The contrast between the Nevada and the Arkansas sequence is seen in the fact that in the Arkansas section, the faunas above and below the Spring Creek limestone are typically Carboniferous and do not present admixture of Devonian forms, while the Nevada and, from present reports also, the California sections from Devonian far into Carboniferous time. present remarkable commingling of Devonian with Carbonif erous types. The 3000 feet of conglomerate in Nevada, reported by Walcott as separating the White Pine shale from the Carboniferous limestone with Rhynchonella Eurekensis, suggests the cause of the appearance of this western fauna in the Arkansas region, i. e. an elevation of the western area sufficient to cause diversion of the ocean currents and shifting of such species as endured the transport into the Mississippi Valley area. In the Appalachian province the Devonian species are still more sharply distinct from the Carboniferous forms than in the Mississippi Valley.

These are the facts in the case: in northwestern Arkansas is found a fauna in the Spring Creek limestone which by its stratigraphical position and general fossil contents is shown to be of Carboniferous age, separated by at least two well defined. and distinct faunas (the Kinderhook-Chouteau and the Burlington-Keokuk) from the latest Devonian fauna of the Mississippi basin province, but containing several species which in the more eastern Paleozoic sections are, both specifically and subgenerically, Devonian forms.

One of the species is known only from a Carboniferous horizon of the western part of the continent, and others are distinctive of more western faunas in which the commingling of Devonian and Carboniferous species is conspicuous.

The recurrence of the Devonian species in the Arkansas Carboniferous rocks is most readily explained as a case of migration of species from a region in which they had continued to live unchanged, into a region from which they had for a long time been absent or into which they had never before entered. It is an example of the living on of species in one locality long after they had become extinct or were replaced by other species in another region. This is not an anomalous fact; the Australian land fauna is a remarkable illustration of the same law. Deep sea dredgings have shown the same fact in regard to some of the abyssal species which are of Eocene or Cretaceous type, ancient characters having been preserved in the ocean depths while they have been superseded by evolved successors in other environments. But the inter

esting point in this Arkansas case is that we have here a suggestion as to the cause of the unchanged continuance of the species. The recurrent Devonian species were evidently asso

ciated with a generalized as contrasted with a specialized fauna. The faunas of the Devonian and Carboniferous in Nevada and California are peculiar for showing a very long continuance of the same general fauna, with changes by the accession of new species as time progressed, but with remarkable persistence of early species unchanged. Walcott called attention to the fact in his Nevada faunas. The great thickness of Devonian rocks shows not only long range of species but such mingling as to bring species supposed to be characteristic of Upper Devonian in the east down at the base of the series in Nevada. In the Appalachian series we find the opposite to be the fact, the faunas are much broken up into distinct zones, with very few species tying together the successive faunas. And this breaking up of the faunas is plainly associated with oscillating levels and general passage from pure marine conditions up to brackish water, and finally by the closing of sedimentation upon elevation of the bottom above the sea-level. This is shown by the passage from limestones of the Lower Devonian into shale, sandstone and conglomerate-in the succession of deposits that followed.

Taking the first full appearance of a characteristic fauna as indicative of common horizon for each of the separate prov inces, we observe that in the Appalachian province the Devonian species are more closely restricted to the Devonian age. In the Mississippi province, after the Devonian stage is sharply closed, this case of recurrence is seen well up in Carboniferous faunal succession, while in the western continental border there was a mixture of Devonian and Carboniferous species all along the way till the characteristic Carboniferous species were present in full force in the prevalent fauna. This later and later removal of the earlier types from the prevalent faunas as we pass westward across the continent is coördinate with the continental expansion occurring at the same time. The conclusion seems evident that the cutting off of the Devonian species was in some way associated with the progress of the continental border westward during later Devonian and the first half of Carboniferous time. The coal measure conditions, were as early as the Warsaw and St. Louis limestones in Pennsylvania, if not still earlier, and limestone had ceased to be a factor in the deposits in northern Pennsylvania and New York, before the Chemung began. In the Mississippi Valley limestones continue up to the Chester, and then the sands, conglomerate and coal measure condition suddenly appear.

Before we reach the central ridges of the Rocky Mountains. the limestone conditions are prevalent till the upper coal measure time, and still further west there are no indications of any

continental lands sufficient to disturb the reign of the marine faunas till near the close of the Carboniferous age. Thus in the Appalachian province diversity and alternation of deposits is marked by numerous successive and distinct faunas; in the western continental border province uniformity of prevailing calcareous sedimentation for long periods is marked by the abnormally long continuance of many of the Devonian species; while the central continental province midway between the two is marked by the unmistakable recurrence of Devonian species far up in the midst of Carboniferous sediments.

While theoretically such a fact might be expected, this unique case of recurrence furnishes us with the evidence connecting the three distinct geological provinces, and makes it possible to correlate, with more than ordinary precision, the chronological horizon of the several widely separated faunas. This series of observations furnishes a demonstration of the hypotheses that the persistence of species without modification is associated with continuance of uniformity of conditions of environment, and that change in the successive faunas of geological time is associated with the change and rearrangment of the conditions of environment to which the fauna is subjected.

ART. IX.-Constituents of the Cañon Diablo Meteorite; by ORVILLE A. DERBY.

A SPECIMEN of the Cañon Diablo meteorite obtained from Mr. E. E. Howell of Washington and stated to be one of the original lot brought from Arizona by Dr. A. E. Foote, has been treated in the laboratory of the Commissão Geographica e Geologica de São Paulo by the fractional method of Prof. E. Cohen with the following results. The chemical analyses accompanying this paper have been kindly made for me by Dr. Guilherme Florence, assistant to the Commission.

The specimen, weighing nearly 200 grams, was a perfect meteoric individual; that is to say, it presented no fractural surfaces, but everywhere the rough pitted surface of meteoric masses. In appearance it suggested a metallic bleb broken or weathered out of friable, or more easily decomposable, material. An examination of a considerable number of specimens of all sizes in Mr. Howell's collection shows this to be a general characteristic of the Cañon Diablo group. Nothing in the shape and aspect of the masses suggests the occurrence of planes of slight cohesion (presumed to be the limits of crystalline

individuals, such as in a paper now in press, I have described under the name of Wollaston Planes, in the Bendegó mass) and which, by facilitating fracture either in the original place of formation, or in the act of falling, have probably produced the approximately plane faces and angular edges that characterize that meteorite. Such faces and edges might be expected on the Cañon Diablo masses on the hypothesis of their being fragments of a single homogeneous mass, which seems to be required by the conditions under which they were found. A rough, jagged and pitted surface is however common to all of them, showing a perfect individualization and suggesting on a large scale, the small, irregular metallic masses scattered through the stony matrix of a mesosiderite. Referring them to a single original mass, the hypothesis may be ventured that, on its arrival in our atmosphere, this was not homogeneous but consisted of a large mesosiderite with unusually large metallic nodules that became separated by the explosions attending the fall, and probably also by subsequent decay and disaggregation of the stony matrix.*

After freeing the specimen as far as practicable from its rust crusts by scraping after a soaking in strong acid, it was treated with cold hydrochloric acid of a strength of 1 to 10. The solution was effected slowly with evolution of gas and a separation of a variety of grains with a metallic aspect and of a light, black residue resembling coal dust. A vein-like mass some three millimeters thick, that showed through the rust crust with the appearance of the pencil-like inclusions of troilite in the Bendegó meteorite, extended for about a centimeter into the mass and not being acted upon by the acid, came away in fragments. After fourteen weeks of treatment with frequent changes of acid, the action almost ceased although a considerable mass still remained undissolved. This had much the shape and appearance of the original meteorite though much. more irregular and jagged, and represents a nucleal portion less soluble than the generality of the mass.

The undissolved residue was separated by screening through fine bolting cloth, sorting under the lens, and with a magnetized knife point, into the following groups: vein matter consisting

*I was also shown in Washington schistose masses of iron oxide found in the same region, whose connection with the meteorite was considered doubtful. These closely resemble the thicker masses of rust crust formed on the Bendegó meteorite and like it, as is well seen in sections prepared by Mr. Diller, show minute particles with a metallic luster which are almost certainly grains of schreibersite, as that mineral has been separated from the rust crust of both Bendegó and São Francisco do Sul. In view of its occurrence it can hardly be doubted that these Canon Diablo specimens are due to secondary alteration of the meteorite. As the iron masses in general have only a thin rust crust, indicating considerable resistance to oxidation, it may be suggested that these thicker masses of oxide may perhaps come from original pyrite as in the case of São Francisco do Sul.

of massive schreibersite with cohenite; irregular jagged fragments resembling the large nucleal piece and bristling with needles of rhabdite (zackige Stücke of Cohen?); taenite; coarse schreibersite and cohenite from the general mass and not from the vein (a considerable part of the schreibersite was free but the grains of cohenite were so charged with it that no satisfactory separation of the two could be effected); fine magnetic residue for the most part schreibersite in the form of rhabdite needles but with fine particles of taenite; granular schreibersite and cohenite, and a black, coal-dust-like residue highly charged with rhabdite. The separation could not be completely made except for the jagged pieces and coarser taenite, schreibersite and cohenite. In the finer material the two last were so lumped together that neither by sorting nor by gravity or magnetic methods could they be satisfactorily separated. An attempt to separate the light, coaly matter by the use of the Thoulet solution was only partially successful, as a small amount of the black particles were carried down with the heavy metallic grains and a larger portion of these were retained by the spongy, coaly particles. The proportions given below, calculated for the dissolved portion after deducting the nucleal piece and the vein matter, are therefore only approximate.

An analysis of 100ce of the solution corresponding approximately to 0.5 grams of the meteorite (except for the copper determination which was made with 500cc) gave:

This result agrees fairly well with the composition of kamacite (Fe 93-11, Ni and Co 6-89) according to the formula Fe, Ni