 1880. Leaves 2-3 mm. wide. Inflorescence 15 cm. long, with a few suberect basal branches one-third as long. cg) Fruit suborbicular, 7 X 9 mm., scarcely inflated. Seed very large, 3-4 X 5 mm., prominently exposed. ne East-central Mexico. In the region of N. Watsoni, Calibanus, and Dasylirion Parryanum and graminifolium., Specimens examined: San Luts Potosi. Vicinity of Sen pe Potosi (Parry & Palmer, 875, 1878,—the type). | N. Watson Hemsley, Biol. Centr.-Amer. 3: 372. 1884; 5. pl. Beaucarnea Watsoni Baker, Journ. Linn. Soc., Bot. 18: 236 6 1880, . Leaves 5 mm. wide, rather concave and unusually rough-ma ‘ Inflorescence 25-30 cm. long, with rather nu strict branches scarcely one-third as long, smooth ¢ %. somewhat scabrid on the short peduncle. Fruit m« Pas lliben or less ovate-orbicular, cordately notched, 8 X 8 mm., inflated. Seed (immature) 2X 3 mm. } East-central Mexico. In the region of N. humilis, etc. Specimens examined: San Luts Porosi. Vicinity of San Potosi (Parry & Palmer, 874, 1878,—the type, 502, 1878; — 261, 1879). Leaves rather thick, linear or narrowly oblong-triangular, scancely over mm. wide, green, more or less concave and unequally keeled on one both faces, raggedly dentate-scabrous in most species and in age fibrous-lacerate at tip. Inflorescence usually about as long as the le peduncled, compound-panicled. Bracts not usually very OEY, B lets more or less lacerate. Fruit small, not inflated, the relatively large seed early exposed prominent (pp. 416-420). ERuMPE! Inflorescence (as in the last preceding species) often ro lines. Pedicels rather thickened in fruit. Acaulescent. — Lower bracts firmly long-attenuate from a somewhat 4d scarious-margined base. Lower panicle divisions much shorter than the subt bracts, with rather weak strongly ascending lets. N. TEXANA Watson, Proc. Amer. Acad. 14: 248. 1879.—Na' Journ. N. Y. Bot. Gard. 6: 48. f. 16. | Beaucarnea texana Baker, Journ. Linn. Soc., Bot. 18: 236. 1 —tgtt.] TRELEASE—THE DESERT GROUP NOLINEZ. 417 Leaves very narrow, 2-5 mm. wide, smooth-edged or slightly roughened, from half-round becoming triquetrous. Inflorescence often much shorter than the leaves, with 9o oblong divisions often 15 cm. long and lower branch- . lets half as long, or subsimple. Fruit somewhat depressed, 4 X 5-6 mm. Seed 3 mm. in diameter—PI. 12, 15. Central Texas. In the region of N. Lindheimeriana and Dasy- _ firton texanum. Specimens examined: Texas. Vicinity of New Braunfels (Lind- : & heimer, 550, 1846, 712, 1847, the types; 1218,1849). Austin (Hall, | 635, 1872). Hamilton County (Reverchon, 967, 1882). Cibolo _ (Havard, 1883). Blanco County (Reverchon, mixed with 1606). Kerr County (Bray, 184, 1899). Davis Mountains (Earle & Tracy, 322, 1902). Gillespie County (Jermy, 327). Comstock (Thomp- son, 1911). Without locality (Buckley). Lower bracts mostly triangular, becoming friable. Lowest panicle division much shorter than the long-caudate subtending bract, with rather weak finally ascending branchlets. N. affinis Trelease. Leaves very narrow, 3-4 mm. wide, sometimes smooth-edged. _ Inflorescence at length with broad divisions 10 cm. long and lower branchlets scarcely half as long. Fruit de- S x pressed, 5 X 6-7 mm. Seed 3 mm. in diameter. ae North-central Mexico. On the outskirts of the range of N. erum- pens, N. microcarpa and Dasylirion leiophyllum. Specimens examined: CH1tHUAHUA. Rocky hills near Chihuahua (Pringle, 1, 2,1885,—the type). Santa Eulalia (Palmer, 139, 1908; Rose, 11672, 1908). N. caudata Trelease. _ ?Nolina sp. Rose, Contr. U. S. Nat. Herb. 20. pl. 46-8. # 3 Leaves very narrow, 4 mm. wide, somewhat rough- | # eS edged. Inflorescence slender, with narrow divisions scarcely 10 cm. long and lower branchlets 2-5 cm. long. Fruit rather depressed, 4 X 5-6 mm. Seed 3 mm. in diameter—PI. 6. PROC. AMER. PHIL. SOC., L, 200 AA, PRINTED AUG. 7, IQII. 418 TRELEASE—THE DESERT GROUP NOLINE:, [April21, Southern Arizona. In the region of N. microcarpa and Dasy- lirion Wheeleri, BE Specimens examined: Arizona. Mule Mountains (Toumey, 1894,—the type). Huachuca Mountains (?Wilcox, 1892, and 257, 1894; Griffiths, 4831, 1903). Dragoon Summit (?Vasey, 1881,— leaves). Nogales (?Brandegee, 1892; Ferriss, 1902; Coville, 1624, 1903; Thompson, 1911). Sierra del Pajarito (?Trelease, 387, 1900). Bounpary Line (?Parry, Bigelow, Wright & Schott, 1443; Mearns, 258, 290, 1892). Lower panicle divisions more or less equaling the atteaisate subtending bracts, with rather stiff spreading branchlets. N. ERUMPENS Watson, Proc. Amer. Acad. 14: 248. 1879. Dasylirion erumpens Torrey, Bot. Mex. Bound. 216. 1859. _ Beaucarnea erumpens Baker, Journ. Bot. 10: 326. 1872. Leaves usually 6-10 mm. wide and very rough- ie edged, exceptionally narrower or smooth-edged. In- seh florescence with pyramidal divisions 15 cm. long and ?) Lane lower branchlets half as long. Fruit rather depressed, ft ee ‘ 5 < 5-7 mm. Seed very large, 4 mm. in diameter. o Western Texas and adjacent Mexico. In the region 6f Dasy- lirion leiophyllum and D. Wheeleri Wislizeni. Specimens examined: TExAs. Western Texas (Wright, 1918, 1851-2,—the type of D. erumpens; 692, 1849). Chisos Mountains (Bailey, 391, 1901). Eagle Mountain (Bigelow, 1852). Eagle Spring (Hayes, 1858). Podrero (?Schott, 1855). CHIHUAHU: Between El Paso and Chihuahua (Wislizenus, 219, 1846). N. erumpens compacta Trelease. Leaves almost as in fexana, sometimes scarcely 5 mm. ais edge either rough or smooth. Inflorescence with very compact ov: divisions scarcely 6 cm. long and branchlets about 1 cm. long. — . Extreme western Texas. Specimens examined: Texas. El Paso (Ferriss, 1902, type). Sierra Blanca (Trelease, 386, 1900). Sanderson ( ?Tho son, 1911). Marathon (Lloyd, 1910). Presidio (Havard, 1880) N. GREENE! Watsoninherb. Greene, Bot.Gaz.5:56. 1880.—Nam only. 1911] § TRELEASE—THE DESERT GROUP NOLINEZ. 419 Leaves 6-7 mm. wide, smooth-edged. Inflorescence with rather narrow divisions scarcely 10 cm. long and lower branchlets nearly half as long. Fruit depressed, B. 4xX6mm. Seed 2X 3 mm. ___ Southeastern Colorado to northeastern New Mexico. The north- _ ernmost species of the group. _ Specimens examined: Cotorapo. Between the Purgatory and _ Apishipa rivers, north of Trinidad (Greene, Jan., 1880,—the type). _ New Mexico. San Miguel County (Brandegee, 1879). Lincoln _ County (Wooton, 656, 1897)- Lower panicle divisions considerably shorter than the sub- tending bracts, with short stiff spreading branchlets. @ N. cespitifera Trelease. _ Leaves 6-10 mm. wide, with dorsal as well as “marginal roughening. Inflorescence very rough from 4 ©) compound tussocks, with narrow divisions Io cm. we long and lower branchlets scarcely one-third as long. Fruit nearly orbicular, about 5 mm. in diameter. Seed ? | North-central Mexico. On the margin of the range of Dasy- _lirion cedrosanum. Specimens examined: CoaHuiLa. Battlefield of Buena Vista _(Wislizenus, 308, 1847,—the type). High dry lands near Saltillo (Gregg, 81, 1847). Inflorescence (as usual in the genus) “essentially smooth. Lower bracts triangular, scarcely equaling the panicle divisions. Pedicels slender. Acaulescent with one exception. Bate a _N. Patmert Watson, Proc. Amer. Acad. 14: 248. 1879. _ Beaucarnea Palmeri Baker, Journ. Linn. Soc., Bot. 18: 235. 1880. Leaves 8-10 mm. wide, serrulate-scabrous. In- aea§ florescence with narrow divisions 15 cm. long and ie} rather stiff ascending lower branchlets scarcely one ee. fourth as long. Fruit depressed, 4X5 mm. Seed : 3 mm. in diameter. - Lower California. Overlapping the region of N. Bigelovii-and 420 TRELEASE—THE DESERT GROUP NOLINEA:, — [April 21, N. Beldingi deserticola—The type locality is given as Tantillas Mountains. Specimens examined: Lower CAtirornia. Pifion district (Or. cutt, 713, 1882,—determined by Mr. Watson). San Pedro — (Brandegee, 1893). Paraiso ( ?Brandegee, 1890). N. Palmeri Brandegeei Trelease. Nolina sp. Brandegee, Proc. Cal. Acad. ii. 2: 209. 1889. — ‘ ?N. Palmeri Brandegee, Zoe. 1: 306. Mae Arborescent. Trunk about 5 m. high, at length few-branche z above. Leaves 7-8 cm. wide, rather glossy, denticulate- ‘abt Ss. Inflorescence with divisions 15 cm. long and lower branchlets « one-third as long. Lower California. Specimens examined: Lower Catrrornia. San Julio (Bra ) gee, Apr. 11, 1890,—the type). Northern Lower California (Or July 3, 1885). Fruit moderate in size, somewhat inflated, the relatively small seed protruding if early exposed. Panicle divisions with rather weak elongated and ascending branchlets (pp. 420-422). pee 0 Lower panicle divisions more or less equaling the friable tri bracts. Acaulescent. Leaves elongated. N. microcarpa Watson, Proc. Amer. Acad. 14: 247. ee Beaucarnea microcarpa Baker, Journ. Linn. Soc., Bot. 8: 1880. Leaves 6-12 mm. wide, raggedly denticulate-scaleallal ‘Inflc cence with often broad divisions 15-30 to even 45 cm. long, and lower branchlets—sometimes again branched at base—half as long or less. Fruit nearly YQ as long as the pedicels, depressed, 5 X 7-8mm. Seed 3 mm. in diameter, attached and exposed after dehiscence of fr ui Fag Rae POs Southeastern Arizona and adjacent New Mexico and M Overlapping the region of N. caudata and associated with lirion Wheeleri. The type locality is Rock Cafion, Arizona. Specimens examined: Artzona. Rocky Cafion (Rothroc 1874). Chiricahua Mountains (Toumey, 1894; Blumer, 1316, 1 gtr.) TRELEASE—THE DESERT GROUP NOLINE2. 421 _ Santa Catalina Mountains (Pringle, 1881, 1882, 1884). Santa Rita Mountains (Pringle, 1882; Brandegee, 1891). Without locality _ (Toumey, 447, 1892). Sun Flower Valley (Girard, 1, 1873). Blue _ River (Davidson, 775, 1902). New Mexico. Santa Rita del Cobre (Greene, 1880). Burro Mountains (Rusby, 413, 1881,—fruit ; Gold- _ man, 1530, 1908). Dog Mountains (Mearns, 294, 1892). Lone - Mountain (Mulford, 427, 429, 1895). Otero County (Rehn & = Viereck, 1902). Round Mountain ( ?Wooton, 1905,—very narrow- a leaved, as in terana). Mogollon Mountains (Rusby, 412, 1881; _ Metcalfe, 232, 1903). Mimbres River (Metcalfe, 1025, 1904). San _ Luis Pass (Mearns, 186, 1892; Wooton, 1906). Twin Sisters _ (?Blumer,1905). Silver City (?Bailey,1906). Big Hatchet Moun- tains (Goldman, 1341, 1908). Bounpary Line (Parry, Bigelow, é Wright & Schott,1442). CurHuaHua. Colonia Garcia (Townsend _ & Barber, 76, 1899). Vicinity of Chihuahua (?Pringle, 159, 1885; Palmer, 355, 1908). N. durangensis Trelease. “a Leaves very thin, 7-11 or even 20 mm. wide, irreg- Y ularly serrulate-scabrous. Inflorescence with broad . divisions at length 15-20 cm. long and chiefly basal AAP branchlets 10-12cm. long. Fruit usually considerably shorter than the rather slender pedicels, more or less depressed, small, 5-6 X 6-7 mm. Seed 3 mm. in diameter—PI. 10. Northwestern Mexico. In the region of Dasylirion durangense and simplex. Specimens examined: DuraNnco. Vicinity of Durango (Palmer, 249, 1896,—the type; Ochoterena, 1911; Patoni, 1911). Tepehuanes _ (Palmer, 329, 1906). CuHrHuUAHUA. Southwestern Chihuahua (?Endlich, 1162a, 1162b, 1906).. _ N. ELEcANs Rose, Contr. U. S. Nat. Herb. 10: 91. f. 6. 1906. Leaves very thin, 12 mm wide, sometimes lanceo- lately narrowed above the base, serrulate-scabrous. Inflorescence with broad divisions 10-15 cm. long and “e0 rather few branchlets scarcely half as long. Fruit about equaling the pedicels, rather large, 7 X 8-10 mm. Seed 3 X 4 mm. | 422 TRELEASE—THE DESERT GROUP NOLINEZ, (Aprilay, Central Mexico. In the region of N. Hartwegiana? Specimens examined: Zacatecas. Sierra Madre Mountains (Rose, 2396, 1897,—the type). Lower panicle divisions considerably longer than the trinsgulat ’ bracts. Shortly caulescent. Leaves much shorter than the in- florescence. N. rigida Trelease. ree Anatis rigida Brongniart, Ann. Sc. Nat., Bot. ii. 14: 320. + 1840. Leaves 4-5 mm., scarcely 10 cm. long, ciliate-scabrous. Inflores- cence much surpassing the leaves, sessile, with broad divisions about 10 em. long and rather few branchlets scarcely half as long Fruit about equaling the slender pedicels, moderate, about 6n mm, in diameter. Seed 2 mm. in diameter.—PI/. 17. Mexico? Known only from the unpublished figures of Sese anc Mogifio and Node-véran, which M. de Candolle has placed in my hands for study, and of which he has furnished for paler) an excellent photographic copy. 4 Leaves relatively or actually thin, 15-40 mm. wide, serrulate-scabrous, usually brush-like at tip. Inflorescence ample, often peduncled, com- pound-panicled. or occasionally decompound. Bracts usually dilated. and papery, often showy. Bractlets fimbriate-lacerate, conspicuous. Fruit large, inflated, the seed not protruded. Trees (with one exc ep tion?) (pp. 422-426). Be Leaves rather thick, little shredded at tip. Pedicels scarcely half as lon as the fruit. N. Parry Watson, Proc. Amer. Acad. 14: 247. 1879. Trunk 1-2 m. high. Leaves almost pungent, rather thick, cave, keeled, 15-25 or even 35 mm. wide, serrulate- scabrous. Inflorescence with rather narrow divisions 15-30 cm. long and spreading densely flowered branchlets scarcely 4 cm. long. Flowers large, with perianth segments 4 mm. long. Fruit very large, orbicular, deeply notched at both ends, 12-15 mm. Seed 3 X 4 mm.—P. 5, 12. Colorado desert. In the region with N. Bigelovit. Specimens examined: CALIFORNIA. Desert east of San nardino (Parry, 1876,—the type). Whitewater (Vasey, 1881, in diam 191.) - TRELEASE—THE DESERT GROUP NOLINEZ. 423 leaf). San Gorgonio Pass (Engelmann, 1880). San Bernardino Mountains (Parish, 1879: 910, 1882; 3145, 3165, 1894). San Felipe (Brandegee, 1894). Pala (Orcutt). San Jacinto Mountains (Hall, 1819, 2432, 1901). Arizona. Fort Whipple (Coues & Palmer, 1865). Between Sandy and Bill Williams Forks (Mrs. Stephens, 1902). N. Brcetovir Watson, Proc. Amer. Acad. 14: 247. 1879. Dasylirion Bigelovti Torrey, Bot. Whipple. 151. 1857; Bot. Mex. Bound. 216. 18509. Beaucarnea Bigelovii Baker, Journ. Bot. 10: 326. 1872. Trunk 1-2 m. high. Leaves almost pungent, scarcely concave or keeled, 15-25 mm. wide, often roughened on the surface, the at | > first rough margin shredding away in brown fibers. Inflorescence with rather narrow divisions 15-30 cm. long and branchlets scarcely 4 cm. long. Perianth segments about 3 mm. long. Fruit large, orbicular, deeply notched at both ends, usually 10-12 mm. but occasionally 15 mm. in diameter. Seed 3 X 4 mm. Western Arizona, across the Colorado desert, and into Lower California. In the region of N. Parryi and overlapping the ranges of N. Palmeri and N. Beldingi. A sketchy picture of it, in the Tinajas Altas, is given by Schott in Emory, Rept. Bound. Surv. I. pl. 50. _ Specimens examined: Arizona. Bill Williams Fork (Bigelow, 1853-4, the type of D. Bigelovii). Union Pass (Palmer, 1870). Havasupai Cafion (Kinner, 1900). Gold Road (Mrs. Stephens, Ig02). Little Meadows (Mrs. Stephens, 1902). CALIFORNIA. Mountain Springs, near the boundary (Parish, 1880; Vasey, 1880; Mearns, 2980, 3015, 3066, 3146, 1894). Lower CALIFORNIA. Can- tillas Cafion (Orcutt, July 8, 1884). Yubay (Brandegee, 1889).— Bounpary Line. Tule (Mearns, 320, 1894).—Sonora. (?Schott, 1441,—with fruit scarcely 8 mm. in diameter.) ane Leaves. rather thin, sometimes shredded at tip. Pedicels nearly or quite equaling the fruit. 424 TRELEASE—THE DESERT GROUP NOLINEZ, [April21, N. NELson! Rose, Contr. U. S. Nat. Herb. 10: 92. 1906. Trunk 1-3 m. high. Leaves 30-40 mm. wide, strongly serrulate-scabrous. Inflorescence with nar- Y bat row ascending divisions 15 cm. long, and branchlets “~~~ —chiefly upwards—scarcely half as long. Fruit? Cet Northeastern Mexico. In the region of Dasylirion longtesimaie a 2 Specimens examined: TAMAULIPAS. Mountains near Miqui- huana (Nelson, 4489, 1898,—the type). N. Betprner Brandegee, Zoe. 1: 305. 1890. a N. Beldingii Brandegee in Bailey, Cycl. Amer. Hort. 3: 1092. My 1901 ; Gard. Chron. iii. 34: 43. f. 18. 1903. | Trunk 3-5 m. high, rather openly branched. Leaves very slightly 3 glaucous, 15-20 mm. wide. Inflorescence long- peduncled, narrow, with narrow divisions 50 cm. long and branches 8-10 cm. long, often again branched ~ with branchlets 1-2 cm. long. Fruit much depressed, _ retuse at base, very large, 8-10 X 15 mm. or more. Seed Gee 4X5 mm. Lower California. The type locality is mountain tops in the Cape a Region. ae Specimeris examined: Lower CaLrrornrA. Sierra de San Fran- cisquito (Brandegee, 583, 1892). La Chuparosa (Brandegee, 1893, 1905). N. Beldingi deserticola Trelease. a Subacaulescent with leaves scarcely 50 cm. long, otherwise re- sembling the type. ; Lower California. In the desert association of N. Palmeri and. N. Bigelovii. a Specimens examined: Lower CALirorNiA. Yubay (Brand. gee, 1889,—the type). N. PARVIFLORA Hemsley, Biol. Centr.-Amer. 3: 372. 1884. Cordyline parviflora HBK., Nov. Gen. Sp. 1: 268. 1815; 7. pl. 674. 1825. Dracena parviflora Willdenow in Schultes, Syst. 7: 348. 1829. 1.J TRELEASE—THE DESERT GROUP NOLINEZ. 425 Roulinia Humboldtiana Brongniart, Ann. Sci. Nat., Bot. ii. 14: 320. 1840. Dasylirium Humboldtii Kunth, Enum. 5: 42. 1850. Nolina Altamiranoa Rose, Proc. U. S. Nat. Mus. 29: 438. 1905. ?Beaucarnea recurvata stricta Baker, Journ. Linn. Soc., Bot. 18: 234. 1880.—As to localities cited. Trunk 2-4 m. high. Leaves 15-20 or 25 mm. Pe wide. Inflorescence with divisions 25 cm. long and G lower branchlets half as long. Bracts very showy, nearly 50 cm. long, caudate-attenuate. Fruit very large, 8-10-12 X 14 mm. Seed 3 X 4 mm. South-central Mexico. The type locality is between Hauhtitlan and Tanepantla. _ Specimens examined: Feperat District. Above Santa Fe (Pringle, 8060, 1899,—the type of N. Altamiranoa; 13620, 1905; Rose & Hay, 5388, 1901; Rose & Painter, 8659, 1905). Rio Hondo Cafion (Pringle, 6787, 1898). Chalchicomula (J. G. Smith, 451, 1892). Guadalupe (Bourgeau, 520, 1865-6). PuesLa. Esperanza (Purpus, 821, 1907). VERA Cruz. Limon (?Trelease, 80, 1905). N. tonciror1a Hemsley, Biol. Centr.-Amer. 3: 373. 1884. _ Yucca longifolia Schultes, Syst. 7: 1715. 1830.—Zuccarini, Allgem. Gartenzeit. 6: 258. Dasylirion longifolium Zuccarini, Abhandl. Akad. Munchen. Cl. II. 3 (=Denkschr, 16) : 224. pl. 1. f. 2.1840; 4 (= Denkschr. 19): 20, 21.—Morren, Belg. Hort. 1865: 321. pl. 20-——Garden. tr: 291. f—Gard. Chron. n. s. 7: 493. f. 73, 567. f- 90.— Fenzi, Bull. Soc. Ort. Tosc. 1890: 112. pl. 6—Rehnelt, Gartenwelt. 11: 14. f—Urban, Gart. Zeit. 3: 66. f. 20—Die Natur. 34: 340. f—Murison, Garden. 24: 433. f.—Garten- flora. 29: 117. f.; 33: 68. f—Roezl, Belg. Hort. 33: 139— Gérome, Rev. Hort. 83: 206. f. 82. Roulinia Karwinskiana Brongniart, Ann. Sc. Nat., Bot. ii. 14: 320. 1840. ?Yucca Barrancasecca Pasquale, Cat. R. Ort. Bot. Napoli. 108. 1867.—See also Zuccarini, /. c., and Rept. Mo. Bot. Gard. 13: 114. 426 TRELEASE—THE DESERT GROUP NOLINEA, (April an Beaucarnea longifolia Baker, Journ. Bot. 10: 324. 1872. oe Trunk 2-3 m. high, swollen at base, at length lo ae branched at top. Leaves 20-30 mm. wide, very long and recurving over the trunk; green. Inflorescence nearly sessile with divisions 30 cm. long and lower branchlets scarcely one-fifth as long. Fruit sub- - orbicular or rather depressed, large, 8 X 10-12 mm. Seed 3 x4 mm.—PI. 3, 8, 13. or South-central Mexico. Inthe region of Dasylirion serratifolium The type locality is given as San Jose del Oro by Schultes, on on authority of Karwinski. Roezl gives its occurrence at about 3,000 m. altitude in Puebla, Oaxaca and Mexico. Specimens examined: Oaxaca. Huachilla (Conzatti). Puree Esperanza (Purpus, 5077, A, ?5076, ?5078). San Luis Tultitlana; (Purpus, 432, 1907, 5079, B, 1908). Curtivatep. Munich Botan- ical Garden, from Karwinski’s seed (Radlkofer, 1901,—semi-typical Palermo Botanical Garden (Trelease, 1, 1905). Bushey Hot Gardens (Blake, 1909). ee J Certain questionable thin- but narrow-leaved forms grown in ga dens under this name, or, in a glaucous form, as var, glauca or as Pincenectitia glauca, appear to be forms of Beaucarnea. 7 CALIBANUS. Rose, Contr. U. S. Nat. Herb. 10: 90. 1906.—Monotypic, basec on the species figured by Hooker for Dasylirium Hartwegianum. Calibanus Hookerii Trelease. Dasylirium Hartwegianum Hooker, Bot. Mag. iii. 15. pi. 5099. ) 1859. | D. Hookerii Lemaire, Ill. Hort. 6. misc. p. 24. 1859. D. caespitosum Scheidweiler, Wochenschr. Verein Beford. tenbau. 4: 286. 1861. D. Hookeri Lemaire, Ill. Hort. 12. misc. p. 52. 1865. ?D. flexile Koch, Ind. Sem. Berol. 1867. Append. 1: 5. Beaucarnea Hookeri Baker, Journ. Bot. 10: 327. 1872. : Calibanus caespitosus Rose, Contr. U. S. Nat. Herb. 10: go. f. pl. 24-5. 1906. 1911.) TRELEASE—THE DESERT GROUP NOLINE. 427 Shortly caulescent. Trunk depressed globose with numerous ¢rowns of leaves. Leaves rather thin, somewhat concave and keeled, narrowly linear, 2-3 mm. wide, serrulate-scabrous on the margin, not brush-like at tip, blue. Inflorescence scarcely 25 cm. long, shorter than the leaves, very short-peduncled, simply 00 panicled with thin spreading branches 6-8 cm. long, a or with exceptional very short and few basal branch- lets. Bracts scarious, much shorter than the subtended branches, _ the floriferous ones and the bractlets inconspicuous, ovate or lanceo- late, little-toothed. Flowers minute. Perianth segments about 1 mm. long. Fruit triquetrously subglobose, 3-ribbed, 4-5 & 5-7 mm. Seed melon-shaped, 3 X 3-4 mm.—P. 6, 8, 9, II, I4. _ East-central Mexico. The type locality is Real del Monte. Specimens examined: Hiparco. Ixmiquilpan (Rose, Painter & Rose, 8954, 1905; Purpus, 1200, 4775, 1905). SAN Luts Porost. ~ San Luis Potosi (Orcutt, 1903; Palmer, 1905). BEAUCARNEA. _ Lemaire, Ill. Hort. 8. misc. p. 57, with plate. 1861—Baker, Journ. Bot. 10: 323. 1872; Journ. Linn. Soc., Bot. 18: 233. 1880, —in both cases including Nolina—Rose, Contr. U. S. Nat. Herb. to: 87. 1906.—Though not monotypic, based primarily on B. ° _ recurvata, and capable of precise definition. s, Leaves with essentially smooth grooves and nearly smooth margins, thin, 'mearly flat, recurved, green. Floriferous bracts rather elongated. Fruit large, rather long-stalked before falling. Slender trees, about 10 m. high, _ moderately enlarged at base. EUBEAUCARNEA, BEAUCARNEA RECURVATA Lemaire, Ill. Hort. 8. misc. p. 61. 1 pl. 1861.—Gard. Chron. 1870: 1445. f. 254; iii. 46: 4. f. 3— Deutsch. Gart. Mag. 1871: 288. p/—Gartenflora. 28: 210. f—Croucher, Garden. 19: 372. f. _ Pincenectitia tuberculata Lemaire, l. c., as synonym. Beaucarnea tuberculata Roezl, Belg. Hort. 33: 138. 1883. Nolina recurvata Hemsley, Biol. Centr.-Amer. 3: 372. 1884.— Rehnelt, Gartenwelt. 11: 78. f—Gard. & Forest. 9: 94. f— Fl. des Serres. 18. misc. p. 26. f—Karsten & Schenck, Vege- tationsbilder. 1. pl. 34—-Gérome, Rev. Hort. 83: 207. f. 83. N. tuberculata Hort. 428 TRELEASE—THE DESERT GROUP NOLINEZ, [April 21, Trunk openly slender-branched above. Leaves 15-20 mm. wide, 1.5-2 m. long. Inflorescence nearly sessile, broadly ovoid-panicled, decompound with divisions 30 cm. long, lower branches nearly half as long and branchlets 5 cm. long. Perianth segments 3 mm. long. Fruit? Southeastern Mexico. Noted by Roezl at Paso del Macho and by Karsten at Sta. Maria, in the State of Vera Cruz.—The type of the genus.—Two garden varieties, intermedia and rubra, are noted by Baker, Journ. Linn. Soc., Bot. 18: 234. 1880. Specimens examined: CuttivaTep. Palermo Botanical Garden (Trelease, 1905). Missouri Botanical Garden. ie B. INERMIS Rose, Contr. U. S. Nat. Herb. 10: 88. f. 2. 1906. Dasylirion inerme Watson, Proc. Amer. Acad. 26: 157. 1891. Trunk rather closely few-branched at top. Leaves 12-15 mm. wide, about 1 m. long. Inflorescence long-stalked, narrowly pyramidal-panicled, somewhat decompound with divisions 30 cm. lorig, slender lower branches half a as long and few branchlets 3-4 cm. long. Perianth 4 y segments scarcely 2 mm. long. Fruit elongated- “~~~ elliptical, 10 X 14mm. Seed (immature) 2 X 3 mm. : East-central Mexico. : Specimens examined: San Luis Potosr. Las Palmas (Pring 3108, 1890,—the type of Dasylirion inerme). San Diegui (Palmer, 644, 1905). VERA Cruz. Zacuapam (?Purpus, 44 1907). East of Huatusco (?Endlich, 1162, 1906). Carrizal (?Goldman, 708, 1901).—The incomplete Vera Cruz material pet haps belongs to the preceding, though short-leaved. B. PLIABILIS Rose, Contr. U. S. Nat. Herb. 10: 89. 1906. ; Dasylirion pliabile Baker, Journ. Linn. Soc., Bot. 18: 240. 188 Trunk openly slender-branched at top. Leaves 15 mm. wi less than 1 m. long. Inflorescence compound-panicle with broad divisions 30 cm. long and few rathe short spreading branches. Perianth segments 3 mm. long. Fruit somewhat obovately round-elliptical. 11 1213-15 mm. Seed 3X4 mm., irregularly 3 lobed, transversely wrinkled.—PI. ro. : tort.) TRELEASE—THE DESERT GROUP NOLINE*. 429 Southeastern Mexico. Specimens examined: YucATAN. Near Sisal (Schott. 892,—the type of Dasylirion pliabile). Progreso (Goldman, 607, 1901). _B. GUATEMALENSIS Rose, Contr. U. S. Nat. Herb. ro: 88. f. 1. 1906. Trunk often with slender multiple stems, variously branched. Leaves 25-30 mm. wide, less than 1 m. long, smooth- Pp Ke) R edged. Inflorescence short-stalked, broadly ovoid- = panicled, decompound with divisions 30 cm. long, 4g . wp rather spreading branches sometimes half as long, and few branchlets 6 cm. long. Perianth segments 3 mm. long. Fruit elliptical-obovate, 13-15 X 15-18 mm., at length openly notched at top and base. Seed 5 mm. in diameter, irregularly 3-lobed, smooth—PI. 7. _ Guatemala. The southernmost species of the group. Specimens examined: GuaTeMALA. El Rancho (Kellerman, _ 4320, 1905,—the type; 5398, 1906; 7015, 1907, and 7029, 1908). CULTIVATED. Guatemala City (Kellerman, 6069, 1907). -B. GorpMantt Rose, Contr. U. S. Nat. Herb. 12: 261. pl. 20. 1909. _. Trunk openly slender-branched above. Leaves 15 mm. wide, scarcely I m. long, essentially smooth-edged. Inflores- cence nearly sessile, compound-panicled with narrow ascending divisions 15-20 cm. long and few strict branches about half as long. Perianth segments about mm. long. Fruit elliptical, very large, 12-15 X 18-20 mm. Seed? Southern Mexico. Specimens examined: Curapas. San Vicente (Goldman, 887, 1904,—the type). ‘Leaves papillate-grooved as in Nolina, rather rough-margined, firm, more or less concave, keeled or plicate, nearly straight, pale or glaucous. Flori- ferous bracts short. Fruit small for the genus, very short-stalked. Trees about 10 m. high, greatly swollen at base. PAPILLAT. B. stricta Lemaire, Ill. Hort. 8. misc. p. 61. 1861. Pincenectitia glauca Lemaire, /. c., as synonym. 430 TRELEASE—THE DESERT GROUP NOLINE. tape * Beaucarnea recurvata stricta Baker, Journ. Linn. Soc., Bot. 8 a 234. 1880. a B. glauca Roezl, Belg. Hort. 33: 138. 1883. nt B, Purpusi Rose, Contr. U. S. Nat. Herb. 10: 89. “1906 Purpus, Mdller’s Deutsch. Gartn.-Zeit. 23: 223. f. Trunk moderately swollen, irregularly rather few-bran hi Leaves more or less keeled or plicate, 8-15 mm. wide, scarcely 1 m. long, the yellowish margin usually minutely serrulate-scabrous. Inflorescence short- stalked, ovoid-panicled, decompound with narrow divisions 20 cm. long and short branches, the lower * with branchlets 3 cm. long. Perianth segments 2 mm. long. Fr i broadly elliptical, 8-10 K 12 mm. Seed 3X 4-5 mm. irregular 3-lobed, smooth.—PI. 8, 14. South-central Mexico. Associated with the next and Dasyl ries lucidum. Specimens examined: Puepra. Tehuacan (Rose, Painter Rose, 10156, 1905,—the type of B. Purpusi; Rose & Rose, 112. 1906; Purpus, 2397, 1907). San Luis Tultitlanapa (Purpus, 5080 1908). Oaxaca, Tomellin Cafion (Rose & Rose, 114274 and b 1906). Almoloyas to Sta. Catarina (?Conzatti, 1644, 1906). B. GRACILIS Lemaire, Ill. Hort. 8. misc. p. 61. 1861. B. edipus Rose, Contr. U. S. Nat. Herb. ro: 88. pl. 23. MacDougal, Publ. Carnegie Inst. 99. pl. 10. ?Nolina histrix Hort. Trunk enormously swollen below, variously and irreg branched. Leaves very glaucous, 4-7 mm. scarcely 50 cm. long, minutely but sharply serre ae 0 scabrous on the paler margin. Inflorescence sh stalked, ovoid- or oblong-panicled, decompound divisions scarcely 30 cm. long and weak branches as long, the lower often similarly branched. Perianth segm scarcely 2 mm. long. Fruit round-elliptical, 7-9 & 10 mm. , 2X 3 mm., smooth.—PI, 4, 1. South- ceatied Mexico. Associated with the preceding. Specimens examined: Pursra, Tehuacan (Rose, Painter 912.1 TRELEASE—THE DESERT GROUP NOLINE. 431 Rose, 10157, 1905,—the type of B. edipus; Trelease, 1, 1903; Pur- pus, 1253a in part, 1905, and 2503, 1907). CuLtivaATED. New York Botanical Garden (Taylor, 25734, 1906). DASYLIRION. _ Zuccarini, Allgem. Gartenzeit. 6: 258, 303. 1838—Plant. Nov. vel Minus Cognit. 4: 221. pl. r (Abhandl. Akad. Miinchen. Cl. II. _3-—=Denkschr. 16). 1840;—Plant. Nov. etc. 5: 19. (Abhandl. _ Akad. Miinchen. Cl. II. 4. = Denkschr.19). 1845—Kunth, Enum. _ Plant. 5: 38. 1850,—as Dasylirium—Baker, Journ. Bot. 10: 296. 1872; Journ. Linn. Soc., Bot. 18: 237. 1880.—Rose, Contr. U. S. Nat. Herb. 10: 89. 1906—Though primarily based on Yucca pitcairniefolia (Hechtia glomerata) and made to include with ques- tion Yucca [Nolina] longifolia, it finally stood with its author for e prickly-leaved sotols, of which D. serratifolium and D. gramini- folium were definitely included in his first publication on the genus d mark its type. ves 2-edged, usually somewhat concave and irregularly keeled, prickly- _margined and usually roughened with minute intervening denticles or -serratures (pp. 431-440). EUDASYLIRION. Fruit small (3-5 mm. wide). Fruit normally with moderately deep notch, narrowly elliptical to obovate, the style not surpassing the wings. Perianth segments about 2 mm. long. Leaves elongated, rather wide (usually 15-20 mm.). sis prevailingly upcurved. Dasylirion cedrosanum Trelease. _ Dasylirion sp. Kirkwood, Pop. Sci. Monthly. 75: 438, 445. f. _ Shortly caulescent. Trunk 1-1.5 m. high. Leaves 20 mm. wide, ais upwards of 1 m. long, slightly brush-tipped, glaucous, wei *() slightly rough-keeled, dull; prickles mostly 10-15 mm. apart: 2-5 mm. long, yellow, becoming red upwards. Be Inflorescence 5 m. high. Fruit very narrowly elliptical, 4-5 X 7-9 mm., the style barely half as long as the narrow deep notch. Seed 23.5 mm. PI. 5, 12, 15. Northeastern Mexico. Overlapping the range of WNolina cespitifera. 432 TRELEASE—THE DESERT GROUP NOLINEAE, [April ar, Specimens examined: Zacatecas. Cedros (Lloyd, 118,—the type, and 82, 1908; Kirkwood, 96, 1908). CoAnurLta. Rancho La Luz (?Endlich, 7, 1905). Saltillo (?Gregg, 78, ios SS stura (?Wislizenus, 307, 1847). D. LtucipuM Rose, Contr. U. S. Nat. Herb. 10: 90. 1906. — Dasylirion sp. Schenck & Karsten, Vegetationsbilder. 1. pl. Shortly caulescent. Trunk 1-2 m. high or sometimes prostrat ly elongated. Leaves 10-17 mm. wide, scarcely 1 m. Ja long, strongly brush-tipped, typically yellowish, smooth and glossy: prickles mostly 10-15 mm. apart, 2-3 mm. long, from yellow passing through red to almost chestnut, the margin often reddish. Inflorescence 2-3 high. Fruit narrowly elliptical-obovate, 4-5 X 7-8 mm., the rat eh slender style about half as long as the rather narrow deep n note . Seed 2.5-3.5 mm.—PI. 2, 10, 12. South-central Mexico. With Beaucarnea gracilis and stricta. Specimens examined: Puesta. Tehuacan (Rose, Painter Rose, 10009, 1905,—the type; Trelease, 1903, 1905; Purpus, 39 7, 1909). Esperanza (??Purpus, 1907). San Luis Tultitlanapa (EP. pus, 5082, 1908). D. Palmeri Trelease. ; Habit? Leaves 25 mm. or more wide, scarcely 1 m. long, som what brush-tipped, green or lightly glaucous, smoot — () dull: prickles mostly 15-25 mm. apart, 3-5 mm. 7 yellow, becoming brown upwards, the interv ——~~~~ margin rather smooth. Inflorescence moderatel Fruit narrowly elliptical- or triangular-obovate, 3-4 x 6 mm. stout style about equaling the rather open shallow notch. Seed mm.—PIl. 12. Northeastern Mexico. | Specimens examined: CoAnurLa. San Lorenzo Cafion ( aln 696, 1905,—the type). D. Parryanum Trelease. Habit? Leaves heron 10-15 mm. wide, scarcely 50 cm. rg1t.] TRELEASE—THE DESERT GROUP NOLINE#. 433 brush-tipped, whitened, minutely roughened and SOK (i) dull: prickles about 5 mm. apart, 2 mm. long, yellow, Our becoming red upwards, the margin very rough be- tween them. Inflorescence moderate, exceptionally sub-simple. Perianth segments scarcely 2 mm. long. Fruit elliptical or somewhat obovate, 4 < 6 mm., the style scarcely equaling the narrow moderately deep notch. Seed? East-central Mexico. In the region of D. graminifolium and Nolina humilis and Watsont. Specimens examined: San Luts Porost. Vicinity of San Luis Potosi (Parry & Palmer, 876, 1878,—the type; Schaffner, 242, 1878,—mixed with D. graminifolium?). Prickles prevailingly recurved. D. leiophyllum Engelmann in herb. Dasylirion sp. Trelease, Pop. Sci. Monthly. 70: 220. f. 14. Shortly caulescent. Leaves becoming 15-20 mm. wide, scarcely I m. long, somewhat brush-tipped, green or at first somewhat glaucous, smooth, rather glossy: prickles usually I0-I5 mm. apart, 3-4 mm. long, yellow, usually be- ie () coming orange or red at least above the middle, the >. 9 ys margin sometimes smooth between them. Inflores- cence rather high. Perianth segments scarcely 2 mm. long. Fruit obovately subelliptical, 4-5 X 6-8 mm., the thick style about equaling the moderately open and deep notch or exserted if the wings have not fully developed. Seed 2X 3 mm.—PI. 12. _ Southern Texas, in the Rio Grande region, passing into New Mexico and reaching or reappearing in central Chihuahua—Ad- joining or overlapping the range of D. Wheeleri Wislizeni and _ Nolina erumpens and affinis. Specimens examined: Texas. Presidio (Havard, 1830,—the type). Eagle Pass (?Havard, 1883). Sierra Blanca (Trelease, 1892, 388, 1900; Mulford, 275, 1895; Rose, Standley & Russell, 12222, 1910). Van Horn (Eggert, 1900). New Mexico. Cen- tral (Mulford, 424, 1895). Florida Mountains (Mulford, 1037, 1895). CHiHuAHUA. Sta. Eulalia Mts. (Pringle, 149, 1885; _ Williamson, 1885). PROC. AMER. PHIL. SOC., L. 200 BB, PRINTED AUG. 7, IQII. 434 TRELEASE—THE DESERT GROUP NOLINEZ:, [April 21, Fruit with very shallow notch, broadly elliptical, the style rather surpass- ing the wings. Prickles prevailingly upcurved. D. TEXANUM Scheele, Linnza. 23: 140. 1850.—Bray, Bull. Univ. Tex. 82. pl. 13; Bot. Gaz. 32: 288. f. Shortly caulescent or with buried trunk. Leaves 10-15 mm. wide, scarcely 1 m. long, somewhat brush-tipped, green, smooth or rough-keeled, glossy: prickles 5-10 START) mm. apart, 2-3 mm. long, yellow, becoming brownish. \ % Zt Inflorescence 3-5 m. high. Fruit elliptical, 4-6 & 7-8 mm., the very short style equaling or surpassing the open shallow notch. Seed ?—PI. 12, 15. South-central Texas. In the range of Nolina texana and Lind: heimeriana. Specimens examined: Texas. Vicinity of New Braunfels (Lind- heimer, 548, 1845,—apparently the type, 549, 1846, 1271-1213, 1849). Blanco Cafion (Reverchon, 1605, 1885). Putnam (Trelease, 1892). Gillespie County (Jermy). Kerr County (Heller, 1920, 1894; Bray, 228, 1899). Hueco Tanks (Mulford, 90,1895). Sanderson (Wein- berg, 1907; Thompson, 1911). Marathon (Lloyd, 1910). Com- stock (Thompson, 1911). Ft. Davis (Blake). D. texanum aberrans Trelease. Differing from the type in its dull somewhat glaucous leaves, 15_ mm. wide. Northern Mexico. Specimens examined: “States of Coahuila and Nuevo Leon” (Palmer, 1315, 1880,—the type). Fruit rather small (5-6 mm. wide), subcordate, the style not surpassing the wings. D. simplex Trelease. Acaulescent. Leaves 7-10 mm. wide, scarcely I m. long, rather sparsely very long-fibrous at tip, green, smooth, glossy: prickles 10-15 or 20 mm. apart, 2-3 mm. long, rather straight, JKEO prevailingly upcurved, yellow, the upper half becom- 12 _O ing brown or finally almost black, the margin nearly — smooth between them. Inflorescence (typically?) small, the staminate with one to three small branches to each division 1911.] TRELEASE—THE DESERT GROUP NOLINE#. 435 and the pistillate simple or its divisions with a short basal branch. Fruit broadly obovate, 5-6 X 7 mm., the thick style about equaling the open moderate notch. Seed (immature) 2 3 mm.—PI. 12. North-central Mexico. In the region of Nolina durangensis. Specimens examined: Duranco. Tepehuanes (Palmer, 310, 1906,—the type). Santiago Papasquiaro (Palmer, 422, 1896). Fruit moderately large (usually 6-8 mm. wide), the style not surpassing the wings. Prickles prevailingly upcurved. Leaves not brush-tipped, glaucous. D. GLAUCOPHYLLUM Hooker, Bot. Mag. iii. 14. pl. 5047. 1858.—[As Dasylirium].—Baker, Journ. Bot. 10: 298. 1872.—?Gard. Chron. iii. 40: 247. f. Tor. D. glaucum Carriere, Revue Hort. 44: 435. f. 1872——Copied in Garden, 3: 23. f., and Florist and Pomol. 1874: 17. f—Gard. Chron. n. s. 13: 82, 205. f. 37—??Roezl, Belg. Hort. 33: 139. ?D. serratifolium Rept. Mo. Bot. Gard. 14: 12. f. Bonapartea glauca Hort. Shortly caulescent. Trunk scarcely 5 m. high. Leaves as much as 12 mm. wide, over 1 m. long, glaucous, nearly or , quite smooth, dull: prickles 5-10 mm. apart, 2 mm. j / long, yellowish white or the tips becoming slightly Te brownish. Inflorescence 4-6 m. high. Fruit sub- elliptical, 6 X Q-10 mm., the thick style about half as long as the closed deep notch. Seed 2.5 X 4 mm.—PI. 12. East-central Mexico. In the region of D. acrotriche and Cali- banus.—The type locality is Real del Monte. Known to me only in cultivation. The fruit description is based on material cultivated at La Mortola (Berger, 560, 1911). Leaves more or less brush-tipped. Leaves narrow (scarcely Io mm.), strongly brush-tipped. D. AcrotRICHE Zuccarini, Abhandl. Akad. Munchen. Cl. II. 3 (= Denkschr. 16) : 226, 228. pl. 1. f. iv. 1840. Yucca acrotricha Schiede, Linnea. 4: 230. 1829; 6: 52— Schultes, Syst. 7: 1716. 1830. Roulinia gracilis Brongniart, Ann. Sc. Nat., Bot. ti. 14: 320. 1840. 436 TRELEASE—THE DESERT GROUP NOLINEZ, [April 21, — Barbacenia gracilis Brongniart, /. c—As synonym. ere Yucca gracilis Otto, Allgem. Gartenzeit. 9g: 123. 1841.—As | synonym. i Bonapartea gracilis Otto, |. c—As synonym. “ Dasylirion gracile Zuccarini, Abhandl. Akad. Miinchen. cl. Il. 4 : (== Denkschr. 19): 22. 1845. er Dasylirium acrotrichum Kunth, Enum. Pl. 5: 40. 1850.—Hooker, Bot. Mag. iii. 14. pl. 5030.—Copied in Fl. des Serres. 14. pl. — 1448.—Schlotthauber, Deutsch. Mag. f. Gart.-u. Blumen- kunde. 1871: 49, 64, 81, 96. 2 pl—Koopmann, Gartenwelt. 3: : 375-6. f. Dasylirium gracile Planchon, FI. des Serres. 7: 6, 10. f. 1851-2. Littea gracilis Verschaffelt, Cat. 1864.—Hansgirg, Phyllobiol- ogie. 422. a Dasylirion acrotrichum Baker, Journ. Bot. 10: 297. 1872.— Regel, Gartenflora. 30: 24. f—Gard. Chron. iii. 19: 204. pl. —Deutsch. Gartner-Zeit. 20: 536. f. | ?D. robustum Hort.—( Perhaps = serratifolium. ) Caulescent. Trunk at length 1 m. or more high. Leaves 6-10 or rarely 15 mm. wide, less than 1 m. long, green _ & ny and glossy or somewhat glaucous and dull, often — / rough on the keels: prickles 5-10 or 15 mm. apart, scarcely 2 mm. long, rather straight, pale yellowish, — with slightly brown tips. Inflorescence 3-5 m. or more high. Perianth segments 2-3 mm. long. Fruit round-cordate, 6-7 X 8-9 mm., the thick style about equaling the shallow noteh. : Seed 3 X 3.5 mm.—PI. 11, 12. 4 East-central Mexico. Collected by Schiede and Deppe on tee | Serro de la Ventana, on the flanks of Mt. Orizaba, by Karwinski at Ixmiquilpan, and by Reppert at Real del Monte. Deppe’s original seed collection, in 1825 (Otto, Allgem. Gartenzeit, 16: 276. 1848) was made between Real del Monte and Pachuca, and most of the — earlier plants of European gardens were raised from this at Berlin - (Bouché, Monatsschr. Verein Beford. Gartenbau. 1880: 481).— Range of the preceding. Specimens examined: Hipatco. Dublan (Pringle, 11196, 1902; & — -3gtt.) TRELEASE—THE DESERT GROUP NOLINE2. 437 q Rose & Hay, 5305, 1901). Metepec (Pringle, rooor, 1904). Tula | (Pringle, 6637, 1897 ; Rose, Painter & Rose, 8280, 1905). Ixmiquil- _ pan (Rose, Painter & Rose, 8969, 9029, 1905). Sierra de Pachuca _ (Rose, Painter & Rose, 5571, 1901 ; 8801, 1905; Rose & Rose, 11484, - 1906). QueRETARO. Cadereyta (Rose, Painter & Rose, 9714, 1905). ~ San Luts Potost. San Luis Potosi (Parry & Palmer, 876, 1878). Leaves wide (rarely under 15 mm.), only moderately brush tipped. Fruit elliptical or obovate. Prickles small. D. GRAMINIFOLIUM Zuccarini, Allgem. Gartenzeit.6: 259, 303. 1833; Abhandl. Akad. Munchen. Cl. II. 3 (== Denkschr. 16): 225: pl. 1, f. 1. 1840—Kunth, Allgem. Gartenzeit. 9: 121. pl. 1. Yucca graminifolia Zuccarini, Cat. Hort. Monac. 1837. Dasylirium graminifolium Kunth, Enum. Pl. 5: 39. 1850. Subcaulescent? Leaves 12 mm. wide, about 1 m. long, green, smooth, glossy: prickles mostly 5-10 mm. apart, I \ (} mm. long, yellowish white or with slightly darkened tips. Inflorescence moderately high. Fruit broadly elliptical, 6 X 8-9 mm., the thick style equaling the rather open shallow notch. Seed ?—PI. 12. East-central Mexico. The type is said by Otto to have been 4g raised from seed sent with that of D. acrotriche by Deppe in 1827.— As the earliest well-characterized and figured species, this may per-_ g haps be accepted as the type of the genus, though it follows D. ser- . ratifolium in position—lIn the region of D. Parryanum, Calibanus 3 and Nolina humilis and Watsonti. 3 Specimens examined: San Luis Potosi. Vicinity of San Luis Potosi (Parry & Palmer, 876, 1878; ?Schaffner, 242, 1878,—mixed | 4 with D. Parryanum). Las Canoas (Pringle, 3746, 1891). a Leaves of D. hybridum of the Botanical Garden of Rome do not 4 differ from this except in being 15-20 mm. wide, nearly dull and _ scarcely brush-tipped. It is said to be a hybrid between D. [Nolina] _longifolium and D. serratifolium, but in foliage shows no characters of the former, and it is unlike the latter as I understand it. Prickles moderate. 438 - TRELEASE—THE DESERT GROUP NOLINEZ®, _ [April at, D. durangense Trelease. | Habit? Leaves 20 mm. wide, scarcely 1 m, long, lightly ack 4 cous, nearly smooth, dull: prickles 5-10 mm. apart, eS ae 2-3 mm. long, yellow, with orange tips. Inflorescence \of tall. Fruit broadly elliptical-cordate, 7-8 & 9 mm., d Suey 4 the thick style scarcely half as long as the rather open ~ == deep notch. Seed 2 3 mm.—PI. 11, 12. ees North-central Mexico. In the region of Nolina durangride EF Specimens examined: Duranco, Durango (Palmer, 557, 1896,— the type). D. SERRATIFOLIUM Zuccarini, Allgem. Gartenzeit. 6: 258, 303,—name only. 1838; Abhandl. Akad. Miinchen. Cl. II. 3 (= Denkschr. 16): 225, 228. pl. 1. f. tii. 1840—Roezl, Belg. Hort. 33: 139. pape Yucca serratifolia Schultes, Syst. 7. 1716. 1830. hes Roulinia serratifolia Brongniart, Ann. Sc. Nat., Bot. ii. 14: 3 1840. | Dasylirium serratifolium Kunth, Enum. Pl. 5: 41. Beek Dasylirion laxiflorum Baker, Journ. Bot. 10: 299. 1872, ?D. robustum Hort. oe s Subacaulescent. Leaves 15-20 or even 35 mm. wide, scarcely gy 1 m. long, whitish, finely roughened on one or both | > () faces, dull: prickles 5-10 or even 20 mm. apart, 2-3 MK mm. long. Inflorescence ample. Fruit quadratel round-obovate, the style equaling the narrow rather deep is Seed 3 X 4 mm. Riise g® Southeastern Mexico. Collected by Andrietx near Oana 1 by Karwinsky at San Jose del Oro.—Region of Nolina longifolia. — Specimens examined: Oaxaca. Las Sedas (Pringle, 6697, 1807); Nochistlan (Conzatti & Gonzales, 1899). D. WHEELERt Watson in Rothrock, Rept. Wheeler. 6: 378. 18 Proc. Amer. Acad. 14: 249. 1879.—Wooton, Bull. N. Me Exper. Sta. 18: 92. pl—Lloyd, Plant World. 10: 254. f. 5I- MacDougal, Publ. Carnegie Inst. 99: 74, pl. 58—De Wild man, Icones Sel. Hort. Thenensis. 6: 91. pl. 225.—Sket habit figures, without name, are given by Schott in Emo: Rept. Bound. Surv. 1. pl. 37, 42. TRELEASE—THE DESERT GROUP NOLINE. 439 Shortly caulescent. Trunk scarcely 1 m. high. Leaves 15-20 or 25 mm. wide, scarcely 1 m. long, glaucous, nearly 3 (7) smooth, dull: prickles 5-10 mm. apart, 2-3 mm. long, 4 yellow, becoming brown upwards. Inflorescence 3-5 a high. Fruit round-obovate, 6-7 * 7-9 mm., the _ style normally about equaling the open moderately deep notch. Seed ' (immature) 3 mm. long—PI. 7, 8, 11, 12. : Southeastern Arizona and adjoining Mexico, New Mexico and _ Texas. Region of Nolina microcarpa and caudata. __ Specimens examined: Arizona. Ash Creek, etc. (Rothrock, 329, _ 655,—the types). Rio Grande to Gila Rivers (Emory, 1846). Sun- flower Valley (Girard, 1873). Dragoon Summit (Vasey, 1881). 4 Sta. Catalina Mountains (Lemmon, 1881; Pringle, 1881; Toumey, 1894). Without locality (Pringle, 1884). Sta. Rita Mountains 3 (Brandegee, 1891). Ft. Huachuca (Wilcox, 208, 264, 1894). Chiri- ' cahua Mountains (Toumey, 1894). San Carlos (Straub, 1895). Nogales (Coville, 1623, 1903; Thompson, 1911). White Tail (Pils- bry, 1906). Benson (Rose, Standley & Russell, 12326,1910). NEw Mexico. Silver City (Greene, 1880; Metcalfe, 637, 1903). Burro “Mountains (Rusby, 413, 1881,—leaves). Pinal Mountains (Toumey, 49, 1892). Las Cruces (Wooton, 72, 1897). Mangos (Metcalfe, 897). Kingston (Metcalfe, ror4, 1904). Alamogordo (Rehn & Viereck, 1902). Organ Mountains (Standley, 1906). Texas. El aso (Evans, 1891). Tortugas Mountains (Rose, Standley & Rus- ell, 12254, 1910). BouNpary LINE (Parry et al.). CHIHUAHUA. Lake Sta. Maria (Nelson, 6393, 1899). 4 Fruit triangular-obcordate: prickles moderate. D. Wheeleri Wislizeni Trelease. . Shortly caulescent. Leaves 15—20 mm. wide, scarcely I m. long, een or slightly glaucous, typically smooth and rather = lossy: prickles 5-10 mm. apart, 2-3 mm. long, red- rown or with yellow base, the intervening denticles ften reddish. Inflorescence ample. Fruit triangular- ROL, bcordate, 6-7 X 8-9 mm., the thick style about equaling the open derately deep notch. Seed (immature) 3 mm. long.—PI. 12. 440 TRELEASE—THE DESERT GROUP NOLINEZE, [April North-central Mexico and adjacent Texas,—apparently grading into D. Wheeleri. Adjoining or overlapping the area of D. Wheel: and Nolina erumpens. | Specimens examined: Curmuanua. Paso del Norte [Juarez] (Wislizenus, 218, 1846,—the type; ?Stearns, 1910,—with smaller, slightly roughened dull leaves). Without locality (Thurber, 1852) Texas. El Paso (Dewey, 1891; Wagner, 985, 1892). Franklin Mountains (Rose, Standley & Russell, 12280, 1910). i ' Fruit large (8-9 mm. wide), the style surpassing the wings. D, BeRLANDIERI Watson, Proc. Amer. Acad. 14: 249. 1879. Habit? Leaves? Inflorescence apparently ample. Bractlets rather long, lanceolate, finely toothed. Perianth seg- ments 2-4 mm. long. Fruit round-elliptical, 7-9 x 7-10 mm., the style rather exceeding the very open moderately deep notch. Seed ?—PI. 12. Northeastern Mexico. ; Specimens examined: Nuevo Leon. La Silla, Monterey (B landier, 3218, June, 1843,—the type). Leaves 4-sided, unarmed. QUADRANGULA D. LtoncisstmMuM Lemaire, Ill. Hort. 3. misc. p. 91. 1856. D. quadrangulatum Watson, Proc. Amer. Acad. 14: 250. 1879 Gartenflora. 36: 280. f. 75.—Bull. Soc. Tose. Ort. 9: 236. 35: 331. pl. 6—Die Natur. 34: 340. f—Hooker, Bot. M iii, 56. pl. 7740. D. juncifolium Rehnelt, Gartenwelt. 11: 77. f. 1906. . Caulescent. Trunk 1-2 m. high. Leaves narrowly linear mm. wide, at length 2 m. long, not brush-tipped, green, dull, rho ‘ ‘4. or square in section, smooth, the edges minutely i o@) ular-roughened or further with very low eleva 10-30 mm. apart representing the prickles of species. Inflorescence 2-6 m. high. Perianth ments 3-4 mm. long. Fruit broadly obovate or elliptical, 5-8 X mm., the style surpassing the open very shallow notch. as mm.—PI. 9, 12. Eastern Mexico. Of wide range, overlapping the regions Nolina Nelsoni, Calibanus and D. acrotriche. tort.) TRELEASE—THE DESERT GROUP NOLINE. 441 Specimens examined: TAMAULIPAS. Sierra Nolas, between San Luis Potosi and Tampico (Palmer, 1878-9,—the type of D. quad- rangulatum). Miquihuana (Nelson, 4480, 1898). San Luts Porost. Minas de San Rafael (Purpus, 5009, 1910,—a form with small fruit, 4X 4 mm., with style and wings abbreviated and equal). H1patco. Sierra de la Mesa (Rose, Painter & Rose, 9007, 1905,—called “junquillo”’). TEXT REFERENCES. * Bouché, Sitzungsber. Ges. Naturf. Freunde, Berlin. 1875: 118—Mon- atsschr. Verein Beférd. Gartenbau. 23: 482. 1880. * Bray, Bull. Torr. Bot. Club. 30: 627. f. 6. 1903——Bull. Univ. Texas. 60: 22-24. 1905. * Bruno, Boll. Soc. Natural. Napoli. 19: 159. 1906. “Christy, New Commercial Plants and Drugs. 6: 42. 1882. *Engler & Prantl, Natiirl. Pflanzenfam. 2 Teil. 5 Abteil. p. 71. f. 51. 1887. * Hansgirg, Sitzungsber. Bohm. Gesellsch. rg01™: 31.—Phyllobiologie. 421. 1903. * Havard, Bull. Torr. Bot. Club. 23: 43. 1896—Amer. Journ. Pharm. 68: 267. 1896. * Hooker, Curtis’s Bot. Mag. iii. 14. pl. 5047. 1858. * Kirkwood, Pop. Sci. Monthly. 75: 446. 1909. * Klebs, Unters. Bot. Inst. Tiitbingen. 1: 568 f. 73. 1885. ™ Lemaire, Il]. Hort. 8. misc. p. 61. 1861.—See also Gard. & For. 9: 94. 1896. * Lloyd, Plant World. 10: 254-5. f. 51. 1907. * McClendon, Amer. Nat. 42: 308. ff. 1908. * Newberry, Bull. Torr. Bot. Club. 10: 123-4. 1883. * Orcutt, Bull. Torr. Bot. Club. 10: 106-7. 1883. * Pirotta, Ann. R. Ist. Bot. Roma. 3: 170. pl. 20, 21. 1888. * Preda, Bull. Soc. Bot. Ital. 1896: 135-141. * Reverchon, Bot. Gaz. 11: 213, 216. 1886. * Rose, Contr. U. S. Nat. Herb. 5: 224, 240. pl. 36, 37. 1880. *Solms Laubach, Bot. Zeit. 36: 69. pl. 4. 1878. * Trelease, Pop. Sci. Monthly. 70: 219. 1907. *Went & Blaauw, Proc. Sect. Sci. K. Akad. Amsterdam. 8: 684; Rec. Trav. Bot. Neerland. 2: 223. pl. 5. 1906. * Wooton, Bull. N. Mex. Agr. Exp. Sta. 18: 92. 1806. * Zuccarini, Allgem. Gartenzeit. 6: 303. 1838; Abhandl. Akad. Miinchen. Cl. II. 3: 224, 228. 180. In addition to those noted in the above papers, histological studies are to be found in De Bary, Vergl.-Anat. 636-640.—Cedervall, Anat.-Fys. Unters.— Cerulli-Irelli, Ann. R. Ist. Bot. Roma. 5: 414.—Falkenberg, Vergleich. Unters. PROC. AMER. PHIL. SOC., L. 200 CC, PRINTED AUG. 7, IQII. 442 TRELEASE—THE DESERT GROUP NOLINEA:, [Aprilar, Monocot.—Giovannozzi, Nuov. Giorn, Bot. Ital. ns. 18; 9, 53. f. 14.—Gre- villius, Bot. Notiser. 1887: 140.—Haberlandt, Ber. Deutsch. Bot. Ges. 4: 223.— __ Hausmann, Beih, Bot. Centralbl. 23. Abt. 2: 43-80. #—Kny, Bot. Wandtafeln, - Abt. 5.—Mobius, Ber. Deutsch. Bot. Ges. 5: 22.—Morot, Ann. Sci. Nat, Bot. vi. 20: 272.—Schoute, Flora. 92: 42, 46. pl. 4. f. 5, 10.—Schwendener, Abhandl, Akad. Berlin. 1882. ai er eet es EXPLANATION OF ILLUSTRATIONS. The distribution map indicates the occurrence of specimens ; actually examined. Half-tone plates are from unpublished photographs by the author unless otherwise credited. Text-cuts are “uniformly reduced from ‘ged camera lucida drawings to natural size except that leaf sections are x2, ‘the finer arming of leaf margins X 20, and the style and wing tips of | Dasylirion — X 6; and a few exceptional details with other enlargement are introduced. — Prates 1-4. Habit of growth: 1, Trunkless (Nolina microcarpa, Arizona, MacDougal) ; 2, with elongated finally erect caudex (Dasylirion lucidum, Tehuacan) ; 3, arborescent (Nolina longifolia, cultivated in the Pa ern he botanical garden) ; 4 arboreous (Beaucarnea gracilis, Tehuacan). Ae reatly — reduced. reduced. Ee Prats 6-7. Inflorescence Getails :—6 A, Nolina caudata 09 size, : Pirate 8. Flowers.—A, Nolina longifolia (Radlkofer); B, ‘Ca i a Hookerii (Rose, 8954); C, Beaucarnea stricta (Purpus, 2397); D, De ] Wheeleri (Toumey).—All X to. oa PLaTE 9. Septal nectar slits as shown on the matured fruit. A, Dasylirion longissimum (Palmer); B, Calibanus Hookerii (Purpus, 1200).—Both > 25. PiaTe 10. Seeds. A, three seeds of Nolina durangensis—the middle one sectioned to show coat, endosperm and embryo; a seed of Beaucarnea lia- bilis; and two seeds of Darylirion lucidum.—All X 3. B, en ‘ Dasylirion lucidum, with “ reserve-cellulose” walls. X 200. C, cross” roe of seed of Nolina durangensis showing embryo cavity with much extn protoplasm and oil—in chlor-iodide of zinc. X 20. D, endosperm of ‘Noting durangensis swollen in chlor-iodide of zinc, with extruded oil. X 200. a PLATE 11. Fruit characters. A, four fruits and a seed of No georgiana; six fruits and two seeds of Dasylirion Wheeleri; six fruits two seeds of Beaucarnea gracilis; and four fruits and a seed of Calibanus Hookerii.—All natural size. B, 1, Dasylirion acrotriche (3 and 4-winged) ; 2, D. acrotriche (4- and 5-winged); 3, D. durangense (2- and 5-winged);— D. Wheeleri (4- and 5-winged); 5, Calibanus Hookerit Can All X 2. PROCEEDINGS AM. PHiILOS. Soc. VoL. L. No. 200 PLATE | — NOLINA MICROCARPA WNGIONT NOIWMASVd {| atv1d 002 ‘ON “J “10A ‘908 "SO1IHd ‘WY SONIGS300Nd PLATE Ill PHiILOs. Soc. VoL. L. No. 200 PROCEEDINGS Am. NOLINA LONGIFOLIA ——————————— PROCEEDINGS Am. PHILOS. Soo. VoL. L. No. 200 PLATE IV BEAUCARNEA GRACILIS peo am e Rect Shy PLATE V PROCEEDINGS Am. PHILOS. Soc. VoL. L. No. 200 A.—Nolina INFLORESCENCE OF NOLINEAE A.—Nolina B.—Calibanus INFLORESCENCE OF NOLINEAE ce aT eR Ae PROCEEDINGS Am. PHILOS. Soc. VoL. L. No. 200 FLaTe VIl A —Beaucarrea B.—Dasylirion INFLORESCENCE OF NOLINEAE PLATE VIII FLOWERS OF NOLINEAE s Am. PHILOS. PROCEEDING Seon anubdanus B.—C PLATE X SEEDS OF NOLINEAE SMETANA as ae Done Boy! se is te tY, PROCEEDINGS AM. PHILOs. Soc. VoL. L. No. 200 PLATE XI A—Normal B.—Teratological FRUITS OF NOLINEAE 3 ee a me 3 ces 3, EE —C PROCEEDINGS Am. PHILOS. Soc. VoL. L. No. 200 PLATE XII A.—Nolina PP eORG OVE 100 O88 Gv y 99 3 B.—Dasylirion FRUITS OF NOLINEAE PROCEEDINGS Am. PHILOS. Soc. VoL. L, No. 200 PLATE XIll GERMINATION OF NOLINA LONGIFOLIA ey cele PROCEEDINGS Am. PHiLOS. Soc. VoL. L. No. 200 PLATE XIV A —Beaucarnea B.—Calibanus GERMINATION OF NOLINEAE PROCEEDINGS Am. Puitos. Soc. Vor. L. No. 200 PLATE XV HAUSTORIUM AND LEAF TIPS OF DASYLIRION oe x WwW = < ad a 4 Zz < oO 8 : : : (e} 4 a ¢ a — _ z 9° al (@) . g Oo () Ww o °o = po a > concerned, but that mountain ranges were at least in part sup- ed by the rigidity of the crust. When Dr. Nansen drifted across the North Polar basin in the am he provided pendulums to determine the force of gravity when the ship was frozen in ice; and the discussion of his observa- é Hons showed that gravity was normal over that basin, or, at least, where his observations were made.* Professor Helmert,’ in Germany, has done much in the discus- sion of gravity measures and Dr. Hecker has made some notable s and has determined the forces of gravity at sea, over the fic, Indian and Pacific oceans, and over the Black Sea, the ults showing that on the whole the force of gravity is normal er these bodies; only in special and limited areas, in the neighbor- d of very steep slopes, was any marked anomaly found.*® But the most important work which has been done along this is the work of Dr. John F. Hayford,* who, while connected with = United States Coast and Geodetic Survey, discussed in a thor- h and able manner the deflections of the vertical at a large num- - of stations in different parts of the United States, and his results w definitely that over this region isostatic equilibrium actually sts. He has concluded that this is true even for areas as small square degree, that is, seventy miles on the side. He believed “The Norwegian North Polar Expedition of 1893-96,” Volume IL, Par VIIL., Results of the Pendulum Observations, by E. O. Schiotz. _*“Hohere Geodesie,” Leipzig, 1880. _** Bestimmung der Schwerkraft auf dem Atlantischen Ozean,” Veréff. s Kénig. Preuss. Geodet. Instit., Neue Folge, No. 11. “Bestimmung der Schwerkraft auf dem Indischen und Groszen Ozean,” Veréff. des Zentral sureaus der Internat. Erdmessung, Neue Folge, No. 18. “ Bestimmung der schwerkraft auf dem Schwarzen Meere,” same, No. 20. -*“The Geodetic Evidence of Isostasy, etc.,” Proc. Washington Acad .» 1906, Vol. VIII., pp. 25-40. “The Earth a Failing Structure,” Bull. . Soc., Washington, 1907, Vol. XV., pp. 57-74. “The Figure of the n and Isostasy,” United States Coast and Geodetic Survey, 1909. “Sup- entary Investigation in 1909 on the Figure of the Earth and Isostasy,” 1910. “The Relation of Isostasy to Geodesy, Geophysics and Geology,” , February 10, 1911. 448 REID—ISOSTASY AND MOUNTAIN RANGES, (April 21, that the earth is not strong enough to sustain an added thickness of more than about two hundred and fifty feet of rock over an area as large as a square degree without slowly yielding. The stations where the observations were made are scattered over various parts of this country, on the eastern coast, in the Appalachian mountain range, in the region of the Great Lakes, near the Gulf of Mexico, in the great plains of the Mississippi basin, on the great elevations of the Rocky Mountains, the plateaux of Utah, the Sierra Nevada mountains and the Pacific coast, regions exhibiting a great variety -_ of topographic forms and differing greatly as to geologic activity. Whatever movements may be going on in the Rocky mountains, and in the region.between them and the Atlantic ocean, are certainly very small; whereas to the west, and particularly in the state of California, the movements seem to be very active. The eastern edge of the Sierra Nevada received additional elevation at the time of the Owens Valley earthquake in 1872, and the comparatively frequent earthquakes in the Sierras and the Coast ranges make it quite possible that these mountains are now being elevated as actively as at any time in their history. In view of the great variety of the country in which the stations were located, both as to topography and geologic activity, in view of the great amount of material being continually eroded from one region and deposited in another, thus tending to overthrow the isostatic equilibrium, and in view of obser- vations in other parts of the world, we are driven, with Dr. Hay- ford, to the conclusion that isostasy is not an accidental condition — existing at the present time within this country, but is due to the fact that the earth yields plastically to the long continued action of even small forces. We feel justified, therefore, in believing that isostatic equilibrium exists in other parts of the world and existed in other geologic ages, and in saying that the whole earth is, and 4 always has been in isostatic equilibrium. ] This conclusion carries with it many important consequences and ~ has a very direct bearing on the theories of the origin of mountain — 4 ranges; for it tells us that every segment of the earth, having an q equal area of surface and with its apex at the center, contains the | same amount of material, which it is impossible either to increase ae 4 ‘4 & pig ity 5ST Ble ace rte a EPR EG ea esd Sp ‘tott.] REID—ISOSTASY AND MOUNTAIN RANGES. 449 or decrease. If by erosional transportation a large quantity of material is removed from a high land and deposited in the oceans, then the increase of weight under the ocean and the decrease under the mountains will, as Major Dutton explained, set up a subter- Tfanean counter flow, which will restore the equality of material in the segments. If by the exercise of tangential forces a portion of the earth’s crust is compressed and folded and the quantity of material in the segment thus increased, the added weight will cause a slow sinking of the region and material will flow out from below and reduce the mass of the segment to its proper value. Indeed, the folding up of the rock by tangential pressure would not elevate a mountain range, but would cause the folded region to sink; not, however, necessarily below its former level. When we consider the origin of the mountain ranges ‘te theory of isostasy requires that all hypotheses, which call for more than ~ the normal amount of material in any segment, be excluded. The folding of rock under tangential forces, and the increase of material by subterranean flow are necessarily debarred. Dana noticed that e great mountain ranges of the world were opposite the great oceans and, in some cases, were opposite the great depths of the eans. The inference was natural that material was taken from the ocean bed, increasing its depth and added to the land increasing its height ; but the theory of isostasy forbids this inference. He also suggested that the segments of the earth forming the oceans were sinking more rapidly, as the earth cooled, than the segments forming the continents and also that they were stronger; so that they com- pressed the continents, folding the rock and making mountain ranges around their borders. Besides other objections to this idea, the theory of isostasy excludes it on account of the increased material required in the land segment. Professor Charles Davison* has sug- gested that the oceans owe their existence to the stretching and ‘consequent thinning of the strata below them, but the theory of isostasy does not permit the withdrawal of material from the ocean *“On the Distribution of Strain in the Earth’s Crust resulting from Es Secular Cooling, with special reference to the Growth of Continents and E the Formation of Mountain Chains,” Phil. Trans. R. S., 1887, Vol. 178(A), pp. 231-242. aaa 450 REID—ISOSTASY AND MOUNTAIN RANGES. [Aprila, | bottoms. Sir George Darwin® has suggested that the continental areas of the earth may be due to elevations caused by the differen- tial retarding effect of lunar tidal action. But the theory of isostasy a tells us that they could not have maintained themselves unless they were especially light; and in this case they would have existed inde- pendently of the tidal forces. Although these elevations, or “wrinkles,” as Sir George Darwin calls them, might have been dis- torted by the different tidal effects in different latitudes, theit sac} inal meridional direction still requires explanation. The foldings and contortions of the rock have been so insicadbaty associated, in the minds of geologists, with mountain ranges, that a low-lying region of folded rock has been looked upon as the remains of a mountain range removed by erosion; but as mountains are not due to rock folding, this inference may be entirely wrong. Only a few of the consequences of the theory of isostasy have been mentioned ; but the principle is of such fundamental importance that it will surely exercise a strong influence over our future theories, and will be applicable in directions not now suspected. Unfortu- nately, it does not tell us definitely what is the cause of the elevation _ of mountains and plateaux; but it limits our inquiries by excluding all theories which assume the addition of matter to a segment. It tells us, quite definitely, that the elevation of mountains, or the depression of the oceans, must be due to vertical forces brought about by a decrease, or increase, in the density of the material under 3 these regions. According to it, the mountains are high because their — material is light ; and as geological history tells us that the mountains _ have not always existed, we must conclude that they were elevated — by an expansion of the material in and under them. And the great deeps of the oceans are deep because the material under them is. dense and they have become deep by an increase in the density of this material. Since all mountain areas are being lowered by active erosion and many of the great ocean deeps are being filled by depo- sitions, the great heights of the former must be due to the fact that they are still in the process of elevation or that they have onl *“ Problems connected with the Tides of a Viscous Spheroid,” P, Trans. R. S., 1879, Vol. 170, p. 589. REID-ISOSTASY AND MOUNTAIN RANGES. 451 ently been raised; and the great depths of the latter to the fact they are in the act of sinking, or have only recently sunk. As centres of the great majority of strong earthquakes are along boundaries of high mountain ranges, or of great ocean deeps, seems most probable that the forces which have produced these ery interesting features of the earth’s surface are still in active A FOSSIL SPECIMEN OF THE ALLIGATOR SNAPPER | (MACROCHELYS TEMMINCKII) FROM TEXAS. (Piates XVIII ann XIX.) By OLIVER P. HAY. (Received May 23, 191i.) The writer has received for examination from Professor Mark Francis, of the Texas Agricultural Experiment Station, at College — : Station, Tex., a nearly complete skull of the great fresh-water tor- toise known as the alligator snapper. This fine specimen was dis- : covered last summer or autumn during some dredging operations in the Brazos River, between College Station and Navasota. After passing through various hands it came into the possession of Pro- fessor Francis, who, on application, kindly transmitted it to me. — With the skull came also a part of a carapace, which doubtless be- longed to the same animal. The skull was found in a mass of — gravel, and had undoubtedly been washed out of the river bank not far away. This proximity of the place of burial is evident from the little damage done to the skull, and is made more probable from the presence of a part of the shell. ei The cavities of the skull, when it came into Professor Francis’ hands, were full of gravel, wedged in very tightly. Some of this | _ gravel was sent with the skull. It was strongly colored with iron oxide; and this oxide served to cement the bits of gravel togeth and to the bone. The bone is also colored with the oxide, and is so thoroughly mineralized that, on being struck, it rings like piece of porcelain. It would be interesting to know exactly the geological age this specimen; but it appears now impossible to determine Professor Alexander Deussen, of the University of Texas, has b engaged in studying the Quaternary and Recent deposits of so of the rivers of Texas; and a part of his results is soon to be pub- lished by the United States Geological Survey. He has kindly i - 452 agit.) THE ALLIGATOR SNAPPER FROM TEXAS. 453 formed the present writer that there occur along the Brazos River three principal terraces. The oldest and highest of these, the Hidalgo Falls terrace, lies at a height of 100 or more feet above the “present water line of the river. In the materials of this terrace have been found remains of Mammut, Elephas, Megalonyx, Equus, etc. About 75 feet below this terrace is found another, the Port Hudson, whose thickness is from 20 to 30 feet. The upper terrace is regarded as older Pleistocene; the Port Hudson, as newer Pleistocene. At a level some 15 to 20 feet below that of the Port Hudson, is a terrace which Professor Deussen considers as of early Recent time. It constitutes the real “bottom lands” of the Brazos and is subject to overflow. It is very probable that the remains described here were derived from the lowest and youngest terrace and that the individual lived ° at some time about the beginning of the Recent epoch. The species probably lives today in the Brazos River. | The skull (plates XVIII and XIX) lacks the lower jaw, a part of the temporal roof of the left side, most of the occipital condyle, and the hinder part of the supraoccipital process. A close exami- nation reveals no characters by which it can be distinguished spe- _ cifically from the alligator snapper. The profile (Pl. XIX, Fig. 2) a is much less concave than in most specimens of the species collected in the rivers of the Southern States; but there is in the United States National Museum a skull of considerable size, no. 3769, from ‘Mississippi, which presents no greater concavity than does the Brazos River specimen. There are two other skulls, the one con- siderably larger than the skull from Mississippi, the other con- siderably smaller, both of which are much more concave than the _ specimen from Mississippi. Hence, the amount of concavity seems _ not to depend on youth or old age. The skull of the fossil is, relative to the length, slightly both _ broader and higher than are two skulls with which it is compared, as is here shown: Snout to Specimen. condyle. Width Height, Brazos River skull.......: SD NER aso eager a I 1.19 2 No. 3769 U. S. N. M., from Mississippi.......... I 1.14 78 No. 3444 U.S. N. M., from Red R,, Ark........ I 1.08 74 454 HAY—A FOSSIL SPECIMEN OF _ Dilay 2 It will be observed that the last two skulls differ from each other about as much as the second differs from the fossil. Liat The same three skulls furnish the following measurements. fae . Brazos R, oie one eee Me*surements, skull, No, 3769. No.3444 Snout to occipital condyle.............eseeeeees 183+ 190) ae te Snout to hinder end of supraoccipital process.. 262+ 296: Least width pterygoid region.............0:0e08s 33 io ae Outside to outside of quadrates................ 187 166 == hee Distance between hinder ends of cutting edges ieee ae ee Of Upper JAWS: ss sia 0 ck dan Owe bd Sebe cerns 142 1D a Width in front of ear cavity, 05.60.60 6. race 218 106 2) Width of temporal arch where narrowest....... ayy) 98 > Orbit to excavation of postorbital arch.......... 87 Se ee Horizontal diameter of orbit.............+..+55+ 32 Regia. - pee Distance between fronts of orbits.............. 55 bie or a a ; cmeaicl the skull. This is reduced to five twelfths the gop Ri size. It consists of a part each of the third and fourth costal plates, | and of a part each of the sixth and seventh peripherals. On these parts are present areas representing the outer and hinder angle of the third costal scute, a little of the third and the whole of the fourth supramarginal scutes, the whole of the eighth marginal scute ay Fic. 1. Section of rim of carapace between sixth and seventh peripherals. _ and a part each of the seventh and the ninth. These structures a almost identical with the corresponding ones of a mounted specimen of the species in the United States National Museum. Fig. 1 represents a transverse section of the rim of the cara taken between the sixth and the seventh peripherals. PROCEEDINGS Am. PHILOS. Soc. VoL. L. No. 200 PLATE XVIII MACROCHELYS TEMMINCKII PROCEEDINGS Am. PHiLos. Soc. VoL. L. No. 200 PLATE XIX 1 j 4 MACROCHELYS TEMMINCKil. XX & | HAY—A FOSSIL ALLIGATOR SNAPPER. 455 EXPLANATION OF THE PLATES. MACROCHELYS TEMMINCKII, fossil. In the figures of these plates the sutures between the bones are repre- ited by narrow white lines; the seams between the horny scutes by wider lines. All the figures are two-fifths of the natural size. Pirate XVIII. Fic. 1. Skull-seen from above. Fic. 2. Skull seen from below. Pirate XIX. Fic. 1. Fragment of right side of carapace. c. p. 4, part of fourth costal , or bone; behind it is a part of the fifth. mar. 8, mar. 9, the eighth mar- horny scute and a part of the ninth. per. 6, per. 7, the sixth and seventh ripheral bones, but only a part of each. s. mar. 4, the fourth supramarginal ny scute; in front of it is a part of the third. The third costal horny > area occupies the portions of the costal plates present, except the hinder of the fifth. 2. Skull seen from the right side. PROCEEDINGS AMERICAN PHILOSOPHICAL SOCIETY HELD AT PHILADELPHIA FOR PROMOTING USEFUL KNOWLEDGE VoL. L SEPTEMBER—I)ECEMBER, 1911 No. 201 AN HYDROMETRIC INVESTIGATION OF THE INFLU- ENCE OF SEA WATER ON THE DISTRIBUTION OF SALT MARSH AND ESTUARINE PLANTS. (Pirates XX anp XXII.) By JOHN W. HARSHBERGER, Pu.D. . (Read April 22, 19I0.) Elsewhere’ I have discussed the general character of the vegeta- tion of the salt marshes of the northern New Jersey coast and the - factors controlling the distribution of marsh plants in that area. . This earlier study was based largely on physiographic and floristic __ considerations, although reference is made on page 379 of that paper to the use of the hydrometer in the investigation of the actual influ- ence of sea water, or salty soil, on the distribution of a limited num- ber of plants. The investigation begun in 1909 has been continued until sufficient facts have accumulated to warrant their publication. The use of a special kind of hydrometer was suggested as a simple but efficient method of investigating the salt content of salt marsh soils and of the estuarine water which, at first strongly saline, becomes largely diluted, as it mingles with that of streams flowing in a seaward direction. This is the first actual use of the hydrometer *Harshberger, John W., “The Vegetation of the Salt Marshes and of the Salt and Fresh Water Ponds of Northern Coastal New Jersey,” Pro- ceedings Academy of Natural Sciences of Philadelphia, 1909, 373-400, with 6 figures. PROC, AMER, PHIL, SOC., L, 201 EE, PRINTED AUG. 25, IQII. 457 458 HARSHBERGER—INFLUENCE OF SEA WATER (April a3, in phytogeographic and phytoecologic investigation. The méthed: is applicable not only to a study of salt marsh soils, but also to an investigation of salt lakes and alkaline soils, which are found in~ many parts of our western arid districts and in other parts of the : world (Fig. 1). The use of the hydrometer supplements, if it does >. duncua Gerardi _. ~~"* Distichiis spicata Temperature oH: Salt MarghMan; -|PElE |... Salt Marsh, Belmar Fic. 1. Special hydrometer and thermometer used in the investigatio of the salt marsh vegetation of the New Jersey coast. The names of p! are arranged along the scale to graphically represent the maximum density o salt water to which these plants are subjected in their marsh environment. Other data are given for comparison. not replace, the employment of the more expensive and cumbersome apparatus which determines by electric means the salt content of © soils. Although this investigation was made in the salt marshes 0 1910) == ON THE DISTRIBUTION OF PLANTS. 459 northern coastal New Jersey, the results obtained are equally appli- cable to the same salt marsh species as they are found distributed along our eastern Atlantic coast. For stretching from the Bay of _ Fundy along the New England, middle and southern Atlantic coasts, as far south as Jupiter Inlet in Florida, are extensive salt marshes covered with a vegetation which consists with minor differences of almost the same and characteristic species. PossIBLE METHODS OF INVESTIGATIONS. 4 There are four possible methods open to the investigator of the 4s salt content of salt marsh soils and estuarine waters. These methods have been used by chemists, soil analysts and plant physiologists. 4 Method of Titration The determination of salt content by volu- 3 metric analysis has been ‘the favorite one of chemists. For this purpose, a tenth-normal silver nitrate volumetric solution is used, prepared as follows: Dissolve 16.869 grams of silver nitrate, which _ previous to weighing has been pulverized and dried in a covered a porcelain crucible in an air-bath at 130° C. (260° F.) for one hour in sufficient water to measure at 25° C. (77° F.), exactly 1,000 c.c. This solution is kept in dark amber-colored, glass-stoppered bottles, carefully protected from dust and sunlight. A tenth-normal potas- sium bichromate test solution is prepared by dissolving 4.8713 grams _ of pure potassium bichromate, which has been pulverized and dried at 120° C. (248° F.) in sufficient water to measure at 25° C. (77° _F.) exactly 1,000 c.c. To a definite volume of salt water, sufficient potassium bichromate test solution is added to impart a yellow tint, then the tenth-normal silver nitrate solution is slowly added from a burette, stirring or agitating until the mixture acquires a permanent tint, due to the formation of red silver chromate. The fluid to be tested must be neutral, as free acids dissolve the silver chromate. The cubic centimeters of silver nitrate solution used must now be multiplied by .oo5850 Fresenius (.005837 Sutton, .005806 National Dispensatory) to give the weight of the sodium chloride, because .005850 Fresenius (.005837 Sutton, .005806 N. D.) grams of sodium chloride is the equivalent of one cubic centimeter of tenth-normal 460 HARSHBERGER—INFLUENCE OF SEA WATER [April 22, silver nitrate volumetric solution. This is the method adopted I am told by the botanists at Johns Hopkins University in studying the salt marsh vegetation of Maryland, the results of which investi- gation have not yet been published. Method of the Electric Bridge—The Bureau of Soils, United _ States Department of Agriculture, adopted some years ago the prin- ciple of the slide wire bridge to the measurement of the salt content of soils. The earlier instruments have been described in various bulletins and the results obtained with them are scattered through various publications of the bureau. Since 1899, when the electric bridge was put first into practical use, various improvements have _ been made, so that the improved instrument is the result of the experience gained by its use in the actual field study of soils. The use of the electric methods for determining the soluble content of - a soil depends on the fact that the electric current is conducted by — the salt in solution and that the conduction of the solution, or con- versely, its resistance to the passage of the current, is largely deter- mined by its concentration. The magnitude of current that will pass e is increased by an increase of salt in solution; or the resistance to the passage of the current decreases with the increase of salt. The experience gained by the use of the modified instrument is embodied _ in the recent bulletin of the Bureau of Soils noted above and its general utility in the study of alkali soils, the salt content of — tion and seepage waters is given. Method of Plasmolysis.—It is a well-known physiologic fact that dilute solutions of potassium nitrate, sodium chloride and cane sugar cause a removal of water from living plant cells, so that the proto- plasm contracts away from the inside of the cell wall. The per- *Consult Hare, Hobart A., Caspari, Charles, Rusby, H. H., “The National Standard Dispensatory,” 1905, 1684; Fresenius, C. Remigius, “ System of Instruction in Quantitative Chemical Analysis,” 1894, 430; Sutton, Francis, “A Systemic Handbook of Volumetric Analysis,’ 1890, 124; Fraps, G. S., see bibliography. *Davis, R. O. E., and Bryan, H., “The Electrical Bridge for the De- termination of Soluble Salts in Soils,” Bull. 61, Bureau of Soils, 1910, where reference is made to previous bulletins; Cannon, W. A., “On the Electrical Resistance of Salt Plants and Solutions of Alkali Soils,” The Plant Worle : II, 10-14. 1910.] ON THE DISTRIBUTION OF PLANTS. 461 centage of substance in solution necessary to cause plasmolysis varies not only with the plasmolyzing substance, but also with the plant used in the experiments. The protoplasm in some plants plasmolyzes quickly ; in other cases with difficulty, so that stronger solutions are necessary to produce a change in the more refractory plants. If we know, therefore, that a certain percentage strength of sodium chloride in solution will produce plasmolysis in say the cells of the staminal hairs of Tradescantia, then if raw or diluted sea water be used and a similar plasmolysis occurs, the percentage of sodium chloride in the sea water must be equivalent to that of the salt solu- tion known to produce similar plasmolytic effects.. A comparative table can be constructed by which the varying percentages of sea water can be ascertained. An extensive literature, part of it noticed in the bibliography, is concerned with such plasmolytic studies.* Hydrometric Method.—The use of the hydrometer in determin- ing the salt content of salt marsh soils suggested itself to me, as a simple but efficient method of making a phytogeographic survey of the vegetation of salt marshes upon purely ecologic lines. The advantage of the hydrometer is that it is light, can be carried easily from place to place and lends itself to immediate use, the record depending upon two simple readings, one of specific gravity and one of temperature. The hydrometer is plunged into a vessel contain- ing the water to be tested. Styles of Hydrometers——Aiter a simple hydrometer had been used ina number of preliminary tests, search was made for a hydrom- eter which would record accurately the density of sea water. It was found that there are many kinds of hydrometers in use to test acids, alcohol, alkali, ammonia, bark liquor (tannometer), beer, benzine, chlorine, cider, coal oil, ether, gasoline, glycerine, milk (lactometer), naptha oil, salt solutions (salimeter), silver solution, sugar, sugar and syrup. Some are constructed with Baume’s scales, others with Richter and Trolle’s scales and those used to test sugar with Balling’s and Brix’s scales. Finally, after testing several different kinds of *Drabble, E., and Lake, H., “The Osmotic Strength of Cell-sap in Plants growing under Different Conditions,” The New Phytologist, October, 1905, 189; Duggar, B. M., “The Relation of Certain Marine Algae to Various Salt Solutions,” Trans. Acad. Sci. St. Louis, XV1., 473-479. 462 HARSHBERGER—INFLUENCE OF SEA WATER _ [April 22, hydrometers, one was obtained which fulfilled all the conditions of — - experimentation most perfectly. The hydrometer purchased of - Arthur H. Thomas Company is one designed to test the specific gravity of liquids heavier than water. The scale reads from 0.995 to 1.065 and is divided into single units and half units (Fig. 1). — For example, beginning at 1.0000, the divisions of the scale read as follows: 1.0005, 1.0010, 1.0015, 1.0020, 1.0025, 1.0030, 1.0035, — 1.0040, 1.0045, 1.0050. The last figure is the next prominent figure ; a on the scale printed in black letters. Altogether 140 separate read- ings can be made from this scale,.and if the observer wished to test the salinity of the water of Salt Lake, Utah, the length of the scale would have to be increased to the point indicated in Fig. 1, and the size of the bulbs would have to be increased correspondingly. There — are three bulbs blown in the hydrometer tube. The lower one is the sinker with metallic mercury. The middle one carries the mer-_ cury of the thermometer, which is inclosed in the third and upper bulb. The thermometer scales reads from —5° C. to + 45° C., and — is divided into degrees with the fifteenth marked in red. With this — instrument temperature and specific gravity can be determined simultaneously. oe Corrections to Readings.—In actual use, the experimenter finds — that the hydrometric readings vary with the temperature of the water, and that to make the results harmonious all of the readings for specific gravity must be reduced to the uniform temperature o 15° C. No table exists by which the reduction can be made directly without calculation. Such a table for all temperature degrees and degrees of specific gravity is a desideratum. A mechanic rule, or sliding scale, might be constructed from which corrected readir might be taken directly by adjusting the movable parts of the s to corresponding degrees of specific gravity and of temperature. — the absence of such a table and mechanic scale after a prolon search through theoretic text-books of chemistry, the following one was discovered, which enabled me to standardize all of the readi made by the hydrometer by reference to the specific gravity 5° C. : 1910.] ON THE DISTRIBUTION OF PLANTS. 463 TABLE FOR THE REDUCTION OF SpeEciIFIC GRAVITY AT ANY TEMPERATURE TO THE Speciric Gravity AT 15° C. (Table B, Landolt-Bérnstein Physikalisch-Chemische Tabellen, Berlin, 1905, page 323.) “Reducktion der Dichte d- auf die Dichte bei 15° nach den Beob. von Dittmar” (Challenger Exped.), Ekman (Vetensk. Handl., 1870), Lenz u. Reszof (Mém. Petersb., 1881), Thorpe u. Riicker (Phil. Trans., 166, II; 1876), Tornoé (Norw. Atlantic Exped., 1880) berechnet von Makarof (J. Russ, Phys. Chem. Ges., 23, 30; 1891). Auszug. 0° 5° 10° | 15° 20° 5° 30° Dest Wasser | 0.99988 | 99979 | 99974 | 99915 | 99828 | 99714 | 99577 Seewasser 1.00077 | 00087 | 00060 | c0000 | gggII | 99796 | 99659 os 1.01130 | OT120 | o1075 | O1000 | 00898 | 00774 | 00630 ee 1.02182 | 02152 | 02090 | 02000 | 01886 | O1751 01600 ae 1.03228 | 03179 | 03102 | 03000 | 02876 | 02732 | 02572 This fact should be noted in connection with the use of the above table, viz., the specific gravity of a sample of sea water is the num- ber representing its weight as compared with an equal volume of pure water at the same temperature. The latter is usually called 1.000 so that the specific gravity of a sample of sea water may be some such number as 1.025. The density is the weight in grams of one cubic centimeter of water at the temperature in situ (#°) com- pared with that of I c.c. of pure water at 40° C. It is usually expressed as D. The salinity is the total weight in grams of the matter dissolved in 1,000 grams of water. Mathematic Calculation—Through the kindness of a graduate student, Mr. John C. Bechtel, to whom my thanks are due, I was relieved of the labor of making the mathematic calculations neces- sary to reduce the hydrometric readings to 15° C. His method of procedure is herewith given in a sample case. To determine density at 15° C. of salt water whose density at 23° is 1.0155. From the table we see that this corresponds to a solution whose _ density at 15° lies between 1.01 and 1.02. We therefore find figures for density of salt water at 23°, if density at 15° is 1.01, and also if density is 1.02. Density at 20° is 1.00808 for first and 1.01886 for second. 464 HARSHBERGER—INFLUENCE OF SEA WATER [Aprila, Density at 25° is 1.00774 for first and 1.01751 for second. For first, the change in density for 5° i 1.00898 — 1.00774 = .00124. For 3° it is # of .00124 = .00740. Hence density is 1.00898 — .00740 = 1.00824 at 23°. For second change in density for 5° 1.01886 — 1.01751 == .00135. For 3° it is 2 of .00135 = .00081. a Hence density at 23° is equal to density at 20° —loss for 3° or E 1.01805. . a We now have densities at 15° as 1.01 and 1.02 as limits and from observations we see that at 23° the density of our solution is 1.0155. We also have this proportion which will give a saa approximate result: If y is the density of this solution at 15° y—I1.0I1 __ 1.0155— 1.00824 1.02—1I.01. 1.01805 — 1.00824” y—I.0I_ .00726 Ol. 30008 y = 1.01 + .0074, y= 1.0174, therefore density at 15° = 1.0174. As the above computation is a rather long one and must be sted for each of the actual readings obtained by the hydrometer, it has been thought advisable to give the entire set of original readings at various temperature and the corrected specific gravity at 15° C. Such a table may enable future workers in the same field to make their corrections at once by omitting the long computation otherwise necessary. The numbers in the first column of the table refer 1 the observations as recorded in the field note book and which have been added as subnumbers to the specific gravities placed on the map of Shark River and Bay which comprises Fig. 4 of the text. 3 1910.) ON THE DISTRIBUTION OF PLANTS. 465 Tame Grvinc HyproMeEtric OBSERVATIONS ON SALT MarsH PLANTS OF NEW ~ JERSEY WITH CorRECTIONS AT 15° BY Mr. Joun C. BEcHTEL. | s ° ’ ~ . . No, | Gras’ | Temp.cc.| SPs | No. | SeeGrt | temp.cc.| SPS 71 1.0155 | 23 1.0174 108 1.0005 14 1.0004 72 I.0160 | 23 1.0179 109 1.0025 24 1.0044 73 1.0180 ae 1.0194 be fe) 1.0025 20 1.0034 75 1.0170 | 20 1.0182 112 1.0120 28 1.0153 76 1.0175 19 1.0184 113 1.0035 25 1.0057 77 1.0260 29 1.02996 114 1.0015 25 1.0036 78 I o150 22 1.0166 115 1.0000 23 1.0016 1.0165 19 1.0174 116 0.9990 25 1.0011 81 1.0160 26 1.0188 117 1.0005 21 1.0016 82 I O105 21 | 4.0117 118 1.0160 27 1.0181 83 1.0090 20 | 1.0102 119 1.0165 26 1.0193 84 1.0020 20 | 1.0029 120 1.0160 27 I.OI1QI 85 I.0010 20 ~—sT.0019 121 1.0415 24 —_ 86 1.0000 18 1.0005 — 1.0150 25 1.0175 87 1.0140 25 | 41.0164 124 I.0155 29 1.0192 1.0030 23 | 31,0046 125 1.0010 22 1.0024 89 1.0000 22 1.0014 126 1.0190 23 1.0210 go 1.0000 20 1.0009 127 1.0110 21 1.0022 91 1.0005 20 1.0014 128 1.0035 22 1.0049 92 1.0140 22 1.0156 129 1.0205 23 1.0225 93 1.0205 20 1.0217 130 1.0255 20 1.0267 94 1.0200 20 1.0212 131 1.0065 20 1.0075 95 1.0215 | 19 1.0224 132 1.0110 19 1.0180 1.0050 19 1.0058 133 1.0250 21 1.0265 97 1.0110 | 19 1.0118 134 1.0205 22 1.0222 1.0165 18 1.0172 135 1.0250 21 1.0265 1.0215 | 18 1.0223 130 1.0240 20 1.0252 100 1.0180 | 18 10187 — 1.0210 22 1.0224 Ior 1.0210 | 21 1.0224 137 1.0245 27 1.0278 102 1.0245 | 20 1.0257 138 1.0215 24 1.0238 103 1.0185 = 20 1.0196 139 1.0180 25 1.0205 104 1.0195 | 21 1.0209 140 1.0055 20 1.0065 107 I.0150 | = 25 1.0175 After having discussed the theoretic methods, we must next con- _ sider the actual study of the vegetation in the field by the use of the hydrometer. _ Aids to Field S tudy.—The equipment which was carried into the | field for the study of the edaphic conditions under which salt marsh _ vegetation grows was accommodated in a light basket and consisted a, © of a meter measure, reading to decimeters, centimeters and milli- meters; a narrow, but deep, glass cylinder to hold the water upon which the specific gravity determinations were made; a tin dipper to collect the water and a field note book. A narrow trenching “Spade was carried in the hand and by this spade it was possible to 466 HARSHBERGER—INFLUENCE OF SEA WATER [April dig deep holes in the tough resisting marsh sod. The water f study was dipped either directly from holes in the marsh or taken from the ocean and open bays along the New Jersey coast. Theh was dug in all cases deep enough to allow the soil water to perco- late into it, and upon this water the specific gravity readings were made. The region especially traversed in this way extended from Manasquan Inlet on the south to Sandy Hook Bay on the north, and thus an insight was obtained of the problems concerned in the distri- bution of the various species of salt marsh plants. FIELD OBSERVATIONS AND DATA. Altogether sixty readings were made with the first style of sali nometer used. This type had such a small range of scale divisions that it was discarded as being too inaccurate for the purposes of the salt marsh investigation where the total salt content of the wa’ increased, or decreased, by almost inappreciable amounts. Although many of these observations are of interest, they are not incorporated here. The second style of hydrometer was like the final one adopted, as to the divisions of the scale, but it lacked a thermometer. — data obtained by this hydrometer are considered here, but they are only of comparative value, because théy lack the accuracy of the later readings which were made for both specific gravity and perature. They are of value because they give habitat relations not included in the more accurate data obtained later. For the above reasons the field observations will be consi under two heads: (1) the readings made by the hydrometer wi the thermometer, and (2) the readings which include both es metric and thermometric measurements. Hydrometric Readings without Thermometer-—The ‘readin which are numbered consecutively from 1-70 inclusive are a geographically as affording more interesting comparative data. stand as follows: Beginning in the north readings were obta along the Shrewsbury River, starting at the railroad bridge cot ing Highland Beach with the Navesink Highlands proper. P Island, where the first measurements were made, is a small is back of the Sandy Hook peninsuula in Sandy Hook Bay. Unde agro.) ON THE DISTRIBUTION OF PLANTS. 467 Be edly, the water of this bay is less strongly saline because influenced _ by large fresh water rivers, such as the Hudson River. a 55. Spartina stricta maritima, association on Plum Island. Sp. gr. 1.016. f° 56. Baccharis halimifolia, association with Salicornia herbacea, Suaeda maritima, water covering plants at high tide two inches deep. Sp. gr. 1.0155. 57. Salt Pond on Plum Island, fringed by Spartina stricta mari- - tima. Sp. gr. 1.016. 2 59. Water from a hole two feet deep in tension strip between _ Spartina stricta maritima and Baccharis halimifolia. Sp. gr. 1.018. 60. Water from hole eighteen inches deep in middle of a Spar- _ tina patens association. Sp. gr. 1.020. 61. Water from a hole eighteen inches deep on the tension line between Spartina patens and Spartina stricta maritima. Sp. gr. 1.019. 62. Water from a hole on the tension line between Spartina patens ~__and Baccharis halimifolia. Sp. gr. 1.0185. 64. Water from a hole eighteen inches deep in the middle of an __ association of Salicornia mucronata, Limonium carolinianum, Spar- ___ tina patens and near by on the same level Atriplex hastata, Suaeda maritima and Baccharis halimifolia. Sp. gr. 1.003. The following observations were made in ascending the Shrews- _ bury River toward Pleasure Bay: 53. Salt water at Highlands Pier. Sp. gr. 1.019.° +66. Water surrounding Spartina stricta maritima fringing beach in front of the Navesink Highlands. Sp. gr. 1.0185. 68. At the confluence of the Navesink and Shrewsbury. rivers _ with a lot of Fucus vesiculosus attached to pilings and also Spartina _ stricta maritima. Sp. gr. 1.0185. 69 and 70. At this point water submerged an association of Limonium carolinianum, Suaeda maritima, Spartina patens, Sali- ‘cornia herbacea, Plantago maritima and Atriplex hastata. Sp. gr. ‘1.018. _ Ascending the Shrewsbury River, the head of navigation is Teached at Pleasure Bay. From here to the head of the bay the : _*¥For comparison, the sea water from the ocean at Belmar read sp. ‘gr. 1.0215 at Temp. 20.6° C. corrected to 15° the sp. gr. = 1.0224. 468 HARSHBERGER—INFLUENCE OF SEA WATER _ [April 22, water becomes gradually fresher and the salt marsh vegetation is _ replaced gradually by fresh water marsh plants. a 32. Water from Pleasure Bay at the head of navigation. — sg gr. 1.010. a 33. Water from ditch two feet deep in middle of Spartina patens a association. Sp. gr. 1.010. x” 35. Water at head of small ditch with Scirpus — Cicwta 4 maculata, Scirpus robustus. Sp. gr. 1.005. a 37. Slue with Baccharis halimifolia and Spartina stricta. mare 4 tima. Sp. gr. 1.010. a 42. Hole in salt meadow on tension line between Juncus Gerara (cut for hay) and an association of Scirpus pungens, Pluchea cam phorata, Atriplex hastata and Spartina patens on the other side. Sp gr. 1.005. | 44. Water from bases of plants of Typha angustifolia and Seir pus pungens. Sp. gr. 1.015.° p 45. Water at third bridge above Pleasure Bay in the middle of — an association of Spartina stricta maritima, Scirpus ites: S. ro- bustus. Sp. gr. 1.005. 46. Above the fourth bridge in middle of a S partial peices: mari- tima association. Sp. gr. 1.0005. 47. Here a pure association of Scirpus robustus. Sp. on eho 48. Association of Zizania aquatica and Scirpus robustus. Sp. gr. 1.0005. 50. Water from inner edge of an association of Typha ang folia (tall), Peltandra virginica and Cicuta maculata. Sp. gr. 1.0¢ 51. Muddy cold water from a hole in an association of Sagittas latifolia (=S. variabilis), Cicuta maculata, Typha angustifolia Polygonum sagittatum. Sp. gr. 1.0015. = 52. Water from channel under last bridge. Sp. gr. 1.0015. — The fact that such salt marsh species as Spartina stricta mari mingles with fresh-water marsh species under almost fresh- conditions is to be explained by the occasional inundation of st plants by more strongly saline water at exceptionally high tides, that the exceptionally high tides enable the salt grass to persist rounded by fresh-water marsh species. The salt marsh species ° Probably due to evaporation. ci arG ee 1910.] ON THE DISTRIBUTION OF PLANTS. 469 withstand fresh water better than the fresh-water species can salt water. These latter plants are able also to withstand an occasional flooding, although normally they are controlled by fresh water. This is probably to be accounted for by the resistance of the leaves that surround the stem, while the roots are in practically fresh water, which saturates the ground and prevents the entrance of salt water into it for some time. The occasional flooding of salt water is not for a sufficiently long time to effect the character of the ground water in which the roots of such plants as Sagittaria latifolia, Cicuta maculata grow. The observations at Belmar began with an estimation of the salinity of the ocean water. The readings from 4-19 are interesting Fic. 2. Basin-like slue along Fifth Avenue, Belmar, N. J., fringed by salt marsh vegetation and backed by forest trees. Several of the stations for hydrometric determinations were chosen along this shore. because they were made while Shark River Inlet was closed to the sea by a sand bar. 2. Sea water from surf at Belmar. Sp. gr. 1.0215 at 20.6° C. (69° F.) ; corrected to 15° C. Sp. gr.==1.0224. 4. Water in Shark River Inlet flooding Spartina stricta maritima association. Sp. gr. 1.015. 5- Water from channel opposite B Street, Belmar. Sp. gr. 1.0185. 470 HARSHBERGER—INFLUENCE OF SEA WATER _ [April 22, 7. Water from seaward end of marsh island in Shark River. Sp. gr. 1.017. With Spartina stricta maritima. * 8. Water submerging Juncus Gerardi. Sp. gr. 1.016. g. Water covering Spartina patens. Sp. gr. 1.017. 12. Water in large slue along Fifth Avenue, Belmar. Sp. gr. 1.016, : 13. Water from bay at Casino Landing, where a rise of eighteen inches was noted after the inlet closed. Sp. gr. 1.0175. These several readings show the condition of salinity when the inlet through which the tidal salt water enters Shark River is closed and the salt water thus inclosed is diluted by rain and river water until the river shows a perceptible rise of eighteen inches above the level of normal high tide. In such rivers the salt marsh vegetation for considerable periods of time is exposed to fresh water, which would ultimately control, if the inlet would remain permanently closed. But when the inlet is reopened the original conditions of salinity are restored by the tidal flow of sea water in and out of the landlocked bay. This is an interesting corroboration of the recent. work of D. W. Johnson,’ who believes that the indications of appar- ent subsidence are due to fluctuations in the tidal level due to change in the configuration of the coast. During the closure o Shark River there was a rise of water level in the river which mi account for the rise in the height of the salt marsh layers. Af the causal influences had been obliterated, an examination of layers of salt marsh soil would indicate, according to the older v a total submergence of the coast line equal to the depth of formed marsh peat. The observations on the salinity of the water at the wane T of Newberry (Stockton) Lake, an arm of Manasquan R of interest as displaying the edaphic conditions which cot distribution of Typha angustifolia. The size of this plant is directly conditioned by the amount of salinity as measurements to be presented will show. However, if we begin near the "Johnson, D. W., “The Supposed Recent Subsidence of the chusetts and New Jersey Coasts,” Science, N. S., XXXIL., 721-723; Ba: H. H., “Botanical Evidence of Coastal Subsidence,” Science, XXXIII., 29-31; Johnson, D. W., Science, XXXIII., 300-302. 1910.] ON THE DISTRIBUTION OF PLANTS. 471 of the lake where the cat-tails occur, the following series of readings are suggestive. 28. (Position I.) Association of Typha angustifolia—base of plant covered by water at high tide. Sp. gr. 1.0145. 27. (Position III.) Typha angustifolia. Sp. gr. 1.014. 26. (Position V.) Typha angustifolia. Sp. gr. 1.014. 25. Position VI.) Typha angustifolia with Atriplex hastata. Submerging water with sp. gr. 1.012. 24. (Position VIIa.) Association of Typha angustifolia, Atri- plex hastata, Salicornia herbacea, Spartina stricta maritima. Sp. gr. 1.0135. 23. (Position VIIb.) Association of Typha angustifolia, Scirpus lacustris, S. pungens. Sp. gr. 1.0125. 22. Outer edge of Typha angustifolia association at the head of the lake. Sp. gr. 1.0115. 21. Head of Newberry Lake at inner edge of dense masses of Typha angustifolia with Hibiscus moscheutos (third lot). Sp. gr. 1.0050. Influence of Saline Water on Typha angustifolia.—Before begin- ning a consideration of the data obtained by using the hydrometer and thermometer combined, it is important to consider the influence Fic. 3. Cat-tail, Typha angustifolia, at the head of Stockton Lake near Sea Girt, N. J. The tall plants are growing at Position III. | ‘ b a: 9 472 HARSHBERGER—INFLUENCE OF SEA WATER (soa of the varying salinity of the water upon the plants which scigal - jected to the different densities of salt water. For this purpose, i have chosen Typha angustifolia because it seems to show ina marked degree the influence of the variation in the saline environment. Six series of measurements were taken at this plant. Three sien: a based on plants from Stockton Lake and three bese: plants from: Pleasure Bay. . te . In all the measurements the height of the plant is meamrell the top of the fertile part of the terminal spike. The upper s : ile and staminate portion is included, but it is only — t. Measurements are metric. First Serres. Typha angustifolia rrom STocKTON LAKE Som. (Pe TI I.) Sp. cr. 1.0145. a No.| Height of Length of Breadth of | Width of Third Plant. Spike, 9. Spike, Q. Leaf from Top. Dry I 929 .087 016 004 6 2 1,030 075 .020 .006 6 3 Broken Broken Broken .004 7 4 1.353 124 .020 .006 6 5 1,015 080 O19 006 8 6 1,119 .090 018 005 7 8 1,100 .098 ,022 .006 | 7 1.124 082 .020 005 9 9 1,008 .076 018 004 6 Te) .922 075 019 005 4 oe. Series of heights: .922, .929, 1.008, 1.015, 1.030, 1.100, —_ 1 134, : Arithmetic mean = 1.066. im «a Length of spikes, 2: .075, .076, .o80, .o82, .087, .090, .098, 14 A rithi mean = .o89. ite Breadth of spikes, 2: .016, .018, .o19, .020, .022. Arithmetic 1 2 II.) Sp. Gr. 1.014. N Height of Length of Breadth of | Width of Third ne Plant. Spike, 9. Spike, 9. Leaf from Top. OE 1.398 .090 023 _ 005 2 1,288 813 023 .006 3 1.430 .O9I O21 .006 + 1.473 -095 025 .007 5 1.545 -145 024 .007 6 1.293 .087 .022 .006 7 1.572 .084 .023 .007 8 1.413 .130 025 .008 9 1,300 126 023 008 10 1,560 .I20 .025 .007 1910.] ON THE DISTRIBUTION OF PLANTS. 473 - Series of heights: 1.288, 1.293, 1.300, 1.398, 1.413, 1-430, 1-473, 1.545, 1.560, ‘1.572. Arithmetic mean = 1.427. Length of spikes, 2: .o84, .087, .090, .0QI, .005, .II3, .120, .126, .130, .145. Arithmetic mean = .108. Breadth of spikes, 2: .021, .022, .023, .024, .025. Arithmetic mean = .023. Turp Serres. Typha angustifolia From StocKToN LAKE SHore (PosITION III.) art Heap or Lake. Sp. GR. 1.005. No. of Leaves. H of Length of Breadth of | Width of Third Sterile Phat. Spike, 2. Spike, 9. Leaf from Top. Dry. | G Part. 2.026 -164 .023 009 8 7 -O16 2.108 -162 025 o10 10 7 -024 1.862 -154 -o18 or! 9 5 -Oo16 1.882 -146 -022 oIo 7 6 .026 1.803 - 169 -022 o1o 9 7 -031 1.789 -141 -025 OIr 8 8 -022 1.668 -161 .020 009 8 6 -027 1.678 -138 .O21 008 10 6 -c28 1.920 -182 -024 009 7 7 -030 1.815 -166 -026 008 6 6 -O12 Series of heights: 1.668, 1.678, 1.789, 1.803, 1.815, 1.862, 1.882, 1.920, 2.026, 2.108. Arithmetic mean = 1.885. _ Length of spikes, 2: .138, .141, .146, .154, .161, .162, .164, .166, .169, .182. Arithmetic mean = a _ Breadth of spikes, 2: .018, .020, .021, .022, .023, .024, .025, .026. Arith- metic mean = .022. If we take the arithmetic means of the plant heights, lengths of pistillate spike portions and breadths of pistillate spike portions of the thirty plants taken from three separate localities along the shores of Stockton Lake, we will appreciate the influence of the saline con- ditions of the soil upon the relative size of the plants of these three sets. MEAN DIMENSIONS OF 30 PLANTS. Height of mgth of Breadth of rs Mare. | Sone o | esta Position I. 1.066 -089 -O19 1.0145 Position IT. 1.427 -108 +023 | 1.0140 Position III. 1.855 158 -022 1.0050 _ This table clearly shows that the cat-tails in fresh water are much ller than those growing under more saline conditions, and this lies not only to the heights of the plants, but to the other dimen- S as well. PROC. AMER. PHIL. SOC., L. 201 FF, PRINTED AUG. 25, IQII. 474 HARSHBERGER—INFLUENCE OF SEA WATER [April 22, The next three series of Typha angustifolia were collected along the shores of Pleasure Bay under somewhat similar conditions to those along the shores of Stockton Lake. Ge Sines FourtH Series. Typha angustifolia 1s Sart MArsH. Sp. GR. LOIS. Lee No. | | Height of Length of Breadth of ress he essai at hoof Plant. Spike, Q. Spike, 9 Tey: i Pate : I .788 110 005 4 7 1025 2 .962 .209 O10 7 2 035... 3 .980 .100 005 5 6 040 — 4 .g10 .119 O10 7 4 ect feng 5 888 135 .009 6 3 3 a7 6 768 .096 .006 5 4 GS. 7 925 .120 007 6 5 ei 8 857 .100 009 6 4 .1008 — 9 1,005 119 006 3 6 104% | 10 .904 .130 006 4 6 Fi ig Series of heights: .768, .788, .857, .888, .904, .910, .925, .962, .98o, 1.005. Arithmetic mean = .808. Length of spikes, 2: .096, .100, .110, .119, .120, .130, .135, .200. Arithmetic mean = .127. Breadth of spikes, 2: .005, .006, .007, .009, .o10. Arithmetic mean = 007. Firtu Series. Typha angustifolia NEAR MippLe Part oF UPPER PLEASURE Bay. SP. GR. 1.005. No Height of Length of Breadth of No. of Teaver of ‘ Plant, Spike, 9. Spike, 9. Dee: Greta: Part. I -910 -102 -O10 5 40 +029 2 1,030 «112 -O12 6 4 : ee 8 3 1.130 -128 -O14 4 6 +034, 4 -879 -090 -OII 5 4 027 — 5 .877 .089 -O10 4 5 042 6 .833 -087 .007 6 4 +045 7 -932 .o81 .009 5 4 +040 8 1,102 115 -O14 5 5 +034 9 1.096 -I14 -013 5 6 +030 10 1.180 -133 a 5 5 034 _ Series of heights: .833, .877,. .879, .910, .932, 1.030, 1.006, 1.102, 1.180. Arithmetic mean = .996. Length of spikes, 2: .o81, .087, .o80, .090, .102, .112, .I14, .115, .128, Arithmetic mean = .105. Breadth of spikes, 2: .007, .000, .O10, .OII, .OI2, .013, .O14, .O15. metic mean = .OII. * Measurements include sterile and staminate part of the spike. mgr.) ON THE DISTRIBUTION OF PLANTS. 475 SrxtH Series. Typha angustifolia Cottecrep aT HEAD or PLEASURE Bay. Sp. GR. 1.000. No. of Leaves. Height of No. Pent. Spike, Q Spike ee NNN Sterile Part. I 1.564 -145 -O15 5 8 .028 2 1.642 -193 -O16 5 6 -040 3 1.430 -147 -020 6 5 -032 4 1.688 -1I7 -o18 6 6 -035 5 1.543 -148 -019 6 6 -031 6 1.467 -134 -0i8 5 5 -052 7 1.615 -144 -O19 6 7 -032 8 1.307 -148 -020 5 6 .030 9 1.632 -173 -O16 7 6 -021 10 1.657 -182 -O15 5 5 -026 Series of heights: 1.307, 1.430, 1.467, 1.543, 1.564, 1.615, 1.632, 1.642, 1.657, 1.688. Arithmetic mean = 1.554. Length of spikes, 9: .117, .134, .144, .145, -147, .148, .173, .182, .193. _ Arithmetic mean = .154. Breadth of spikes, 2: .o15, .016, .o18, .o19, .o20. Arithmetic mean — .o18. | Constructing a table of means of the last three series, we dis- _ cover that the heights of the cat-tails and the dimensions of the spike increase with the decrease in the salinity of the water. MEAN DIMENSIONS OF 30 PLANTS. Height of Length of Breadth of raed Eaoiage : Spike, 9 . | = Position IV. 898 .127 .007 | 1,015 Position V. | 996 105 OI 1,005 Position IV. 1.554 154 018 1,c0o _ Now, if we combine the two tables which demonstrate the mean dimensions of the sixty measured plants collected from six widely _ diverse positions, we will see at a glance that Typha angustifolia when found in soil with saline conditions, as indicated by the specific gravity of the soil water, is reduced in size compared’ with other plants growing under more, or less, fresh-water conditions. All of the dimensions of the plants are influenced, but not in corresponding proportions, and it is also noteworthy that the cat-tails in a more saline soil are not only smaller in size, but show a more yellowish- green appearance than the taller, darker green plants controlled by fresh water. 476 HARSHBERGER—INFLUENCE OF SEA WATER _ [April 22, TABLE SHOWING INFLUENCE OF SALINITY OF WATER ON THE DIMENSIONS OF Sixty PLants or Typha angustifolia. MEASUREMENTS IN METERS. Habitar, | Hoel) Gosone | Height of Height of| of Spies, Lent of ot Spite, neath Most aioe | yd] 1045 | "966 | gon | 089 | son | 088 | oe Medium saline. a uore an 1.211 a -106 pa ory Fresh water. oo spe. ee 1.710 on -156 “O18. 020 Having presented the results obtained by using the hydrometer without the attached thermometer, it next concerns this paper to discuss and tabulate the results obtained by the hydrometer so con- structed as to combine with the hydrometer scale a thermometer, whereby density and temperature can be estimated at the same time. This enables us then to reduce all of our specific gravity determina- tions to the uniform temperature of 15° C., so that the second series — of observations are far more accurate as giving the actual salinity of the water which bathes the roots of a number of typic salt marsh © plants. In all of the following data, the corrected specific gravity — determinations are placed within brackets. OBSERVATIONS WITH HyDROMETER AND ATTACHED THERMOMETER. ‘ The numbered data given below were collected at three localities convenient to Belmar, N. J., easily reached by trolley, viz., Manas- quan Inlet, Wreck Pond and Shark River. The same plan was adopted of working from the most saline conditions of environment to the least saline conditions and the gradual change of the vegeta- tion will be noted, if we follow the sequence of the numbered sta- tions at whith hydrometric readings were made. i 81. Salt water in north arm Manasquan River. Thoroughfare fringed with Spartina stricta maritima and Salicornia herbacea. Sp. gr. 1.016; temp. 26°. [Sp. gr. 1.0188.] a 71. Salt Creek at bridge back of Manasquan Life Saving Statio Meadow sod is here 45 cm. deep, with sand below. Sp. gr. 1.01 temp. 23°. [Sp. gr. 1.0174.] 1910.1 ON THE DISTRIBUTION OF PLANTS. 477 | 72. Salt Creek, nearer Manasquan Inlet, below the bridge. Here is Spartina stricta maritima associated with Salicornia herbacea. _ Sp. gr. 1.0160; temp. 23°. [Sp. gr. 1.0179.] 73. Hole dug in middle of Spartina patens association. Water _ reached at 82cm. At same level of the marsh, but in a slightly dif- _ ferent position were found Salicornia herbacea and Limonium caro- ‘linianum. Sp. gr. 1.018; temp. 21°. [Sp. gr. 1.0194.] 74. Hole dug in the middle of a patch of Salicornia herbacea, sur- _ rounded by Distichlis spicata, Limonium carolianum. No free water _ obtained after digging to a depth of 82 cm. 75. Water from ditch cut through Spartina stricta maritima, Spartina patens, Salicornia herbacea. Sp. gr. 1.017; temp. 20°. _ [Sp. gr. 1.0182. ] _ 76. Hole 56 cm. deep in association of Spartina stricta maritima, _Salicornia herbacea, Distichlis spicata. Sp. gr. 1.0175; temp. 19°. [Sp. gr. 1.0184.] _ 77. Small marsh pool (7 cm. deep) with Spartina stricta mari- _ tima, Salicornia herbacea, Spartina patens. The high specific gravity of the water in this pool due to strong evaporation. Sp. gr. 1.026; temp. 29°. [Sp. gr. 1.02996. ] : 78. At head of drainage ditch with Spartina patens. Sp. gr. 1.015; temp. 22°. [Sp. gr. 1.0166.] 79. Hole in Juncus Gerardi association which fringes Spartina patens inwardly and touches an association of Baccharis halimifolia, _ Panicum virgatum, Solidago sempervirens. 80. At head of drainage ditch with Juncus Gerardi as in 79). Sp. gr. 1.0165; temp. 19°. [Sp. gr. 1.0174.] 82. Water from a drainage ditch in Juncus Gerardi association. Soil 49 cm. deep. Sp. gr. 1.0105; temp. 21°. [Sp. gr. 1.0117.] The observations at Wreck Point were made on August 13, 1909, six days after the inlet, which had been closed for some time, _ was opened. The first three tests were made of the water from the pond proper without relating them to the nearby vegetation. ’ 83. Water at trolley bridge. Sp. gr. 1.0090; temp. 20°. [Sp. gr. 1.0102.] : _ 84. Water at railroad bridge. Sp. gr. 1.0020; temp. 20°. [Sp. gr. 1.0029. | 478 HARSHBERGER—INFLUENCE OF SEA WATER _ [April 22, : 85. Water at carriage bridge. Sp. gr. 1.0010; temp. 20°. [Sp. gr. 1.0019. ] 87. Water in Spartina stricta maritima association at high tide, just above the railroad bridge. Sp. gr. 1.0140; temp. 25°. [Sp. gr. 1.0164. ] on 88. Water in Spartina stricta maritima association at high tide, at carriage bridge. Sp. gr. 1.0030; temp. 23°. [Sp. gr. 1.0046.) 86. Stream entering Wreck Pond at tension line between salt — marsh and fresh-water marsh at low tide. Here were found Spar-— tina stricta maritima in broken patches being gradually replaced by Scirpus lacustris, Scirpus pungens and Spartina polystachya. on “ gr. 1.0000; temp. 18°. [Sp. gr. 1.0005.] 89. Water in Spartina stricta maritima association slong high bank fronted with Panicum virgatum. Sp. gr. 1.0000; est aa [Sp. gr. 1.0014.] The observations begun on Shark River were delayed by a severe 2 northeast shifting to southeast storm, August 17, 1909, so that the — tides were exceptionally high and all of the typic salt marsh plants — along Shark River were submerged. Unusual opportunities were — presented, therefore, to determine the salinity of the water which flooded the salt marsh species. 93. Frontal association of Spartina stricta maritima near open- ing of the inlet. Sp. gr. 1.0205; temp. 20°. [Sp. gr. 1.0217.] . 94. Somewhat back from inlet water covering Spartina stricta maritima, Solidago sempervirens. Sp. gr. 1.020; temp. 20°. [ gr. 1.0212.] 95. All of the salt marsh associations of plants on the Belmar side of Shark River, such as Spartina patens, Juncus Gerardi, Sali- cornia herbacea, including Atriplex hastata and Myrica carolinensis, — submerged excepting the tops of Spartina stricta maritima and the low sand dunes on which grow Ammophila arenaria, Baccharis halimifolia, Solidago sempervirens. Sp. gr. 1.0215; temp. 19°. Sp gr. 1.0224. ] 97. Some distance back from the inlet along Fifth Avent Belmar, the following plants were found submerged: Scirpus gens, Cicuta maculata, Hibiscus moscheutos, Panicum virg — a910.] ON THE DISTRIBUTION OF PLANTS. 479 Baccharis halimifolia. Sp. gr. 1.011; temp. 19°. [Sp. gr. 1.0118.] It will be seen by reference to the above observations that even the least typic of the salt marsh species which usually grow subjected to the influence of fresh water are placed occasionally under more __ trying conditions during exceptionally severe storms, when they are _ subjected to the action of almost pure sea water. On August 19, 1909, the storm having subsided, the normal flow of the tide in and Fic. 4. Map of Shark River and Bay, N. J., showing stations at which hydrometric determinations of the salinity of the water were made. Numbers indicate specific gravities at the points directly under the first figure and the subnumbers indicate the observational stations as noted in the paper and in the original field note-book.” out of the inlet was reéstablished and the following series of obser- _ vations were then made. § 126. Water from lagoon inside jetty at Shark River Inlet at high tide. Sp. gr. 1.0190; temp. 23°. [Sp. gr. 1.0210.] 127. Hole dug 20 cm. deep in an association of Scirpus pungens, Solidago sempervirens, Spartina patens, Atriplex hastata, Suaeda _ maritima. Water rising from a sandy gravel. Sp. gr. 1.0010; temp. 21°. [Sp. gr. 1.0022.] 130. Water from a hole 46 cm. deep in an association of Juncus 480 HARSHBERGER—INFLUENCE OF SEA WATER _ [April 22, Gerardi, Limonium carolinianum, Sp. gr. 1.0255; temp. 20°. [Sp. gr. 1.0267. ] . 131. Water from hole 28 cm. deep in an association of Distichlis — 4 spicata, Salicornia herbacea. Sp. gr. 1.0065; temp. 20°. Lo i ‘ 1.0075. ] 133. Water from a hole 50 cm. deep in a Spartina stricta mari- tima association. At 20 cm. a hard pan of gravel stones was reached, then a layer of sand was passed and at the bottom a hard gravel layer. Sp. gr. 1.025; temp. 21°. [Sp. gr. 1.0265.) 134. Water from Shark River. Sp. gr. 1.0205; temp. oll [Sp. gr. 1.0222.] . 135. Water from a ditch along back of Shark River Belmar side, lined with Spartina stricta maritima, Spartina Salicornia herbacea. Sp. gr. 1.0015; temp. 20°. [Sp. gr. 1.002 136. Water from hole in marsh 61 cm. deep at the base o! clump of Baccharis halimifolia associated with Distichlis . Aster tenuifolius, Salicornia herbacea, Limonium carolinianum soil brown and loose, subsoil sandy. Sp. gr. 1.024; temp. 20°. : . gr. 1.0252.] | 137. Water covering surface of the marsh in middle of a : Dies tichlis spicata association with Spartina stricta maritima, Salicornia herbacea, Limonium carolinianum. Sp. gr. 1.0245; temp. 27°. [Sp. gr. 1.0278. | ie. 138. The last test recorded here was to determine if there was any difference in the salinity of the water between ebb and flow. Tide receding rapidly. Sp. gr. 1.0215; temp. 20°. [Sp. gr. 1.0238.] 101. Shark River Bay water in channel through salt marsh bounded by S$ a stricta maritima and Spartina juncea. Sp. gr. 1.021; temp. 21°. [Sp. gr. 1.0224.] 102. Hole jae in marsh 30 cm. deep in an association of Spartina stricta maritima, Limonium carolinianum, Salicornia herbacea. Sp. m gr. 1.0245; temp. 20°. [Sp. gr. 1.0257.] ‘103. Hole dug in ard 30 cm. deep in an association of Dis-— tichlis spicata and Limonium carolinianum. Sp. gr. 1.0185; temp a 20°. [Sp. gr. 1.0196. ] 4 104. Hole dug 20 cm. deep in a pure association of Juncus Gerardi. Sp. gr. 1.0195; temp. 21°. [Sp. gr. 1.0209.] 1910.) ON THE DISTRIBUTION OF PLANTS. 481 107. Water at head of north arm of Shark River Bay (see map, Fig. 4) in middle of an association of Spartina stricta maritima and Spartina patens. Sp. gr. 1.0150; temp. 25°. [Sp. gr. 1.0175.] Fic. 5. View of salt marsh island fringed with Spartina stricta maritima and covered by associations of Spartina patens, Juncus Gerardi, Distichlis Spicata, etc. Shark River, New Jersey. 109. Contracted portion of the north arm of Shark River Bay, where Spartina stricta maritima breaks up into patches between which grow Scirpus pungens, Spartina patens and N ymphea odorata. Sp. gr. 1.0025; temp. 24°. [Sp. gr. 1.0044.] ' 110. Water at base of a patch of Spartina polystachya. Sp. gr. 1.0025; temp. 20°. [Sp. gr. 1.0034.] 112. Water from hole 20 cm. deep at the base of an association of Phragmites communis (see Fig. 4). Sp. gr. 1.0010; temp. 28°. [Sp. gr. 1.0153.] 113. Water in upper portion of the south arm of Shark River Bay. Sp. gr. 1.0035; temp. 25°. [Sp. gr. 1.0057. 114. Water from middle of Spartina stricta maritima associa- 482 HARSHBERGER—INFLUENCE OF SEA WATER _ [April 22, tion, south arm of Shark River Bay at carriage bridge. Sp. gr. 1.0015; temp. 25°. [Sp. gr. 1.0036.] 115. Water from last extensive patch of Spartina stricta mari- tima merging with Typha angustifolia, Sp. gr. 1.0000; temp. 23°. [Sp. gr. 1.0016. ] 116. Water tested at the head of the south arm of Shark River Bay (Fig. 6), where the vegetation becomes continuous and the patches of Spartina stricta maritima are divided by narrow lines of Scirpus fluviatilis, Typha angustifolia, Zizania aquatica, touching a forest growth of Nyssa sylvatica, Sassafras variifolium, Pinus Fic. 6. Association of Spartina stricta maritima in upper part of the south arm of Shark River Bay, showing blending and transition of salt-water and fresh-water vegetation. rigida, Quercus prinus, Q. alba. Sp. gr. 0.9990; temp. 25°. [Sp. gr. I.001I.] 119. Water 60 cm. deep at a point along south shore of Shark 1910. ] ON THE DISTRIBUTION OF PLANTS. 483 River Bay with Vallisneria spiralis abundant. Sp. gr. 1.0165; temp. } p- § : 26°. [Sp. gr. 1.0193.] 120. Hole dug in a marsh at base of a clump of Peltandra vir- gimica. Sp. gr. 1.016; temp. 27°. [Sp. gr. 1.0191.] 124. Water from a marsh lagoon at base of a steep bluff sub- Fic. 7. Clump of Panicum virgatum growing along shore of Stockton Lake controlled by fresh water. jected to evaporation between the daily tides. Lagoon surrounded by Spartina stricta maritima. Sp. gr. 1.0155; temp. 29°. [Sp. gr. 1.0192. | The above observations give the geographic data upon which the study of the distribution of the salt marsh species has been based. It will be seen that proceeding from the ocean up the various bays and inlets there is a general decrease in the saltiness of the controlling water as revealed by the use of the hydrometer and the amount of 484 HARSHBERGER—INFLUENCE OF SEA WATER _ [April 22, : salt which controls in general the habitats of the several species is graphically shown in the sketch map until a point is reached where the salt marsh vegetation mingles with that of the fresh water marshes until it is gradually replaced by vegetation controlled by fresh water (Fig. 7). a SEQUENCE OF HypROMETRIC READINGS. It is important to give now in sequence the various specific gravi- ties corrected to 15° C. to which readings are appended, the name ~ of the plant or the names of the plants subjected to that specific density of salt water. The subnumber indicates in all cases: the : number of the observation in the series previously given. 1.02996,, Spartina stricta maritima, Spartina patens, Salicornia herbacea. herbacea, Limonium carotbiavian 1.02670,,, Juncus Gerardi, Limonium carolinianum. . I — S partina stricto maritima, Limonium carolinianum. cornia herbacea. 1.02520,3, Baccharis halimifolia, Aster tenuifolius, Sala bacea, Limonium carolinianum, Atriplex hastata. / ie 1.022419, Spartina stricta maritima, Spartina patens, ocean wa t from surf at Belmar. ; 1.0224,3;, water in thoroughfare. 1.02170,, water at high tide. Shark River Inlet, covering Sauk Spartina stricta maritima. 1.02090,9, Juncus Gerardi. ot 1.01960,), Spartina patens, Salicornia herbacea, Limonium car linianum, 1.01930,,, Vallisneria spiralis. 1.01920,., Spartina stricta maritima. I.01910,,, Spartina stricta maritima, Scirpus pungens. 1.01910, Peltandra virginica. 1.01880,, water in channel connecting Manasquan River and § ton Lake fringed with Spartina stricta maritima, Salico herbacea. * ON THE DISTRIBUTION OF PLANTS. 485 1.01820,, Spartina stricta maritima, Spartina patens, Salicornia herbacea. : 1800,,. Spartina stricta maritima. 1.01790,, Spartina stricta maritima, Salicornia herbacea. 1.01750,,; Spartina stricta maritima, Spartina patens, Distichlis Spicata. i 1740,, Juncus Gerardi. 1660,, Spartina patens. .01640,, Spartina stricta maritima. _ 1.01530,,. Phragmites communis. 1.01170,, Juncus Gerardi. 00750,;, Distichlis spicata, Salicornia herbacea. .00650,,, Scirpus maritima, Pluchea camphorata. .00490,., Spartina stricta maritima, Spartina patens. _ 1.00460,, Spartina stricta maritima. 1.00440, Spartina stricta maritima, Spartina patens, Scirpus pun- gens, Nymphea odorata. [.00360,,4 Spartina stricta maritima. 1.00340,,. Spartina polystachya. 1.00240,,, Spartina stricta maritima, Spartina patens, Salicornia _ herbacea. 1.00240,., Spartina stricta maritima. 1.00220,,, Scirpus pungens, Solidago sempervirens, Atriplex has- tata, Spartina patens, Suaeda maritima. 1.00160,,, Spartina stricta maritima, Typha angustifolia. 1.00160,,,; Phragmites communis. 1.00140,, Panicum virgatum, Spartina stricta maritima. 1.00140,, Scirpus lacustris, Scirpus pungens, Spartina polystachya. 1.0118,, Hibiscus moscheutos: I.00110,,, Scirpus fluviatilis, Zizania aquatica, Typha angustifolia. 1.00050,, Spartina polystachya, Scirpus lacustris, Scirpus pungens, Spartina stricta maritima. .00040,,, iron-sulphur spring water. 486 HARSHBERGER—INFLUENCE OF SEA WATER _ [April 22, SEQUENCE OF SALT MarsH PLANTS ARRANGED ACCORDING TO MAXIMUM DeENsitTy OF SALT WATER. In order to make what follows more general and intelligible, the specific gravity of saline solutions at 15° C. and the corresponding percentages of sodium chloride in solution is displayed in the fol- lowing table for converting specific gravities of salt solutions into per cent. of sodium chloride taken from Landolt-Bérnstein, “ Physi- kalisch-Chemische Tabellen,” p. 322. Per Cent. d a Per Cent. d - 0.5 1.0034 3-0 1.02 1.0 1.0064 3.5 ae 1.5 1.0100 4.0 1.0282 2.0 1.0137 4-5 1.0319 2.5 1.0173 5.0 1.0355 Now, if we place the salt marsh plants according to their ability | to withstand degrees of salinity of water, we can appreciate better the factors which control their distribution in the bays and estuaries s of the New Jersey coast. The first figures show the greatest degree a of salinity to which the various species are subjected and the second number indicates the limit toward the fresh water end of the series. — The range varies in the different species to a marked extent. Spartina $tricta MOretmd nges bs Osm Pressure and Temperature. Publications of the University of fornia, II.: 1906, No. 8. 1906-7. On the Importance of Physiologically Balanced Solutions for Pla The Botanical Gazette, XLII. (1906) : 127-134; XLIV. (1907) : 25 Osterhout, W. J. V. 1906. Extreme Toxicity of Sodium Chloride and its Prevention be ot Salts. Journal of Biological Chemistry (1906) : 363-360. 1910.) ON THE DISTRIBUTION OF PLANTS. 495 1900. Die Schutzwirkung des Natriums fiir Pflanzen. Jahrb. Wiss. Bot., XLVI. (1900) : 121-136, figs. 3. 1909. On the Similarity in the Behavior of Sodium and Potassium. The Botanical Gazette, XLVIII. (1909) : 98-104. Oswald, W. 1905. Versuche iiber die Giftigkeit des Seewassers fiir Siisswassertiere. ; Pfltiger’s Archiv, CVI. (1905) : 568-508, pls. 2-7. Scofield, Carl S. 1905. The Salt Water Limits of Wild Rice. Bull. 72, U. S. Bureau of Plant Industry, 1905. _ S§mith, John B. _ ‘1902. The Salt Marsh Mosquito. Special Bull. T, Agricultural Experimental Station, N. J., July 8, 1902. 1907. The New Jersey Salt Marsh and its Improvement. Bull. 207, N. J. Agricultural Experiment Station, November 14, 1907. - $tevens, F. L. 1908. The Effect of Aqueous Solutions upon the Germination of Fungous _ Spores. The Botanical Gazette, XXVI.: 377-406, December, 1908. _ Transeau, Edgar N. _ 4908. The Relation of Plant Societies to Evaporation. The Botanical Gazette, XLV.: 217-231, April, 1908. True, Rodney H. 1898. The Physiological Action of Certain Plasmolyzing Agents. The Botanical Gazette, XX VI.: 407-416, December, 1808. — and Oglevee, C. S. 1905. The Effect of the Presence of Insoluble Substances on the Toxic Action of Poisons. The Botanical Gazette, XX XIX.: 1-21, June, 1905. Wahl, Robert, and Henius, Max. 1902. Comparative Table of Beaumé Degrees and Specific Gravity according to Bourgougnon. American Handy Book of the Brewing, Malting and Auxiliary Trades (second edition), 1902: 1156. _ Warming, Eug. 1897. Halofyt-Studier. Mem. de l’Acad. Roy. Sci. et des Lettres de Danemark. Copenhagen 6me Ser. Sect., t. VIII., No. 4, 1897. 1908. Dansk Plantevaekst. I., Strand vegetation. 8vo, pp. vi + 325, figs. 154," Copenhagen and Christiania, 1906. (Review in Bot. Gaz., XLV.: 55- 56, January, 1908.) West, G. S. 1904. The British Fresh-water Alge. 1904: 55. Willis, Clifford. ‘Torr. Alkali Soils. Bull. 126, South Dakota ‘Agricultural Experiment Sta- tion, April, ror. Umno 4 ics = ek es eA ee ee 496 and low sand dunes covered with red cedar, "Janibers distance. Fic. B. Salt marsh near Avalon, N. cs inieadotert by at head ofa bay blending with i deciduous gcc in eae PLATE XXI. growth of pee 80 i har bet pre hs sien! Mex . Pinus rigida, eee PROCEEDINGS Am. PHiLos. Soc. VoL. L. No. 201 PLATE XX Fic. A : : Fic. B Salt Marshes near Avalon, N. J. a ati, ae ; PLATE XXl 201 ‘f °N ‘JuUlog Sszour0sg 7 Yysreyy IVS S - PROocEEDINGS Am. PHiLos. Soc. VoL, L. No. 2 ae: ¥ 2; & at an . THE COST OF LIVING IN THE TWELFTH CENTURY. By DANA C. MUNRO. : . (Read April 20, 1911.) s As yet it is impossible to make any statement of the average cost _ of living in the twelfth century in any country of Europe. Much material is accessible in the Pipe Rolls and similar accounts, in the _ charters and other legal documents, of which so many thousands _have been preserved ; but no one has attempted a careful statistical study for this period. Thorold Rogers began his work on the prices _ in England with the year 1259; Curschmann collected some items _ __ for Germany during the years 1190-1225 ; Lamprecht gathered some 4 data on prices for France in the eleventh century; and there are ; a some other partial statements. Whether it will be possible to make ‘ 4 an accurate estimate can only be ascertained after minute and ex- 3 @ tended examination of the accessible material. q But it is possible to gather some examples which are illuminat- @ ing. In 1181 the former mistress of Henry II., and the mother of Geoffrey, was receiving an annual pension of 20 marks or 13 pounds _ 6s. 8d. In the same year the “ Archbishop of Norway,” who was _ then visiting in England, was allowed by the king tos. a day for _ the expenses of himself and suite. The same amount was allowed in 1180 to the Abbot of Glastonbury. Evidently tos. a day was con- _ sidered sufficient for the expenses of a high church official and his attendants; probably the pension of 20 marks, or a little over 8d. a day, was sufficient for the expenses of a lady and her servant. _ This is rendered more probable by the fact that Richard the Lion- _ hearted, when he hired vessels for his crusade, had to pay only 2d. a day to sailors and 4d. a day to the captains. In 1201 the French ambassadors made a treaty with Venice, by which the latter agreed to carry the crusaders across the sea and furnish them with pro- 497 498 MUNRO—THE COST OF LIVING [April 20, visions for a year on the payments of 2 marks for each man and 4 marks for each horse.’ The prices current at the time throw some light upon the above: in Lincolnshire, in 1181, a goose cost a penny; a sheep, 4d.; an ox, — 3s.; a farm horse, 5s.; a pig, 1s.; scarlet cloth, 6s. 8d. an ell; fine green cloth, 3s.; gray, 1s. 8d.; blankets, 3s. an ell. Thus, if Geof- frey’s mother had expended her pension in buying live stock, she could have bought 25 horses, 25 oxen, 25 pigs, 25 sheep, and 100 geese; or if she had preferred, she could have bought 50 yards of scarlet cloth, say, enough for four or five dresses in the fashion of the day. The difference in cost between the necessities and the luxu- ries is very noticeable. While it is impossible to state the exact cost of living, it is certain that this cost was increasing rapidly for the upper classes, and prob- ably for the middle classes. The rise was due to a variety of causes, and it would be easy to make out a long list, including war, famine and pestilence; but two appear to have been especially important. First, there was a change in the standard of living. Acquaintance with the east through the crusades led to a desire for the luxuries which were produced at Constantinople and in Asia. Before the first expeditions to the Holy Land spices had been used only to a slight extent in the west of Europe. At the capture of Caesarea, in 1101, the Genoese received over 16,000 pounds of pepper as a por- tion of their booty. This, and other spices, soon came into general use and were imported into western Europe in great quantities. The references in the literature of the day point conclusively to the wide- spread use of spices and their great popularity. The costly fabrics of the East were also in great demand, a the heroines of the poems are frequently described as clad in the stuffs made in Constantinople, or farther eastward. No lady was considered well-dressed by the poet unless she had garments im- — ported from the East. Oriental rugs became so fashionable that a manufactory for them was established in Paris. Glassware, sugar, — *De Wailly estimates a mark as equivalent to 52 francs at the present — day; that is, two marks would be equivalent, roughly, to 104 francs, or $20. — Of course this is entirely misleading, as it would be impossible to furnish — transportation and food for a year for $20 per individual. P tgtt.] IN THE TWELFTH CENTURY. 499 dye-stuffs, and other oriental products were coveted and secured as far as possible. Life as a whole became more luxurious. In Germany four meals a day supplanted the three of an earlier period; and the ideal hero was a mighty trencher-man. According to the Pseudo-Turpin, Charles the Great ate “a whole quarter of a lamb, two fowls, a goose, or a large portion of pork; a peacock, a crane, or a whole hare” at a meal. Luxury in dress, at least among the middle classes, was not confined wholly to the oriental products. Fashion began its despotic sway for Germany and other parts of western Europe in the twelfth century, and those who could not afford the Byzantine stuffs might in their domestic weaves imitate the prevailing styles of long trains and full sleeves almost sweeping the ground. Shoes for both men and women changed in style almost every year ; some- times the toes were long and pointed, extending up toward the knees; at other times, short and broad. Other items of extrava- gance might be mentioned, such as the enormous head-dresses, wigs and other false hair; but enough has been indicated. Another great source of expenditure was building. The mon- archs spent large sums on their castles and residence halls, and the nobles and citizens followed their lead. Palaces, cathedrals, for- tresses, country houses, town halls, hospitals and other edifices were going up in all the leading centers. The cost of building was greatly increased by the general substitution of stone for wood, and by the frequent use of lead for the roofs. Great quantities of this metal were exported from England to various places in France, and even to other parts of Europe. The second cause of the rise in the cost of living was the increase in the amount of money available. Western Europe was just chang- ing from Natural- to Geldwirtschaft. The author of the “ Dialogus de Scaccario,” who wrote about the beginning of the last quarter of the twelfth century, says that he had been told of the former custom by which all payments to the treasury were made in kind, and that he had seen a man who had witnessed the bringing in of the provisions from the various parts of the country. In fact, in the reign of Henry I. of England the sheriffs obtained their receipts 500 MUNRO—THE COST OF LIVING [April 20, for so many fowls, eggs, ducks, hogs, oxen, etc., or so much beer, wool, corn or other grain. But this practice had not wholly gone out under Henry II., in spite of the statement of the author of the “Dialogus.” In the Pipe Roll of 1181-82, for instance, there is the record of the payment by Cheshire of forty cows in addition to their money dues. On the whole, however, all through western Europe payments in money were superseding payments in kind, and this was due mainly to the increase in the amount of the circulating medium. Large numbers of coins were brought home from the East. In the Scandinavian lands it is said that more than 25,000 Arabic coins have been dug up in recent times. In the literature of the twelfth century, Arabian gold is very frequently referred to and is contrasted with the lighter-colored gold of the West. At the capture of Casa- rea in 1101 (when the pepper was obtained) the Genoese secured over 400,000 solidi of Poitou, and they received only one third of the booty. The crusaders were always keen for gold. Whenever they won a victory they sought anxiously for the precious metals; frequently they cut open the bodies of the slain enemies, because they believed the latter had swallowed their coins; sometimes they made great heaps of the bodies and burned them in order to obtain the gold which had been secreted. Many similar facts might be cited which would illustrate the enrichment of the West by the coins brought in from the East. Far more important, probably, was the coinage and use of the precious metals which had previously been hoarded, especially in ornaments and works of art. Until about the close of the eleventh century, there had been comparatively little occasion for a large stock of ready money, but when the crusaders made their prepara- tions for their long expeditions they needed large sums of money, both for their equipment and for their journey. Even the partici- — pants in the so-called Peasants’ Crusade took enough money with : them to pay all the expenses for several months, when they marched under the leadership of Walter the Penniless and Peter the Hermit. — Because of the demand for coins, the mints of the West were very active in the twelfth century. Under Henry I. of England, 94 minters were busy. In 1125 all the 94 were called up for punish- ~ r91t.] IN THE TWELFTH CENTURY. 501 ment on the ground that they had debased the coinage, and each one had his right hand struck off. Under Henry II. there was a great amount of coining, of which the details may be followed in the Pipe Rolls, as far as they are accessible. In addition, instruments of credit came into use, especially bills of exchange, which greatly in- creased the amount of capital. The Templars in their house at Paris received deposits, and gave orders upon their house in Jerusalem. In doing this, they were probably imitating the example of the Jews, who had long used such papers; and we find the example of the Templars, or of the Jews, imitated by others, so that, e. g., by 1188 bills of exchange had become very common in Hamburg. The extravagance of the age is well depicted in the literature. The knightly hero is always lavish in his gifts and entertainments, as well as in his attire. Sumptuous banquets, where the boards lit- erally groaned under the weight of the dishes, were the fashion. Large stone castles were built and richly adorned, and in these the number of attendants increased greatly. The armor became more costly; the legal expenses, from which the nobles were never free, mounted up; but the main source of out-go was the necessity of keeping up the style of living demanded by the fashion of the day. Consequently the knight had to spend much more, and the minstrels sang only of those who were generous. Even the fathers and mothers in their advice urged their sons to give freely and never to be niggardly. There were great opportunities to acquire wealth. One of the men who improved his chances to the best advantage was Suger. He was of peasant stock and was educated at the Monastery of St. Denis, where he became intimate with the prince who later was known as Louis VI. of France. The intimacy always continued, and after the death of Louis VI., Suger, who was then Abbot of St. Denis, acted as regent of France during the absence of Louis VII. on the crusade. During the time of the king’s expedition, Suger paid all the expenses of the kingdom of France out of his own for- tune. He had previously restored and beautified the church of St. Denis at his own expense. And he still had enough wealth, so that in the last year of his life he planned to equip and finance a crusade 502 MUNRO—THE COST OF. LIVING [April 2, wholly from his own money. Suger was able to acquire this enor- mous fortune because of his great ability, and because he understood the economic conditions of the time. The average noble had no genius for acquiring wealth, and his feudal income, which was fixed mainly by custom, appears to have been stationary or even declining. With the establishment of better order and the increase of the royal power, the nobles had lost both their opportunity to plunder and the right of private coinage, which greatly lessened their income. One feature of the Pipe Roll for 1181-1182 is very significant in this connection. About 300 debtors to the king were listed from various parts of England, most of whom had disappeared or were destitute of means, so that these debts could not be collected. Apparently most of the individuals came from the lesser nobility. The only resource for men of this class was to borrow at usury. The usurers formed one of the two classes of wrong-doers against whom the preaching of the twelfth century was especially directed. They were evidently very numer- ous, and they preyed chiefly upon the nobles. The merchants and the peasants seldom had to resort to the usurers. There were many Christians engaged in this business, but more Jews, and the latter were to suffer severely as the result of the economic conditions. The rate of interest in England when the security was good was 2d. on the pound each week, compounded once in six weeks, or about 52 per cent. a year.?, Consequently if a knight borrowed 40 pounds, — a sum frequently in excess of the annual income of a knight, and — was unable to pay the interest, in a year he would owe 60 pounds ~ and 16s.; in two years, over 92 pounds; in three years, over 140 — pounds; in five years, over 324 pounds, and the interest then would — be over 3 pounds a week. Probably the ill-feeling against the Jews — was due very largely to the anger of the borrowers who found them- selves hopelessly involved in debt. There is a very decided change in the attitude toward the Jews in the twelfth century, and it is sig- nificant that the preparations for the crusades, when ready money was especially needed, were so frequently accompanied by a perse- cution of the Jews; e. g., in 1096, in 1147, in 1189. Their great * Much more was demanded when the security was not good. 1911.] IN THE TWELFTH CENTURY. 503 wealth is shown by the fact that the Jews of England contributed 60,000 pounds towards the crusade of Henry II., and all others only 70,000. There is no estimate of the number who contributed this 60,000 pounds, but there had been a great increase since the begin- ning of the reign of Henry II. At that time all Jews who died in England had to be buried in the cemetery near London. At the end of Henry II.’s reign almost every great town had a Jewish cemetery in the suburbs. The peasants both in town and country gained in prosperity during the twelfth century. The agricultural laborers profited by the opening of more markets for their products. They were some- times able to hire the demesne land and even to rent the mill or the whole manor, because the lord of the manor was in need of ready money. In France many villeneuves were established which offered special privileges in order to attract tenants. Suger’s example in emancipating his serfs was followed more and more frequently by the kings and by the lords. In England many individuals escaped to the towns, and if they were able to remain there unmolested for a year and a day, they were free from all possibility of pursuit. The merchants in the towns profited most. The Lombard cities of Italy gained great wealth by the carrying both of crusaders and of wares. The trade extended widely in western Europe. Fairs were established where the commodities of the whole known world were offered for sale by the merchants from the various countries, who travelled about from place to place. The increase in the dues which the lords received from these fairs bears witness to their prosperity and to the enlarged trade, of which they were the scene. Gross states that the gild merchant first appeared in England about 1100, and that the craft society first appeared on the continent, as in England, early in the twelfth century. If we connect these _ statements with Ashley’s dictum, “Trade, as an independent occu- pation, grew up first in the service of luxury,” the importance of the change in the standard of living will be apparent. The establish- ment of uniform weights and measures, and the universality of certain standards of money, such as the Cologne mark, the Venetian ducat, or the bezant also indicate the rapid advance in commerce. 504 MUNRO—THE COST OF LIVING The fabliaux, or “laughable stories told in verse,” the especial literature of the merchant class, began about the middle of the — twelfth century. In these tales class-consciousness is very evident. They ridiculed the knights and the clergy, while always depicting the latter as wealthy. Some of these fabliaux, which were written for the merchants of the twelfth century, sound curiously modern, as if they might have been told in the nineteenth century in our own western states. They are frequently irreverent, and show an inde- pendence of thought which is very noteworthy in this early period. Their attitude toward women is entirely at variance with that of the courtly literature of the age. In fact, the merchants were think- ing for themselves, and were no longer willing to be subservient to the nobility and the clergy. They were rapidly becoming important political factors, and were winning recognition from the monarchs. They were vying in comfort and luxury with the nobles, and fre- quently ineffective sumptuary laws were enacted to restrict these nouveaux riches. . As yet too little attention has been paid to this change in the standard of living and its effects. In this paper an attempt has been made to set forth only a few of the facts, merely to indicate the nature and importance of the problem. Every one of the subjects here discussed is susceptible of elaboration, and needs to be worked out in detail for each country of western Europe and each period in — the twelfth century. The material is voluminous: as indicated above, the legal documents should be utilized for the definite state- ments which they contain, and the literature of the age should be laid under contribution for its information as to the character, cus- toms and points of view of the various classes. The chronicles un- fortunately will furnish comparatively little, because they generally — give only the unusual events; statements about prices drawn from — them are frequently of little value, because the figures are given account of their extreme highness or lowness. This field, as a whol offers a good opportunity for many monographs, and such work essential before we can understand the economic history of century which was most important in the advance of western Eur UNIVERSITY OF WISCONSIN, Mapison, WIs. AN ANCIENT PROTEST AGAINST THE CURSE ON EVE. By PAUL HAUPT. (Read April 22, 1911.) 8 ‘Tn the Biblical Legend of the Fall of Man, which symbolizes the first connubial intercourse,’ the Lord pronounces a curse on Eve, saying, I will greatly multiply thy sorrow and thy sighing ;? in pain thou wilt bear children; nevertheless thy desire is* to thy husband and he will rule over thee (Genesis, iii., 16).* The great pessimistic philosopher ARTHUR SCHOPENHAUER says that the story of the Fall of Man contains the only metaphysical truth found in the Old Testament; it is the acme of Judaism, der Glanzpunkt des Judentums; but it is an hors d@ ewuvre: the pessi- mistic tendency of this legend has no echo in the Old Testament which, on the whole, is optimistic, whereas the New Dispensation is pessimistic, at least so far as this world is concerned.® We all know what the forbidden fruit® in the midst of the Garden’ of Eden*® means: he who eats of it loses his childlike inno- cence; his eyes are opened, just as Adam and Eve perceived that they were naked. Not to know good and evil, that is, what is wholesome and injurious, means to be like a child.® In the eight- eenth book of the Odyssey (v. 228) Telemachus says to his mother Penelope, I am intelligent and know good and evil,’° I am no longer a child.* In the Bible this phrase is used also of the second child- hood: Barzillai of Gilead answered David, when the king asked him to follow him to Jerusalem, I am this day fourscore years old and can no longer discern between good and evil, that is, my intellect is impaired by old age, I have become again like a child. The explanation of the Fall of Man as the first connubial inter- _ course was given by the celebrated English philosopher THomas _ Hoszes in his Leviathan (London, 1651) and it may be traced back _ to CLEMENT of Alexandria in the second century of our era? But older than this philosophical explanation of original sin** is an PROC. AMER. PHIL. SOC., L. 201 HH, PRINTED SEPT. 6, 1gI1I. 505 506 HAUPT—AN ANCIENT PROTEST [April 22, ancient protest against the curse on Eve, which we find in the fol- lowing chapter of the Book of Genesis, containing the legend of Cain and Abel. The story of Cain and Abel is an institutional legend.’* cae as the narrative of Jacob’s wrestling at Peniel (Genesis, xxxii., 24-32) explains why the Jews do not eat the great sciatic nerve, so the story of Cain and Abel shows why the Cainites, or Kenites,® had the mark of Cain,'® that is, a tattooed tribal mark which warned every man not to slay a member of that tribe. The murder of a Kenite was avenged sevenfold: if a Kenite was killed, the Kenites would slay — seven fellow-tribesmen of the slayer. The tribe of Lamech avenged even the slightest scratch by the death of a youth of the tribe to which the assailant belonged. Lamech and Cain represent tribes, not individuals.17 The Lamechites guarded their tribal honor even more jealously than did the Kenites: if a Kenite was slain, seven fellow-tribesmen of the slayer were slain to avenge his blood; a ee Lamechite, however, was not avenged sevenfold, but seventy-seven- fold; even a wound inflicted on a Lamechite was punished by the death of a fellow-tribesman of the assailant, and a boy of the hostile tribe had to pay with his life for the slightest scratch received by a Lamechite. Therefore an ancient tribal poet addressed the women of Lamech: Ss O Adah and Zillah, attend to my voice! Ye wives of Lamech, give ear to my utterance: A man, if they hurt us, we slay; a boy, if they scratch us, we kill; If sevenfold Cain be avenged, then seventy-sevenfold, Lamech!* — The Kenites were a nomadic tribe in the desert south of Judah They came to Canaan with the men of Judah from the Palm City. that is, the port of Elath,?° at the northeastern end of the Red § Moses’ father-in-law is said to have been a Kenite.2* The Keni were worshipers of Juvu,”? but their offerings were different ft the sacrifices of the sheepmen of Judah,”* represented in the st of Cain and Abel by Abel, that is herdsman, herder.* Cain brow to JHvH offerings of the fruit of the ground, but Abel brought the firstlings of his flock and of the fat thereof; and Abel’s sacri was more acceptable to JHvH than Cain’s bloodless offering. agit.) AGAINST THE CURSE ON EVE. 507 Kenites may have been a semi-agricultural tribe settled near Elath*® before they emigrated with the Edomite ancestors of the Jews to Canaan. Afterwards there may have been some religious differ- ences: the Kenites clung to their vegetable offerings,”® whereas the men of Judah** sacrificed lambs. This led to an expulsion of the Kenites from the region of Judah. The introductory verse, connecting Cain and Abel with Adam and Eve, is a subsequent addition. The name Cain is explained there (Genesis, iv., 1) as being connected with the verb gandh, to produce.27_ When Eve bare Cain, she said, I have produced a man as well as JHvH:** just as JHvH fashioned me from the rib He took from Adam, so I have produced now a new human being.— Some people think that, when the Lord created Eve, He did not take a rib from Adam, but his backbone. Most of us have all our ribs. At any rate, woman is not a side-issue. 3 The story of Cain and Abel was originally simply: Abel was 2 keeper of sheep, and Cain a tiller of the ground. Cain offered _ vegetable offerings to JHvH, whereas Abel sacrificed the firstlings of his flock. Abel’s sacrifices were more acceptable to JHvu. This _ displeased Cain, and Cain said to Abel, Let us go into the field ;?” and when they were in the field, Cain rose up against his brother Abel, and slew him. The field was a tribal battle-ground where the Cainites smote the Abelites, but afterwards they were overpowered and expelled from the territory of the sheepmen of Judah.*° A later theologian has inserted two verses (Genesis, iv., 6, 7) which are translated in the Authorized Version as follows: And the Lord said unto Cain, Why art thou wroth? and why is thy counten- ance fallen? If thou doest well, shalt thou not be accepted? and if thou doest not well, sin lieth at the door. And unto thee shall be his desire, and thou shalt rule over him. There is no connection between this last clause and the preceding one, and the translation sin lieth at the door is impossible. The Ancient Versions show that the text of this theological in- terpolation was corrupt at an early period, and the rendering given in the Greek Bible echoes the tradition that the feud between Cain 508 HAUPT—AN ANCIENT PROTEST [April 22, and Abel was due to some ritual differences. The Septuagint ren- ders: Is it not so? If thou offerest rightly, but doest not cut in pieces rightly, thou hast sinned? Be still!—The Syriac Bible has: Behold, if thou doest well, thou receivest; and if thou doest not well, at the door sin croucheth——We find the same rendering in the Vulgate: Nonne si bene egeris, recipies; sin autem male, statim in foribus peccatum aderit—The Targum paraphrases: If thou doest thy work well, thou wilt be pardoned ; but if thou doest not thy work well, for the day of judgment the sin is laid up, ready to take vengeance upon thee, if thou doest not repent; but if thou repentest, thou shalt be forgiven.*‘—All these explanations are untenable. — The original text seems to have been: If thou art good, I shall receive thee graciously ; but if thou art a sinner,*? I shall not accept thy offering.** The final clause, And unto thee shall be his desire, and thou shalt rule over him, has no connection with the preceding theological interpolation, but is a gloss protesting against the state- ment in the preceding chapter: Thy desire shall be to thy husband, and he shall rule over thee*®* Genesis, iii., 16, states: Unto the woman He said, I will greatly multiply thy sorrow and thy .sigh- ing ;? in pain thou wilt bear children; nevertheless thy desire is* to thy husband, and he will rule over thee. Some one—possibly a woman,** or a man under the influence of a woman, a species of the genus Homo, which is common—added to this statement in the margin: His desire is unto thee, and thou wilt rule over him.®* The story of the Fall of Man and the legend of Cain and Abel may have been written in two parallel columns.** The glossator, who added the theological interpolation in the legend of Cain and Abel, and the author of the polemical gloss to Genesis, iii, 16 may have written their remarks in the space between the two columns. Afterwards these two marginal glosses crept into the text, the “suffragettic” gloss to Genesis, iii., 16 being appended to the theological interpolation after Genesis, iv., 5. a The word desire or longing is used also in the Biblical love- songs, commonly known as the Song of Solomon, where the maiden says of her lover: = BEM Ve a, em pe RA ge Ay My dear one’s am I; he is mine, too; for my love he is longing.* 191t.] AGAINST THE CURSE ON EVE. 509 _ The corresponding word in Arabic (shauq) means passionate love. q If man eats his bread in the sweat of his face till he returneth unto the ground, and if women bring forth children born to suffer, it is due to the forbidden fruit. SCHILLER says,** the fabric of the world is held together by hunger and by love.*° 4 Notes. = *See my paper Some Difficult Passages in the Cuneiform Ac- _ __ count of the Deluge in the Journal of the American Oriental So- ciety, vol. xxxi., fifth page of the article, 1.2. Cf. below, n. 13. 2 Instead of hérénék, thy conception, or thy pregnancy, we must read hagigék, thy sighing; cf. Psalms, v., 2; xxxix., 4. The Greek _ Bible has tov ocrevayydv cov. Hegyénék would have a different meaning, and yégonék or caératék could not have been corrupted to hérénék. * Not shall be or will be; see my remarks in the Journal of the American Oriental Society, vol. xxv., p. 71, n. 1; vol. xxxi., fourth _ page, below, of the article cited in n. 1. The last two clauses may __ represent an observation of the narrator; cf. below, n. 36. *The preceding verse, the so-called protevangelium or proto- gospel, should be rendered: J will put enmity between thee and the woman, and between thy seed and her seed; tt (that is, her seed, the human race) will crush (lit. tread down, tread under foot, Assyr. Sépu) thy head, and thou wilt snap at its heel. There will be per- petual warfare between snakes and the human race; all human beings loathe snakes. The Messianic interpretation of this passage is unwarranted. See my Note on the Protevangelium in the Johns Hopkins University Circulars, No. 106 (June, 1893), p. 107; cf. my remarks in the Nachrichten of the Royal Society of Gottingen, April 25, 1883, p. 102; also GUNKEL, Genesis (1910), p. 20. ® See my remarks in the Journal of Biblical Literature, vol. xxi., p. 55, 1. 8; p. 66, n. 21; Haupt, Biblische Liebeslieder (Leipzig, 1907), p. 66. ®* We use this term now especially of illicit love. In Ceylon the fruit of Ervatamia dichotoma is called forbidden fruit or Eve’s apple. The forbidden fruit in the legend of the Fall of Man is, it 510 HAUPT—AN ANCIENT PROTEST [April 22, may be supposed, the orange-colored berry of the mandrake which | is still regarded as an aphrodisiac and supposed to promote concep- tion; see my paper on Jonah’s Whale in vol. xlvi. of these Proceed- ings (Philadelphia, 1907), p. 152, n. 4. In Genesis, xxx., 14, the mandrakes are called in Hebrew: diida’im, that is, love-apples. The fruit of the mandrake is quite round and of the size of a large plum; it resembles a small tomato. The largest berries have a diameter of 1% in. (nearly 4 cm.). The idea that the forbidden fruit was a fruit from which an intoxicating drink was prepared is untenable; contrast CHEYNE’s article in the eleventh edition of the Encyclopedia Britannica, vol. i., p. 168°. In the article on mandrake, vol. xvii., p. 566°, there are five misprints in the five letters of the Heb. word didaim; similarly there are two misprints in the three letters of the Arabic name for Egypt, vol. ix., p. 41°. The new edition is — marred by a great many misprints and inaccuracies, not only in Oriental words, but also in the English text. 7 Garden is often used for pudendum muilieris; see Haupt, The Book of Micah (Chicago, 1910), p. 62, n. 9. , gee 8 Eden means pleasure, delight; Heb. gan-‘edn denotes a pleasure- ground. Damascus, the earthly paradise of the Arabs, is called in — Amos, i. 5: Bét-‘edn, House of Pleasure; see my remarks in — PeIser’s Orientalistische Literaturzeitung, June, 1907, col. 306. The Greek Bible has for Heb. gan-‘edn in Genesis, iii., 23, 24: 6 Tapddeacos THs Tpudys; the Vulgate: paradisus voluptatis. The reading a garden in Eden in Genesis, ii., 8 seems to be a subsequent modification introduced by some one eis connected Heb. ‘edn with : the Babylonian edinw= Sumerian edin, desert; he may have re- : garded Paradise as an oasis in the desert like Damascus; of. 2 Pincues’ note in the Proceedings of the Society of Biblical Arche. ology, London, June 14, 1911, p. 161. Damascus means seftl in a well-watered region; the original form of the name was | masqi; see my remarks in the American Journal of Semitic guages, vol. xxvi., p. 26. r * See Deuteronomy, i., 39; Isaiah, vii., 16; cf. the translation Isaiah, in the Polychrome Bible, p. 11, 1. 25; p. 141, n. 16. *°To know good and evil has about the same meaning as phrase to cut one’s eye-teeth. 1911.] AGAINST THE CURSE ON EVE. 511 11See my paper on Midian and Sinai in the Zeitschrift der _ Deutschen Morgenlandischen Gesellschaft, vol. \xiii., p. 519, 1. 25. _ 12 Compare above, note 5. 18 The serpent symbolizes carnal desire, sexual appetite, con- cupiscence. This is the original sin which has been transmitted to all descendants of Adam; only the innocents are free from it. CoLeERIDGE (Aids to Reflexion, 1825) held that Adam’s fall was a typical experience repeated afresh in every son of Adam. Mutato nomine, de te fabula narratur ; see Hastincs’ Dictionary of the Bible, vol. i., p. 842°. In the well known Assyrian relief from Nimrid, representing the fight with the dragon, the penis of the monster is a serpent; see the plate in Gro. SmituH, The Chaldean Account of Genesis, edited by Sayce (London, 1880). The serpent in the story eof the Fall of Man is a later addition; in the original form of the _ legend Eve was the sole seductress; Eve means serpent (Heb. Hawwah—= Aram. hiwydé, snake, Arab. hdyyah). See n. 29 to my paper cited above, n. I. 1* This legend explains the institution of tattooed tribal marks ; and the institution of blood-revenge (cf. nn. 15 and 17). It illus- 4 __ trates also the superiority of nomadic animal sacrifices compared cE: with agricultural bloodless offerings (cf. n. 26). a * Kenite means descendant of Kain or Cain; Cain is the eponym ancestor of the Kenites. a 16 See Genesis, iv., 15; cf. Haupt, The Book of Canticles, p. 41; _ Biblische Liebeslieder, p. 61. ** Cf. our Uncle Sam, John Bull, Columbia, Germania, &c. A Bedouin tribe Cain (Qain) dwelt in the desert of Sinai and the neighboring districts about six centuries after Christ ; see NOLDEKE’s article on Amalek in the Eucyclopedia Biblica, col. 130. 18See Genesis, iv., 23, 24; cf. my paper on Moses’ Song of Triumph in the American Journal of Semitic Languages, vol. xx., p. 164. 19 Cf. 1 Samuel, xxvii., 10. The Kenites lived with the Amale- kites, but they were on friendly terms with the men of Judah, whereas the Amalekites were perpetually at feud with the Judahites, cf. t Samuel, xv., 6 and Judges, i., 16 (see below,'n. 21). In the 512 HAUPT—AN ANCIENT PROTEST [April 22, Book of Esther, Haman is called an Agagite, that is, a descendant — of Agag, the king of the Amalekites, who had been spared by Saul, — but was hewn in pieces before Javu by Samuel, whereas Mordecai is introduced as a descendant of the first king of Israel; see Haupt, Purim (Leipzig, 1906), p. 12, 1. 30. The Amalekites were Edomites- who had invaded southern Palestine before the Edomite ancestors of the Jews, after their exodus from Egypt, conquered the region afterwards known as Judah (see n. 23). In Numbers, xxiv., 20 Amalek is called the first (that is, oldest) of the nations. The Amalekites, however, had intermarried with other (non-Edomite) tribes ; in Genesis, xxxvi., 12, therefore, Amalek is introduced as a son of Esau’s first-born, Eliphaz, by a concubine, just as the sons of Jacob’s concubines, Bilhah and Zilpah, were tribes with foreign elements; see my paper on Leah and Rachel in the Zeitschrift fiir die alttestamentliche Wissenschaft, vol. xxix., p. 285. The identi- fication of Amalek with the cuneiform Meluha (Orientalistische Literaturzeitung, June, 1909) is untenable. According to 1 Chron- icles, ii., 55, the Rechabites (cf. Jeremiah, xxxv.; 2 Kings, x., 15, 23) were descendants of the Kenites; but this can hardly be correct. The Rechabites resembled the ancient Kenites in that they were ardent worshipers of JuvH, and that they continued to live in tents after the men of Judah (see n. 23) had settled in Canaan. 20See p. 360 of my paper on The Burning Bush and The Origin of Judaism in vol. xlviii. (No. 193) of these Proceedings (Philadelphia, 1909) and my paper on Midian and Sinai (cited above, n. II), p. 506, 1. 12; p. 512, Il. 15 and 33; p. 513, |. 2. In Genesis, iv., 17 we read that Cain built a city. "1Tn Todesen: iv., If the words mib-béné hébdb Maséh are secondary gloss (or variant) to mig-Qain, and hétén is a tert gloss to hébdb. The original text of Judges, i., 16 seems to h been: wé-Qain ‘alah me-‘ir hat-témarim et-Yéhiidah midbdr ‘A wai-yélek wai-yéseb et--Amaléq, Cain went up with Judah from Palm City to the wilderness of Arad, and went and lived Amalek. The words béné ... hétén Méséh and Yéhidah ban-négeb are glosses. See the translation of Judges, in the F chrome Bible, pp. 8 and 2; also p. 49, n. 15; p. 62, 1. 55; ef. 1gtt.] AGAINST THE CURSE ON EVE. 513 paper on Hobab= father-in-law in the Orientalistische Literatur- seitung, April, 1909, col. 164. 22For JHVH see p. 355, n. 2 and p. 357 of my paper The Burning Bush, cited above, n. 20. *8 Judah is the name of the worshipers of JHvH, who were united under the leadership of David about 1000 8. c. David was not an Israelite, but an Edomite. See n. 18 to my paper The Aryan Ancestry of Jesus in The Open Court, Chicago, April, 1909; cf. p. 358 of my paper The Burning Bush, cited above, n. 20, and my paper on Midian and Sinai (see above, n. 11), p. 506, 1. 2; p. 507, 1. 36; also Erpt’s remarks in Orientalistische Literaturzeitung, July, I9II, col. 298, 1. 19. For the sheepmen of Judah see p. 284, n. 5 of my paper on Leah and Rachel, cited above, n. 19; cf. my paper on the five Assyrian stems /Ja’u in the Journal of the American Oriental Society, vol. xxxi. 24In Syriac, habbalta (or hébalta, ébadita) means herd, drove, especially of camels; cf. Obil, the name of David’s keeper of camels, I Chronicles, xxvii., 30 (see Encyclopedia Biblica, col. 6). Hebel, the Heb. form of Abel, may be connected with hébdil, to lead. The name of Jabal, the father of such as dwell in tents and of such as have cattle, Genesis, iv., 20, may be derived from the same root; cf. Hastincs’ Dictionary of the Bible, vol. i., p. 5*. The original form of Jabal seems to have been Jébil; the Greek Bible has IwferX (and IQOBHA for IOQBHA). Hebel may be a subsequent modification of Hobil, due to a popular etymology combining the name with Heb. hébel (for hdabil) breath, transitoriness; see below, n. 27. For Jébil = Hébil cf. my remarks on Jair—= Me’ir, p. 513, 1. 24 of my paper cited above, n. 11. The name Moses, Heb. Méséh, may have had originally an ‘Ain at the end so that it would be equivalent to Joshua ; see /. c., 1. 26, and for the vanishing of the final laryngeal, Op. cit., p. 522, 1. 47; also Haupt, The Book of Esther (Chicago, 1908), p. 74, 1. 14. *° Cf. p. 528, 1. 38 of my paper cited above, n. II. . 7° In Canaan a bloodless offering smacked of Canaanite heathen- a ism; cf. the remarks on p. 44 of the translation of Judges in the Polychrome Bible. SKINNER says on p. 106 of his new commentary 514 HAUPT—AN ANCIENT PROTEST [April 22, on Genesis (1910): It is quite conceivable that in the early days of the settlement in Canaan the view was maintained among the Hebrews that the animal offerings of their nomadic religion were superior to the vegetable offerings made to the Canaanite Baals. - 27 Cain may be connected with the Ethiopic tagénya which means __ to till the ground; cf. the Pachomian rules in DitLMANN’s Ethiopic chrestomathy, p. 60, 1. 4. Tagdnya means also to worship God; cf. Arab. gdnata (quntit) and Lat. colere. Stems tertie y and medie y often interchange; cf. Ethiopic gandya, to sing, and Arab. qdinah, songstress, Heb. ginah, elegy. For Ethiopic qgéntiy, servant, we have in Arabic: gain, plur. giyén. In Arabic, gain means also — smith, metal-worker, Syr. qainéya. Some scholars, therefore, be- lieve that the Kenites were a tribe of wandering smiths. SAyce says (in Hastincs’ Dictionary of the Bible, vol. ii., p. 834”) that the Kenites resembled the gipsies of modern Europe as well as tne traveling tinkers or blacksmiths of the Middle Ages. SKINNER states (on p. 113 of his commentary on Genesis) that there are some low-caste tribes among the Arabs, who live partly by hunting, partly by coarse smith-work and other gipsy labor in the Arab encamp- ments; they are forbidden to be cattle-keepers and are excluded from intermarriage with the regular Bedouins, though on friendly : terms with them; they are the only tribes of the Arabian desert that are free to travel where they will, ranging practically over the: whole peninsula from Syria to Yemen. ; The legend of Cain and Abel may have connected the name Cain with the allied stem ginné, to be jealous, envious, passionate, just as the name Abel (see n. 24) was combined with habl (for hdbil) breath, transitoriness. The saying of Ecclesiastes, Vanity of v ties (that is, How utterly transitory is everything!) is in Hebi habél habalim; see Haupt, Koheleth (Leipzig, 1905), p. 1; . clesiastes (Baltimore, 1905), p. 34, n. 2. 8 Lit. with JuvH. Also we use with in the sense of like, alogously to. SHAKESPEARE says, As if with Circe she would ch my shape. Cf. the Critical Notes on the Heb. text of in the Polychrome Bible, p. 118. My interpretation of this cult passage has been adopted by CHEYNE, Encyclopedia B rgtt.] AGAINST THE CURSE ON EVE. 515 | col. 619, n. 3: J have created a man even as Yahweh; but we must not substitute /é-‘ummdt. Nor can we read is 6t Yahweh, the man of the mark (cf. above, n. 14) of JuvH, or i§ et‘awwéh, a man _ whom I desire. The prediction of the serpent that Eve and her - husband would be like God, if they ate of the forbidden fruit, im- __. Plied that they would be able to create new human beings, and this would make the race of Adam immortal. C f. the fourth page, below, of my paper cited in n. 1. a 2° This clause is preserved in the Samaritan Pentateuch and in the Ancient Versions. The Vulgate has Egrediamur foras. 3 %° Cf. my explanation of the story of Judah and Tamar in n. 26 __ to my paper cited above, n. 11. a 31 Cf. G. J. SpuRRELL, Notes on the Text of the Book of Genesis (Oxford, 1896), p. 2. 3 82 Contrast the blood of righteous Abel in Matt. xxiii., 35; see also Hebrews, xi., 4; 1 John, iii., 12. 4 83 We must read: Halé, im tétib, essa panéka; wé-im hété atta, = 16 eqqah qorbanéka. In the received text hété atté is mispointed and misplaced: it appears as hattat between lap-pétah and rébé¢ which are corrupted from 16 eqgadh qorbankd. The Greek Bible read /é-nattéh instead of lap-pétah, and rébdc for rébéc. The read- ing of the received text, im 16 tétib, if thou doest not well, is a later substitution for the original im hété atta, if thou art a sinner. We might read also 16 ercéh minhatéka, but this could not have been corrupted to lap-pétah rébéc. In 16 eqqah qorbankdé one of the Alephs in 16 egqgah was omitted; q of gorbankd dropped out after the final h of eqqaéh, and n was omitted after the b of gorbankd; the letters for m and b are similar in Hebrew; for g=—hA see Crit. Notes on Kings, in the Polychrome Bible, p. 187, 1. 20. For eqqah gorbanké ci. Psalm vi., 10: Yahweh igqgah tépillati, Javu will re- ceive my prayer, and Assyr. telégi téméqsu and leqat unnéni, &c. (see Detitzscu’s Assyr. Handwérterbuch, p. 384°, d). GUNKEL’s reconstruction of the text (in Die Schriften des Alten Testaments tibersetzt von GRESSMANN, GUNKEL, &c., part 5, Gottingen, 1910, -p. 69) does not commend itself. %* Cf. Ephesians, v., 22; Colossians, iii., 18; Titus, ii., 5; 1 Peter, 516 HAUPT—AN ANCIENT PROTEST [April 2a, 8° Like Deborah, Esther, Judith, &c. eu *° Cf. the observation of the narrator (see n. 3) in Genesis, 3 it, * 24: Therefore a man leaves his parents and clings to his wife. The rendering shall leave (Matt., xix., 5; Mark, x., 7) is incorrect; it is. not a prophecy, nor is it an old saying dating from remote times when the husband went to the tent of the wife and joined her clan, although it is noteworthy that Eve, not Adam, names the child in Genesis, iv., 1 (cf. above, n. 28). We may compare the line in the - Biblical love-songs (Canticles, viii., 7) where the poet says of Love: If one should resign for it all his possessions, could any man therefore contemn him? This means, from the Oriental point of view: If a man should sacrifice all his possessions to buy a beautiful girl; see Haupr, Biblische Liebeslieder, p. 111. Tuomas Drxon, Jr., says in his novel The Leopard’s Spots of Simon Legree: They say he used to haunt the New Orleans slave-markets when he was young and owned his Red River farm, occasionally spending his last dollar to buy a handsome negro girl who took his fancy. r 87 Cf, the remarks in n. ** to my paper Isaiah’s Parable of the Vineyard in the American Journal of Semitic Languages, vol. xix., p. 194. 88 See Haupt, The Book of Canticles, p. 5; Biblische Liebeslieder Pp. 4. 8° SCHILLER says in the last stanza of his poem Die Weltweisen ; Doch weil, was ein Professor spricht, Nicht gleich zu allen dringet, Es tibt Natur die Mutterpflicht Und sorgt, dass nie die Kette bricht Und dass der Reif nie springet. Einstweilen, bis den Bau der Welt Philosophie zusammenhilt, Erhalt sie das Getriebe Durch Hunger und durch Liebe. 40 As a striking illustration of the manner in which some of ow leading newspapers occasionally mislead their readers, I will join here the “report” of my paper, which appeared in The Pr 1911.) AGAINST THE CURSE ON EVE. 517 _ Philadelphia, April 23, 1911, under the caption Education and Race _ Suicide: “ Declaring that race suicide is due to an increase in intelli- gence, and theorizing that the human emotions become fewer as human beings become better educated, Dr. Paut Haupt, professor of Semitic languages at Johns Hopkins University, spoke at the "session yesterday morning. Contrary to the hope of many members of the Society, Dr. Haupt advanced none of his religious opinions in the course of his address. He spoke upon ‘An Ancient Protest against the Curse on Eve’ and confined himself wholly to observa- ___ tions on race suicide.”—The abstract which I had placed at the dis- _ posal of the press was printed in the Philadelphia Ledger, the North _ American, &c., April 23, IQII. PROCEEDINGS AMERICAN PHILOSOPHICAL SOCIETY HELD AT PHILADELPHIA FOR PROMOTING USEFUL KNOWLEDGE VoL. L OcToBER—I)ECEMBER, 1911 No. 202 THE FORMATION OF COAL BEDS. IL* By JOHN J. STEVENSON. SoME ELEMENTARY PROBLEMS. (Read November 3, 1911.) It is necessary, first of all, to consider some matters to which ref- erence is made by every one who has endeavored to explain the formation of coal beds. Too many seem to have been content with acceptance of current statements respecting the apparently common- place phenomena and too few have thought essential a careful study of work done in recent years. The indefinite and often contradic- tory assertions contained in discussions, published within the last : decade, compelled the writer to study the available literature and, as far as possible, to make examination of the phenomena in place. This study has led him to reject some of his cherished beliefs while it has confirmed others. At the same time, it has increased his respect for the problem, which he has undertaken to solve. The topics to be considered in this portion of the work are: The effect of floods upon a cover of vegetation, The phenomena of peat deposits, The buried forests. * Part I appeared in these Proceedings, Vol. L., pp. 1-116. PROC. AMER, PHIL. SOC., L. 202 II, PRINTED NOV. I5, IQII. 519 520 STEVENSON—FORMATION OF COAL BEDS, _ [November 3, Each of these topics will be considered only in its relation to the main problem. Tue EFrrect oF FLoops UPON A COVER OF VEGETATION. A torrent dashing through a narrow gorge is the poet’s symbol of resistless force; a great river in flood, bearing on its surface } houses, trees and other floating materials in prodigious quantity, | seems possessed of almost illimitable power for destruction. These conditions, so familiar to all, have led to the conception that, in the eroding and transporting power of streams in flood, one can find explanation for the origin not only of sandstone and other inorganic deposits but also for that of coal and lignite beds interstratified with them. So much importance has been assigned to this explanation by several authors, that the phenomena must be considered in detail. The Work of. Torrents—The torrent in full flood is an interest- ing spectacle but its importance in this connection is confined chiefly to its bearing on the origin of inorganic sediments. ; Long ago De Luc’ described deposits made by torrents as resem- bling a much flattened loaf of sugar and he employed the term “cone” to distinguish them from the talus at foot of cliffs. When a stream ceases the cutting down of its bed, the formation of a cone ends and vegetation takes possession of the surface, even of steep slopes alongside of the stream. Eventually the surface is covered with a thin coat of soil and men settle upon it. Such a cone was seen by De Luc on the right bank of the river Arc, en route from Mount Cenis into Italy. It extends from Aigue-belle to Saint-Jean de Maurienne and the line around its base is nearly three miles. It is high against the rocky wall and the surface is comparatively steep. Its history is the same with that of many others. The stream cuts a channel in the cone; occasionally, during a great flood, the banks are overflowed, but the injury is not enough to drive away the inhabi- tants. A fall from the walls of the gorge forms a talus against which the stream flows; water finds its way through the sands and *J. A. DeLuc, “ Lettres physiques et morales sur l’histoire de la terre et de ’homme,” Vol. II., 1780, pp. 67-68. 118 191t.] STEVENSON—FORMATION OF COAL BEDS 521 renders the mass almost jelly-like. A winter of heavy snows, fol- lowed by an unusually warm spring, leads to abrupt swelling of the water and the soft mass is swept out over the surface of the cone, flowing like a flood of lava. Such were the conditions here. A village, la Chapelle, had been built; in 1752, the torrent washed out a vast mass of rock and débris, which covered everything to a depth of 15 to 20 feet, completely burying the hamlet. In a broad valley, the effects of the torrent’s work are confined to the cone, but conditions are different where the valley is narrow. A few years prior to the disaster at la Chapelle, another cone was formed in the Arc valley between Saint-Andre and Saint-Michel. 4 The river was dammed and a lake was the result. The inhabitants of a drowned hamlet endeavored to make a canal through the dam, but the great fragments foiled them and the river was left to make its own canal. De Luc visited the locality several times. When he last saw it, the river had filled the lake with débris, through which it flowed gently, while above and below it foamed amid masses of rock. De Luc notes with some interest that on the other side of the valley, opposite la Chapelle, a torrent was still at work, building its cone. Each spring saw the surface covered by a flood, bringing down new material. Yet the trees, which had gained place and had sent their roots down amid the rock-fragments, maintained them- selves against the rushing waters, laden with rocks and débris, and inhabitants of the region gathered fagots there. Similar phenomena are familiar to all students of mountainous areas. Dejection cones, some of them enormous, are numerous along the Rhone from Viége to Martigny. The flood additions are distinct, several of the cones being deposited by streams with fall _ of more than 250 feet per mile. Except where new channels had been cut, the trees growing on the surface of the cones are appar- ently uninjured, though in some cases they have been bent by blocks of rock, whose advance they had checked. The valley of the Adige _ between Botzen and Rovereto presents many illustrations of similar _ type; in the mountain areas of the western United States such con- _ ditions are merely commonplace. 119 522 STEVENSON—FORMATION OF GOAL BEDS. [November 3 McGee’s? observations in the Sonoran district of Arizona and Mexico are equally instructive, though dealing with a somewhat dif- ferent type of phenomena. That district is within the arid region and the valleys are sand wastes, with shallow channel ways which are dry during all but, perhaps, five days in each year. The moun- tains are scarred by barrancas or stream-worn valleys, which end abruptly at the plain. The streams during the greater part of the year are mere threads; but, during thunder gusts or cloudbursts, the old channels are filled and new ones digged, though the flow may last but a few minutes or at the most a few hours. The stream gathers loosened rock masses in the mountains, hurls them down slopes into the barrancas, dashing them to fragments and carrying the débris to the edge of the plain. There the coarser materials are dropped, but the finer stuff is transported beyond as a sheetflood until the water disappears by absorption or evaporation. The incli- nation of the channel ways is not far from 300 feet per mile in the mountains, decreases to 200 feet at the edge of the plain and to 50 feet at the end of the torrential area, several miles away. The flooding of the plain has, in some cases, a width of ten or more miles. McGee saw one of the floods in 1894. It came abruptly, a mass of water thick with sand, foaming and loaded with twigs and dead leaves. It advanced at first with “racehorse speed” but the velocity diminished quickly owing to evaporation and the flood died out in irregular lobes. The depth was not more than 18 inches on the lower border, where the width of the muddy flood was about half a mile. The noteworthy feature in this connection is that the mass of slime, moving with so great speed, had no injurious effect upon the clumps of mesquite bushes, scattered here and there over the slope. When the flowing mud reached those clumps, its speed was — checked momentarily ; its course was diverted and it moved along- side at twice to thrice the ordinary rate. After the flood was over, the most striking observable effect was the accumulation of twigs and branches against the clumps of shrubbery and other obstacles. Russell’s notes on the Yahtse river of Alaska, to be considered — *W J McGee, “Sheetflood Erosion,” Bull. Geol. Soc. America, Vol. VIII., 1897, pp. 87-112. 120 1911.] STEVENSON—FORMATION OF COAL BEDS. 523 in another connection, are equally in place here. That river, issuing from the Malaspina glacier as a swift flood, 100 feet wide and 15 to 20 feet deep, has invaded a forest and surrounded the trees with sands and gravel to a depth of many feet. That current, rapid enough to carry a great load of very coarse material, was not strong enough to uproot the trees and evidently it could not break off the trunks until decay was well advanced. Russell’s photograph shows the conditions distinctly. Smyth? has described conditions in the Adirondack mountains of New York, which are of no little importance in this connection and they will be utilized in the sequel, as the character of the region is very like that imagined for some limnic basins in Europe. Many small lakes exist in that region, varying from a few rods to three or four miles in the longer diameter. They are of post-glacial origin and, in many cases, are surrounded by high hills of metamorphic rock, whence streams with rapid fall flow in comparatively narrow valleys. Big Rock lake is typical. It is a mile anda half long, three fourths of a mile wide and is fed by streams which carry much sedi- ment and by their deposits are changing the outline of the lake. The new area is a level marshy meadow, about a foot above the water, covered with a heavy growth of grass and carrying some small bal- sams and tamaracks. The lakes show every stage from pond to meadow and one of them has been changed throughout into meadow, through which its stream meanders on the way to Big Rock lake, one mile away. After heavy rains, water flows over the meadows to a depth of a foot or more, leaving a sediment of varying thickness; but the torrential streams feeding the lake, though flowing through gorges, whose steep walls are more or less densely timbered, rarely bring down trees or other vegetation. Torrents carrying no débris do as little injury to vegetation as to the rocks over which they flow. The writer has recognized this many times in the Rocky and other mountain areas of the western United States. Clear creek in Colorado, formed by the union of streams from Gray, Torrey and other high peaks of the Front range, *C. H. Smyth, “ Lake Filling in the Adirondack Region,” Amer. Geolo- gist, Vol. XI., 1893, pp. 85-90. 121 524 STEVENSON—FORMATION OF COAL BEDS. _ [November 3, is subject to abrupt rise in the early summer, when hot days cause rapid melting of the snow on those peaks. The upper streams pass over hard rocks and among huge masses which choke the channels, so that, below the union, the creek, following a rocky, mostly narrow gorge and with rapid fall, carries even in great flood comparatively little sand or silt. A rise of 10 feet within 3 or 4 hours is of fre- quent occurrence. The low narrow “bottom,” though so often overflowed by rapidly moving water, is grassy and bears some shrubs. In some places, where the canyon widens and the stream, under ordi- nary conditions, flows less rapidly, one finds petty islands of detritus covered by shrubs, some of them more than 15 feet high. Clearly these plants are torrent-proof, they have maintained their place from birth. Similar conditions were observed on many other streams of like character. A Cover of Vegetation Protects against Erosion—Within any given district of moderate size, some streams are turbid while others are limpid even in flood time. The water may be limpid or turbid in different portions of the same stream. As the rainfall is the same throughout there must be local causes for the variation. The process of erosion begins with the impact of a raindrop; but that impact is ineffective if the drop fall on more or less elastic material. Thus it is that a cover of decayed vegetable matter, a coating of humus, protects a slope against the erosive power of rain- fall; it protects existing vegetation from removal and it may enable plants to regain hold on spaces bared by fire or other agents, even though the slope be abrupt. ey Ashe* studied an area which has great variety of soils, as the section extends from the Archean to the Quaternary. Soils from partly metamorphosed sandy shales or from mica schist offer little resistance to erosion; in some cases they cannot nourish a sod but each supports its own type of trees. Denuded of forest, the surface — is gashed quickly by rains, but when forested it loses little. The accumulation of litter or humus within the Potomac area is small *“W. W. Ashe, “Relations of Soils and Forest Cover to Quality and Quantity of Surface Water in the Potomac Basin,” U. S. Geol. Survey, Water Supply and Irrigation Paper, 192, 1907, pp. 290-335. 122 1911.] STEVENSON—FORMATION OF COAL BEDS. 525 because of climatic conditions, but it suffices to protect these soils on all except the steepest slopes. Limestone soils, occupying much of the area, are very apt to “wash” when under cultivation, but where covered with forest even the steepest slopes retain their cover of humus and the run-off water is never turbid. Sandstone soils vary much in resistance, when bared, but where they are protected by a thin cover of humus the waste is insignificant. The water of small streams flowing from forest mountain-sheds is clear and pure. The great resistance offered by humus is apparent from the figures given by Ashe. Pines growing on poor soils, rarely yield more than 2 inches; yet this protects all except the steepest slopes. Chestnut oak and white oak give but 3 inches; they too grow on poor soils, which, when exposed, are torn away rapidly. Other woods give from 5 to 6 inches of litter, which is so absorbent that for several days after a rain one can squeeze water from it as from a sponge. Ashe’s observations show that this vegetable litter, in the semi- decomposed condition, is so interwoven that it not only protects the underlying soil but also itself resists removal as does a well-rooted sod. The streams coming from the humus covered area are free from vegetable matter, aside from occasional twigs and, at times, some soluble matters leached from the humus. The White mountains of New Hampshire illustrate well the incompetence of rainfall to remove living vegetation. The rock in that region is mostly granite and the soil, formed since the glacial period, is very largely humus. The slopes are abrupt and the walls of gorges frequently show more than 50 degrees; but most of the area below timber line has a dense cover of vegetation, largely spruce. Yet rains have always been frequent and many times almost deluge- like. The covering of humus is undisturbed by those rains; even where lumbermen have cut away the forest and left their litter and the humus exposed to the fury of storms, one finds little evidence of removal. Cloudbursts or extraordinary downpours of rain have occurred many times within this area. C. H. Hitchcock has described the flood in the Flume at Franconia, which washed away the great boulder which had been dropped by the retreating ice and had re- 123 526 STEVENSON—FORMATION OF COAL BEDS, _ [November 3, mained suspended in the Flume. That huge mass has never been found. Yet, aside from a landslide or two, the terrific rainfall left the vegetation on the steep slopes unscarred. In June of 1903, a cloudburst of unusual severity broke on the northern part of the White mountains. The roads were gullied and rendered impassable ; bridges, large and small, were swept away throughout the region as the streams were filled beyond the high water mark of spring freshets ; sheets of water poured down naked rock surfaces in many portions of the abrupt spaces and landslides of limited extent were produced where the slope was covered with loose material. But this vast flood of water did practically no injury to the forest-covered slopes; even débris left on the mountain side by tree-choppers was almost undisturbed.® But the most noteworthy evidence in this region is found on the areas which have been burned over. When a forest fire destroys the soil near the top of a divide or on a very abrupt slope, the residue is removed quickly by rain and the granite is exposed. But if the organic matter has not been destroyed, the soil resists ordinary rains even on steep slopes. If drenching rains be delayed for a few weeks, the surface gains a cover of fireweed (Erechtites hieraci- folia) and rain is powerless. This growth is succeeded in the fol- lowing season by a dense cover of raspberry, fern and other plants, among which a cherry takes root to become the characteristic form in the third season. Birches, maples and poplars are prominent during the next season and within five years the spruces make their appearance. If drenching rains follow quickly after a forest fire, the process of restoration is merely retarded, it is not prevented. Glenn® studied the problem throughout the southern Appa- lachians, an area of 400 by 150 miles, and his studies were extended to another area farther north, 200 by 50 miles. The examination was continued westward for a long distance down the Tennessee — river, so that the investigation embraced every type from the bold *Communicated by C. A. Snell of Malden, Mass., who examined the ; whole area within four days after the disaster occurred. e *L. C. Glenn, “Denudation and Erosion in the Southern Appalachian — Region,” U. S. Geol. Survey, Professional Paper, 72, 1911, pp. 15-18, 23, 24, 59, 93, 96, 99. 124 191t.} STEVENSON—FORMATION OF COAL BEDS. 527 mountains, cut by canyon-like gorges, to broad river valleys with _ wide bottoms in which the streams mearder. This study concerns also some matters to be considered hereafter, but they are included here for convenience of reference. Glenn asserts that, in forested areas, erosion is at its minimum, for the soil is protected by the litter from impact of raindrops. As drops move down the slope, they are checked by the litter or are absorbed by it, and the rainfall moves so slowly through the mass that for hours after rainfall, the cover is full of water. Even such gullies as were seen have their bottoms covered with litter and plants, showing that the erosion, by which they have been produced, is very slow. Streams flowing from the forested regions rise grad- ually during heavy rains and fall to normal more gradually, because the litter retards flow. Such streams, even when highest, are, as a rule, but slightly discolored and that discoloration is caused in great part by macerated fragments of leaves and decaying plants, for they carry little mineral matter in suspension. Some of them remain wholly clear even when swollen to far beyond their normal stage. But removal of the forest brings about an abrupt change. The pro- tective efficiency of even a root-matted soil is evident, for when a tree is uprooted or a road is cut, so as to break the continuity, erosion begins at once. The contrast between forested and denuded areas is so striking that no argument is needed. Grass-covered slopes may be destroyed by breaks made when a cow crosses them after pro- longed rainfall, but erosion can be checked by covering the surface with litter, held in place by brush; weeds and bushes spring up quickly. The writer adds his testimony in confirmation of these observations by Glenn, for he has seen many thousands of acres of __ cleared land, which had been abandoned after a few years of culti- __ vation and which now are covered by a dense growth of hard wood— and this on the steep slopes of the Virginian Appalachians. Glenn’s volume is a commentary on the protective influence of vegetation and on its resistance to erosion. The changes in the rivers since the removal of forests from their headwaters, the increased _ erosion, the increased destructiveness of floods owing to the greater load of inorganic matter are set forth clearly on almost every page. 125 528 STEVENSON—FORMATION OF COAL BEDS, [November 3, This increased load has led the formerly almost limpid streams to 4 aggrade their lower reaches, to convert once fertile bottoms into 4 marshes or to cover them with sand and gravel. This aggrading, in | many instances, forced the streams to cut new channels or a network of channels through the plain. But this lateral cutting is prevented now by planting willows, aspens, balm of Gilead and other rapidly-growing plants on the river banks down to the water’s edge. The silt-laden flood does little injury to these plants and the plain itself is injured only by drowning-of crops when floods come in the growing season. Glenn contrasts conditions in the Coosa and Chattahoochee basins. The former river rises in an area still forested and its waters in flood carry little inorganic matter to cause destruction; but forests have been removed from the headwaters of the latter, floods are more frequent and the accumulation of sand is very great—a condition wholly unknown one hundred years ago. Rixon’ gives similar testimony to the ability of humus to protect itself as well as underlying material from erosion. ‘The litter and underbrush among the alpine timber are very heavy, having accu- mulated for ages. One class of timber, having reached maturity, decays, dies and falls, only to be supplanted by another growth, which in time follows its predecessor.” This is in a region where rainfalls are infrequent but extremely violent. Tuomey® studied the influence of forests on surface run-off. His observations were made on four small catchment areas in south- ern California, where wet and dry seasons are well-marked. In December, 1899, the rainfall was 18 inches. This was at the close — of the long dry season, when litter and soil alike were desiccated and each absorbed a large part of the rainfall. The percentage of run-off is given in the first column: ‘ 1. Forested 3 35 2. Forested 6 33 3. Forested 6 43 4. Non-forested 40 95 *T. F. Rixon, “Forest Conditions in the Gila River Forest Reserva- tion,” U. S. Geol. Survey, Prof. Paper, 39, 1905, p. 18. Be! *J. W. Tuomey, “ The Relation of Forests to Stream Flow,” Year Be 0. of U. S. Dep’t of Agriculture, 1903, pp. 279-288. 126 tgtt.] STEVENSON—FORMATION OF COAL BEDS. 529 But in January, February and March, when the absorbed moisture in the litter was great, the contrast still remained, as appears from the second column, where the run-off from the forested areas aver- ages only three eighths while that from the non-forested area was nineteen twentieths. The great dunes of Bermuda have their advance checked by vege- tation. A network of vines creeps over the surface and breaks the force of the wind. Clumps of grass take root in the open spaces and, within a brief period, the heavy rains can do little more than to move the sand a few inches to be piled against the obstacles. Vegetation holds its place on the loose materials until, at length, a dense growth of oleander and cedar render the deluge-like rains wholly ineffective. The same condition exists along railroads within the drift covered areas of the United States. Many of the through cuts are in drift gravels, with no trace of consolidation, yet their walls show the steady advance of plants in spite of rain and the steep slope. The resistance which vegetation offers to erosion is manifest on a grand scale in the tropics, where growth is luxuriant and the rain- _ fall extreme. The writer has had opportunity to examine at close _ range fully 200 miles of the Venezuelan and Colombian coast, much of Trinidad, about 50 miles of the Jamaican coast, as well as much of the Pacific coast of Central America. There are some localities where the rock is not consolidated and vegetation cannot maintain itself. Such as gains rooting toward the close of the wet season is killed during the dry season and rain finds only the unprotected sur- _ face on which to act. But such areas are of limited extent. The slopes along the coast are usually quite steep and the stratified rocks commonly dip at a high angle. Landslides, owing to this structure, are not rare and they leave a scar on the surface which persists for years; but aside from those merely temporary interruptions, vege- tation is practically continuous on even the steepest slopes. The Jamaican conditions are especially instructive. Where vegetation was destroyed by fire in some extensive areas, Guinea grass has _ taken possession of even the steepest slopes, giving great spaces of bright green, which are notable features of the scenery—and this in 127 530 STEVENSON—FORMATION OF COAL BEDS, _ [November 3, spite of the excessive rainfall. During November of 1909, the rain- fall in the mountains of Jamaica was of unprecedented volume, there being at one locality 120 inches in eight days, while in others there were 20 to 30 inches within one day. Banana plantations, with unprotected soil, were washed down the hills and the plants became projectiles with which the flood destroyed vegetation on the low- land; but the forest remained almost uninjured and the litter cover- ing the surface around the trees was practically undisturbed. Where the land was protected by trees, damage was confined to gullies digged by fallen trunks pushed forward by the water. These gullies widened in soft materials and trees, tumbled into the torrent, were carried to the lowland, where they were deposited, péle méle, with mineral matter on the cultivated land. Nowhere in the whole area was there evidence that rainfall did any serious injury to ar forests or the forest litter. Cornet’s® observations in the Congo region are to the same effect. Where the dry season is prolonged, plants are practically dried by desiccation, so that the first rains do great damage; in such localities, this is so serious that vegetation cannot re-establish itself. But, near the equator where rains are almost constant, the forest quickly reoccupies areas which man has cleared. Even in regions with a long dry season, the bottoms of the valleys, owing to dampness, be- come forested and that puts an end to the action of the wild waters —it may cause even diversion of streams. Clearing of forests lays the humus open and it is carried off to be spread elsewhere, there to enrich the soil. This actually occurs in many valleys, giving what Dupont has termed terrenoir; but in the broad alluvial valleys, where humidity prevails throughout the year, vegetable detritus accumu lates on the surface and gives a formation of humus sur place. It matters not where one looks, the conditions are the same Geikie,!° familiar with the Highlands of Scotland, where bogs in the heath stage cover great areas, says that the surface of a district pr *J. Cornet, “Les depots superficiels et l’erosion continentale dans | bassin du Congo,” Bull. Soc. Belge de Géologie, Vol. X., 1897, Mem., pp. 44-116. ” A. Geikie, “ Textbook of Geology,” 3d Ed., London, 1893, p. 475. 128 191t.] STEVENSON—FORMATION OF COAL BEDS. 531 tected by a thin layer of turf, is denuded with extreme slowness except along the lines of its watercourses. Indeed, the evidence is wholly clear to every one who has crossed Scotland by way of the _ Caledonian canal, which utilizes a chain of small lakes, fed by streams rising in the Highlands and descending with rapid fall. The lakes ___ are not turbid, they rarely show blocks or chunks of peat where the streams enter, the only evidence of vegetable matter being coloration of the water by salts of organic acids leached from the peat. The same condition exists elsewhere in Scottish lakes. Many years ago, Marsh*! wrote elaborately respecting the pro- tective influence of vegetation and the disastrous consequences fol- _ lowing removal of forests. He recognizes that humus can absorb almost twice its weight of water, which it surrenders to the under- ___ lying soil and becomes ready to absorb more. Twigs, stems, fallen 4 _ trunks and the rest oppose the rush of water and break into small __ streams any larger ones formed by union of petty rivulets. He cites many works, reporting official as well as private studies—all record- ing the same results. 3 In the French Department of Lozére, which was among those most seriously injured by the inundation of 1866—caused by rains, not by melting snow—it was remarked everywhere that “grounds covered with wood sustained no damage even on the steepest slopes, _ while in cleared and cultivated fields the very soil was washed away and the rocks laid bare by the pouring rain.” Marsh cites Foster, _ who describes an area with slope of 45 degrees, which consisted of _ three sections: one, luxuriantly wooded, with oak and beech from summit to base; a second, completely cleared; a third, cleared in the upper part but retaining a wooded belt for one fourth of the height from the bottom. The surface rose 1,300 to 1,800 feet above the stream flowing at the foot. The first section was wholly un- scarred by the rains; the second showed three ravines, each increas- ing in width from summit to base; while the third, of same superficial extent, had four ravines widening from the summit to the wooded belt, in which they became narrower and soon disappeared. He *G. P. Marsh, “ The Earth as Modified by Human Action,” New York, 1874, pp. 232-238. 129 532 STEVENSON—FORMATION OF COAL BEDS. _ (November 3, — refers to his own observations that, in primitive regions, running streams are generally fringed with trees and that even now in for- ested areas of the United States trees come almost to the water’s edge, so that the banks are but slightly abraded by the current. He cites Doni respecting the Sestajone and Lima, two streams rising in the Tuscan Appenines and flowing into the Serchio. In rainy weather the volume of the former is only about half as much as that of the latter and its water limpid; whereas the water of the latter is turbid, muddy. The drainage areas are almost equal, but the Sestajone winds down between banks clad with firs and beeches, while the Lima flows through a cultivated, treeless valley. The writer had opportunity in 1910 to observe the effect of heavy rainfall on the steep wooded slopes in central France, where the rocks are resistant gneisses and granites—a condition much like that of the White mountains. The rainfall during that summer was not merely in excess, it was extraordinary. The showers came suddenly and often resembled the cloudbursts of mountain areas within the United States. In many parts of the area, the gorges are deep, with © walls often exceeding 35 degrees, at times exceeding 45 degrees. Many gorges have densely wooded walls; many others have a some- | what scanty growth, scattered over the rocky slope with plants grow- 5 ing here and there in decomposed material occupying clefts or accu- mulated behind projecting craggy points. During some showers, the — water ran off exposed places not in rills but often in broad continuous sheets and the streams were converted into roaring torrents. More than once, after one of these almost cataclysmic rains, the writ : passed through some of the gorges and was surprised to find that, apparently, no injury had been done to vegetation on even the steepest slopes. Tender plants, growing in handfuls of loose material on projections, seemed to be unharmed. The streams were follow: for many miles, but they had received only rare stems of trees from undermined banks and the eddies showed no accumulation of plat material. Trees, lining the streams and in many cases growin down to almost the low water line, gave no evidence of having been subjected to the force of a dashing torrent. The conditi differ from those, with which every one is familiar, only in that the 130 191t.] STEVENSON—FORMATION OF COAL BEDS. 533 are on a larger scale. The almost vertical walls of railroad cuts _ through hard rock are adorned by small plants growing in clefts or even by trees in similar position. These have grown in spite of rains, which threatened to wash away the little soil on which they depend. But the rains are as powerless against plants in railroad cuts as they are against plants growing in like conditions on the walls of Alpine gorges or of canyons in the Sierra and the Rockies. River Floods——The floods of rivers have much in common with those of torrents, for most rivers are more or less torrential in their upper reaches ; but there are noteworthy differences, aside from those due to volume. The topographical conditions required for torrents are wholly unlike those amid which great rivers exist. Torrents flow, for the most part, in narrow valleys with here and there some wider portions in which are insignificant floodplains; but rivers usually flow in broader valleys, have-less rapid descent and are bor- dered frequently by extensive floodplains. Rivers entering the At- lantic along the eastern coast of the United States empty in most cases into estuaries, which occupy the drowned lower portion of the valley and conceal the floodplain ; but the condition is different in the vast interior basin where many great rivers find discharge through the Mississippi channel. Each important tributary of that stream flows for long distances through broad lowlands, which fuse with those of the Mississippi, extending from above Cairo to the Gulf of Mexico and constantly increasing in width toward the south. The coast and the interior types must be considered separately. Illustra- tions of river floods will be selected mostly from those of the United States, partly because the conditions seem to be unfamiliar to many, and partly because the topographical relations of the central Missis- Sippi region are much like those supposed by some to have existed during the coal-forming periods. Rivers of the Atlantic Coast—Shaler,** in describing several northward flowing streams of eastern Massachusetts, says that the floodplain is in direct communication with the present margin of the river, so that a very slight rise sends water over the whole of it. “WN. S. Shaler, “ Fluviatile Swamps of New England,” Amer. Journ. Sci., 3d Ser., Vol. XXXIII., 1887, p. 203. 131 534 STEVENSON—FORMATION OF COAL BEDS. _ [November 3, The streams, though draining comparatively small areas, carry an enormous amount of water in flood time. At low water, the river extends for some distance through reedy flats on each side of the flowing stream. The swamps, which are without Sphagnum, may be divided into three classes: those, formed in areas so frequently overflowed and so penetrated with water that they cannot afford a site for perennial shrubs, are occupied by rushes in the lower por- tions and by grasses in the upper; those, occupying a narrow belt in which the grasses give place to various bushy and low growing plants, among which alders are the prevalent forms; then, in some places, a third class, a wide field of swamps, really very wet woods, covered with water not more than twice a year and usually two or three feet above the ordinary inundations: The vegetation is con- tinuous from the lower bench to the wet woods and it is able to resist the flood, though the mass of water is very great and the current very rapid. During flood these streams are almost torrential. The rivers of Maine tell the same story. The Androscoggin, Kennebec and Penobscot are all liable to sudden floods and the fierce rush of water is reinforced by logs cut for timber. But the banks of those streams are covered with bushes and trees to within a foot and a half of the August stage of water; the flood, though aided by — the logs, has not succeeded in tearing out these trees, but the trees — have seized the logs, which may be seen for long distances entangled in the bushes. Islands in the Androscoggin have trees 40 feet high, - against which the floating timber has lodged. The Connecticut river, draining a great part of the White moun- tains as well as of the Massachusetts highlands, flows for nearly 200 — miles in a broad valley, rising in terraces. It is subject to great . floods, for much of the rugged region around its headwaters has been cleared. The writer has ridden several times for a distance of 150 miles along the banks soon after high floods, which had over- flowed the second bottom, 15 to 20 feet above ordinary low water. Loose material, twigs and fallen branches, which had become d y but not decayed, had been removed to be deposited in eddies or the bottoms. But trees and bushes growing on the lower bottom or on the banks down to within a foot of low water, were not removed. 132 1911.] STEVENSON—FORMATION OF COAL BEDS. 535 Many of those are old trees which had withstood floods for more than a century, others were very young ; but the age mattered nothing, the sapling resisted as well as did the older tree, provided only that it was rooted in material that would not soften during the flood. One great flood had poured over the second bottom in the late sum- mer when the maize had attained its height. But it did not tear the plants from the soil; pressure against the broad leaves sufficed only to prostrate the plant; none was removed. At the same time, the effect of the flood was shown by trees on the lower bottom, for those 25 or more feet high, if slender, were bent down stream. Those with broad spreading crowns were affected ___ by pressure at the surface of the current. No doubt, if the flood had been repeated at intervals of two or three days, not a few of those trees would have been overturned ; but, once overturned against __ their neighbors, they would tend to protect the others by increasing _ the density of the mass and so acting as breakwaters to divide the _ flow. The flood had no effect where the vegetation was dense, the _ close growth evidently reducing the current to gentle movement. - Croppings of peat bogs, 1 to 3 feet thick, appear at many places in j 3 the banks. Such bogs suffer no injury except by undermining; in which case, a floating log occasionally tears off a piece. . The floods of the Passaic in New Jersey and of the Susquehanna have been described in several publications. They are more disas- trous than those of the Connecticut, from a pecuniary point of view; but those rivers in flood are no more effective than the Connecticut - in the struggle against vegetation. The Potomac river, though of rather rapid fall, flows in a broad shallow channel, an anomaly due in great degree to the relation _ between its normal stage and its freshets. The flood of June 1 and 2, 1888, the greatest on record, was described briefly by McGee.** The es height of water at Washington was no greater than during freshets 1 4 caused by ice jams, but, above the limit of tidal influence, the volume _ of water and height of rise exceeded any previously recorded. The *W J McGee, Tenth Ann. Rep. U. S. Geol. Survey, 1890, “ Administra- 4 tive Report,” pp. 150-152. « PROC. AMER. PHIL. SOC., L. 202 JJ, PRINTED NOV. I5, IQII. 133 536 STEVENSON—FORMATION OF COAL BEDS. _ [November % discharge was thirty-eight times as great as that during the abnor- mally wet summer of 1889; five hundred and seventy-nine times that — of the average low water discharge; it closely approximated that of the Mississippi in ordinary years and was two fifths of the discharge by that river during the flood of 1858. At Great Falls, the torrent was one third of a mile wide and 150 feet deep. This was a flood — of unprecedented extent, such as might not be repeated in centuries. — It should afford full opportunity for determining the ability of floods to remove vegetation. As McGee entered into no detail in his admin- istrative report, the request was made that he would give such infor- mation as seemed proper. His letter’ is in complete detail and the following citations are taken from it. “The most impressive river flood I ever saw occurred in the Potecnat several years ago, when during June a series of rains occurred in such : order about the headwaters as to raise the river far above any high stage previously recorded—indeed I inferred from its effect in bending smaller trees in connection with the undisturbed attitude of the older trees that it far exceeded any flood of the preceding 150 years. The discharge was not accurately measured, because the flow was too swift to get a weight into the water, but approximate measurements gave a discharge comparable w that of the Mississippi at ordinary stages. After the water subsided I went over the flooded ground with care; and this is what I found—the bottom being irregular, chiefly wooded, partly in field and pasture; in the w trees of less than, say, a foot and a half in diameter, were bent down stream and largely robbed of foliage, and a few were broken off, leaving s' the higher trees had generally lost branches and most of their foliage (the — water having risen forty to sixty feet, or well toward the tops of the highest — trees) ; here and there, especially near the channel, a tree or clump of t had been uprooted and swept away, though not more than say one or per cent. of the wood in tree or branch was gone. Here and there in woods, where the current was concentrated by rocks or large trees, a gul generally two or three feet deep, as many yards wide and as many ° long, had been cut out; elsewhere, especially where rocks and trees slackened the local current, there were bars and banks of sand; and ge erally throughout the woods there was a layer of silt, of course, left ch by the subsiding waters overspreading the soil—which usually was unmod: otherwise. From a little field, previously on the bottom, a short di above Georgetown, the entire crop and the soil to plow-depth or more removed; and in a sloping and somewhat rocky tract of pasture land, u stream from the field, the sward was irregularly furrowed by ordinarily a few feet deep and as many yards wide—the number being | * Of December 6, 1910. 134 STEVENSON—FORMATION OF COAL BEDS. 537 _ that perhaps a quarter or perhaps a third of the sward was removed. The furrowing in this pasture, by the way, represents the most extensive flood _ femoval of sward that I have ever seen. Now considering the translocation _ of material generally by the flood, it is clear that despite the favorable con- _ ditions due to abundant vegetation and to a higher declivity of the flood _ than that of the normal stream, the ratio of organic matter moved to the inorganic sediment was trifling. . . . What is true of that flood is, I am con- vineed, true of river floods generally—while the flooded river generally has its transportative capacity greatly increased, the material transported is chiefly ' inorganic, so that the resulting sediments are mainly mud, silt or sand, rather than organic accumulations.” a The writer rode through much of the area two months after the . flood had subsided. The chief evidence of great flood presented by 4 the vegetation consisted of somewhat inclined trees, deposits of q débris in branches of trees at a distance above the stream and an _ occasional furrow in the sod. These furrows were produced when the water in swirling around a projecting rock worked under the "sod and, soaking the materials below, burst the cover, so opening a the way for making a gully. In the forested portions, the litter seemed to have suffered very little injury beyond, as noted by _ McGee, receiving a cover of inorganic sediment. . 4 Murphy** has described a flood on Willow creek in Morrow a county, Oregon, a stream combining the features of a torrent with ; those of a river. The creek, 30 to 40 feet wide and enclosed in _ banks 10 to 15 feet high, has a fall of 38 feet per mile, but, unlike most of such streams, it flows through a fertile valley, 500 to 1,500 _ feet wide. The storm causing the flood of 1903 was brief, a cloud- _ burst, and the flood had passed in less than an hour. The water _ came down as a mass, 20 to 25 feet high, with a slope in front of about 30 degrees, and it was 500 feet wide. It swept away a great _ part of a town which was in its path. No details are given respect- _ ing the damage done to vegetation, but some incidental remarks make _ the matter sufficiently clear. Referring to methods of determining _ the high-water level of floods, he says that trees are the best marks; small trees are often bent over and silt or light drift is deposited on them. When the water pressure is removed, the trees straighten up i oe *E. C. Murphy, “ Destructive Floods in the United States in 1903,” _ U.S. G. S. Water Sup. and Irr. Paper, 96, 1904, pp. 9-12. 135 538 STEVENSON—FORMATION OF COAL BEDS, _ [November 3, and the drifted material is raised above the high level; but rings of silt left on trees, all on approximately the same level, show the true waterline. In this way he determined the extent of the flood. The houses, made of lumber, were lifted from their foundation and were dashed to pieces against rocks or trees. Wilkes’® made use of the same method. In speaking of foode on the Willamette river of Oregon, he says that the sudden rises of the stream are remarkable, the perpendicular height of the flood — being at times as much as 30 feet, the limit being marked very dis- 4 tinctly on trees along the banks. In New South Wales “near the — source of streams, grass is to be seen attached to the trunks of trees thirty feet above the present level of the water, which must hay been lodged there by very great floods.” This is a commonplace condition ; the writer observed it at the head of Sacramento bay in California almost forty-five years ago. He saw many bunches of drift stuff entangled in branches of trees at 10 to 15 feet above t water level, and he was astonished by the fact that so great a flood had done no injury even to the shrubs growing among the trees. — Rivers in Great Interior Basins —Excellent descriptions of floc within the Mississippi-Missouri area are given in reports of United States Weather Bureau, those by Morrill and Fae being the most comprehensive. ; Man’s skill has brought about great changes in the lowlands the Mississippi. The fertility of that region from the mouth of the Ohio to the Gulf of Mexico early led to settlements at many place But the periodic floods of the river rendered agricultural opera precarious and levees were constructed for protection. Eventually construction of such levees was assumed by the Federal goverr and they now protect a vast area from overflow. The region, exposed to devastation under ordinary circumstances, is very small but, during abnormal floods, the levees sometimes give way af ' ™%C. Wilkes, “Narrative of the United States Exploring ped 1845, Vol. IV., p. 358; IL. p. 260. ; a}, Morrill, “ Floods ‘of the Mississippi River,” Rep. Chief of Wi Bureau for 1896-7, Washington, 1897, pp. 371-431. H. C. Frankenfield, Floods of the Spring of 1903 in the Mississippi Watershed,” Bureau Bull., 1904. 136 agit.) STEVENSON—FORMATION OF COAL BEDS. 539 crevasses are formed—at times half a mile wide—through which a stream pours with amazing velocity. The conditions are materially different from those prior to settlement of the region, when the floodwaters spread over an area of 100,000 or more square miles; the energy of the flood stream, when it bursts through a crevasse, is much greater than when there were no levees. This, however, is unimportant, for if the later floods are incompetent to inflict serious injury upon lands protected by vegetable cover, the incompetence must have been more marked when the natural conditions existed. The protection afforded by levees is shown by constant decrease in extent of the flooded area; the flood of 1887 overflowed almost 30,000 square miles below the mouth of the Ohio; that of 1897 cov- ered somewhat more than 13,000, while the area was reduced in 1903 to somewhat less than 7,000—and in this year the extent would have been much less if the new levees at critical localities had been com- pleted so as to resist the very high water. Rivers carrying much detritus and subject to flood build low levees in their passage through _ the lowlands. The Mississippi constructed such ridges for long dis- 3 tances, thus preventing return of the floodwater, much of which is q _ ponded in swamps and gradually finds its way to the river farther _ down. This secondary drainage complicates the problem of recla- mation. The Mississippi floods, unlike those of the Nile, are very complex, for below the mouth of the Ohio the river receives great tributaries from the east and the west, whose floods rarely coincide; while the | : upper Mississippi, receiving the Missouri and other rivers, has its own periods of flood. The source of floodwaters is in the conti- _ nental storms, arising in the west or southwest and moving toward _ the east-northeast. The effects are felt first in the lower Missis- _ sippi, which is filled by streams entering from the west; the storm . if advances to the western ridges of the Appalachian where rise streams _ forming the Ohio, Cumberland and Tennessee rivers. The heavier ; z rains on the Appalachians pour out chiefly through the Ohio but the other streams contribute a great mass. Important floods in the eastern tributaries occur in the spring months, when heavy rains are _ reinforced by melting snow. The upper Mississippi is not an impor- 137 540 STEVENSON—FORMATION OF COAL BEDS. _ [November 3, tant factor in respect of quantity, but its swell, coming later than the others, often prolongs the stage of high water. The western rivers, entering below the mouth of the Ohio, are the Arkansas, Red, Ouachita and Yazoo, all of which descend into lowlands, where they meander for a long distance before reaching the Mississippi. The condition in this drainage area is that of rapidly flowing streams emerging from highlands on an immense area of lowland, most of which, unless protected, is subject to overflow. 7 Both Frankenfield and Morrill emphasize the oradedt rise of 4 floods within the open area. Frankenfield gives the record for 1903. _ 7 The gauge showed at Feet. Feet. Feet | Cairo, Jan. 28, 17.5 March 8, 45 March 15, 50.6 «| Memphis, Feb, 1, 10.8 Feb. 22, 33 Mar. 20, 40.1 4 Vicksburg, Feb. 2 21.0 Mar. 3, 45 New Orleans, Feb. 8, 9.1 Feb. 26, 16 Apr. 6, 20.4 The advance was deliberate, the first wave requiring four days a for passage from Cairo (at the mouth of the Ohio) to Memphis and seven days thence to New Orleans. The rise was gradual at Cairo, a being a foot and a half daily for 39 days to March 8—which was thought to be remarkably rapid—and much less thereafter to the a crest ; at Memphis, it was one foot for 21 days and only one fourth © ’ of a foot for each of the remaining 28; at Vicksburg, barely nine — tenths of a foot during each of the first 27 days; while at New Orleans, the daily rise averaged little more than one fifth of a foo throughout the whole period. The great mass of the water came from the Ohio, but the Red and Ouachita, entering from the west, were abnormally high; at New Orleans, the water was at or above danger line for 85 days. When one studies the reports of local observers, as given in the publications from which this synopsis is taken, he is surprised by the nature and extent of damage within the flooded areas. Artificial protection is almost unknown along the upper Mississippi (abe Cairo) as well as along the Missouri and its tributaries. Floo have free course in the low-lying prairie regions of Illinois, Towa : and Kansas as well as in portions of Missouri, and there one sho expect to find record of the greater disaster. Morrill has compat 138 1gtt.] STEVENSON—FORMATION OF COAL BEDS. 541 the flood of 1897 with its predecessors as far back as 1858 and he has given details in all parts of the drainage area for that of 1897. In 1897, the Ohio river was out of its banks everywhere from Pittsburgh to Cairo and the tributary streams, also in high flood, were miles wide for long distances, the “bottom,” at times, being covered with 20 feet of water, while the overflow reached into the upper portions of cities along the banks. At the mouth of the river, the lowland was flooded for 4 to 6 weeks and the city of Paducah in Kentucky was flooded for 7 weeks. The river rose 50 to 60 feet along the whole distance of more than 600 miles from Pittsburgh to Cairo. Similar conditions prevailed along the Tennessee river, which for 60 miles was 2 miles wide, reaching to the hills on both sides. In the upper Mississippi region, the river spread from bluff to bluff, 3 miles wide for 147 miles along the lowa border, and a great area of farming land in that state was inundated. Imperfect levees gave way and along the Illinois river an area of 500 square miles was flooded, making a continuous body of water from the Illinois to the Mississippi. Central Arkansas was submerged for long distances along the Arkansas river ; while below Cairo, several levees gave way and the flooded district in that region embraced more than 13,000 square miles. When one comes to sum up the effect of this disastrous flood, as given by the local observers, he discovers, that as far as the geolo- gist is concerned, they were comparatively insignificant. The damage to manufacturing interests by destruction of machinery and by de- posits of mud in mills was very great ; the railroads lost much through washing out of embankments, the ruin of bridges and the removal of ties and lumber; but loss to the farming population was only mod- erate because Weather Bureau warnings led them to transfer movable property to higher land. Small houses, barns, lumber and other loose material were floated off to be used as battering rams against bridges ; but, for the most part, farms overflowed by the rapid cur- rent were little injured. Where wheat had come up, it was drowned, not removed; where seed had been sown, it rotted; where the flood became sluggish, it left a deposit of sand, which made the land worthless, but elsewhere, as soon as the water withdrew, the farmer 139 542 STEVENSON—FORMATION OF COAL BEDS. _ [November 3, immediately set about replanting. The great flood had done little injury, had hardly disturbed the soil of cultivated fields. : Frankenfield tells the same story for the flood of 1903. The sunken area of New Madrid was filled and the water, being more or less ponded, left deposits of sand. In the lower Mississippi area, crevasses permitted great overflow, but there was no injury to farms, aside from drowning of the crops, for which there was ample com- pensation in the form of a rich alluvial deposit. The Ohio river was more than two miles wide in many places between Cairo and Louisville. Near Evansville, Indiana, 300,000 acres of maize and 30,000 acres of wheat were covered, but the only loss was that of 3,000 acres by drowning. The local observer at Evansville reported that the damage would have been much greater if the water had not remained in constant motion. At Topeka in Kansas, the flood was diverted from the river by obstructions piled against a railroad bridge and the water, loaded with sand, swept over a wide area. — Crops were ruined and the nursery fields near Topeka were covered with sand which buried the young trees. These instances are merely illustrations of conditions prevailing throughout the whole area. The reports contain no reference to the disastrous effects of such’ floods upon areas covered with forest or otherwise protected by close vegetable growth, which at first glance seemed strange, be- cause wooded areas occupy much of the lowland or bottom regions. But the omission was due not to neglect but to absence of anything to record. Reproductions of photographs given by Frankenfield and Morrill show that trees and even shrubs were undisturbed amid the rush of water and coarse sand. The writer asked the former for information respecting the matter. The reply was “During the Mississippi floods no forests are uprooted and no bogs are torn away. A considerable quantity of sand is sometimes carried down and deposited when the velocity of the water decreases, either by contact with obstruction or by reason of decrease in inclination of the floodplain. It is o course conceivable that the mass of water rushing through a crevasse carri away a quantity of vegetable matter and perhaps some trees, but the would necessarily be limited. The true Mississippi flood moves along v sedately, carrying only the enormous amount of alluvial matter in suspension, but very little indeed of foreign matter. Previous to the era of levee 140 1gtt.] STEVENSON—FORMATION OF COAL BEDS. 543 struction, the forests do not appear to have been seriously disturbed by floods.” An observation by McGee*® is in place here. The Mississippi, as it flows past northeastern Iowa, meanders through a densely wooded floodplain, four or five miles wide, now in one main and half a dozen subordinate streams and yet again in numerous large and small channels. But this plain is flooded each year; according to writers already cited, the river at times covers the whole plain from bluff to bluff as a rapid stream. Lyell,*® in referring to the 1844 crevasse near New Orleans, says that the water poured through at the rate of ten miles an hour, inun- dating the low cultivated lands and sucking in several flat boats, which were carried over “the watery waste” into a dense swamp forest. He mentions that the great Carthage crevasse was open during eight weeks and that nothing was visible above the flood except the tops of cypress trees growing in the swamp. Humphreys and Abbot”? state that the bottoms of the IIlinois river are two to ten miles wide and raised only a few feet above the usual level of the river. The greatest part of this swampy country is included in the “American bottom.” The Kaskaskia flows with © crooked course through a heavily wooded alluvial bottom, over- flowed eight or ten feet by freshets. These authors emphasize the fact, too often ignored, that lowland areas are usually well soaked by rains preceding the floods and the swampy areas become covered with water, so that when the overflow comes, it finds everything prepared for resistance. Lyell** had the weird experience of descending the Alabama river in time of high flood. At night the passengers were startled by crashing of glass and partial destruction of the steamer’s upper *W J McGee, “Pleistocene History of Northeastern Iowa,’ Eleventh Ann. Rep. U. S. Geol. Survey, 1891, p. 204. *C. Lyell, “Second Visit to the United States of North America,” Lon- don, 1850, Vol. IL., p. 1609. » A. A. Humphreys and H. L. Abbot, “Report upon the Physics and Hydraulics of the Mississippi River,” Reprint, Washington, 1876, pp. 38, 66, 76, 82. *C. Lyell, “Second Visit,” etc., Vol. II, pp. 51, 141. 141 544 STEVENSON—FORMATION OF COAL BEDS. _ [November 3, works. The boat had “ got among the trees.” The river banks are fringed with canes over which deciduous cypresses tower, while farther back is the evergreen pine forest. During floodtime, the actual channel is very narrow, as the branches of the high trees stretch far over the water, so that, when the stream has risen 40 or 50 feet, much skill is required to keep the way between them. At that time, the adjoining swamps and lowlands are inundated far and wide. But this flood does practically no injury to the forest directly in the path of its strongest current or to that farther back where the current is less rapid. Lyell found the same condition on the Mis- sissippi delta, where the flood waters, though laden with silt, have not injured even the willow saplings. The floods of great rivers in other lands exhibit the same asi nomena. According to Humboldt,”* the floods of the Orinoco begin soon after the vernal equinox and attain their maximum in July. The water remains at practically the same height until August 25, after which it falls more slowly than it rose. Its bounding region is much like that of the lower Mississippi and the flooded area is as large as England though less than the exposed region along the Mis- sissippi. The delta area is always wet except in some petty eleva- tions, which are dry for brief periods. The surface is completely inundated during several months each year. But it is covered with a dense growth of Mauritius palm in which the inhabitants construct raised platforms, on which they reside. Wallace** has described the broad level area extending to 20 or 30 miles from the main stream of the Amazon and extending for long distances along the main tributaries. This is flooded at every time of high water. It is “ covered with a dense forest of lofty trees whose stems are every year, during six months, from ten to forty — feet under water.” Much of the flooded area at the mouth of the — Amazon is covered with the mirite palms, Mauritia fleruosa and M. vinifera. =A. Humboldt, “ Personal Narrative,” Bohn Eng. Ed., 1852, Vol. IIL, : p. 8. 2A. R. Wallace, “A Narrative of Travels on the Amazon and R Negro,” London, 1853, pp. 419, 436. 142 1911.] STEVENSON—FORMATION OF COAL BEDS. 545 Kuntze* sailed along the Lourengo river through the vast wooded swamp occupying three degrees of latitude. The remarkable fea- ture, for him, was the absence of transported vegetable detritus. Rare fragments of the swamp are torn off during high water, but these consist not of detritus but of living plants; and these frag- ments become stranded elsewhere or go to sea broken into bits. The river water is brownish. Everywhere on the Paraguay as well as on its tropical and subtropical tributaries, one finds the dense forests coming down to the river, which are overflowed during floodtime; yet there is no outgoing organic detritus except frag- mentary driftwood from trees, which tumbled in from undermined banks. Kuntze followed all the great streams above the mouth of the Parana. Livingstone** has given admirable illustrations of the resistance offered by vegetation. Many times he encountered the rivers in flood, when water spread far over the plains. The conditions within the area of the Chobe and in that of the Lecambye are of the familiar type. He reached the Leeba river at the beginning of the rainy season. The river is bordered by a plain, at least 20 miles wide, where, at that time, the water was already ankle deep in the shal- lower parts, while on the Lobale plain it was thigh deep and impass- able. This flooding was not due to the river, for that had not over- flowed its banks. The condition was the same as that observed prior to the coming of great floods in the Mississippi lowlands. The Lobale plains are too nearly level to permit the rain water to flow off rapidly ; while the thick sward, so dense as to conceal the water, prevents furrowing of the surface and formation of rivulets. On approaching the Kasai river, he crossed valleys, half a mile to a mile or more in width, with clear fast-flowing water almost chin deep. One, half a mile wide, was deeper and the men crossed it by seizing the tails of their oxen. The extremely rapid current “soon dashed them against the opposite bank.” The middle of the flood was where a rivulet exists during most of the year. Boggy places are extensive * ©. Kuntze, “ Geogenetische Beitrage,” Leipzig, 1895, pp. 67, 68. *D. Livingstone, “ Missionary Travels and Researches in South Africa,” New York, 1858, pp. 191-195, 234, 235, 333. 363, 364, 392. 143 546 STEVENSON—FORMATION OF COAL BEDS. _ [November 3, on both sides of the river, but “even here, though the rapidity of the current was very considerable, the thick sward of grass was ‘laid’ flat along the sides of the stream and the soil was not abraded so much as to discolor the flood.” In his later work,®* he offers an explanation of the conditions. “The shallow valleys, along the sides of which the villages are dotted, have, at certain seasons of the year, rivers flowing through them, which at this time formed only a succession of pools, with boggy and sedgy plains between. When the sun is vertical over any part of the tropics on his way south, the first rains begin to fall and the effect of these, though copious, is usually only to fill the bogs and pools. When on his way north he again crosses the same spot, we have the great rains of the year, and the pools and bogs, being already filled, overflow and produce the great floods which mark the Zambesi, and probably in the same manner cause the inundations of the Nile. The luxuriant vegetation, which the partial desiccation of many of these rivers annually allows to grow, protects these bottoms and banks from abrasion, and hence the comparative clearness of the water in the greater floods.” Darwin and Mrs. Agassiz tell a similar story respecting the Parana and the Amazon; Cameron and Stanley have shown the con- ditions in the region of Lake Tanganika and the Sudd of the Nile has been described by Baker, Willey and other travellers. Every- where, the conditions are the same; living vegetation and even humus are practically proof against the action of floods. The Plant Materials Transported by Rivers—While it is true that a vegetable cover is an almost complete protection against ero-_ sion and that neither rain nor floods have much ability to remove — rooted plants or to take off the superficial coat of decayed or decay- ing plant-stuff, still the fact remains that rivers do carry away great quantities of plant materials in one form or another. The quantity brought down by a single torrent may be insignificant; even that borne by a river of considerable size may not impress an observer i as important; but when one reaches the lower Mississippi, draining an area of not less than 1,250,000 square miles, fed by tributaries ; from the Rockies at the west and the Appalachians at the east, which flow for long distances in broad alluvial plains, he finds a — *“ Narrative of an Expedition to the Zambesi and its Tributaries,” ‘ New York, 1866, p. 554. 144 rott.] STEVENSON—FORMATION OF COAL BEDS. 547 mass which seems to be almost inconceivably great. On some streams he finds or learns of huge log barricades, apparently afford- ing ample confirmation of Ochsenius’s barricade theory ; along others the river bed is set with “snags,” impeding navigation ; while along the main stream the casual observer is apt to regard the floating trees and other débris as an almost continuous mass. It is equally certain that a vast amount of finely divided vege- table matter, derived from chafing of logs and trunks during their voyage as well as from partial decay of the floating plants, is carried by all rivers. It is true that studies of the Mississippi, in flood and in ordinary stages, have shown that the quantity in the silts is utterly insignificant when compared with the inorganic materials, but it suffices, during decay, to give off a notable discharge of gas in the outer area of the delta. The suggestion has been made that vege- table matter, minutely divided, may explain the fertility of the Nile deposits. According to Reclus, cited by Marsh,?* it has been com- puted that the Durance river, fed by torrents of great erosive power, carries down annually enough solid matter to cover 272,000 acres with a deposit two fifths of an inch thick, containing more available nitrogen than 110,000 tons of guano and more carbon than could be assimilated by 121,000 acres of woodland in one year. The black waters of the Scottish lakes, of several rivers in Florida, of great rivers like the Congo in Africa, the Negro and others in South America prove that an enormous amount of vegetable material is leached from peaty deposits. When one considers the mass of transported timber, the content of organic matter removed by solution, and reads the more or less crude estimates of organic stuffs in the detritus carried by rivers, the mind is staggered and he is almost ready to concede that in this transportation there is the process fully competent to bring about the accumulation of coal beds. It is important, then, to ascertain, if possible, what becomes of this material. The trees and shrubs carried by the rivers were not uprooted by the torrents; they come not from abrupt slopes but from lowlands where meandering streams undermine their banks and the plants ~**G. P. Marsh, “ The Earth as Modified by Human Action,” p. 245. 145 548 STEVENSON—FORMATION OF COAL BEDS. _ [November 3, tumble in with the rest. Gibbs,?* who saw the spring flood of the | Yukon river in Alaska, relates that “During the high stage of water, which lasts for perhaps two or three weeks, great sections of the heavily wooded banks are undermined and swept away. The majestic spruce trees and tamaracks and birches, which covered them, topple over and are swept down by the current along with immense quantities of drift wood from the forest beds. The entire accumula- tion, amounting to thousands of cords of wood, is discharged into Bering sea, whose restless waves and shifting winds scatter this fuel and pile it on barren shores, hundreds of miles distant.” The conditions along the Mississippi-Missouri and their tribu- taries are the same; when the weakened banks cave, the forest, with its fallen trunks and litter, finds its way into the water. The masses of drifted wood in the channels of the Mississippi and some of the streams entering that river from the west have been mentioned in nearly all textbooks on geology during the last seventy-five years; but in most instances the descriptions have been incomplete, while in some cases they were sufficiently inaccurate to be misleading. In the early days, great numbers of waterlogged trees were held back by their roots and were moored in the silt with their usually branchless stems pointing down stream. These were the “snags” — which rendered navigation perilous. Fewer of them are encoun- tered now because a very great part of the drainage area is under cultivation, but enough are added annually to necessitate the services of several snag-removing boats along the line of nearly 2,000 miles. Most of the floating stems find their way to the Gulf, but some are stranded on the delta during floods. At one time, however, they were diverted, in chief part, into the Atchafalaya, the first great arm of the river at the head of the delta. : Darby”® has told us that the vast number of trees brought down by the Mississippi were thrown into this arm, through which they were carried with tremendous speed. The Atchafalaya raft began to form in 1778, when practically the whole drainage area of the 8G. Gibbs, “ The Break-up of the Yukon,” Nat. Geog. Mag., Vol. XVIL, 1906, pp. 268-272. ae ”W. Darby, “A Geographical Description of the State of Louis “ 2d Ed., New York, 1817, pp. 131-133. 146 tgtt.] STEVENSON—FORMATION OF COAL BEDS. 549 river had still its virgin forest, there being only a few, insignificant settlements west from the Alleghany mountains. By 1816, the head of the raft was within 27 miles of the Mississippi. Darby examined it in that year and reported that it was 20 miles long, 220 yards wide and perhaps 8 feet deep. As it was not continuous, but showed many open spaces, he was convinced that a length of 10 miles would be nearer the truth, thus giving about 4,000,000 cubic yards of loose material as the total accumulation during almost 40 years, practically the total supply of floated timber from the area of more than 800,000 square miles. He says that “the tales which have been related respecting this phenomenon, its having timber of large size and in many places being compact enough for horses to cross, are entirely void of truth. The raft, from frequent change of position, renders the growth of large timber impossible. Some small willows and other aquatic bushes are frequently seen among the trees but are too often destroyed by the shifting of the mass to attain any consider- able size.” The channel was opened by the state authorities in 1840 and the raft disappeared. Details respecting variation of position and duration of material were not given by those who described the Atchafalaya raft. But no such lack of information exists respecting the more celebrated raft of the Red river in Louisiana. That stream, formed by tribu- taries rising in the higher lands of Texas and Oklahoma, flows for a long distance through a region of yielding materials, which, in many places, is densely forested. According to Veatch,*° this raft began to form in the fifteenth century and by the beginning of the sixteenth, its head was near the present town of Alexandria, some- what more than 60 miles from the Mississippi river. It consisted of a series of complex logjams, each filling the channel. These ponded the river, which found a new outlet above the raft, so that this, by additions, gradually moved stream, becoming a great irreg- ular accumulation of logjams and open water about 160 miles long. * A.C. Veatch, “Geology and Underground Water Resources of North- ern Louisiana and Southern Arkansas,” U. S. G. S. Professional Paper, 46, 1906, p. 60. 147 550 STEVENSON—FORMATION OF COAL BEDS. [November s, At the time of the early settlements, the foot of the raft was at Natchitoches, 120 or 130 miles from the Mississippi. Humphreys and Abbot,** writing in 1861, reported that the raft was an enormous accumulation of drift logs, some floating, some sunken. The rotting of the logs at the lower end and fresh acces- sions at the upper end led to advance up stream at the rate of from one mile and a half to two miles each year, while the retreat at the lower end was about equal to the gain above. At one time, the lower end was at Natchitoches but in 1854 it was at 53 miles above Shreveport, a retreat of about 150 miles. At that time, it was nearly 13 miles long. Above Shreveport to the raft, the river bed was strewn with logs, stumps and other vegetable débris. The river is very shallow, 3 to 4 feet deep at low water, and it was about 220 | - yards wide at the head of the raft, 405 miles from the Mississippi. Rotting would cause the 13 miles, reported by Humphreys and — Abbot, to disappear within seven or eight years. The notes given by Humphreys and Abbot were from a casual examination by their assistant, the matter lying outside of the scope of their work. The raft was a serious obstruction to navigation, cutting off the region above from communication with the Missis- — sippi. Congress in 1871 ordered a complete survey, which was made by Lieutenant Woodruff, whose report, rendered in 1872, gave the — first exact information respecting the actual conditions. Captain | Howell,?* in transmitting the report, remarks that the facts have been misapprehended even by engineers. “The ‘great raft’ itself — dwindles to a mere pigmy in comparison with the popular notion of — its extent and composition.” Woodruff, in 1871, proved that the raft extended along the river for about seven miles, but that through- out that distance the channel is only partially obstructed. The whole mass of floating raft was but 290 acres and the whole area of “ tow- heads” or raft resting on the bottom was 103 acres. The towheads are formed during freshets by accumulations of logs and drift around a “snag.” As the water falls, the pile rests on the bottom and a ** A. A. Humphreys and H. L, Abbot, “ Physics and Hydraulics,” ete., 1876, pp. 21-23. *C. W. Howell, 42d Congress, 2d Session, Exec. Doc., 76, p. 1. 148 igtt.] STEVENSON—FORMATION OF COAL BEDS. 551 rapid deposition of mud takes place around it. The surface left _ above water produces willows, which, growing rapidly and binding the mass together by their roots, protect it from the washing by sub- sequent freshets. Woodruff advised removal of the raft, but, to prevent renewal, he recommended that the narrow part of the river, in which the raft was forming, be cleared of the willows lining the banks, which obstructed the passage of large bodies of floating drift. li this were done, the banks, no longer protected by the vegetable growth, would cave readily, the river would be widened and the formations of raft would cease. The removal of the raft was com- pleted in 1872 and Woodruff afterwards gave a full history of the obstruction, to which the reader is referred for other details.** Franklin** found much floating timber on the Athabasca river in northwestern Canada. “The river carries away yearly large por- tions of soil, which increases its breadth and diminishes its depth, rendering the water so muddy that it is hardly drinkable. Whole forests of timber are floated down the stream and choke the channels between the islands at its mouth.” It is clear that, on the Athabasca as on the Mississippi, caving banks yield the supply of drifting timber. In the same volume, Richardson** describes conditions ob- served along some rivers and lakes within the region traversed by -Franklin’s expedition. His statements have been cited by tuany writers but so far as the present writer has seen, not in full. They are as follows, with omission only of some details which are irrele- vant here. Peace river brings much large drift timber into Slave river “and as the trees retain their roots, which are often loaded with earth and stones, they readily sink when water-soaked and, accumulating in the eddies, form shoals which ultimately augment into islands. A thicket of small willows covers the new-formed island as soon as it appears above water and their fibrous roots serve to bind the whole firmly together. The trunks of the trees gradually decay until they are converted into a blackish substance *E. A. Woodruff, in App. Q, Ann. Rep. Chief of Engineers for 1873, Separate, pp. 45-61. * J. Franklin, “ Narrative of a Journey to the Shores of the Polar Sea,” London, 1823, pp. 192, 357, 364, 374, 38I- * J. Richardson, Ibid., p. 518. PROC. AMER. PHIL. SOC., L, 202 KK, PRINTED NOV. 16, IQII. 149 552 STEVENSON—FORMATION OF COAL BEDS, [November 3, — resembling peat, but which still retains more or less of the fibrous structure of the wood, and layers of this often alternate with layers of clay and sand, — the whole being penetrated to the depth of four or five yards or more by the long fibrous roots of the willows. A deposition of this kind, with the aid of a little infiltration of bituminous matter, would produce an excellent imitation of coal, with vegetable impressions of the willow roots. The same operation goes on in a much more magnificent scale in the lakes. A shoal of — many miles in extent is formed on the south side of Athabasca lake by — the drift timber and vegetable débris brought down by the Elk river; and — the Slave lake itself must be filled in process of time by the matters daily conveyed into it from Slave river. Vast quantities of drift timber were © buried under the sand at the mouth of the river and enormous piles of it — are accumulated on the shore of every part of the lake. The waves washing © up much disintegrated vegetable matter, fill the interstices of these entangled — masses and in process of time a border of spurious peat is formed sround the various bays of the lake.” In a later work,** referring to the drift timber of Slave river, he describes the trees as “partially denuded of their branches and wholly of their bark.” The absence of all-important details in Rich- ardson’s account is due to the fact that he was not a geologist, but he was an acute observer, as is evident from the general tenor o his reports. McConnell*’ has supplied many of the details omitted by Richardson. The sandy beaches and islands along the low reaches of Slave river owe their origin to drift timber, which lodges and soon has the growth of willows noted by that author. But those islands cause currents, which either destroy them or move them down stream. Beds of drift timber alternate with clays and sands — on many of the islands and in some instances, constitute a consid- erable portion of the whole mass. The east end of Big island Great Slave lake is fringed by a wide margin of drift timber. W the interstices have been filled by gradual deposition of sand and the decay of the wood, a dense growth of willows covers it. The cc of the main shore show the same features in many places. A’ basca and Slave lakes are inland seas, larger than Lake Ontario. — Islands, such as those described by Franklin, are not unusual i * J. Richardson, “ Arctic Searching Expedition,’ London, 1851, bec e p. 142. * RG. McConnell, Ann. Rep. Geol. Survey of Canada, N. S., Vol... 1890, pp. 63, 64, 74 D. 150 1ptt.] STEVENSON—FORMATION OF COAL BEDS. 553 4 the Mississippi river. Humphreys and Abbot** have described the process of formation and destruction: “Driftwood is lodged upon a sandbar. Deposition of sediment follows. A willow growth succeeds. In high water, more deposition is caused by the resistance thus presented to the current. In low water, the sand blown by the wind lodges among the bushes. An island thus rises gradually to the level of highwater and sometimes even above it, sustaining a dense growth of cottonwoods, willows, etc. By a similar process the island becomes con- nected with the mainland, or, by a slight change of direction of the current. the underlying sandbar is washed away, the new made land caves into the river, and the island disappears.” : When in the temperates such an island, which in spite of current _ and flood, had grown and had become coated with trees, disappears through undermining, the vegetation floats away piecemeal; but in _ the tropics the whole mass is bound together firmly by climbing and other plants, whose roots are interlaced and whose stems embrace the trees; so that, when the underlying loose material has been re- _ moved, the entangled vegetable matter floats away to be broken up 4 4 gradually. Many travellers have referred to floating island of plants - and plant material. Humboldt saw them on the Orinoco; Mrs. e Agassiz was astonished by their size on the Amazon, where some were like “ floating gardens, sometimes half an acre in extent.” _ Kuntze saw patches of moderate size floating down the Parana sys- _ tem. Miss Kingsley*® relates that during high water, the Congo and Ogowé tear away their banks in the region above brackish water, ____where there is no network of mangrove to protect them. Along the _ Ogowé, the banks are of “stout clay” and the blocks hold together, so that they often go sailing out to sea and are seen far from land __ with shrubs or even trees on them. Not all reach the open water, _ for many are stranded in the delta region, where they collect débris from the flood water and become matted with floating eee Eventually they all go to pieces. De la Beche*® cites Tuckey’s Expedition to the Zaire (Cougs) * Humphreys and Abbot, op. cit., pp. 97, 98. *M. W. Kingsley, “ West African Studies,” 2d Ed., London, 1901, p: 87: “H. T. de la Beche, “ The Renews Observer,” siesignsetih 1851, 151 554 STEVENSON—FORMATION OF COAL BEDS. [November 3, as containing the statement that Professor Smith had seen a floating mass about 120 feet long and probably washed out of the Congo, consisting of reeds resembling Donax and a species of Agrostis, among which branches of a Justicia were still growing. Powers* saw a floating island in the Gulf Stream in July of 1892. Its area was estimated at about 9,000 square feet and it carried trees, 30° feet high. It was seen again in September, having travelled more than 1,000 nautical miles. This, first seen in latitude 39° 5’, may have been torn off from the Florida coast. In every case the float- ing islands are of small extent and their rarity makes them objects of curious interest. : Driftwood.—Great rivers carry immense quantities of trees from the undermined banks. Where the course of the stream is inter. rupted by extensive lakes, such as Great Slave or Athabasca, much ; of the floating timber becomes scattered and is cast on the shore to be mingled with the mineral material, which eventually buries it. When the stream is continuous, some of the drift is cast ashore _ in eddies, more is stranded during flood time on the delta or in shal- lows at the mouth; but by far the greater part is swept out to sea, there to be battered by the waves or carried by currents to per distant shores. Nordenskiold*® relates that driftwood in the fe of small branches, pieces of roots and whole trees with adheri portions of roots and branches, occurs in such quantity at the botton of two well-protected coves near Port Dickson, that the seafai may provide a sufficient stock of fuel without difficulty. The g mass of driftwood carried down by the Yenesei floats out to Some of it is drifted by currents to Nova Zembla, the north coast Asia, Spitzbergen, perhaps to Greenland. Some of it becomes w soaked and sinks before reaching those shores. But not all goes sea, for some sinks in the river bed, upright as though rooted in | sands. A bay off Port Dickson was found barred by a palisade driftwood. “'S. Powers, “ Floating Islands,” Pop. Sct. Monthly, Vol. LXXIX,, PP. 303-307. “A, E. Nordenskiold, “The Voyage of the Vega around As Europe,” New York, 1882, pp. 152-154. 152 191t.] STEVENSON—FORMATION OF COAL BEDS. 555 That the amount of driftwood accumulated on shores has been estimated in exaggerated style is amply evident from Brooks’s*® observations. In the north and northwest part of Alaska from Norton bay to the mouth of the Mackenzie river, the shore at one time was abundantly supplied with driftwood. The Eskimos, who have been using this wood for generations, are very economical in the matter of fuel and, until the coming of the white man, the proba- bilities are that the wood accumulated more rapidly than it was con- sumed. This driftwood is brought down from the interior by the larger rivers, whose banks are forested. The cutting of wood along the banks of the Yukon has already diminished the contribution to the northern Bering sea. The north Arctic coast, eastward from Point Barrow, which is thinly populated by natives and rarely visited by the white man, has some driftwood, but according to Franklin, the quantity is unimportant along the coast visited by him; the mate- rial is brought down by the Mackenzie and is carried eastwardly by a strong current. McClure found driftwood on the northeast coast of Banks land, where it is often at a considerable distance above sea level. Low** found that the prevailing winds and the force of the waves have determined the accumulation of driftwood on the shores of Hudson Bay. During great storms, the older, the lighter portions of the mass are thrown to a considerable height above mean tide. The driftwood deposits on the northern side of Spitzbergen, on Jan Mayen and Iceland are mentioned in most of the current text- books of geology. Some of them, such as the Suturbrander of Ice- land and the deposit on the New Siberia islands are clearly not drift- wood, at least not the driftwood under consideration, as they contain fruits and tender portions of plants and belong to the Tertiary. As to the others, of undoubted modern origin, the common conception is simply misconception. The descriptions, in many cases, would lead one to suppose that the mass is closely packed and almost con- “A. H. Brooks, “The Coal Resources of Alaska,” Twenty-second Ann. Rep. U. S. Geol. Survey, 1902, Part III., p. 570. “ A. P. Low, Ann. Rep. Geol. Surv. Canada, N. S., Vol. III., 1889, p. 33 J. 153 556 STEVENSON—FORMATION OF COAL BEDS. _ [November 3, tinuous. Potonié*® has given reproductions of two photographs, one from Jan Mayen and the other by Nathorst from Amsterdam island. The drift material is scattered irregularly, as one should expect, with here and there a considerable pile. The fragments, some of which are very large, are thoroughly battered—the whole resembling very closely what one sees on the gravel flats of rivers subject to flood. Crantz*® was among the first to describe the driftwood deposits of Greenland and adjacent regions, and his statements have been cited again and again, acquiring the increment which usually comes with frequent repetition. In the driftwood on those shores he saw many great trees, which had been torn out by the roots and which, through driving and dashing amid the ice for many years, had been deprived of bark and branches and had been bored by worms. A small part of this débris consisted of willows, alders and birches from bays in southern Greenland ; with these were aspens from some more distant region; but the predominant forms are pine and fir with great abundance of a fine grained wood, with few branches, which he took to be larch. With these is a reddish wood of agree- able odor, resembling the Zirbel of the Alps. The grouping shows — that the trees came from a fertile but cool or alpine region. The drift could not have come from the American coast at th southwest as the current is contrary; it must have come with the icy current. It is most plentiful on the coast of Iceland and, on the southwest side of Jan Mayen in N. L. 75°, there are two bays in which there is so much wood, driven in by the ice, that a ship might be loaded with it. He thinks the wood may have come from Siberia, where pine and cedar abound, though it may have come ont west coast of North America by way of Bering strait. : Crantz may be right or wrong in his suggestion respecting thi source of the material; that is unimportant. - His description shows that the mass, though considerable, is comparatively insignifi nt The accumulation on the Jan Mayen bays, instead of being a clos “TH. Potonié, “ Die Entstehung der Steinkohle und der Kaustobio ith iiberhaupt,” Berlin, Funfte Aufl. 1910, p. 126; Naturwiss.-Wochinscht Vol. IV., Part 4. ee “TD. Crantz, “The History of Greenland,” Eng. Trans., London, Vol. I., pp. 35-37. 154 to1t.] STEVENSON—FORMATION OF COAL BEDS. 557 packed deposit, was merely enough to load a Danish merchantman of one hundred years ago—say a vessel of 400 or 500 tons. The description shows also that the wood had floated long; it matters not whether it came by the slow drift from Siberia or Norway to north- ern Greenland and thence southward, or followed the north coast eastward to Davis strait, it is certain that the voyage was very long and the wood showed the effects. This long continued buoyancy of the wood is equally evident in the Gulf of Mexico, where one finds the Mississippi drift wood stranded on the shores along hun- dreds of miles. In all extensive deposits of driftwood, the trees are battered, stripped of leaves, bark and often of branches; they are scattered on the strand or piled in irregular loose heaps, where, exposed to the air, they decay; or if in more favorable conditions, the inter- stices become filled with sediment, the trees become merely logs in shale or sandstone, even their genus being unrecognizable except by microscopic study of the structure. The quantity of finely divided organic matter transported by rivers is minute in comparison with that of inorganic. Pourtales*’ examined sediment from samples of Mississippi water collected at Carrollton, Louisiana. The first series, taken in March during a flood from the Red and Ohio rivers, yielded no matter of organic origin aside from some spicules of sponges and rare vegetable fibers. A second series, collected in June, during a Missouri flood giving the most abundant sediment of the year, contained in water from the surface some indistinct vegetable fibers and wood cells, but no re- mains of vegetables were found in the other samples taken at various depths. A third series, in August, contained a vegetable scale or a leaf of moss; while a fourth series, in October, contained no organic matter. A fifth series, collected in the following January, showed on nearly all the filters minute black bodies which may have been pollen or spores. But the absolute quantity of organic matter carried out is con- siderable. The “mud lumps” off the mouth of the Mississippi, are masses of tough clay, occasionally forming islands of several acres, “L. F. Pourtales, in Humphreys and Abbot, Appendix 8, p. 651. 155 558 STEVENSON—FORMATION OF COAL BEDS. _ (November 3, which rise 3 to 10 feet above the water and give off much inflam- mable gas. They were studied by Sidell** for the Mississippi survey of 1838 and by Humphreys and Abbot at a later time. Sidell be- lieved that in outflow of the river the finest materials, organic matter and silt, went farthest and, after deposition, were covered with coarser materials. Decomposition of the organic matter generates gas, which lifts the overlying deposits. According to Humphreys and Abbot, the mud lumps are formed on the crests of bars and their activity ceases when the gas is exhausted. In 1858, during operations for removal of obstructions, some mud lumps on the bar of Southwest pass were broken by an explosion of gunpowder. Strong ebullition of gas continued over a wide area for 20 minutes after the explosion and the surface of the bar, in a space, 100 feet diameter, sank, assuming the form of an extinct crater. Hilgard*® gives as composition of gas from one of the mud lumps, marsh gas, 86.20, carbon dioxide, 9.41, and nitrogen, 4.39, which closely resem- bles the average composition of gas fromswamps. Oxygen is absent. Very little of the vegetable material is stranded on the banks of rivers, comparatively little is deposited on the deltas. Most of the great accumulations in the Mississippi delta, supposed at one time to be driftwood, have proved to be buried forests in place. Terrace deposits along the Monongahela, in Pennsylvania and West Virginia, contain only here and there a woody fragment in the mass of sand, clay, gravel and blocks of rock. The same is true of terrace deposits generally, away from the lines of abandoned curves. In this con- nection, Brown’s®® observations along the Amazon are instructive. Long lines of cliffs, now on one side, now on the other, are composed of bright-colored deposits, contrasting with the monotonous clay banks of the river. The elevated plateau of these old river deposits, as well as the alluvial plain, is covered with luxuriant forest. Ex- cept a narrow strip along the bank, the whole plain is overflowed “ Humphreys and Abbot, op. cit., pp. 485, 486; W. H. Sidell, Ibid, Ap- pendix A, p. 490. “E. W. Hilgard, “The Exceptional Nature of the Mississippi Delta,” Science, N. S., Vol. XXIV., 1906, p. 865. *C. B. Brown, “On the Ancient River Deposit of the Amazon,” Q. J. Geol. Soc., Vol. XXXV., 1879, pp. 763-777. 156 tgtr.] STEVENSON—FORMATION OF COAL BEDS. 559 during the periodical floods. The cliff crests are from 10 to 160 feet above the floodmark and sections show white, yellow and red sands with bluish or variegated clay beds. Ten feet of red clay was seen at the top of one section, while in another are white clays resting on bright red clays. The area studied by Brown is about 400 by 1,000 miles. Within this he saw four insignificant exhibitions of vegetable matter in the deposits, which seem to have been laid down in an estuary. It is evident that very little of the drifted material is deposited anywhere within the river region or immediately beyond; the drift- wood deposits on the northern shores, though vast in the aggregate, clearly represent but a small part of the timber brought down by the great northward flowing rivers; the writer has followed the Gulf Stream for more than 3,000 miles but he has rarely seen floating logs; in the central part of the so-called Sargasso sea of the north Atlantic, he saw no floating timber and captains of steamships familiar with that area have assured him that driftwood is of rare occurrence ; the Orinoco brings down great numbers of trees which should be caught by the westward current, but the writer, during two voyages between Trinidad and Colon, saw not one. What, then, becomes of the vegetable material carried down by the rivers? Deep sea soundings in the Atlantic ocean give no answer to the question. Material brought up by the trawl seems to be practically free from vegetable matter. It has been suggested that the constant “creep” at great depths maintains the supply of oxygen, which under the pressure would be greater than at the surface, so that organic matter would be destroyed. Whether or not the explana- tion be in this suggestion, it is certain that where different conditions of movement and pressure exist, one finds an accumulation of vege- table material on the ocean bottom. The observations by Agassiz** in the Gulf of Mexico and the Caribbean as well as in the Pacific _ Off the coast of Mexico are final in respect to this matter. His ___ Statement is that | * A. Agassiz, “Three Cruises of the Blake,” Bull. Mus. Comp. Zool. Vol. XIV., p. 291; “General Sketch of the Expedition of the Albatross, 1901,” Vol. XXXIIL., p. 12. 157 560 STEVENSON—FORMATION OF COAL BEDS, [November 3, “While dredging to the leeward of the Caribbean islands, we could not fail to notice the large accumulation of vegetable matter and of land débris brought up from deep water many miles from the shore. It was not an un- common thing to find at a depth of over one thousand fathoms, ten or fifteen miles from land, masses of leaves, pieces of bamboo and sugar cane, dead land shells and other land débris, undoubtedly blown out to sea by the prevailing trade winds.’ We frequently found on the surface masses of vegetation, more or less waterlogged and ready to sink.” The violent hurricanes of the Caribbean, as described by Maury, must contribute very largely to the mass of vegetable material. Agassiz found similar conditions at the bottom of the Pacific during the cruise of the Albatross, when he was surprised at the distance to which land-derived material had been carried. Along most of the distance between Acapulco and the Galapagos islands as well as all along the coast from Acapulco northward to within the Gulf of Cali- fornia, there is a very sticky mud covering the bottom and inter- fering seriously with the work of dredging. His description of con- ditions between Acapulco and the Galapagos is “ Nearly everywhere along our second line of exploration, except on face of the Galapagos slope, we trawled along a bottom either muddy or com- posed of Globigerina ooze, more or less contaminated with terriginous deposits and frequently covered with a great amount of decayed vegetable matter. We scarcely made a trawl which did not bring up a considerable amount of decayed vegetable matter and frequently logs, branches, twit! seeds, leaves, fruits, much as during our first cruise.” The conditions were similar along the continental coast. The trawl was ordinarily well filled with mud along with the usual supply of decayed vegetable matter. Observations of like character hav been made by others elsewhere in the Pacific. In all probability, a great part of the vegetable matter swept out to sea disappears by oxidation at the surface or in the depths; bt in favorable localities another great part is deposited in fragmentary, condition with fine muds on the bottom, there eventually to forr beds of carbonaceous shale. : Conclusions—The grouping of facts presented in the pre din; pages, proving similarity of conditions in all parts of the worm seems to justify the following conclusions: 1. If torrents carry clear water, they produce little effect 4 158 agit.) STEVENSON—FORMATION OF COAL BEDS. 561 either the rock over which they flow or the vegetation growing on the floor or on the walls. If they carry sand, clay or even fine gravel, vegetable growth on islands, formed by aggradation, resists the flow and causes increased local deposit without material injury to itself. Even fierce mountain torrents sweeping their load of very coarse materials over a dejection cone do not clear the surface ffom trees. 2. Water of rainfall has practically no ability to remove a cover of living vegetation, even on steep declivities, except indirectly—as by finding access to unconsolidated material below, which may be rendered semi-fluid, so as to move down as a landslide. The heaviest rainfall barely disturbs the,cover of litter in a forest; that material breaks the force of the falling drops, it absorbs much of the water, it obstructs the formation of rivulets and protects itself as well as the underlying soil from erosion. Forests are practically uninjured by the heaviest rainfall; even tiny plants, growing in clefts of rocks are equally undisturbed. Rain does not remove the petty deposit of soil on a projecting point of rock, if a tuft of grass has thrust roots into it. 3. River floods in great lowland areas rise slowly, as is shown by the floods of Mississippi, Amazon, Orinoco, Nile and the rivers of central Africa. In passing through forests, those floods lose speed and become merely a quiet overflow with sluggish movement, so that they disturb neither the living growth nor the decomposed litter on the surface. They move the loose dried twigs and leaves, but even those in great part, are transported only a short distance, unless swept into the stream from the bank. The main current itself cannot uproot trees; it cannot even tear loose a floating tree which has lodged against a sandbar directly in the line of strongest flow; but, unless the sandbar be washed away and the tree be set free, that will remain as a “snag,” to become more firmly fixed during each succeeding flood. In most cases, the areas subject to these vast floods are prepared beforehand by heavy rains, whereby the humus cover is soaked and, so to say, cemented to resist the moving water. A dense growth of vegetation forms in the channels of tropical rivers and offers such resistance that even the mightiest 159 562 STEVENSON—FORMATION OF COAL BEDS, _ [November 3, floods are checked as surely as though dammed by a mountain. Logjams are not swept out by the greatest floods, though the accu- mulation is merely superficial; in spite of floods, they accumulate at the upper end while decaying at the lower. If floating débris enter a lake of moderate size, it finds its way directly to the outlet; but if the lake be large and disturbed by waves, the débris is scattered along the shores. Finely divided organic matter carried out by floods is in minute quantity compared with the mineral matter and it is deposited along with the fine mud. 4. The trees as well as the humus swept out by floods were not uprooted by the running water. Their presence in the current is due to the undermining of banks composed of unconsolidated mate- rial, so that trees, humus and the rest fall together into the water to be carried away. The damage thus done is very small—the drift carried by the lower Mississippi being the accumulation from the soft banks of more than 20,000 miles of river, the length of tor- — rential and semi-torrential streams being neglected, as they contribute Zi an insignificant proportion of the mass. The quantity of driftwood in all is unimportant, when one considers the area into which it is swept. Comparatively little lodges on its way down the larger 4 streams and most of that is buried in muds as “towheads” or 2 “snags” or accumulates in rafts to disappear by rotting. A small a part is stranded on deltas, far less than has been supposed, for in the Mississippi delta the deposits, supposed to be of driftwood, are now known to be old swamps and forests buried im situ. Much finds its way to shores more or less distant where, after having been for long time the plaything of winds and currents, it is cast in bat- tered condition, scattered here in clumps or in individual pieces to decay or to be buried in sands. But the greatest part floats until, in half-rotted condition, after long exposure, it finds its way to t depths of the sea, to litter the bottom in areas, 1,000 to 1,500 fathoms below the surface, where, mingled with remains of marine animals it will become part of bituminous shales, with here and there a of impure coal. 5. The conception that moving water, under any known or corded conditions, can uproot forests and sweep off peat bogs f 160 1911.J STEVENSON—FORMATION OF COAL BEDS. 563 even moderately extensive areas is wholly without basis in fact. One must regard it as originating in medieval descriptions of the devastating force of the Noachic deluge, which became an integral part of religious and romantic literature, so that the conception was accepted as a fundamental truth, needing neither investigation nor proof. Tue PHENOMENA OF PEAT DEPosITs. Peat or turf is familiar to those living in the temperate zone. It is an accumulation of vegetable matter undergoing special chem- ical change because protected from atmospheric oxygen by an excess of moisture. If one examine an old peat bog, he finds the surface covered by plants of various kinds growing on more or less decayed material. This, in its uppermost portion, is brown or yellowish brown, but the tint deepens downward until in the ripe peat it is almost black. At the top, one finds the vegetable structure distinct, but downward that becomes more and more obscure until in the mature peat it cannot be recognized by the unaided eye, and there seems to be only a vegetable mud containing, at times, fragments of slightly changed wood. Peat-making Plants—Sphagnum is regarded by many authors as the all-important agent in production of peat ; and this supposed con- dition has been utilized more than once to fortify arguments against the suggestion that coal beds originated from growth im situ. The prevalence of this misconception is strange, for evidence to the con- trary has been presented in many works during the last century. Darwin® says that in the Chonos archipelago, S. L. 44° to 46°, every piece of level ground is covered with Astilia pumata and Donatia magellanica, “ which by their joint decay compose a thick bed of elastic peat.” In Tierra del Fuego, the former is the chief agent. Fresh leaves appear constantly around the growing stem, while the lower ones decay. Tracing a stem downward into the peat, the old leaves can be seen in all stages of decomposition until the whole has become blended into a confused mass. Every plant =C. Darwin, “Journal of Researches,” New York, 1846, Vol. II, pp. 24-26. 161 564 STEVENSON—FORMATION OF COAL BEDS, _ [November 3, in the Falkland islands becomes converted into peat. He saw no moss peat anywhere in South America. Thomson** notes that peat in the Falkland islands is very different from that of northern Eu- rope, cellular plants being almost wanting. It is formed for the most part of roots, matted foliage and stems of Empetrum rubrum, a variety of “crowberry” common on Scottish hills; Myrtus num- mularia, a creeping myrtle; Caltha appendiculata, a dwarf species of water marigold; with some sedges and sedge-like plants. The roots, preserved almost unaltered, may be traced downward in the peat for several feet, but finally all structure is obliterated and the whole is reduced to an amorphous structureless mass. Mrs, Brassey’s® description of accumulated decayed and decaying vegetation at Borja bay in the Magellan region is in place here, as showing the origin of peat from forest material. “To penetrate far inland, however, was not so easy, owing to the dense- ness of the vegetation. Large trees had fallen and rotting where they lay under the influence of the humid atmosphere, had become the birthplace of thousands of other trees, shrubs, mosses and lichens. In fact, in some places, we might be said to be walking on the tops of the trees and first one and then another of the party found his feet slipping through into unknown = depths below.” 4 But long prior to Darwin, Al. Brongniart®® asserted that the presence of Sphagnum palustre is not necessary to the formation of _ peat. One finds on the banks of the Meuse, below Maestricht, some peats containing only leaves of resinous trees. He contents himself with the observation that all are agreed that, for formation of peat, the essential condition is stagnant water, covering the surface con stantly and never completely dried up. Lesquereux*® defined peat as a mass of woody plants whose fermentation and, consequently decomposition were retarded by the presence and the temperature °C. Wyville Thomson, “ The Atlantic,” New York, 1878, Vol. IL, p. 1 “Mrs. Brassey, “ Around the World in the Yacht Sunbeam,” New Y 1883, p. 128. * Alex. Brongniart, “ Traité élémentaire de minéralogie,” Paris, 1807, Vol. IIL., p. 41. *L. Lesquereux, “Quelques recherches sur les marais tourbeux,” M, Soc. des Sci. Nat. Neuchatel, Vol. IIl., 1845, p. 26. 162 1911.] STEVENSON—FORMATION OF COAL BEDS. — 665 of water. Vogt,*" while giving pre-eminence to Sphagnum, notes that under favorable circumstances any vegetable substance can be converted into peat, for some peats are formed of grasses and reeds. Heer®*® assigns to mosses only a subordinate part and asserts that peat originates partly from mosses, partly from water-plants, from swamp plants, especially from grasses and sedges, and partly from woody plants. A. Winchell in 1860 and Grand’Eury in 1882 made the conditions equally clear, while Frith®® showed that Sphagnum is a late arrival in accumulation of peat. Peat in the Tropics—tThe belief prevails that peat is not pro- duced in the tropics. Jameson,® long ago, asserted that peat is peculiar to cold climates, but not wholly so, for Anderson had re- ceived some from Sumatra. It is quite natural to find peaty sub- stances in warm climates for peat at the bottom of a mountain is more decomposed than that at the top; that of southern England more than that of north Scotland; that of France is more coaly than that of England, while no peat is found in the French lowlands except under cover. All of which shows that decomposition in- creases toward warm climates, until, in the tropics, it is so rapid that masses of peat cannot form. Frih* was unwilling to believe that true peat forms in the tropics. He discusses a great number of reported occurrences in tropical and sub-tropical areas. For him, the observations are in- complete and his conclusions are that, so far as known, there is no important deposit of true autochthonous peat in the lowlands of the tropics; that, within the tropics, formation of peat is in elevated regions, where the climate is that of the temperate zone; that the supposed peat layers, bored through in the alluvium of great trop- *C. Vogt, “ Lehrbuch der Geologie,” 2d Aufl., Braunschweig, 1854, Vol. IL., p. 110. "0. Heer, “Die Schieferkohlen von Utznach und Diirnten,” Zurich, 1858, pp. 1-4. * J. J. Frith, “Ueber Torf und Dopplerit,” Trogen, 1883. @R. Jameson, “An Outline of the Mineralogy of the Shetland Islands, and of the Island of Arran,” Edinburgh, 1798, pp. 151-153. “J. J. Frith und C. Schréter, “Die Moore der Schweiz,” Bern, 1904, PP. 134-143. : 163 566 STEVENSON—FORMATION OF COAL BEDS. _ [November 3, ical streams, are prevailingly allochthonous. In this last his refer- ence is to the record cited by Lyell,®* of a boring in the Ganges delta, which passed through a deposit, having certainly the characteristics of a peat bed. In view of what has been said on preceding pages respecting the accumulation of drifted vegetable matter and of the fact, that the great deposits in the Mississippi delta, formerly sup- posed to be of drifted material, are of in situ origin, one is justified in saying that the testimony of that and other borings cannot be waived aside lightly by the mere assertion of allochthony. The onus of proof is on the one making the assertion. It is difficult to understand why the a priori reasoning that trop- ical heat should prevent peat formation is thought so conclusive as to make worthless the testimony, which would be accepted as prov- ing the presence of peat in Michigan or north Germany. The con- ditions of temperature during summer in much of the United States are decidedly tropical; yet peat accumulates. It is true that vege- table matter exposed to moist air must decay more rapidly where the temperature is constantly high than in the temperates, where the hot period is of brief duration; but that has nothing to do with the 4 matter under consideration, for one is concerned with decay of vege- table matter protected from atmospheric action by a plentiful cover 4 of water. A priori, one should expect to find peat accumulating in those tropical regions, where the conditions are such as encourage 3 peat-making in the temperates—with only 'the difference that, owing to the continudus high temperature, complete decomposition shoul be more rapid and the bog should have the vegetable mud near th surface. But one is not dependent on a priori reasoning. Harper notes that it is an error to suppose that peat is confined to cold climates, since high temperature does not prevent its forma- tion if humidity and topography favor. Peat is abundant on the very border of the tropics in Florida, where tropical temperat prevails throughout the year. Sphagnum does not occur south f = C. Lyell, “Principles of Geology,” Eleventh Ed., New York, Vol. L, p. 476. *R. M. Harper, “ Preliminary Report on the Peat Deposits of Florid 3d Ann. Rep. State Survey, 1910, pp. 214, 274, 287, 292. 164 iz at Tim me fe Carr ee i we Pe we ae ry 19tt.] STEVENSON—FORMATION OF COAL BEDS. 567 N. L. 29°, but the peat deposits within southern Florida, as far south as lat. 25° 30’, are of great thickness, one dense cypress swamp showing 20 feet. The Everglades of Florida embrace about 7,000 square miles, which must be regarded as considerable. Even be- tween lat. 30° and 31°, there are thick deposits in which Sphagnum is wanting or very rare. Cypress, grasses, fern, myrtle make up most of the vegetation and provide material for peat. The water hyacinth, recently introduced into one of the rivers, is now a peat- maker and, in some localities, the peat is composed almost wholly of this plant. The temperature conditions in the Bermudas are somewhat less severe in summer than in southern Florida, for the summer heat rarely exceeds 84° F. The southwest wind, blowing off the still warm Gulf Stream, prevents low temperature and the humidity is always high. On the main island there are two great swamps, one of which has at least 50 feet of peat. The climate is such that plants of Carboniferous type could grow well, for the banana thrives while palms and the India rubber tree attain great size. The literature bearing on tropical swamps is very limited. Those swamps, often of vast extent and covered with dense forest, are regarded, rightly or wrongly, as malarial to the last degree, so that they do not invite close examination on the part of travellers. Yet, even in the limited literature, one finds ample proof that, when the necessary condition of topography and continuous moisture prevail, peat does form. Wall and Sawkins™ estimate the swamp area of Trinidad at six percent. The long dry season is not favorable to the accumulation of peaty materials; yet the Nariva swamp, drained by streams flow- _ ing 12 to 15 feet below the general level, has a black soil which after desiccation at 300° F. still yielded 35 per cent. of organic matter. Hartt® found peat in the state of Sao Salvador, Brazil, S. L. 10°. He states distinctly that he found peat. “A quarter of a mile south *G. P. Wall and J. G. Sawkins, “ Geology of Trinidad,” London, 1860, pp. 62, 63. *C. F. Hartt, “Geology and Physical Geology of Brazil,” Boston, 1870, PP. 365, 509. PROC, AMER. PHIL. SOC., L. 202 LL, PRINTED NOV. 16, IQII. 165 568 STEVENSON—FORMATION OF COAL BEDS. _ (November 3, — of the Imbugahi is such a grass-covered area, and here excavations by the side of the railroad show that a bed of peat has accumulated, which is two feet thick in some places.” It is difficult to understand on what grounds a recent writer feels justified in asserting that this is probably bituminous shale. Hartt knew peat and he knew bitu- minous shale when he saw them. On the authority of the engineer who constructed the railroad in Sao Paulo, Hartt states that near Tumanduathy the land spreads out between the hills, level as a lake and about two miles wide, covered with deep layers of black soil “fibrous and woody like peat.” The railroad was built over the surface of this bog, but no effort was made to determine the thick- — ness of the deposit. Mrs. Agassiz, cited in another connection, gives evidence respect- ing present conditions in Brazil. Kuntze®* has described the for-— ested swamps on the divide between the Amazon and the Paraguay. High water makes ponds on the broad alluvial plain, in which Ponte- deria and other plants settle. The ponds become filled with silt and — humus and a swamp flora, rich in palms, takes possession of the sur- face. The Lourenco or Cuyaba river, rising in the low divide, unites with the Paraguay at about S. L. 18°. This river, for about three degrees of latitude, flows through the vast wooded Guat swamp, which Kuntze thinks is, at least in part, a floating bog. He gives no estimate of thickness, but his brief statement leaves no doubt that the mass is very great. The material is true peat, for in another part of his work he speaks of the fragments of peat occa- sionally torn off from the mass. The invasion of streams by grass is rapid and complete throughout this region. Kuntze notes this. Morong,” in speaking of his attempt to reach the head of Pileomayc river in S. L. 22°, says that his progress was stopped in a lagoon, miles from the Paraguay river, by a dense growth of grasses weeds, one species of the former attaining the height of 5 meters. The great accumulations in the lowlands of Nicaragua and C Rica have been mentioned by several authors and the writer kr “©. Kuntze, “ Geogenetische Beitrage,” Leipzig, 1895, pp. 5, 67, 70. * T. Morong, Ann. N. Y. Acad. Sci., Vol. VII., 1892, pp. 45, 260. 166 1911.] STEVENSON—FORMATION OF COAL BEDS. 569 that great swamps with notable depth of vegetable mud abound on the isthmus of Panama. Livingstone, Cameron and others, already cited, have given abun- dant evidence that peaty accumulations are numerous and extensive in the wet regions of interior Africa. Lugard,®* in describing the region near Albert Edward lake, refers to interminable river swamps and bottomless quagmires, choked with papyrus, which abound in Uganda and Ungoro and cease below about half a degree south from the equator. Miss Kingsley®® says that she encountered three types of bog in west Africa. The broad deep bog was the least difficult, as it makes a break in the forest, and the sun’s heat bakes a crust over it, on which one may go—if he go quickly; the shallow, knee- or waist-deep bog is little more difficult as one can wade through it; but the deep narrow bog, so shaded that the sun cannot form a crust over it, is the most abundant and the most difficult. “These re- quired great care and took up a great deal of time. Whichever of us happened to be at the head of his party, when we struck one of these, used to go down into the black, batter-like ooze and try to find a ford, going on into it until the slime was up to the chin.” Chevalier*® has described an interesting type of peat formation observed in an extended area between the Gulf of Guinea and the sources of the Niger, N. L. 5° to 9°. This great region has many. granitic peaks rising to 1,200 or 1,400 meters, but in great part it is a peneplain, 200 to 400 meters above sea level. The southern por- tion is covered with a dense forest but the northern portion is a savannah with only scattered trees and shrubs. The whole region would be naked rock, were it not for the role played by a sedge, Eriospora pilosa Benth., which at times covers the rocks to the exclu- sion of all other forms. It attaches itself to granite and gneiss; the seeds germinate in minute fissures and the roots spread in clusters between thin plates of the rock, altered or decomposed by atmos- pheric action. When the thin plate of rock has been worn away “F. D. Lugard, Scottish Geog. Mag., Vol. VIIL., 1892, pp. 636-639. @M. W. Kingsley, “ Travels on the Western Coast of Equatorial Afr‘ca,” Scott. Geog. Mag., Vol. XII., 1806, pp. 119-120. ™ A. Chevalier, “Les tourbiéres de rocher de l'Afrique tropicale,” C. R., Vol. 149, 1909, pp. 134-136. 167 570 STEVENSON—FORMATION OF COAL BEDS. _ [November 3, by water, heat or the activity of the plant, this is ready to resist the rains and winds, for its adhesion is complete on the steepest slopes. After spreading for sometime, the Eriospora lifts its rhizomas into the air and sets off branches, each terminated by a bouquet of grass-like leaves, and each year, before the rains, abundant rosettes of leaves and flowering twigs are developed at the end of these rhi- zomas. At the beginning of the dry season, these leaves wither and soon afterward they are consumed by fires lighted by men or perhaps by lightning; only the bases of the leaves remain; blackened, half- carbonized, this coat thickens around the rhizoma. The growth of the Eriospora tufts is apparently very slow, but, as they may live for several centuries, they attain notable dimensions. On the border — of the virgin forest, some were seen more than a meter high and — half a meter thick at the base. The stem divides midway into ver- — tical branches, themselves dividing, the last division having a diam- — eter of 2 to 3 centimeters. ig The tufts are not always in contact, there being at times an inter- _ val of even 50 meters, but in these intervals on gentle slopes, one — finds a fibrous network, very humic, constituting a veritable bed of — peat, 5 to 30 centimeters thick. This peat is formed not only of roots and rhizomas, but also of young colonies of Eriospora, killed soon after origin by fire or by lack of light. True mosses appear at high altitudes. On the humid flank of Mount Momry, Chevalier found a Sphagnum at 850 to goo meters above sea level. The Erio- spora peat covers tens of thousands of hectares in French west Africa. The condition described by Chevalier is not wholly unfa~ miliar in the temperates, where mosses and other plants cover irreg- " ular rocky surfaces and form a coating of peat at times several inches thick. This is seen frequently in the southern Appalachians. It is the Rohhumus or Trockentorf of the Germans. It was reserved for Potonié to present the final evidence. oe ing no available literature giving details respecting moorfo ion in the tropics, he applied to Koordes, botanist of the Dutch pie dition across Sumatra in 1891. Koordes informed him that in old "H. Potonié, “Die Entstehung der Steinkohle und der Kaustobiol iiberhaupt,” 5th Aufl., Berlin, 1910, pp. 152-160. 168 19t1.] STEVENSON—FORMATION OF COAL BEDS. 571 Javan and Sumatran forests, where hard woods grow, fallen trees, many decades old, lie in great numbers and are still in good condi- tion for export. Termites and fungi are effective agents for de- struction of wood in the tropics, but the moisture is almost equally effective in preventing the decay. During the Sumatra expedition, Koordes saw a great, always green Flachmoor, with a 30-meters high mixed forest, extending along the Kampar river at more than go kilometers from the coast. As Koordes had made no close study, definitive evidence was wanting to prove that this moor had a true peat floor. But Larive made the necessary sounding at Potonié’s request and discovered that the peat in this tropical moor is 9 meters thick. Examination with the microscope proved the presence of pheno- gams ; spores or pollen; occasional brown threads belonging possibly to fungi; some resin-like bodies, etc. The high content of silica in the ash explains absence of diatoms in the microscopic preparations. Chemically, the material is a true peat and German experts pro- nounced it a good fuel. The ash in the dry material is 6.39 per cent. while good north German Flachmoor peat has 5 to 7 per cent. The ash of the Sumatran peat contains 74 per cent. of silica. Koordes estimated the area of the freshwater swamp on the left bank of the Kampar at 80,000 hectares. At both camps within this swamp, the water was stagnant, dark brown and slightly astringent. Walking over the swamp was possible only because roots of trees covered the surface with a dense network. The character of the growth, as shown in Fig. 52 of Potonié’s work, is a clear instance of adaptation such as is seen in the Ta.rodium of the southern United States; for the roots are widespread horizontally just below the surface, uniting into “broom-shaped air roots” and “asparagus- shaped pneumatophores.” The trees of the forest are mostly ever- greens, 25 to 30 meters high and closely set. The underbrush con- sists, for the most part, of the same species but its growth is slow, owing to the dense shade. The forms are all dicotyledonous and the flora is wholly of inland type. Grasses, sedges and mosses are prac- tically wanting; it is a forest moor. The stagnant pools, poor in 169 572 STEVENSON—FORMATION OF COAL BEDS, _ [November 3, phenogams because of the dense shade, are comparatively em in confervae. It is sufficiently evident that there is nothing in tropical condi- tions which would prevent the accumulation of peat. Where there is a long dry season, the vegetable matter is exposed as is that in the ordinary upland forest in the temperate regions. The accumu- lation of the humus cover advances rapidly for a time but at length the waste by oxidation about balances the additions, so that the thick- ness does not increase, though the trees shed a greater quantity each year. But in a lowland area where the moisture is great, the chem- ical changes are modified, the loss is diminished and increased supply brings about increased accumulation. Swamps arise, when from any cause the drainage is impeded. Even along the flow of small springs, — peat forms when the water is held back from any cause. On exten- sive areas, such as the coastal plain on the Atlantic side of the United States or the delta of a river like the Mississippi, where at best the drainage is imperfect, the streams being sluggish and often serpentine, the drainage has been hindered still further by vegeta- tion, the moist area was enlarged and swamps of vast extent origi- nated in post-glacial times. The important condition is the constant supply of water; the drainage must be impeded on the surface and through the bed. In the northern part of the United Stat probably the greater part of the swamps rest on an impervious bed of glacial clay, an underclay ; but it is not necessary that the imme- diately underlying bed be of normally impervious material, for —— large swamps have a floor of fine sand. Harper™ penetrated 10 or 12 miles into the Okefinokee owas of southern Georgia and discovered that the material on which peat rests is a few feet of Columbia sand overlying the clay, 1c or coarse sand of the Grand Gulf formation (Lower Miocene « Upper Oligocene). In some places a “hard pan,” colored by ve table matter and cemented by iron underlies the sand. Sanfor = R. M. Harper, “ Okefinokee Swamp,” Pop. Sci. Month., Vol. LX x 1909, p. 596. ™S. Sanford, “ Topography and Geology of Southern Florida,” 2d Rep. Geol. Survey of Florida, 1909, p. 193. 170 1gt1.] STEVENSON—FORMATION OF COAL BEDS. 573 reports that the peat of the Everglades in southern Florida rests on sand, rock or marl. Grisebach, cited by Frith, endeavored to ex- plain this apparently anomalous condition for north Germany by the suggestion, that, in very wet years, peat may have been formed in that region even on sands and, being itself practically impermea- ble, it may have prepared the way for a Hochmoor. Be that as it may, the fact remains that a swamp may begin on an apparently permeable surface; the Everglades are at little above sea level, but Okefinokee is 50 miles from the ocean and 115 feet above mean tide—and its mucky peat contains 85 per cent. of combustible mate- rial. In this latter case, one must believe that the underlying sand is not far from an impermeable stratum and that it is saturated with moisture or that by absorption of humic acid the sand itself has been rendered impermeable. Peat and Peaty Materials—Russell™* describes the Alaskan tun- dra as a swampy, moderately level country having a cover of mosses and lichens with some ferns and many small flowering plants. Below this dense carpet of vegetation is dark humus. Ponds and lakelets abound, surrounded by banks of moss, and occasionally one finds groves of alders and dwarf willows on their borders. The under- lying black humus shows few indications of its vegetable origin. It is 2 feet thick at St. Michaels but is 12 feet at a mile farther east. He saw 15 feet on the Yukon and a depth of 150 to 300 feet is assigned to it at the head of Kotzebue sound. The flora of the tundra is essentially cryptogamic, but two species of Egquisetum flourish with rank luxuriance in great spaces along the Yukon. So vast is this accumulation, in both area and thickness, that Russell ventures to suggest that some coal seams may have had similar origin. If the tundra coast of Alaska should subside, its peat would be covered with sediments and be ready for transformation into lig- nite or coal. Its associated plants and animals would indicate the climatic conditions, but the overlying sandstone and shale might con- tain leaves and tree trunks, floated in by rivers from warmer regions. But where the swamp is forested, especially if the wood resist “TC. Russell, “ Notes on the Surface Geology of Alaska,” Buil. Geol. Soc. Amer., Vol. I., 1890, pp. 125-128. 171 574 STEVENSON—FORMATION OF COAL BEDS, _ [November 3, rapid decay, the change of that material may advance so slowly that long after the softer parts have been reduced to pulp, the more com- pact materials may remain almost unchanged. The white cedar logs in the New Jersey swamps are so well preserved in many localities as to be serviceable in manufactures, while the thick peat of many swamps in Florida is without commercial value, because it is crowded with cypress stumps and fallen stems. But decay occurs in the softer woods, so that, as v. Giimbel*® relates, flattened trunks are found even at a depth of only a meter in loose peat. The flattening was due to rotting, not to pressure. The newer peat shows distinct felting and Lesquereux” states that this condition is marked even when decomposition is far advanced. In peats formed above the original water-plane, the “emerged peats” of that author, the layers are characteristic, one inch thick at the top but decreasing downward to less than one eighth of an inch. While the older or ripe peat shows no trace of organic structure to the unaided eye, the microscope proves that it is com- posed of fragments of plants embedded in an amorphous material consisting of humic or ulmic acid, or a mixture of those acids and their salts. He observed that, whenever the growth of the peat was checked by dryness or other causes, “the upper surface of the peat becomes crusted, hardened and transformed into a thin coating, quite impervious to the entrance of any kind of foreign matter: and it is upon this hard upper crust that the boggy humus forms; or wherever the land becomes resubmerged, a new peat vegetation begins. In which case, such a crust remains as a parting layer between two beds ¢ peat, like the well known clay partings between two coal benches.” v. Giimbel, in the work just cited, asserts that the minute frag- ments of plants are not only intimately mingled and felted but also, in the denser portions, are bound together and more or less cemented _ by a humus-like substance, which is soluble in a dilute solution of — caustic potash. Peat, treated with this reagent and afterwards dried, — 7 C, W. v. Giimbel, “ Beitrage zur Kenntniss der Texturverhaltnisse der Mineralkohlen,”’ Sitz. Berich. d. k. bayer. Akad. d. Wissenschaften, Math- — Phys. K1., 1883, p. 126. nas Lesquereux, 2d Geol. Surv. of Pennsylvania, Ann, Rep. for Bie: p. 118. 172 1911.} STEVENSON—FORMATION OF COAL BEDS. 575 frequently falls to powder. v. Giimbel found also, in many peats, deep black coaly parts of plants as small fibers, a form which he terms Torffaserkohle and regards as a thoroughly characteristic type. Peat always contains much water, often 95 per cent., when freshly removed; but a great part of this evaporates on exposure, there remaining in the air-dried material from 10 to 30 per cent., the denser peat retaining the larger quantity. When first taken out, it is plastic but after thorough drying the plasticity is lost. Peat is very porous; v. Gtimbel subjected Sphagnum peat to a vertical pressure of 6,000 atmospheres and reduced 100 centimeters to 17.7. The compressed material was apparently homogeneous, the streak was lustrous and lamination was distinct on the fractured surface. The reduction was due wholly to compression, obliteration of the pores, for, when moistened with water, the mass swelled to practically the original bulk. This condition, however, may not be constant. The writer has some briquetted peat, made under great pressure and moderate temperature, which has no tendency to swell when moistened. It has lost all plasticity and in sixteen years it has shown no change on the brilliant surface at each end. Peat, then, consists, aside from introduced sand, clay or calca- reous materials, of more or less changed plant tissues, whose organic texture is still recognizable, and of an enclosing substance derived from complete decomposition of plant tissues, which is originally soluble in water but which, on drying or perhaps on oxidation, becomes insoluble. _ Fuel peat has from I to 25 per cent. of ash. The purest peats contain less mineral matter than is found in the plants whence they are derived; while on the other hand a peat deposit may pass from pure peat into carbonaceous mud and thence into muds almost wholly without trace of carbon. Mills and Rowan™ have given ultimate analyses of surface and dense peats from two localities in Ireland, which represent the extremes of high grade fuel. In each case, the ash is excluded in calculating the other constituents. The same authors give twenty-seven analyses of the ash found in "FE. J. Mills and F. J. Rowan, “Chemical Technology,” Philadelphia, 1889, pp. 15-20. 173 576 STEVENSON—FORMATION OF COAL BEDS. _ [November 3, different fuel peats from Ireland, which show, as one should expect, extreme variations, due to local conditions. Potash and soda are in small proportion, varying from 0.146 to 1.667 of potash and from 0.446 to 2.883 of soda, the greater quantity being in the dense peat. Phosphoric acid is present in all but rarely exceeds 2 per cent., whereas sulphuric acid varies from 10 to 44 per cent. Hydrochloric acid is present in small proportion, but approximately that required by the soda. Lime and magnesia are always present, in some cases the former makes up nearly one half of the ash and in others the latter is one sixth. Ferric oxide varies from 6 to 30 per cent. Silica occurs as sand or as soluble silica and alumina is always pres- ent, though at times in small quantity. The ash in the — analyzed varies from 1.120 to 7.808. pF H. oO. N. Ash. 1. Surface, Phillipstown ..... 58.69 6.97 32.88 1.41 1.99 2. Surface, Wood of Allen .. 59.92 6.61 32.20 1.25 * 2.74 3. Dense, Phillipstown ...... 60.47 6.09 32.54 0.88 3.30 4. Dense, Wood of Allen .... 61.02 5.77 32.40 0.80 7.89 The content of alkalies rarely exceeds 4 per cent. of the ash in New Jersey peats and ordinarily it is less than one and a half per cent.; but calcium carbonate and sulphate are always present in notable quantity, making up from 20 to 30 per cent. of the ash.” Julien” has given a synopsis of the available information respect- . ing the proximate composition of peat. The various organic con- — stituents seem to be of rather indefinite character and their study is — attended with serious difficulty. Julien cites an analysis from Her- — mann, giving composition of a peat obtained near Moscow: Muck-carbon, nitrolin, plant remains ...............-. 77.8 Humic -a¢id nea c ks bbac ee cuca sree es eae nee cee 17.0 Humug Cut OCt och 5 dca cy eset ue sees ween eee ene 4.0 AMMONIA See os sacs ke cone eb naan eth el eee 0.25 Crenic acids uss cakes -s.ske 0d heeaahe melee ee Trace ASH we 6k cin s:0'0'p seis wins Wes os bc site Wee oo wn Acc 1.25 *W. E. McCourt, “A Report on the Peat Deposits of Northern New Jersey,” Ann. Rep. Geol. Surv. of New Jersey for 1905, p. 227. * A. A. Julien, “On the Geological Action of the Humus Acids,” Proc. A. A. A. S. for 1879, pp. 314-324, 329, 331, 353- 174 1911.] STEVENSON—FORMATION OF COAL BEDS. 577 But in peat from another locality, the process of change was different : DRMCK-CREDOGL, OO pie ius hentia dasse pec ack wecues 80.0 : Weaerenics MOR as oe nes Sea oie ewes 17.0 OTN CMO is ce dae cats oe EA oo A wen ie eR 1.0 DAME oo 5 one ede aie ie a da Us wi Patio cs eee kine 2.0 Julien notes the difference in ash between peat and the plants whence it is derived. Sphagnum has from 3 to 4 per cent. and the peat varies from I to 25 per cent. Vohl found only 1.25 per cent. of ash in a Hochmoor or Sphagnum peat. In the ash of living _ plants he found 20 per cent. of alkalies and 42 per cent. of silica, but _ only 3 to 4 per cent. of each in the peat or in the soil. On the other hand, there was concentration of alumina, ferric oxide and calcium carbonate as well as of phosphoric and sulphuric acids. Pyrite occa- "sionally abounds in peat and its decomposition gives a basic ferric sulphate to the bog iron ores. Nitrogen is always present in peat, sometimes as much as 3 per cent. Many suggestions have been made to explain its occurrence; but Julien thinks that most probably it has been derived from animals living in the peat or in the soil. The vast number of insect cases found in peat bogs is well known and Scudder has proved that insects were very abundant in the coal period. The nitrogen content _. is due very largely to the exuviz of insects, and its frequent concen- tration in the lower layers of bogs may be due to the survival of those exuviz as chitin. In the humus one finds as inert substances, nitrolin (rotten wood) and humin, which is black and forms the chief constituent of humus; but it is so mingled with nitrolin that its exact composition cannot be determined. Mulder studied humus and humic acid from the black peat of the Haarlem sea; he obtained ulmic acid from rotten wood as well as from the light brown Frisian peat. The formulas of the several acids obtained seemed to be aint ACI oa ee hk ce eee eke CeHzO0. Case SC a es ee oe C»H:.0; TR NR ee ee ede sp CxwHuOs Stern, however, thought humic and ulmic acids isomeric, with the 175 578 STEVENSON—FORMATION OF COAL BEDS. _ [November 3, formula C;,H,,O0,, while Ditmer thought them the same thing, the latter being produced by drying the former. Humic acid is a colloid and is not absorbed by plants, thou its oxidized product, crenic acid, seems to be taken up. Humic acid is very slightly soluble in water at 6° C.; if dried at 120° C. it is much less soluble and if completely dried at a high temperature it is insol- uble. Its alkaline salts are readily soluble but those with alkaline earths and metallic oxides are insoluble or nearly so in water, though readily in aqueous alkaline solutions. Calcium humate dissolves in 3,125 parts of water and ferric humate in 5,000 parts, but these form soluble double salts with ammonia. Ulmic and humic acids are rarely free except in bogs. A noteworthy property of humic acid is that, as a colloid, it renders sand impermeable to water. These feeble acids yield others upon oxidation. Humic gives crenic, which is present in all waters, in rotten wood, in peat and in cultivated soil. Julien has found it, as well as its oxidized product, apocrenic, in American peat. Crenic acid, pale yellow and trans- parent, is readily soluble in water; in drying, it becomes opaque and blackens when exposed to the light. Alkaline crenates are very soluble but the calcareous salts are only slightly soluble. Those of iron and aluminium are insoluble, but, according to Bischoff, the iron salt is soluble in ammonia, so that it may be dissolved in the presence of decaying nitrogenous substances. The apocrenates have same distribution as the crenates but they are less soluble. These organic acids bleach clays and have solvent effect on silica; the most efficient being the brown or ulmic constituents. Liebig,*° writing soon after Mulder published the results of his investigations, stated that a solution of caustic potash blackens in contact with vegetable mould. Dilute sulphuric acid precipitates — from the solution a light, flocculent brown or black substance which absorbs oxygen rapidly. After drying it is not soluble in water. Cold water dissolves only one ten-thousandth of its weight from vegetable mould and the dissolved material is chiefly salts; but boil- ing water extracts several substances, yellow or yellow-brown. On * J. Liebig, “ Chemistry in its Application to Agriculture and Physiology,” Philadelphia, 1843, pp. 112, 113. 176 19tt.) STEVENSON—FORMATION OF COAL BEDS. 579 exposure, the solution becomes darker and a flocculent deposit is produced. If the yellowish solution be evaporated to dryness and the residue be heated to redness, this becomes black and, when treated with water, yields potassium carbonate. Evidently, boiling water extracts a substance which owes its solubility to alkaline salts con- tained in plants. Leibig says, on authority of Sprengel, that humic acid becomes insoluble when dried in air or when frozen in moist condition. Hunt™ has remarked that organic matters in solution acting on insoluble peroxide of iron form the protoxide, which is soluble in carbonic acid and in excess of the organic (acid) matter. In this way, great quantities of iron may be removed and white clay or sand- stone may be produced. The iron salts become oxidized and go down as hydrated peroxide. Manganese deposits are formed in similar fashion. He is inclined to believe that hydrated alumina may originate in the same way. Organic matter dissolved by sur- face waters reduces sulphates to sulphides and these, decomposed in turn by carbonic acid, yield alkaline and earthy carbonates as well as hydrogen sulphide. One finds in bogs some types of peat to which the descriptions thus far given do not apply. Examined in detail, these in some cases suggest original differences due to mode of accumulation or to character of material, while in others they appear to be due to sec- ondary processes. Long ago, Caspary described the Lebertorf obtained at Purpes- seln, near Gumbinnen in east Prussia. This material was studied very carefully by v. Giimbel,®? his specimens being from the type locality. The deposit is 5 feet thick and at 10 feet below the sur- face. When damp, it is liver-brown in color and dense, but when dry it divides into paper-like laminz. Under the microscope it proves to be composed of very fragmentary parts of plants within a felt-like, flocky mass, in which are insects, recognizable grass and moss, scattered black wood cells, many spores and an immense quan- "T. Sterry Hunt, “Chemical and Geological Essays,” Boston, 1875, pp. 97-99. =C. W. v. Gimbel, of. cit., pp. 131, 132, 133. 177 ¢ 580 STEVENSON—FORMATION OF COAL BEDS. _ [November 3, tity of pollen. Two specimens from other localities agree in that the cross-section shows a comparatively dense mass of boghead-like material, deep brown in color. One contains well preserved remains of leaves and other organs and the other contains some freshwater mollusks. The granular, felt-like mass, treated with reagents, breaks up and then one sees fragments of woody material, seeds, mosses, and above all pollen grains, several thousands to the cubic milli- meter. This substance bears remarkable resemblance to cannel. — Blattertorf, so named from its foliated structure, is closely allied to Lebertorf. v. Giimbel has described a specimen from the kurishen lowland south from Nidden. This mass, in extraordinarily thin laminz, is composed of numerous bright lamellae alternating with dull layers, recalling by their luster, pitch and glance coals. The bright material comes from ribs and the harder parts of plants, the grass leaves, which compose the chief mass. Along with those are bits of moss, bast fibers, etc., in the felt-like mass, as well as an astonishing number of pollen granules. The results of Frith’s** studies were published in the same year with those of v. Giimbel. His conclusions respecting the composi- tion of Lebertorf differ somewhat from those reached by v. Giimbel. One rarely finds in ordinary peat any remains of freshwater alge. _ But Frith finds that those algze do not decompose so readily as one might imagine; yet in the ordinary peat they are only rare and acces- sory constituents, never occurring in such quantity as to be impor- — tant elements. At the same time there are types of which they are — essential constituents. 3 ; The Lebertorf, found in ponds within Prussia, as the basis of the Rasenmoor at Purpesseln and as basis of a Hochmoor at Gum- binnen, is a liver-brown gelatinous mass. That from Jakobau, in west Prussia, consists chiefly of alge, there being more than 60 spe- cies of Chroococcacee, Hydrodicteee and Diatomacee, with which are found indefinite remains of mosses along with pollen of Pinus and Corylus. The Torfschiefer of E. Geinitz from Gustrow has a. very similar composition. The typical Lebertorf from Purpesseln has recognizable colonies of Macrocystis, while that from Niederwyl “J. J. Frith, “ Ueber Torf und Dopplerit,” Trogen, 1883, p. 20. 178 1911.) STEVENSON—FORMATION OF COAL BEDS. 581 in Thurgau shows alge as the chief constituents, with pollen and chitin remains, all felted and embedded in a gelatin-like mass. These Lebertorfs originated in quiet waters or on damp soil, through con- tinuous deposition of gelatinous alge. Lebertorf is the same with Faulschlamm or Sapropel of Potonié,™ an accumulation of stagnant water organisms, animals as well as plants, a formation characteristic of pools in swamps. The fresh- water alge multiply with such rapidity that eventually a great mass may be deposited. Potonié says that there are lakes in south Ger- many so filled with Sapropel that they cannot be navigated. Cas- pary, cited by Frith, conceives that there is no peat-filled lake, on whose bottom this material does not exist. He found it about 9 meters thick at one locality. Frih has given a synopsis in his later work of studies by the students of northern Europe which show the wide distribution of this material. But Lebertorf or Sapropel, so closely resembling cannel in appearance, is not the mass of peat; it is wholly local, originating in open ponds or lakes. The gelatinous alge are of comparatively rare occurrence in true peat, which owes its origin to plants of wholly different type. The substance, known as Dopplerite, was described by Haidinger in 1851 and was studied in great detail by v. Gumbel in 1858. Its similarity, in some respects, to coal led the latter author to give it the name of Torfpechkohle. It occurs at many localities, so many that it may be regarded as a normal constituent of peat. The first reference to material of this type in America is in a paper by Fair- child,** who obtained some from a bog at Scranton, Pennsylvania. It is described as bright, resembling a firm but brittle jelly and as occurring in branching masses through the ripe or older peat. In drying, it shrinks more than the peat and the color changes from yellowish brown to almost black, finally becoming brown. In struc- ture it resembles coal. Julien,** discussing Fairchild’s communication, asserted that the physical features of the substance, as described, are those of apo- Ms er Te ee tigate URN: enc ian Sr ° m Me “H. Potonié, “Die Entstehung,” etc., p. 20. “H. L. Fairchild, Trans. N. Y. Acad. Sci., Vol. 1., 1881, p. 73. "A.A. Julien, Trans. N. Y. Aead. Sci., Vol. 1., 1881, pp. 75, 76. 179 582 STEVENSON—FORMATION OF COAL BEDS, _ [November 3, crenic, humic and other organic acids. He was inclined to believe that this material had been produced by the leaching out of soluble salts of organic acids, in part crenates, from the upper part of the bog and their concentration in the denser portions below, where they filled cavities in the peat. The rapid change in color is not the trifling change due to drying, but is a characteristic reaction of crenic acid, due to oxidation and to partial change into apocrenic acid—a feature observed in the acid and in its salts, both in nature and in the laboratory. Lewis*’ described the material with more detail. It occurred in swamp muck underlying 8 to 10 feet of peat. Near the bottom and confined wholly to the muck, are irregular veins filled with a black jelly-like elastic substance, in quantity varying from mere stains to streaks, two or three inches wide. When first taken out it is jelly- like, with conchoidal fracture, but on exposure it becomes tougher and more elastic. Under the glass it is brownish red and nearly homogeneous. It is tasteless and odorless, burning slowly and with- out flame, when fresh. It is insoluble in water, alcohol and ether but is dissolved by caustic potash. Completely dried, it is brittle and coal-like, resembling jet; it burns with a clear yellow flame and no longer softens in water. In composition, it differs from the typical dopplerite in that it contains little more than half as much carbon and very much more oxygen. Kaufmann, cited by Lewis, regarded dopplerite as a mixture of — humus acids and believed that the portion of peat, soluble in caustic potash, is identical with dopplerite. Compact peat contains minute — black particles of dopplerite. Peat is merely a mixture of partly — decomposed plants with dopplerite, the latter being a homogeneous — peat in which all organisms have been decomposed. Kaufmann — found that the proportion of material soluble in caustic potash in- creases with age, a recent peat giving from 25 to 30 per cent., while © | an old compact peat gave 77 per cent. But in coals, the proportion decreases, from a diluvial brown coal, with 77 per cent., to anthracite ‘in which no portion is soluble. His conception is that, in the forma- *H. C. Lewis, “On a New Substance Resembling Dopplerite, from a Peat Bog at Scranton,” Proc. Amer. Phil? Soc., Vol. XX., 1881, p. 112. Be 180 tgit.] STEVENSON—FORMATION OF COAL BEDS. 583 tion of coal from peat, the first step is the formation of dopplerite and the second is a gradual transformation of the latter into a mate- rial less soluble in alkalies and richer in carbon. The peculiarities of the Scranton mineral, its low percentage of carbon and its mode of combustion led Lewis to suggest that it may represent an earlier stage in transformation than that of dopplerite. v. Giimbel,®* in giving the results of his later studies, described dopplerite as a yellow brown homogeneous mass without trace of organic structure and enclosing only separated parts of plants. It burns with a sooty flame, thus differing from the Scranton mineral which burns with a clear flame. It dissolves in acid with efferves- cence; the calcareous matter seems to be combined chemically with the humus-like material. He is inclined to see in dopplerite a sub- stance originating in mere segregation from plant material as the silica of flints is separated from limestone. He looks upon dop- plerite as possessing great importance, since in most peats, one finds cementing substances which, optically, physically and chemically, resemble it closely. Friih,®* in his earlier work, already cited, gives an elaborate dis- cussion and reaches conclusions differing very much in some respects from those just given. He asserts that dopplerite exhibits the wholesale formation of ulmin compounds and gives detailed descrip- tion of its physical and chemical properties to prove that it belongs to the ulmin group. Owing to the large proportion of calcium, he thinks the material pre-eminently a Rasenmoor deposit. On wholly fresh profiles of Rasenmoor at Gonten, Schwantenau and Rothen- thurm, he saw in the red-brown peat, brown flakes, one centimeter to one decimeter, so mottling the mass that he termed this type Marmortorf. Very frequently the flakes are associated with a frag- _ ment of root or twig, along which water would flow. At Rothen- thurm he found the dopplerite first along a root. These brown flocks are always rich in water; the Rasenmoor is always rich in water, a condition which favors homogeneous ulminification of the *y. Giimbel, op. cit., pp. 129, 130. ; - “J. J. Frith, op. cit., pp. 64, 68, 69-72. PROC. AMER. PHIL. SOC., L, 202 MM, PRINTED NOV. 16, TOL. 181 584 | STEVENSON—FORMATION OF COAL BEDS, [November 3, peat. The brown flocks are sources of dopplerite. There is no sharp separation between dopplerite and the surrounding peat— there is always a passage zone, an intermingling of peat and rp : plerite. The mode of occurrence is variable. In many places, he saw veinlets, one to two meters long and one to five centimeters wide; here and there a vein spreads out from a root—one passed over a thin sandstone and was prolonged horizontally for several meters as a little bed, at most two centimeters thick. At the same place, he observed some wedge-shaped veinlets penetrating the glacial drift — to a depth of 3 to 4 centimeters, where it filled cracks in the clay, a binding the fragments into a breccia. There were no plant remains in the clay, so that the fine gelatinous dopplerite must have been deposited in already existing cavities. The presence of abundant — water being essential to the ulminification, the mineral is found espe- _ cially in the lower part of the peat. As every plant can become ulminified, dopplerite may occur in any moor, where the temperature — and moisture are in proper relation. He has found the mineral — derived from Sphagnum at the contact between Rasenmoor and Hochmoor, where the water-rich condition existed. Kaufmann believed that with the point of a knife he ease particles of dopplerite from good peat; Frith did this with Marmor torf, but he thinks that even the best Rasenmoortorfs are not usuall so far advanced as that. The microscope detects little flakes pro- duced by the flowing together of very tender ulmin material, if the peat be ripe; but one cannot determine whether these are ulmic humic acid—the quantity is too small. At the same time, he méz tains that it is an error to identify with dopplerite the caustic p extract from peat, as Kaufmann and Muhlberg have done, for pot: combines with ulmic and humic acid alike. Dopplerite is a h member of the ulmin group. Kinahan® often observed that, when peat was taken out on t hills near Dingle bay, little streams of tar, which had filled tube made by decay of roots, oozed and trickled out from the newly n ” G. H. Kinahan, Geol. Surv. of Ireland, Explan. of sheets 182, 18 1861, p. 33. 182 1911.] STEVENSON—FORMATION OF COAL BEDS. 585 surfaces. This is clearly the younger stage of dopplerite referred to by Frih. The Schieferkohle, studied in detail by both v. Giimbel and Frith, is a Quaternary deposit observed at several places in Switzerland. It will be described on a succeeding page. v. Giimbel’s™ type speci- mens come from Morschwyl, but he studied also specimens from other localities. The mass is partly loose, like peat, partly dense, like pitch coal, containing remains of conifers, birches, etc. It is undeniably peat-like in the less dense portions, where one can recog- nize mosses and grasses as the predominating constituents. The denser portions are changed by caustic potash into an opaque mass. The microscope shows great quantity of deep brown shell-like splinters of an amorphous textureless substance, which acts as dop- plerite. In many parts of plants, the same dark brown material fills the cell spaces. He thinks it not doubtful that the denser condition of this portion of the coal comes from richer accumulation of the amorphous filling material, which he terms Carbohumin. This Schieferkohle contains vast numbers of pine cones, not deformed, and of flattened pieces of wood. In many of the latter, he found an inner woody zone, composed of a soft yellow substance, like rotten wood, while the bark zone had been changed into a shining pitch coal. Friih,®*? after studying Schieferkohle from many localities, con- firmed the view of Heer, Kaufmann and others that the deposits agree with peat in microscopic character. They are peats more strongly ulminified. He often found the interior of rootlets appar- ently little changed, but after a few minutes exposure, they began to change and at length became brown like the Marmortorf. With ’ regard to the wood fragments, he thinks that the outer portion was ulminified early, perhaps before the bog was covered with drift, whereas the inner portion was merely peated. At the same time he does not recognize dopplerite in the Schieferkohle. It is sufficiently evident that the difference between Frih and the other observers is merely respecting nomenclature. There is agreement on all matters which concern the questions at issue here. ' “vy. Giimbel, op. cit., pp. 136, 137. “J. J. Frith, op. cit., pp. 83, 84. 183 586 STEVENSON—FORMATION OF COAL BEDS, _ [November 3, This is placed beyond doubt in his later work,®* where he modifies the broad statements made in his earlier work and shows that the difference is formal rather than real. He says that dopplerite origi- nates, as does the peat, out of a varying mass of colloid substances, free humus acids, salts of humus acids, inorganic substances and some nitrogen. So one may regard dopplerite as an ulmate, a hu- mate, a crenate or a mixture of them all, with in addition some inor- ganic salts. The essential point is that, during the process of peat- making, a greater or less portion of the vegetable material is brought into a condition admitting of flowage, so that it may remain distrib- uted throughout the mass or may be collected into cavities. When the pores of the peat are filled, farther drainage is possible only to a limited degree and the material will find its way to the tissues, be- coming the Carbohumin of v. Gitimbel. To this absorption of Car- bohumin is due the different effect of pressure upon peat and brown coal; in peat the porosity is very great, in brown coal it is small. Variations in structure or appearance of the peat have been observed in recent bogs, which are as notable as those found in the Schieferkohle. Griffith®** in describing the Irish peat bogs, said that bases of the bogs consist of clay covered with a layer of peat, which is composed of rushes and flags. Above this is another bed of peat, closely resembling cannel coal, with conchoidal fracture and hard enough to be worked into snuffboxes. It yields 25 per cent. of ash and much oxide of iron. This, in turn, is covered with black peat — containing twigs and’ branches of fir or pine, oak, yew and hazel, only the bark remaining. Where whole trees were found, the roots had disappeared. Lesquereux®® relates that on the border of the valley of the Locle, a considerable mass of marl covers a bed of peat, which has become converted into lignite, hard, fragile and with brilliant frac- ture. The thickness on the border is barely 3 inches. Farther — downward toward the bottom of the valley, the marl is only 4 feet . " J. J. Frith, in “Die Moore der Schweiz,” 1904, pp. 164, 165, 166, 167. “Griffith, cited by S. S. Haldeman in Introduction to 2d Ed. of R. C. © Taylor’s “ Statistics of Coal,” Philadelphia, 1855, p. 166. * TL. Lesquereux, “ Quelques recherches sur les marais tourbeux,” Neu- chatel, 1845, p. 95. 184 r9tt.] STEVENSON—FORMATION OF COAL BEDS. 587 thick and the underlying peat, though showing some change, still retains its peaty character and is a passage from the lignite of the border to the peat now worked in the open valley. One is left to conclude that the deposit is continuous from the border to the uncov- ered peat. The Characteristics of Peat Accumulations—Swamp or marsh accumulations of vegetable matter consist essentially of remains of land plants, including the many water-loving types. Locally, as in the Lebertorfs or Sapropels, one finds freshwater alge and remains of mollusks, while in many swamp peats the exuvie of insects abound, often associated with land mollusks. Some of the older books refer to marine peat. Macculloch®* mentioned a peat found in Scotland, which was composed of Zostera marina, and several authors have cited this as a marine peat. But Zostera is the ordi- nary “eel-grass’’ of estuaries and is a land plant able to endure salt water. Al. Brongniart,®* under the title “ Tourbe marine,” states that De Candolle saw in the dunes of Holland some peats which burn well and are composed of seaweeds, notably of Fucus digitatus. He, himself, had seen, opposite the rock of Calvados, some extensive beds of brown material, soft and spongy, which had all the external appearance of peat, but it could be burned only with difficulty. That seaweeds may accumulate on a strand, there to form a considerable deposit, is placed beyond doubt by Potonié’s description of such an accumulation on Heligoland, of which he gives a photograph. But such deposits are wholly local and possess no importance. Muck,°* in the first edition of his work, referred to the occurrence, at several places along the North sea, of peat evidently marine in origin. Samples of the material were sent by him to Friih, who submitted them to microscopic analysis. One consisted almost wholly of de- caying seaweed; when dried, it burned with small flame and foul odor, but it showed no characteristic of peat. Another, a brown substance washed up on the shore at Blankenberghe, contained no “J. Macculloch, “A System of Geology,” London, 1831, Vol. II., p. 330. * Al. Brongniart, “ Traité élémentaire de minéralogie,” Paris, 1807, Vol. II., pp. 41, 46. *“F. Muck, “ Die Chemie der Steinkohle,” 2te Aufl., 1891, p. 164, footnote. 185 588 STEVENSON—FORMATION OF COAL BEDS. _ [November 3, trace of alga, but consisted wholly of fragments from land plants. The same is true of a specimen from the Dollart. After examining samples from all localities of alleged marine peat, Frith felt himself justified in the positive assertion that thus far no marine peat has been discovered. The classification of peaty deposits has received much attention from many authors. The literature in America is somewhat limited, as, until very recent years, peat seemed likely to remain indefinitely without economic importance. Among the earliest attempts at clas- sification is that by Shaler,®® whose grouping was much in detail. He divided the forms into marine marshes and freshwater swamps; the former including grass marshes and mangrove marshes, grow- ing above tide, as well as mud banks and eel-grass marshes, growing below mean tide; the latter including river, lake and upland swamps, each with two subdivisions. The grass marshes are along the coast where salt water bathes the roots of the plants, while freshwater swamps are above tide. Davis,’° in discussing the freshwater de- posits of Michigan, employed the terms bog, marsh and swamp; a bog is an area of wet, porous land, whose soil is mostly decayed or decaying vegetable matter, loosely consolidated and containing so much water as to tremble when one walks on it; the vegetation varies, but usually consists of mosses, sedges or grasses, or a combi- nation of them along with shrubs and even small trees; a marsh does not shake readily when one walks over it, though it may be very soft and wet; the vegetation is mostly grass-like, though shrubs may _ be present in thickets; a swamp soil is firm, but wet, even to flood- — ing at times, and bears trees and shrubby plants as the most impor- a tant part of the vegetation. This grouping is not absolute, for the — types may all be found in a single basin, the passage from one to f the other being very gradual. . In Europe, where peat is of great economic importance, ome students have expended great ingenuity in efforts to classify the *” N.S. Shaler, “ General Account of Freshwater Morasses of the United “f States,” Tenth Ann. Rep. U. S. Geol. Survey, 1890, pp. 261 et seq. Bs. ™C. A. Davis, “ Peat,” Ann. Rep. Geol. Survey of Mich., 1907, pp. 10 100. 186 r91t.] STEVENSON—FORMATION OF COAL BEDS. 589 deposits, which are, practically all of them, freshwater, marine marshes being unimportant economically. Lesquereux, in 1845, rec- ognized two general types of bogs, which he termed supraaquatic or emerged and infraaquatic or submerged, the former being above the waterline and the other at or below it. The prevailing classifi- cation in Germany recognizes the Hochmoor, equivalent to the Heathermoors of Scotland, and, in great part, to the supraaquatic of Lesquereux; the Wiesenmoor, Griinlandsmoor, Niedermoor, or Rasenmoor, equivalent to the bogmeadows of other lands; and the Waldmoor or forested bog. These are the Lyngmose, Svampmose, Hoermose; the Kjaermose, Engmose; and the Skovmose of the Danish authors. A similar division is that of Hochmoor, Flach- moor and Zwischenmoor, these being the terms employed by Potonié and some recent authors. Potonié*™ has described a great moor in east Prussia on the delta of the Memel and Nemonien rivers, which shows the relations of the several types. Going eastward from the shore, one finds first the mud, which on the border is held by water lilies and other plants, referred to as “landmakers” because they are outposts. Higher plants, especially canes, occupy water areas, behind which there develops a meadow Flachmoor of sedges, where frequent floodings prevent growth of trees. Beyond that, foresting begins, and one reaches a moor of black alder, several kilometers broad. The sur- face is occupied by swamp plants, such as Jris, Sium, sedges, which endure well the periodical floodings of this zone. If the area were one of gradual subsidence, equal to the accumulation, the condition would continue for a long period. The surface rises gently and one comes to another flora, accustomed to somewhat drier soil, with alders, hops and nettles. Thus far, one has followed the Flachmoor or Niedermoor ; but at a little distance beyond, swamp birches are seen among the alders. The latter soon disappear and the birch zone is reached, beyond which is a zone of forest with Pinus sylves- tris and Picea excelsa. These form the Zwischenmoor or Waldmoor. In this passage zone, the peat has risen so high that the surface is dry; the forest is here, but as one advances the trees become ™ H. Potonié, “ Die Entstehung,” etc., pp. 35-40. 187 590 STEVENSON—FORMATION OF COAL BEDS, _ [November 3, smaller because increasing accumulation of peat deprives them of their nutriment. Even exceptionally high floods cannot bring dis- solved nutriment to them and they are dependent on rain, dew and dust. At length, the trees are displaced by Sphagnum, able to store away dew and rain, to remain moist on even a dry bed, to keep the © area wet though it may be several meters above the water level. So one, in going eastward, is still on wet land. This is the Hochmoor, swelling as an hour glass—whence its name. But there is a still higher stage. On the Hochmoor, one’s foot sinks deeply into the sphagnum-peat as he advances. At length a pond is reached; the rain collects in pools or small lakes, whence it flows to moisten the __ surrounding area. Plants thrive here because the changing water gives them nutriment. Reeds and sedges are seen and even Pinus: sylvestris is present, though much smaller than on the borders of the Flachmoor. This great bog rests on a sandy deposit, with which are _ mingled the vegetable muds of the kurischen Haff. As the problem of formation of coal beds is world-wide in scope, the essential features of those beds being practically the same in all lands, the study of peat accumulations must be as broad as possible, if the conclusions are to possess any worth for or against any theory. In the pages to follow, the results of studies by observers in many regions will be presented in detail. -This may involve some repeti- tion, but that will serve only to emphasize the importance of certain conditions, which have been overlooked or ignored in some contri-— butions to the discussion. Peat Deposits in the United States of America—Marine marshes exist in extensive areas along the Atlantic coast from Maine to Florida, a region believed by nearly all observers to be subsiding. North from Florida, the tidal marshes are grass-meadows, ordinarily treeless. They are covered with grasses, reeds or coarse sedges and the upper surface is near the level of high water. Cook’? has described those of New Jersey, which are typical of the whole coast — from Georgia northward. Alongside of streams crossing the marshes there is a narrow ridge of dry land, but within a few yards one 2G H. Cook, “Geology of New Jersey,” 1868, pp. 24, 231, 233, 238, 347-350, 361. 188 1911.] STEVENSON—FORMATION OF COAL BEDS. 591 reaches the permanently wet area. Immediately below its sod, is mud or soft earth, which varies greatly in composition. Near the creeks, it is usually fine clayey mud with embedded roots, the whole evidently transported material; at a little distance, it is black earth or muck, formed in a swamp; while at a greater distance one finds only a mass of fibrous roots and vegetable matter with no admixture of earth or mud. The last twoare of in situorigin. The “meadows” along the Passaic» and Hackensack rivers, emptying into New York harbor, show an extreme thickness of 32 feet of “mud” resting on 8 feet of blue clay, while farther up the stream are great marshes resting on fine sandy material. One sees on the surface of these meadows great numbers of white cedar stumps and the mud is crowded with remains of cedar timber. The condition is due to encroachment by the sea, whereby the treeless marshes advance inland and overrun the white cedar swamps along the streams; one finds at many places the old cedar forest buried in the tidal marsh, while the cedar swamp still exists at a little way beyond. The salt water kills the freshwater grasses and the trees on the border. In many places trees flourished 80 years ago, where one finds now only salt marsh muck. The white cedar is a very durable wood; trunks of trees killed by the salt water are still standing in localities where several feet of muck have accumu- lated around them. Lyell’ observed the effects of this encroachment in Georgia. In coming down to the coast, he found the trees becoming dwarfed and at length disappearing to be replaced by reeds ; but in the marshes he saw the stumps and stools of cypress, still retaining the erect position in which they had grown. He quotes Bartram, who stated that when planters, along the coast of the Carolinas, Georgia and Florida, as well as westward to the Mississippi, bank in the grassy tidal marshes for cultivation, they “cannot sink their drains above three or four feet below the surface, before they come to strata of cypress stumps and other trees, as close together as they now grow in the swamps.” “CC. Lyell, “Second Visit to the United States of North America,” London, 1850, Vol. I., pp. 334-336. 189 592 STEVENSON—FORMATION OF COAL BEDS, _ [November 3, When one reaches southern Florida, he finds a different type of tidal marsh. Northward, grasses and rushes are the plants which advance the land seaward, but at the south the mangrove is the agent. That plant abounds on coasts in tropical America and is found northward in Florida to lat. 30°, though it is not abundant above lat. 26°. The eccentric mode of growth exhibited by the special type under consideration, long ago attracted the attention of botanists. Bancroft? says that it rises from several strong woody roots which emerge from the ground for two or three yards before they unite at the trunk. Tough woody shoots, about three inches in circumference, descend from the trunk to take root and, as the tree increases in size, the shoots increase in number. These by their strength compensate for the looseness of the soil. The tree grows in a low, wet soil by the side of running water. The Florida mangrove flourishes only in contact with salt water, being stunted by brackish water. Vaughan’ describes it as attain- — ing the height of 10 to 20 feet and as growing in water or so near it | that the soil is saturated. The long seeds take root in water not — more than one foot deep, leaves being put forth as soon as the sur- ; face is reached. Besides the tufty roots given off at the base, there are others originating at higher levels from the stem, which grow downwards and embed themselves in the soil. Shaler says that these can descend through 8 feet of water in order to take root. Each becomes a new tree to be multiplied in similar manner. Th a tree may advance 20 or more feet in a century, the advance being checked only when the water is too deep or the waves prevent root- ing. These growths, as described by Shaler and Vaughan, fo dense thickets, a fringe, which is made denser by litter from trees; so that débris from the land eventually fills up the sp behind and the trees are killed. But, in the interval, a new fri has been formed. In the moist area behind the growth, freshy types displace the saltwater forms and a swamp results. Sh ™ FE. Bancroft, “An Essay on the Natural History of Guiana in America,” London, 1779, pp. 76-79. * TW. Vaughan, “ Geologic Work of Mangroves in Southern Florida, Smithson. Misc. Coll. Quart. Issue, Vol. V., 1910, pp. 461-464 190 1gtr.J STEVENSON—FORMATION OF COAL BEDS. 593 conceives that the Everglades of southern Florida, with an area of about 7,000 square miles, owe their origin to outward advance of mangroves on shallows of the coast. The freshwater swamps of the Atlantic and Gulf coasts are, for the most part, sharply distinct from the tidal marshes, even where the latter have encroached. Cook*®* has shown the relation in New Jersey by a section extending from Dennisville to Delaware bay, a distance of about one mile. The cedar*** swamp begins at the edge of the low upland and gradually deepens to 15 feet. Like most of the cedar swamps in New Jersey, it has been cleared, but clusters of young trees up to 100 years old remain here and there on the surface, which is only a few feet above high tide. The cedar grows densely but slowly. Old stumps show more than 1,000 annual rings, but those near the bark are as thin as paper and the stumps rarely exceed 3 feet diameter, though some have been seen which were 7 feet. The swamp soil is black, peaty, 13 feet thick at Dennisville and, when dry, burns. It shows no admixture of foreign material and contains only 3.35 per cent. of ash, the water in the dried peat being from 12 to 16 per cent. It is very loose and porous, always full of water; the roots of the trees run through it in every direc- tion near the surface, but do not penetrate to the solid ground. Where the peat cover is thin, the roots do pass through to the under- lying soil, but, in that case, the wood is inferior and it cannot be utilized in manufactures. Trunks of trees are buried at all depths and are so numerous that one has difficulty in thrusting a sounding rod to the bottom. Some had been blown over when rotten; others were merely uprooted. Some, blown down, lived for a considerable time afterward. The prostrated trunks lie in all directions and the conditions are precisely the same as those now seen on the surface of the swamp. Large stumps have been found, which grew over logs, now enveloped by ™ G. H. Cook, op. cit., pp. 301, 302, 355, 356, 360, 361, 484. ™ The cypress or white cedar of New Jersey is Chamecyparis thyoides, which is found in swamps from New Hampshire to Florida and westward to the Mississippi. The bald cypress is Taxodium distichum, a form surviv- ing from the middle Tertiary, which extends from southern Delaware along the coast to Texas and up the Mississippi to southern Illinois. 191 594 STEVENSON—FORMATION OF COAL BEDS, _ {November g, their roots, and at the bottom are found worthless logs of cedar belonging to trees which were rooted in the solid ground below. Shaler™®* has given a general description of the Dismal Swamp, — an area of about 500 square miles, at only a few feet above tide level. It was much larger, but a great part has been reclaimed by draining. The peaty deposit rests on Pliocene sands, of which 10 to 14 feet are exposed on the border; but this is not wholly certain as the bottom has been reached at only one place within the swamp. C. A. Davis has stated recently that the peat is at least 15 feet thick — and of good quality. On the western border is Drummond lake, — 6 feet deep and somewhat more than 2 miles wide. Shaler says — that, liere and there within the swamp, one comes to drained areas of considerable size, one of which, embracing about 2 square miles, A has long yielded fine crops of maize. He notes that Sphagnum has a very small place in this swamp and that it is an unimportant factor everywhere south from the Potomac and Ohio rivers, where the greater heat and decreased rainfall prevent its luxuriant growth. The most important peat-making plants in the region south from those rivers are canes, a grape, the bald cypress and the juniper with, in some localities, the dwarf palmetto—among these, he assigns the chief place to the common cane. The greater part of the Dismal Swamp is under water during most of the time, but there are elevations rising not more than 3 feet above the general level; yet the drainage due to this slight ele- vation suffices for growth of pines belonging to the common southern species. Inthe main area, water-covered, one finds three trees, Taxo- dium distichum, the bald cypress; Juniperus virginiana, the juni- per; and Nyssa sylvatica, the black gum. The juniper occupies usually somewhat desiccated during the dry season, but the oth being provided with special appliances, live where their roots covered with water during even the growing season. The fores very dense and passage through it is rendered difficult by projec knees of cypress and the arched roots of Nyssa, while everywh is a profusion of other plants. The surface is covered with a lit *8N. S. Shaler, Tenth Ann. Rep. U. S. Geol. Survey, pp. 293, 313, Pl. 8, 9, 10, Fig. 209. 192 tgrr.] STEVENSON—FORMATION OF COAL BEDS. 595 of fallen trunks, twigs and leaves. Shaler’s plates from photographs taken in this forested swamp show the conditions thoroughly. Shaler*®® has described the peculiar modification of structure characterizing the bald cypress. This is the greatest of the conifers east from the Rocky mountains and it is the most stately of all the trees on the eastern half of the continent. On dry ground or where there is no water during the summer half of the year, it shows no peculiarities ; but where it lives in swamps, flooded during the grow- ing season, the roots give off excrescences which project above the water, their height depending on the depth of water. These “knees” are subcylindrical and are crowned by a cabbage-shaped expansion of bark, rough without and often hollow within. Whenever these __ knees become permanently submerged during the growing season, 4 the tree dies ; as was proved in the New Madrid area, where, during 4 the 1811 to 1813 earthquakes, the land sank permanently. In Reel- foot lake, within Kentucky and Tennessee, thousands of these long cypress boles still stand in the shallow waters, though 70 years have passed since the slight submergence of their knees. The effect of * drowning is shown on a plate in the work previously cited. Many : 4 _ dead stems of cypress rise above the surface of Drummond lake, ____ which is only a few feet deep. Lesquereux thought that these were once part of a floating forest. Okefinokee swamp in southern Georgia is not wholly a forested swamp. It is larger than Dismal Swamp and more difficult to study. Harper**® succeeded in penetrating it to a distance, all told of about 18 miles. Here and there are islands, raised a little above the swamp level, at times not more than 2 feet, often less. On those the slash pine and the black gum grow, while all around are sphagnous bogs in which are slash pine, as well as swamp cypress, with sedges, ferns, sundews, and pitcher plants. Pines are wanting where the muck is more than 4 feet deep, but the cypress grows densely until the depth exceeds 6 feet. Where that depth is exceeded, no trees are found WN. S. Shaler, “ The American Swamp Cypress,” Science, O. S., Vol. IL, 1883, pp. 38-40. 2°R. M. Harper, “ Okefinokee Swamp,” Pop. Sci. Monthly, Vol. LXXIV., _ 1909, pp. 596-613. 193 596 STEVENSON—FORMATION OF COAL BEDS. _ [November 3, and the surface is a “prairie.” This type has an area of 100 square miles in the western part of the swamp, covered everywhere by water in wet weather, so that one may go in any direction in a canoe. Canes, pickerel weed and water lilies abound but Sphagnum is absent, as in this latitude it can grow only in shaded places. Stumps — of cypress are abundant and the peat is about 10 feet thick. The Florida swamps, described by Harper and others in the official re- ports, show all types from the open marsh to the forested swamp. The cypress swamps of the Lake region have grass marshes near the water, which are separated from the dense cypress growth by a narrow belt of small willows. The peat in these deposits is worth little commercially, as it is crowded with logs and woody roots. The great Everglades area belongs to the stagnant water type. The cypress swamps of the Gulf coast are like those of the Atlantic coast. Lyell’ relates that, in excavating for the founda- tions of the New Orleans Gas Works, the contractor soon discovered — that he had to deal not with soil but with buried timber; the diggers were replaced by expert axemen. The cypress and other trees were — “superimposed one upon the other, in an upright position, with — their roots as they grew.” The State Surveyor reported that in — digging the great canal from Lake Ponchartrain, a cypress swamp was cut, which had filled gradually, “ for there were three tiers of stumps in the 9 feet, some of them very old, ranged one above the other ; and some of the stumps must have rotted away to the level of the ground in the swamp before the upper ones grew over them.” Conditions in the cypress swamps are the same throughout, whether the prevailing tree be bald cypress or white cedar. The peat is formed by accumulation of litter in the dense forest and, fe é the most part, the swamps are due to impeded drainage on an level surface. The trees are rooted in the swamp material, which at times is of great thickness, more than 150 feet of muck, carrying cypress trees on its surface, being reported from Florida. Suc trees find ample nutriment in peat containing less than 4 per cen of mineral matter and they do not send their roots down to the s ground. One sees growing amid such conditions not merely sh ™ C, Lyell, “Second Visit,” etc., Vol. II., pp. 136, 137. 194 tort.) STEVENSON—FORMATION OF COAL BEDS. 597 but also majestic trees, such as cypress and gum, which, as well as the less imposing juniper, yield wood of great importance to the artificer. The inland swamps of the northern states differ in many ways from the coastal swamps. They occur along river borders or in lakelet areas of the drift-covered region. In great part, the former are “wet woods” covered more or less deeply with water during several months of each year, but they show considerable stretches of true swamp. The swamps and marshes of the drift region are less extensive, but they afford better opportunity for studies bearing on the mode of accumulation. They have been investigated by C. A. Davis, H. Ries, N. S. Shaler and others, but the most comprehensive and most recent description is by Davis. Davis’ notes that very few highly organized plants can grow wholly submerged in water, and those are mostly endogens; 10 feet of depth seems to be the limit, although Potamogeton has been found rooted in 23 feet; other types, low forms such as Chara and the floating algz are indifferent. Some plants, burweeds, arrowheads, reed grass, pickerel weed and water lilies can grow when partially submerged; while some land plants, shrubs and trees can endure long exposure to water about the roots. The surface growth on swamps is important. Elm and black ash swamps are of common occurrence and have, besides those plants, tamarack, spruce, willows, alders, with various heaths and mosses. They do not always show much peat, but what there is is well decomposed and is apt to contain much mineral matter. The greatest thickness of peat in these swamps is reported to be 10 feet. Tamarack (Larix laricina) and white cedar (Chamecyparis thyoides) indicate the presence of peat, the latter growing densely on the surface of a deposit, 20 feet thick. Spruces (Picea mariana and P. brevifolia) also grow on thick peat; willows, poplar and alders grow on the thickest peat and in wet places; but the mosses, sci and Sphagnum, grow only in advanced swamps. OC A. Davis, “ Peat, Essays on its Origin, Uses and Distribution in Michigan,” Rep. Mich. Geol. Survey for 1906, pp. 121-125, 128-134, 136-141, 153, 154, 157-159, 160-166, 203, 204, 208, 213, 260,.275, 279, 291. 195 598 STEVENSON—FORMATION OF COAL BEDS, [November 3, — Peat deposits fill depressions but, in some cases, are formed on almost level areas. Depressions more than 25 feet deep may be filled by alge, by floating species of seed-bearing plants, by sedi- — mentation, by plant growth from the sides or by a combination of these processes. A frequent succession is Chara-marl, on which | rests a peaty soil in which plants take root; the land marsh moves out and tamarack advances on the deeper peat of the shore. As the water becomes shallower, each shore type moves out and is suc- ceeded by the type behind—the water growing warmer and more aérated. Formation of peat on a flat space is much under the same — conditions as those on the surface of a filled depression. When the drainage is poor, liverworts or some mosses take possession; if not — too wet, rushes, sedges and grasses appear. Accumulation makes — the place wetter and only the hardier plants remain. Sedges are the chief peat-producers under these conditions. . The process of filling a depression is often very complicated. In southern Michigan, the early stages are shown in many lakes, which are surrounded by zones of aquatic plants. More or less detritus, organic and inorganic, finds its way into the lake. Where the process is more advanced one can trace the whole succession, __ The lowest deposit is formed of Chara and floating alge. — This is succeeded in the shallower water by the Potamogeton zone an that by the water lilies. Just beyond this one comes to the floating mat of sedges, extending on the water surface to a considerable distance from the shore and buoyant enough to support a consider- able weight. The earlier stages may provide soil for rooting of the sedges at the shore line, but the mat itself is wholly unsupported for — a considerable distance and is often 18 inches thick. Finely divide material from the undersurface of the mat increases toward: shore, where it becomes dense and the mat is no longer fl Thus is built the solid peat, structureless, decomposed and nearly black. The surface rises gradually after grounding of the mat an at each level, new plants appear. Shrubs and Sphagnum ad to be overcome in turn by tamarack’and spruce, which in their t are overcome by the marginal flora from behind. Tamarack acco panied by ferns grows far out on the bog. 196 1gtr.] STEVENSON—FORMATION OF COAL BEDS. 599 The final stage is where the sedge mat closes over the surface and the underlying peat has become firm. Sedge is usually the chief factor in the later stages of lake destruction. At times, the mat is pressed down by the weight of trees growing on it. In one case it was found 6 feet thick, resting on semi-fluid peat. A section at one locality showed Feet. Inches. Ei. SOMRENOUR HORE Coe Sis ks ce hee ws ee sis oe ° 6 =. Moss peat and shrubs <.... 265s oh. ccs 2 ° i BARE DORE SS coe eeaake aussie o 3 4. Coarse brown peat, stumps and roots... 2 6 ay mneine: Of SNEUDE. ooo ses oc os owe o 2 6. Dark peat rich in sedge remains........ 2 oO It was impossible to determine the condition farther down as the peat was very wet, but sedges were recognized. Similar conditions were observed in other sections. These all show that the trees were rooted in the mat of pure vegetable material, even when it reposed on the water surface and that, while the trees were growing, the accumulation of peat was continuous. After the mat has been grounded, Hypnum hastens outward from the shore, associated occasionally with some Sphagnum. When the surface rises 2 inches above the water level, ferns appear and they are followed by Sphagnum, which persists even when the sur- face is flooded. It is much hardier than Hypnum and, for that reason, it has been regarded as chief factor in the production of peat. But it is often absent, having been found in less than 30 per cent. of the localities examined by Davis. The first tree is the tama- rack, which grows densely on the level of shrubs, but isolated trees are scattered over the open bog. Chara-marl occurs frequently in southern Michigan but it was not seen anywhere in the northern portion of the state, where the _ general succession differs somewhat from that already given. The _ Chara-stage is wanting; pond weeds, pond lilies and rushes are of ; - irregular occurrence and the sedge-zone is all important. Owing, _ probably, to absence of fragments belonging to the higher plants, _ the work of freshwater alge is more apparent than in the southern PROC. AMER. PHIL. SOC., L. 202 NN, PRINTED NOV. 16, Igit. 197 600 STEVENSON—FORMATION OF COAL BEDS. [November peninsula. Algal lake, now covering only a few acres, is surrounded by a great wooded swamp, extending northeastward to a large lake - and coming down almost to the water at the north end of Algal lake. The swamp loosestrife (Decodon verticillatus) forms the marginal zone. The bottom of the lake is covered with soft flocculent ooze, composed of unicellular alge with diatoms as well as pollen from conifers. Davis conceived that peat of this type would be like cannel and he thinks that freshwater alge may have been more abundant in Carboniferous times, when all types of plant life were lower than now. A similar material was found in a mature bog, where the section is Feet, Coarse peat, with stumps, roots and fallen stems..........-... 5 2. Brown peat, good texture, quite plastic ................. rich Be 3. Soft, light-colored peat, like that at Algal lake ........ Laie ae = . These are the only localities in the United States whence this type of peat has been reported. Ehrenberg, Frith and Potonié have d scribed the felt or Meteorpapier, as Ehrenberg termed it, which remains on swamps after floodwaters have been drained off. Potonié calls it Sapropel carpet, and he has given a photograph showing thi material covering land plants of a swamp. But the phenomenon of by no means rare occurrence in the eastern part of the Uni States. Davis has communicated by letter that he saw it in 191 near St. Augustine in Florida, where the water of a swamp been lowered ; the felt was conspicuous on the tussocks, ete. I Everglades of the same state, he found the felt about the grass sedge stems in the level swamps. Here and there it contain considerable quantity of calcareous matter, due perhaps to acti of Cyaphanacee present in the algal association. The same type felt-like development is seen during springtime in marshes o northern states, where the water drains off slowly. Spirogyra other filamentous algze sometimes cover the temporary ponds ; are left as a felt-like cover when the water has been withdr; This felt breaks into small pieces as it dries and is added to the p The writer has observed it in very small patches on the New Jer marshes; he has seen patches more than 10 feet square at 198 19tt.] STEVENSON—FORMATION OF COAL BEDS. 601 places in Rhode Island and Massachusetts. But in every case, the quantity is insignificant as compared with the mass of other vege- table material and this algal contribution must be wholly unimpor- tant. At the same time, one can conceive of conditions which could render it important. Shaler expressed the prevailing opinion when he asserted that the presence of moisture determines the distribution of plant life in swamp areas. Advance of swamp destroys the forest. He had seen many places on the coast of Maine as well as in northern Mich- igan and Wisconsin, where invasion by Sphagnum made the surface so wet that even the most water-loving trees of those regions could not maintain themselves. Davis, in his work on Michigan peats, has discussed the causes leading to the succession of vegetation in swampy areas. The shrubs growing at the water level are drought plants, though living where water is abundant; their leaves are linear or even scale-like ; the cuticle is dense and the leaves are pro- tected by a waxy or at times resinous coating—all contrived to prevent too rapid evaporation. The explanation of the condition is complex, but it depends mostly on the difficulty with which moisture _ can be extracted from peat. Once thoroughly air-dried, peat is a almost impervious to water, so that plants growing on peat or a peaty soil suffer more from drought than those on other soils. Even when wet, it has little water for plants growing on it. A noteworthy fact in this connection is that some plants, growing near water level in southern Michigan, are found growing only on dry soils in north- ern Michigan. They find their drought-resisting ability equally _ essential in both regions. The distribution of these plants is ex- _ plained by the fact that they have fleshy fruits, which birds eat _ during their southward migration and the seeds are scattered over - moist areas. While the plants must be able to resist drought, they _ must be able to endure excess of moisture in some localities. Davis _ saw Betula pumilla and some willows living in places where their roots had been covered with one foot of water for several years. The conditions of advance described by Davis are familiar in other states. They exist even on high swamp areas, as appears 199 602 STEVENSON—FORMATION OF COAL BEDS. _ [November 3, from Bradley’s'’** notes on the disappearance of meadows which were used as camping places in the Sierra Nevada. Fifteen years ago, these were open and covered with abundant grass. Originally, they were ponds or lakes which became filled with peat, on which grass thrived. As the material became less wet, tamarack seeds, blown in from the border, took root, but the young shoots were killed by the frequent fires. Since protection against fire has be- come complete throughout the region, the tamarack has advanced so as to occupy much of the surface, while pines are encroaching, which eventually will crowd out the tamarack and will occupy the whole area. The trees are rooted in the peat. a Bates"*® has shown that swamp conditions and luxuriant growth of trees are not incompatible. In describing the forests of Para, he says that one swampy area was covered with trees more than 100 feet high, all of second growth. In another swamp, the air was — marked by a mouldy odor, the trees were lofty and the surface was | 4 carpeted with lycopodiums. Farther down in this area, where the ground was more swampy, wild bananas, great palms and exogens grew luxuriantly and were covered with creepers and parasites; while the surface was encumbered with rotting trunks, branches, leaves, and the whole was reeking with moisture. Kuntze, already a cited, states that the tropical swamps are densely wooded. Obser- vations by other authors will be referred to in another connection. — Peat Deposits in Europe.—The importance of peat as fuel in Europe has led to thorough investigation of that material from every conceivable standpoint. The literature is so extensive and, in great _ part, so excellent that one, compelled by limits of space, finds him- self embarrassed in selection of authors as well as of matter. | Lesquereux™* long ago proved that Sphagnum is not the impor- tant factor in peat-making ; he recalled attention to Ad. Brongniart’s ' ™*H. C. Bradley, “The Passing of Our Mountain Meadows,” Sierra’ Club Bull., Vol. VIIL., ro11, pp. 39-42. “* H. W. Bates, “ The Naturalist on the River Amazons,” London, 186 Vol. L., pp. 44, 47, 50, 51. ™ L. Lesquereux, “ Quelques recherches sur les marais tourbeux,” PP. 32 III, 121, 137; 2d Geol. Surv. Penn., Rep. for 1885, pp. 107-121. 200 1911.] STEVENSON—FORMATION OF COAL BEDS. 603 observation that evaporation from that moss is proportionately less than from other plants; and he showed that growth of the moss is checked by freezing and that the plant cannot live in deep shade or under forest trees such as oaks, pines or beeches. He seems to be the first to note that marls covering peat bogs contain impressions of plants. Lesquereux’s conception of the mode of filling depressions from the sides differs somewhat in detail from that given for the United States. Shallow ponds are invaded by vegetation, which forms a mould in which water plants take root. The basin is filled by their decay, the surface becomes humus in which plants of other types _ grow, giving meadows or forests. The filling is rapid in the early stages. Pools of quiet water are invaded by conferve, mingled with _ infusoria, microscopic plants and small shells, which by decay cover the bottom. At times, 6 to Io inches of this deposit may accumu- late in a year. When the water is deep, the same result is reached by another process—the prolonged growth of certain floating mosses, especially of some species of Sphagnum. Those, pushing out from the sides, form a thin cover, in which grasses, sedges and other water-loving plants grow. Eventually, this becomes compact enough to bear the weight of trees, even of dense forest ; until, becoming too heavy, it either breaks or is pressed slowly to the bottom and covered with water. This, he asserts, is no hypothesis but the statement of actual fact. The lac d’Etailléres, near Fleurir in Switzerland, is open water in an extensive series of peat bogs. Prior to the year 1500, it was the site of a forest; but in that year, according to legend, the forest __. disappeared and it was replaced by two lakes. The lakes still exist __and in quiet water one can see the prostrate trees on the bottom. _ But a newcarpet has already spread over much of the surface, which in turn will become forested and will sink. Thus one may find superimposed beds of decomposing vegetable matter, each consisting of remains of small plants below but of forest remains above. An analogous condition exists in Lake Drummond of the Dismal Swamp, where the bottom consists of a forest cover, once at the top but now 201 604 STEVENSON—FORMATION OF COAL BEDS. _ [November 3, under water, while vegetation is encroaching from the sides. It is quite possible that this explanation of the Lake Drummond condi- tion is correct, but that lake is shallow, only 6 feet, and the trees are erect; in the deeper lac d’Etailléres, the trees were prostrated by breaking of the mat. To illustrate the succession in such a case, he gives the section of a bog in Denmark: Feet, Inches. 1. Fibrous yellow peat with undecomposed mosses....... is ae 2. Oak layer, wood still sound, trunks 2 to 3 ft. diameter. 4.° Peat, yellowiay Sick cde ces dacs occ reek eee 6 4. ‘Birches, prostrate, Betula alba i... 000.00 is ee eee 3 °o S, Rlack peat je 3cc sc 5 cb cws mie niecee alae send aeeo eee 4 ° 6. Pines, 6 to 10 inches diameter, most of them pointing toward center of the basin, retaining their branches, embedded in a mass of leaves, cones, etc. ..........+. 8 a 7, Black ‘comtpact peat: 5..c< igcicss oe onnaves ervey eee 4 ° and the bottom not reached. This peat was mined for fuel, the works being extensive. The general description by Lesquereux shows that the conditions are not wholly the same in his localities as in many areas within the United States. They suffice to show that Sphagnum is a late arrival, though in Switzerland, as in some other portions of Europe it is more important than in this country, where Sphagnum-peat rarely exceeds 3 feet. As illustrating this, one may cite Vogt’s'® description of a Hochmoor at the Ponts of the Canton Neuenburg. This lies be- tween two villages built on limestone benches on opposite sides of a valley. In the middle ages, each village was visible from the other, but that is no longer the case. The bog has raised itself, hill-like, growing most rapidly along the middle line. This mass is Sphagnum and its mode of growth shows well the ability of that moss to retain water, so as to thrive at considerably above the water level. Heer’® says that life on land began with minute forms and few types. So, in the water, alge begin the work. Even pure fresh- * C. Vogt, “ Lehrbuch der Geologie,” 2te Aufl., Braunschweig, 1854, Vol. Ef... 0. 110: ™*O. Heer, “Die Schieferkohlen von Utznach und Diirnten,” Zurich, 1858, pp. 1-4. 202 rgtt.] STEVENSON—FORMATION OF COAL BEDS. 605 water, exposed to air and light, is full of minute plants, with bound- ~ less capacity for multiplication, forming in vast legions, which sink and form a layer of organic material, the basis of formations com- posed of higher organisms. These are followed by floating mosses, which, in spite of their small size, soon produce a great mass of organic material. The bladderworts, water milfoils follow and the water lilies spread their leaves over the surface; reeds press out from the shore and sedges of various kinds form a wickerwork of roots, which gradually spread over the whole depression and water is no longer visible. Meanwhile the peat has been growing denser, drawing water from below and keeping the bed moist. In it nestle the milfoils and heaths. The lake closed, woody plants encroach, Betula and then Pinus sylvestris. But the latter does not grow high, breaking off after attaining a certain height and weight, sinking into the underlying soft material, there to be destroyed and converted into peat as are the shrubby plants. These trees are readily over- thrown by the wind and the peat is crowded with the overturned trunks of birch and fir. The harder parts offer prolonged resistance to chemical change and are embedded in a pulp-like mass derived from the softer parts. The conditions in all stages are recognizable in Swiss deposits. The succession may be varied by climatic changes, whereby a Waldmoor may be converted into a Torfmoor and that in turn into a Waldmoor again. Frith’s** descriptions of conditions in Switzerland and Germany are much like those given in later years for localities in the United States, though the succession of events may differ somewhat in detail. At the same time, the Hochmoor or Sphagnum deposit seems to be built up on the Rasenmoor, composed of Cyperacee, Phrag- mites and Hypnum; islands of Hochmoor were seen occasionally in a Rasenmoor. Lorentz is cited as having examined 57 moors, of which 31 were Hochmoors developed on Rasenmoors. Frith inves- tigated Hochmoors in Steiermark, the Bavarian highlands and in Switzerland, all of which showed that Sphagnum is a late arrival in ae rs nent 7 le a ** J. J. Frith, “Ueber Torf und Dopplerit,” pp. 5, 7-0, 15, 18, 20. 203 606 STEVENSON—FORMATION OF COAL BEDS. _ [November 3, * the peat. In the great Digenmoors of the Bavarian highlands he found Meters, 1. -Black peat, with SPhGgnewe is cc vcsnseesacdvescpabene ir t6t2 2. Homogeneous black-brown, compact, plastic peat, with layers of crushed birch stems; a few specimens of Sphagnum, but 90 per cent. of the mass consisting of roots Of CUPOIACOde 4s soso ico k ee cd nee ORR ae I to Ls 4, Wood lager of conifers 3.66055. icvsceas sere 0.4 to 0.6 4 Glacial drift. He gives measurements from fourteen localities in Switzerland, only one of which failed to show the succession observed in the section. The exception is a Hochmoor without Rasenmoor foundation and resting directly on a layer of wood remains. One group seems to contradict Sendtner’s generalization that Hochmoor accumulates only in localities where the water is not calcareous. This, the “ Todte Meer,” is a typical living Hochmoor, near Willerszell, bearing on its surface many hummocks nearly equal in height and basal diameter, and bordered by a mountain stream, whose drainage area is in a limestone region. It shows 1. Hochmoor, Sphagnum ............ Ne buiesemewseneoene 0.2 to 0.3 2. Felted Rasenmoor, upper part consisting of Carex and Arundo, with scattered alge; lower part with ae daa URN a Darya ar are ete OM res IE rear Presto 3 3. Almost pure well-preserved Hypnum. 4. Clay and gravel. He finds a simple explanation in the fact that the stream, at high water, does not wet the Sphagnum. It may be well to note here that in Michigan, according to Davis, Sphagnum is indifferent to the character of the water, the presence of calcium carbonate in no wise affecting its growth. Frith reports 48 Hochmoors in the Alpine region as originating on Rasenmoors. V. Bemmelen and Staring are cited as having proved the same relations for the provinces of Orenthe, Friesland and Gottingue in Holland. The Rasenmoor does not require hard water, for the vast moors of the Rhine and Maas area are watered by those streams, which contain only 65 and 41 millionths of calcium — and magnesium compounds. The relation between Hochmoor and 204 i9tt.] STEVENSON—FORMATION OF COAL BEDS. 607 Rasenmoor is not always apparent as either one may be very thin and the other very thick. In his later, great work on the Swiss moors, Frith has described with much detail all the Swiss deposits 4 and he has offered generalizations which will be considered in 4 another connection. 4 It had been suggested by some observers that the tree trunks _ found in the bogs had been drifted into the depressions, but Frith . a asserts without qualification that they are in place. The condition is wholly normal. A. Geikie,*** after noting the differences in phys- ical structure as well as in vegetation shown by successive portions of a bog, says that remains of trees are common. Some are em- bedded in soil underneath the bog; others are in the heart of the a peat, proving that the trees lived on the mossy surface and finally were enclosed in the growing peat. This is illustrated by a sketch "__ of a peat-moss in Sutherland. J. Geikie*?® has given much informa- E tion respecting the Scottish bogs but it will suffice to cite only his t a later work. The bogs have yielded many species of trees, all of ; them indigenous. The trees are im sttu, each rooted in the kind of | % soil preferred by living examples. There are few acres of lowland bog in which trees have not been found. They occur even in the Hebrides, where trees now are practically unknown. Occasionally, more than one forest bed is present. At Strathcluony, three tiers of Scotch fir were seen, separated by layers of peat. Several tiers were exposed in a railway cutting across the Big Moss; one of stand- ing fir trees with branching roots at 6 feet below the surface, a second at 12 feet and a third at 4 feet lower; so that, counting the surface growth, four different forests have existed there since the bog began. Aher,**° in the Bog reports, says that trees in the Irish bogs “have generally 6 or 7 feet of compact peat under their roots, which are found standing as they grew, evidently proving the formation of the peat to have been previous to the growth of the trees.” On ™ A. Geikie, “ Text-book of Geology,” 3d Ed., London, 1893, pp. 478-480. ™ J. Geikie, “ The Great Ice Age,” 3d Ed., London, 1895, pp. 286-293, 303. ™ Cited by S. S. Haldeman, in 2d Ed. of R. C. Taylor’s “ Statistics of Coal,” Philadelphia, 1855, p. 160. 205 608 STEVENSON—FORMATION OF COAL BEDS. _ [November 3, the same page Haldeman notes that it is a remarkable fact, although very common, that successive layers of trees or stumps, in erect position and furnished with their roots, are found at distinctly dif- ferent levels, at small vertical distance from each other. Grand’ Eury,’*? noting that the plants, active in peat-making, are not the same in all cases, maintains that a distinction must be made between peat, properly so-called, and peat of the marais. The former is supraaquatic, covers high plateaus and is formed chiefly by Sphagnum, with some other water-loving mosses. Unaccom- panied by these, other plants in similar conditions give only soil. Such peat is rarely transformed into a compact charbon and it is obscurely stratified. The peat of marais is formed on low grounds, along the borders of rivers, lakes or the sea, often in extensive areas. In such places, Arundo grows rapidly along with Scirpus palustris and reeds as well as with Hypnum, Nymphea and other semi-aquatic plants. This peat may be divided by sandy deposits and at the bottom one finds a muddy peat, almost without structure. It occurs in Holland and on the shores of the Baltic, the marshes being of great extent in both regions. Fossil peat occurs at Utznach in Switzerland. Still different are the peats of wooded swamps and swampy forests. In depressed areas, where the forests have been killed by swamp plants, the peat, formed of herbaceous plants and prostrate stems, accumulates rapidly. He refers to the wood at Kidgge near Copenhagen, which the Danish naturalists had regarded as due to transport; but Lesquereux had shown that it is in place, the trees having been overturned by the wind—a condition observed in the present forests near by. The mass is composed almost wholly of birch and the upper part consists of empty barks entangled in a mud or half liquid paste, coming from decomposition of the wood. — Grand’ Eury examined in the Ural a peat of swamp-forest origin, | a mass of herbaceous plants and débris of trees. Stumps rooted in the mass were seen at two horizons in the upper part and others were scattered below. Many stems and branches lie prostrate and, ™ C. Grand’ Eury, “ Memoire sur la formation de la houille,” Ann. des Mines, 8me Ser., Tome I., 1882, pp. 197-202. 206 3911.) STEVENSON—FORMATION OF COAL BEDS. 609 at the bottom, a considerable portion is formed of barks, wood, leaves and other débris, transported and deposited in the water. Roots can be seen penetrating the gray clay on which the deposit rests. On the borders, the peat has not been changed in position and it is felted and herbaceous. Inone part it seems to be composed exclusively of transported plants, there being barks of flattened birches ; some laminated portions are formed of humefied epidermis material. No reasons are given for assigning a great portion of the mass to transported material, the matter being taken apparently as beyond dispute ; but one may surmise that the presence of stumps rooted in the peat, the prostrate trunks and the fragmentary condition of the enclosing material may have been for him convincing. Grand’ Eury did not believe that trees would grow in peat and the fragmentary condition of plant remains was proof that they had been washed in. i The conditions, described by him, are precisely those which are ; familiar in bogs, for which no conception of transport is admissible. The Danish swamps were studied by Steenstrup*** long ago; his grouping resembles that employed by the German students. The most important is the Waldmoor or Skovmose type occupying de- pressions in Quaternary deposits, often more than 30 feet deep. Where the area was small, the sides were abrupt and the trees growing on them eventually fell into the bog, where they have been preserved. In depressions of great extent, one finds an exterior wooded zone surrounding an interior or central bog zone. The latter resembles the Lyngmose, the heather or Hochmoor stage. The central area of the Skovmose is very regular. It rests on clay derived from the borders ; above which one finds ordinarily one and a half to even four feet of amorphous peat, becoming pulpy in water and containing indeterminable plant remains. The peat is very pure in normal bogs, but layers of calcareous or silicious matter are not unknown. A layer of hypnum-peat rests on the amorphous deposit, 3 to 4 feet thick, containing Pinus sylvestris, which grew on the spot, at times forming a forest on the swamp. The trees were ™ Steenstrup, as summarized by Morlot, Trans. in Ann. Rep. Smithsonian Inst., Washington, 1861, pp. 304 et seq. 207 610 STEVENSON—FORMATION OF COAL BEDS, _ [November 3, stunted and grew slowly amid unfavorable conditions, there being 70 annual rings to the inch; yet the trees lived for several centuries. In the larger swamps, two or even three layers of pine stumps are found, im situ, with their bases and roots well-preserved. As the surface became higher, and drier, the earlier mosses gave place to others; Sphagnum appeared and, at length, heathers. The pines yielded to the birches and those to alders, hazel bushes and Corylus. This succession is found only in the central zone; the deposit is too thin on the border. Weber’ after prolonged study of peat areas in northern Ger- many, grouped the peat producing plants into (1) those which form the moor; (2) those which grow on the peat; (3) those which love peat or are bound to it. The best illustrations of the relations of these groups are in moors which began in post-glacial time and have continued until now. As the result of his examination, Weber suc- ceeded in determining the stages in development of the bog and in determining the part played by the several groups of plants. He presented a classification which has been accepted by many of the later students. This will be given in detail as applied to the Scandi- navian deposits. Somewhat earlier, Blytt'** had discovered that in western Nor- way the typical succession is Feet, Sphagnous peat, about sso. 66. s Sie c ivan scab cesesasess eee sy Forest bed, chiefly of Scotch fir. Peat more compressed than that of No. 1, about.............. 5 Forest bed with oak stumps and myriads of hazel nuts. Glacial deposits. btn wate det A, But in eastern Norway, there are four peat layers alternating with three forest beds. In Denmark he finds equally distinct evidence for successive wet and dry periods. In summing up the conditions observed in Norway, Sweden and Denmark, he finds record of the following climatic changes: *8 C. A. Weber, “ Aufbau und Vegetation der Moore Norddeutschlands,” Engler. Bot. Jahrb., Vol. 40, 1908, Beiblat., No. 90, pp. 19-34. ™ Blytt, cited by J. Geikie, “Great Ice Age,” p. 495. 208 rg1t.J STEVENSON—FORMATION OF COAL BEDS. 611 1. Arctic freshwater beds, containing Salix polaris, S. reticulata, Betula nana, etc. A semi-continental climate. 2. Sub-glacial stage, with Betula odorata, Populus tremula, Salix, etc. The moors were wet, the climate humid; equivalent to the Danish “ birch or aspen period.” 3. Sub-Arctic stage, drier, many bogs became dry and were overspread by forest growth; Scotch fir (Pinus sylvestris) makes its first appearance. 4. Infra-boreal stage, climate again humid; the flora of Denmark is still of true northern type; Pinus sylvestris the common tree. 5. Boreal stage, climate drier and forests overspread the bogs, forming a root bed; Corylus and oak abundant. 6. Atlantic stage, climate mild and humid; Quercus sessiflora abundant in Denmark and southern Sweden; this is the Danish “oak period.” 7. Sub-boreal stage, drier than the last; many peat bogs dried up and became forested. 8. Sub-Atlantic stage, bogs again wet and the youngest peat layer was formed; this is the Danish “beech or alder period.” 9. Present stage, the bogs are drying and are becoming forested. Stages 1 to 4 are wanting in the low level bogs of the Scandinavian coast as that region was still submerged. The peat deposits of Sweden have been studied by H. and L. von Post, Andersson, Sernander and others, and those of Finland by Andersson. It suffices for the present to present only the salient facts as recorded by L. von Post,?*> reference to the work of some others being deferred to a later portion of this work. Von Post’s studies were made in the province of Narke, southern Sweden. His grouping is essentially the same as that offered by Weber but he gives details, necessary to the present discussion, not noted by other students. He finds the following types of deposits: Limnische. I. 1. Allochthonous mineral deposits made in open water; here are clay, with diatoms, poor in plankton, and clay-gyttja, which is clay with much plankton and diatoms. 2. Allochthonous organic sedi- ments, including (a) plankton-gyttja, in open, comparatively deep water, gray to green, more or less elastic, composed of plankton, algae abounding; (b) detritus-gyttja, in comparatively shallow water, from Potamogeton and Nymphaea, red-brown to yellow-black, granular, mostly plant débris with some plankton; (c) Schwemmtorf, composed of plant detritus; (d) Ufertorf, like the last and formed very near the line of low water. It contains lenses of Lake and of Swamp peat. ™L. von Post, “Stratigraphische Studien iiber einige Torfmoore in Narke,” Geol. Foren. Forhandl., Bd. 31, 1909, pp. 633-640, 644, 647. 209 612 STEVENSON—FORMATION OF COAL BEDS. _ [November 3, II. Autochthonous organic deposits. The Lake peat including (a) Phragmites peat, clear yellow, composed of fibrous roots with reeds and some gyttja; (b) Equisetum peat, like the last in structure, but the color is coal black. Telmatische. I. Swamp or Niedermoor peats, including (a) Magnocari- cetum peat, consisting of sedges with Amblystegium as accessory, yellow to yellow-brown; (b) Amblystegium peat, consisting of stems and leaves of that plant with some sedge constituents; (c) Bruchpeat, red to black, amorphous humefied peat detritus, in situ, with identi- fiable roots of sedges. II. Hochmoor peats, (a) Cuspidatum peat, bright colored Sphagnum cuspidatum and other water-loving mosses, with remains of Scheuch-— seria, Carex and Eriophorum. Semi-Terrestrische. I. (b) Vaginatum peat, Sphagnum with Eriophorum vaginatum roots and stalks, these often making up one half of the mass, humefied and dark colored; (c) Sphagnum peat in lenses with Cladina remains between clear brown layers of Spheme with Eriophorum. II. Forest peat, (a) Alder forest peat, red-black, amorphous, consists of in situ deposited detritus of an alder swamp forest. Remains of alder are recognizable; Cenococcum geophilum abundant. Terrestrische. (b) Birch forest peat, like the last, but commonly dark colored, deposited in a birch swamp forest; (c) Forest peat, rich in Eriophorum and Sphagnum, as a rule, dark colored, almost always with stumps and other remains of Scotch fir; (d) Forest mould, dark, composed of wood detritus and grains of humus, with stumps. All of these types from Lake peat down are autochthonous. The upper limit of the basin or limnic deposits is at the normal line of — low water; the shore or telmatic deposits are in the space covered at high water, while the terrestrial are on forested areas, rarely cov- ered with water. The alder swamp is the passage zone to the ter- restrial. Von Post confirms Blytt’s conclusions respecting the alter- nation of dry and humid periods, and shows how, during the less — humid times, forests invaded the peat deposits and in some cases — covered the surface of pure peat with a dense growth. He presents — sections from a number of localities. One fromthe Asta moor shows — A. Sphagnum peat, 85 centimeters, with, at 80 centimeters, a mass of — fir stumps rooted in the peat 4nd with coaly matter between the stumps. B. Strongly humefied cuspidatum peat, 10 centimeters. C. Sedge peat, 30 centimeters, has much Sphagnum above. D. Alder and birch swamp forest peat, with small stumps of alder, birch willow and a great quantity of Cenococcum geophilum, 15 centimeters 210 1911.] STEVENSON—FORMATION OF COAL BEDS. 613 E. Shore peat, like transported peat, 25 centimeters, roots of Carex, Equisetum and Phragmites. F. Plankton-gyttja, 40 cm. with remains of infloated Phragmites, Equisetum, etc., some pollen of Picea in upper portion. G. Clay, 50 cm. rich in saltwater diatoms. As interpreted by Von Post, one has here at the bottom, a deposit of plankton material or Sapropel. It was invaded by the shore peat, on which a forest of birch and alder grew for a short time amid unfavorable conditions, as the swamp was overflowed at times; this condition became more marked and a sedge swamp followed, in which Sphagnum gradually gained control. Still later, for a short period, during which accumulation of peat continued unchecked, the moor was covered with a dense growth of firs; but as the moisture increased, the non-water-loving elements disappeared and a Calluna- Eriophorum moor occupied the area. Sections in Skarby lake com- plex show the same general features as those observed elsewhere in this region. Though there are differences in detail, the story is prac- tically the same throughout. The open water deposits, gyttjas rich in plankton material, form the lowest stratum resting on clay or sand; on this is the shore peat, which gradually passed across the basin. Then came the time of decreasing moisture; alders advanced on the peat surface, now subject to only occasional overflows; they were succeeded by birches, which were rooted in the alder peat ; and finally came the great forests of Scotch fir growing in the birch and alder peat, to be succeeded by Sphagnum-Hochmoor peat in the moist Sub-Atlantic stage. Peat-making was continuous in the for- ests and each type of forest peat has its own group of minor plants. Buried Peat Deposits—Some authors have contended that peat deposits on the land are not likely to be preserved because, exposed to air, they must be affected by atmospheric conditions and eventu- ally must waste away. Under such conditions, it is certain that only such accumulations of vegetable material as are deposited in water-filled basins would be preserved. But the supposed condi- tions are purely hypothetical and are not in accord with those exist- ing in nature. Indeed, one looking at a peat deposit, many feet thick, would have difficulty in conceiving how there could be uni- formity of conditions for a period long enough to permit wastage 211 614 STEVENSON—FORMATION OF COAL BEDS, _ [November 3, of so great a mass, almost impermeable to water after having be- come thoroughly air-dried. But a priori reasoning is unnecessary ; for, as Lesquereux recognized long ago, burial of peat bogs is part of the normal sequence of events. Dawson**® has described an early Quaternary bog which he saw in Nova Scotia. It underlies 20 feet of bowlder clay and pressure has made the peat almost as hard as coal, though it is tougher and more earthy than good coal. When rubbed or scratched with a knife, it becomes glossy; it burns with considerable flame and ap- proaches the brown coals or poorer varieties of bituminous coal. It contains many roots and branches of trees apparently related to spruce. . Areas of peat buried under glacial drift are numerous in the New England states as well as in New Jersey and some of them will be mentioned in a succeeding section. Newberry,’** many years ago, collected all the observations then available for states west from the Alleghany mountains. In Montgomery county of Ohio, E. Orton found a bed of peat, 15 to 20 feet thick, the surface covered with Sphagnum, grasses and sedges. It contains coniferous wood with bones of elephant, mastodon and teeth of giant beaver ; and it under- lies 90 feet of gravel and sand. At many places in Highland county of the same state, wells have reached a stratum of vegetable matter and, at Cleveland, a “carbonaceous stratum” has been found at 20 feet below the surface. A similar condition exists at Lawrenceburg, Indiana, as well as at many places along the Ohio; and J. Collett reported that, throughout southwestern Indiana, there is an ancient soil, 2 to 20 feet thick, with peat, muck, rooted stumps, branches and leaves, at 60 to 120 feet below the surface. This deposit is known locally as “ Noah’s cattle yards.” The same condition is reported from a portion of Illinois. The great forest bed of Iowa, discoy- ered by McGee at a later time, is in part a buried bog. Leverett, Taylor, and Goldthwait have described autochthonous peat bogs buried under glacial drift at many localities within the Missis- sippi area. ; *° J. W. Dawson, “ Acadian Geology,” 2d Ed., London, 1868, p. 63. ™* 7. S. Newberry, “Surface Geology of Ohio,” Geol. Survey of Ohio, 1874, Vol. II., pp. 30-32. 212 rgtt.] STEVENSON—FORMATION OF COAL BEDS. 615 In America, observations as recorded are very few and, for the most part, they are merely incidental, as until very recently the geo- logical importance of peat was not recognized; but in Europe the case is very different; one finds there such a wealth of illustration as to cause surprise that any student should entertain doubts respect- ing preservation of peat deposits by burial under sediments. A few citations must suffice. J. Geikie’** says that peat bogs often pass below the sea. In the harbor of Aberdeen, trunks of oak are brought up and at a little distance away, peat was seen below the sea level covered with 10 to 12 feet of sand. This bed, enclosing trees, is known to extend for some distance into the bay. In the Carse lands, the river Tay has cut down to a peat bog, now forming the river bed and under- lying about 17 feet of alluvial material, which near the top contains cockles, mussels and other marine forms. This extensive peat de- posit of the wide Carse area rests in part on alluvial sands and in part on marine clays. The peat is highly compressed and splits readily into laminez, on whose surfaces are small seeds and wing cases of insects. As a rule, but not always, it is marked off sharply from the overlying clay and silt. That it represents an old land sur- face is certain but it is equally clear that, in great part, the vegetable débris on top was drifted in from localities higher up in the valley, for the upper part of the peat contains, at times, layers of silt and twigs, while branches as well as trunks are scattered through the lower 3 or 4 feet of the overlying silt. The conditions are the same in Carse lands on both sides of Scotland and they exist in the Hebrides. Prevost and Reade’*® have described a peat bed covered by a thick deposit of sediments. The exposed portion is a dark-brown peaty mass, containing large and small branches, roots and rootlets, the latter passing into the underclay. Some large boles and an occa- sional stump were seen on the upper surface. The authors note as a remarkable fact, that this bed resists erosive action by the river ™ J. Geikie, “ The Great Ice Age,” 1895, pp. 290-293. ™ E. W. Prevost and T. M. Reade, “ The Peat and Forest Bed at West- bury-on-Severn,” Proc. Cotteswold Nat. Club, Vol. XIV., 1901. PROC. AMER. PHIL. SOC., L, 202 00, PRINTED NOV. 17, IQII. 213 616 STEVENSON—FORMATION OF COAL BEDS, [November 3, as well as by the more energetic bore, so that it projects as a prom- ontory. Strahan’*® measured the section exposed during excava- tions for docks on Barry island. The succession is 1. Blown sand, Scrobicularia clay, sand, shingle, with strong line of erosion below. 2. Blue silt with many sedges. 3. Upper peat bed, 1 to 2 feet thick. 4. Blue silty clay with many sedges. 5. Second peat bed, thin. 6. Blue silty clay with sedges. 7. Third peat bed, with many logs and stools, roots in place underneath. 8. Blue silty clays with reeds, willow leaves and freshwater shells. 9. Fourth peat bed with large trees and roots in place and numerous land shells. 10. An old soil with roots and land shells. 11. Rock in place, at 35 feet below the Ordnance datum. ; Here as in the Carse area of Scotland, the peat underlies a deposit containing marine shells. : Lesquereux™? cites a French author, who found at many places in the Department of Nord alternations of peat and sand, the latter containing marine shells. He notes that when the growth of peat is checked by dryness, a crust forms, which is a parting between the old and the new peat. In the valley of the Somme, he found, under- lying 8 feet of clay and concretionary limestone, 23 feet, 4 inches of peat in 15 layers, with the partings distinct and the layers differing in character. Alternations of clay, peat and calcareous concretions are not rare. Geinitz,1*2 more than twenty-five years ago, studied the dune-— covered bogs near Rostock. At a later period he had opportunity for more detailed examination and his observations are important — from several points of view. At the bathing station near Graal, the section shows at the bottom, sand of the Rostock plain, on which : rests a one-foot layer of peat, containing stumps of trees which grew on it. The dune formerly covering this deposit has been removed — for some distance, exposing the peat, but it still remains at a little way landward. Beyond the dune, one finds a forest of great beeches and oaks, with the peat bed covering the surface between them. ” A Strahan, Mem. Geol. Survey, “Geology of the South Wales Coal Field,” Part III., 1902, pp. 87-93. ; ™ 1. Lesquereux, Ann. Rep. 2d Geol. Survey of Penn. for 1885, pp. 116- 118. ™ FE. Geinitz, “Nach der Sturmflut,” Aus der Natur, Vol. IX., 1908, pp. 76-83. 214 i911.] STEVENSON—FORMATION OF COAL BEDS. 617 When he looks at the dune surface, he sees, as it were, shrubs rising out of the sand, some short thick stems of beech and oak; but they are not shrubs, they are the still living parts of trees, the same in age and growth as those standing in the open forest. They have been buried by the advancing dune. A mighty storm flood, tearing away the sea wall and removing part of the dune, will expose vertical trees standing in the sands as in the Coal Measures sandstones. At present, one sees advancing masses of sand burying the trees, which grow on low-lying moors. At another locality, storms, during re- cent years, have exposed an older peat deposit, underlying the sands of the Rostock plain. The outcrop extends hundreds of meters along the shore and shows that the peat is a moss peat, which bore a forest of Scotch fir. There, as also near Graal, the waves have torn off fragments of the peat and have worn them down into elliptical form similar to that of the beach pebbles. Barrois*** has referred to similar origin of peat pebbles on the shore of the British channel, where some neolithic deposits of peat are exposed to the waves. The fragments of peat are rolled, rounded and eventually transformed into true ellipsoidal pebbles. Lorie,’** in his fifth contribution to the surface geology of Hol- land has gathered together all the available information oe the buried recent peat deposits of that region. In all probability the Zuyder Zee was filled with peat prior to the catastrophe of the middle ages, but the only vestige is on the island of Schalkland, where one finds 5 to 7 meters of peat covered with a meter or more of marine clay. The same condition exists on the river Y near Amsterdam and in the province of Zeeland as well as in the west part of North Brabant in Belgium. The peat bed near Oudenbosch, in the latter province, is 0.75 meter thick and _underlies 0.65 meter of sediment. It is readily traceable from that village across Zeeland into western Flanders of Belgium, and thence to the coast at Ostend in Belgium and Dunkerque in France, a dis- **C. Barrois, “Observations sur les galets de cannel-coal du terrain houiller de Bruay,” Ann. Soc. Geol. du Nord., Vol. XXXVIL., 1908, p. 7. ™ J. Lorie, “Les dunes intérieures, les tourbiéres basses et les oscilla- tions du sol,” Archives Mus. Teyler, 2me Ser., Vol. III., 1800, pp. 424-427, fae P12 215 618 STEVENSON—FORMATION OF COAL BEDS. _ [November 3, tance of more than 60 miles. Lorie cites Belpaire pére, who says that it is one to 3 or even 4.5 meters thick and that it rests mostly on blue clay, though in some localities on fine sand. It is double near Ostend, where the lower bed is black, compact, with roots of reeds, while the upper bed contains no reeds but has woody fibers, apparently roots of heath plants. The peat and its overlying clay are sometimes continuous under the dunes and shore, as is also the case on the island of Walcheren in Zeeland. Trees, rooted in the subsoil, occur frequently in the peat. Belpaire fils says that the thickness of the peat and that of the overlying clay vary from 1 to 3 meters and that the clay level is never above high tide. On the left bank of the Escaut (Scheldt) as it flows from France across Belgium the peat is almost a meter and a half thick, but the clay, 2 to 3 meters, decreases as it recedes from the river. Lorie says that Rutot found a divided peat near Blankenberghe in Belgium. Reference to Rutot’s**> publication shows that the section is Meters, t,. SHOVE SON. as 5. eicad Genus snes ces ts bins 000 bayer 2.30 2. “Gray ‘sandy Clay... 6.500 os ee ive ses oe nee ese ee . 0.60 3. Gray sand, with bed of Cardium midway ..............+++- . 1.10 4. -Pure peat. i ssciin ss beicss oe cue pleaneeinen eteeas eee eee 2,00 s. Gray sand, slightly argillaceous . 00.1. .5..5ss