‘ nes aan tare ent eee eri er wo 2 . @ flet c00 tg er nae Henne wee ary (Me ee nae yaaa BENE eee a wee , eran . a . sent . a ohm te re aed reer) Sana tree e Pa ae in Sate me We ev deatboe gat Ox Sr erararer racine Sorin) era ' veseee ' Vee tren ai te Peruri air ver ar ie ear Or rn fe ee ee oe oe. / oa ope alee et ae VYA-tw A Pr THE AMERICAN JOURNAL OF SCIENCE. Epiror: EDWARD S. DANA. ASSOCIATE EDITORS Proressorns GEORGE L. GOODALE, JOHN TROWBRIDGE, W. G. FARLOW anp WM. M. DAVIS, or Camsringe, Proressors ADDISON E. VERRILL, HORACE L. WELLS, L. V. PIRSSON anp H. E. GREGORY, or New Haven, Proressor HENRY S. WILLIAMS, or Irwaca, Proressor JOSEPH S. AMES, or Battimmore, Mr. J. S. DILLER, oF Wasuineton. FOURTH SERIES VOL. XXXI—[WHOLE NUMBER, CLXXXI.] WITH FOUR PLATES. NEW HAVEN, CONNECTICUT. iG etenlige 28101 BF ge . MEE iM agit 1) } : frets +3 are av CONTENTS TO VOLUME XXXI. Number 181. Page Arr. I.—Gravity Determinations at Sea; by L. A. Bauzr.- 1 II.—Stratigraphy of a Deep Well at Waverly, Ohio; by R. She LSU NSS EHTEL <2 pS a ee ree III.—Solid Solution in Minerals with Special Reference to Nephelite; by H. W. Foorr and W. M. Brapuny--.--- 25 IV.—Fossil Evidence of the Age of the Virginia Piedmont Slates; by T. L. Watsow and S. L. Powntt _--------- 33 V.—Native Gold from Gold Harbour, Queen Charlotte lisllewngls; @ loyr dies Lei, 1B), Cor/g IN eieced ee Ses cee ee 45 VI.—Natramblygonite, a New Mineral ; by W. T.Scaatiter 48 VII.—New Emission Theory of Light; by J. TRowsripecE. 51 VIII.—Origin and Peopling of the Deep Sea; by J. DNV WAGTISTSETINRCm Meee eee a armalner a anit 2 Ls SS ud rete A ee 55 IX.—Camels of the Harrison Beds, with Three New Species ; loge 15 18, ILO IS e eee ate ee ane eee Se ee 65 NVAIETANT EDN RM mRE ee 2 ke lh ase oo (Al SCIENTIFIC INTELLIGENCE. Chemistry and Physics—Preparation of Metallic Radium, KE. EBLER: Reduc- tion of Oxide of Iron by Solid Carbon, CHARPy and BoNNEROT, 70.— Action of Light upon an Hlectrie Cell, H. Panason : Sterilization of Large Quantities of Water by Ultra-violet Rays, HreLBROoNNER and RECKLING- HAUSEN, 76.—Refrigeration by Mixtures of Liquids, J. Ductaux: The Relations between Chemical Constitution and Some Physical Properties, S. Suites: Positive Rays, W. Wien, 77.—Deflections by Electrostatic and Magnetic Fields of Radium B after Recoil from Radium, A. S. Russ, etc.: Energy Distribution of Diffraction Gratings, A. TRow- BRIDGE and R. W. Woop: Modification in Magnetic Fields of Lines of the Light emitted by the Hlectric Spark, G. A. Hrmsatece: Electric Motors, H. M. Hopart, 78. Geology and Natural History—Grundziige der Paldontologie, K. A. v. ZirreL, 78.—Notes on Ordovician Trilobites, 11, Ill and IV, P. HE. Ray- MOND: Preliminary List of the Fauna of the Allegheny and Conemaugh Series in Western Pennsylvania, P. E. RaymMonp: The British Carbonif- erous Orthotetine, I. THomas, 79.—Geological Excursion in the Grand Canyon District, D. W. Jounson: Aufbau des Gebirges in der Umgebung der Strassburger Hiitte an der Scesaplana, W. v. SEIDLITZ: Granites of the Southeastern Atlantic States, T. L. Watson: The Volcanic Rocks of Victoria, E. W. Skrats, 80.—Analcite Rocks, G. W. TyRRELL : Morganite, a Rose-colored Beryl, G. Kunz, 81.—Tables for the Determination of Minerals by Physical Properties : The Subantarctic Islands of New Zea- land ; British Nudibranchiate Mollusca, with Figures of the Species ; Part viii, Supplementary, 82.—Medusae of the World, A. G. Mayzr: An Introduction to Zoology, R. W. Heener, 83.—Animal Study: with Directions for Laboratory and Field Work, W. H. D. Murer: Methods of Attracting Birds, G. H. TRarton: Second Report on the Hymeniales of Connecticut, EH. A. Waite, 84. iv CONTENTS. Number 182. Page Arr. X.—Adjustment for the Plane Grating Similar to Rowland’s Method for the Concave Grating ; by C. Barus and M,(BAmugieeesse peso. lon). eee 85 XI.—Determination of the Hardness of Minerals, II; by bc Way A Gu ee ees ce ey 1g Ae Yee gen 96 XII.—Photographing Fossils by Reflected Light ; by L. D. Buntang Johan NG itches ae 99 XIII.—Synthesis of the Paleogeography of North America ; ; by -E. Swass 222.3. Sees ee ee 101 XIV.—Estimation of Silver by Electro-Deposition from an Ammoniacal Solution of the Oxalate; by F. A. Gooca and Jo PS Winisnr 2k ee) See ee eee a eS ne 109 XV.—Notes on the Armored Dinosauria; by G. R. WiELAND 112 ie Manne Gneissoid Structure in the Cortlandt Series ; Dy Guts ROG WR hg peter ones ay etn ane es Men 125 XVII.—Thaumasite from Beaver County, Utah; by B. S. Bouriur and- W.."PAScHAmiEME 2205 eee ten eer e 131 XVIII.—Nomenclature of the Lower Paleozoic Rocks of New) York. byiH. PCusnine 2-95. 952- = 2 eee 135 SCIENTIFIC INTELLIGENCE. Chemistry and Physics—Determination of Copper as Sulphate, Rrcoura : New Method for Determining Boiling Points and Vapor Pressures, SMITH and Menzies, 146.— Reactions of Nascent Hydrogen in the Dry Condition, Vournassos: Synthetic Sapphire, Verneutnt, 147.—Influence of Tem- perature on the Compressibility of Metals, E. GRUNEISEN : Ionization of the Atmosphere due to Radio-active Matter, A. S. Eve: Thomson Effect, P. Cermak: Velocity Measurement of Rontgen Rays, E. Marx, 148. Geology—Osteology of Pteranodon, G. F. Eaton, 148.—The Age of Mam- mals, H. F. Osporn, 150. —Tertiary Faunal Horizons in the Wind River Basin, Wyoming, with Descriptions of New Eocene Mammals, W. GRANGER, 151.—Geological Survey of New Jersey, Annual Report of the State Geologist, H. B. Ktmuet, 182. Astronomy—Transactions of the Astronomical Observatory of Yale Uni- versity, 152.—Determination of the Solar Parallax from photographs of Eros, 153.—Les Déterminations des Longitudes et l’Histoire des Chro- nométres, J. Mascart: Project for the Reform of the Calendar, C. A. Hesse, 154. Miscellaneous Scientific Intelligence—Report ot the Secretary of the Smith- sonian Institution, 155.—Report of the Librarian of Congress and Report of the Superintendent of the Library Buildings and Grounds : Academic and Industrial Efficiency, M. L. Cooxes, 156. CONTENTS. Vv Number 183. Page Art. XIX.—Transmission of Light through Transparent Inactive Crystal Plates, with Special Reference to Observations in Convergent Polarized Light ; by F. E. WitiGht? 3... 2a9e See See Se ere Eee aera 157 XX.—Separation and Estimation of Barium Associated with Calcium and Magnesium, by the Action of Acetyl Chlo- ride in Acetone Upon the Mixed Chlorides ; by F. A. Gooner andeCr NisBOWNTON 6-2 5-5 so ce one secs =n 212 XXI.—Feldspar Aggregate Occurring in Nelson Co., Vir- Pinay Me WetORNTON,, Iie J o22 2 epen yee aie ee 218 XXII.—History of the Coconut Palm in America; by O. F. (CO CHE o e eee, ee e XXIII.—New Mink from the Shell Heaps of Maine ; by F. booming: 2o5 see ange i aie ie eee reset Mena epee 2247 SCIENTIFIC INTELLIGENCE. Chemistry and Physics—Mesothorium, W. MarckwaLp, 230.—Combustion of Hydrocarbons, W. A. Bonn: Supposed Chemical Distinction between Orthoclase and Microcline, VARNADSKY and REvouTsKY, 231.—Preparation of Argon, G. CLaupE: Die Stellung der neueren Physik zur mechanischen Naturanschauung, M. Puanck: History of the Cavendish Laboratory, 1871-1910, 232.—The Principles and Methods of Geometrical Optics, J. P. C. SouTHALL, 233.—Chemische Krystallographie, P. Groru, 2384. Geology and Natural History.—United States Geological Survey, 234.—Pub- lications of the U. S. Geological Survey, 235.—Bureau of Mines, J. A. Hotmes : Florida State Geological Survey, 236.—The Badland Formations of the Black Hills Region, C. C. O’Harra: West Virginia Geological Sur- vey, I. ©. WurtE: New Zealand Geological Survey, 237.—Geological Survey of Western Australia, 238.—Palzontological Contributions to the Geology of Western Australia, 239.—Report of the Vermont State Geolo- gist for 1909-1910, G. H. Perkins: Contribution to the Geologic History of the Floridian Plateau, T. W. VAUGHAN: Recent Discoveries Bearing on the Antiquity of Man in Hurope, G. G. MacCurpy, 240.—Fossil Faunas of St. Helen’s Breccias, H. S. Wintiams, 241.—Palzontologia Universalis, 242.—Botanische Tropenreise, Indo-Malayische Vegetationsbilder und Reiseskizzen, G. HABERLANDT: Plant Anatomy, W.C. Strvens: A. Text- Book of Botany and Pharmacognosy, H. Krammrr, 243.—Biology, general and medical, J. McFaruanp, 244. Miscellaneous Scientific Intelligence—Carnegie Institution of Washington _ Year-Book, No, 9, 1910, 244.—Publications of the Carnegie Institution, 245.—Annual Report of the Board of Regents of the Smithsonian Institu- tion, 246.—Publications of the Allegheny Observatory of the University of Pittsburgh, F. Scauesincer: Bref och Skrifvelser af och till Carl von Linné; Seismological Society of America, 247.—Das Hlectrokardiogramm des gesunden und kranken Menschen, F. Kraus und G. Niconar: Plane Trigonometry, E. R. RopBins : Shop Problems in Mathematics: Ostwald’s Klassiker der Exakten Wissenschaften, 248. Obituary—Sir FRANcIs Gatton; Dr. M. WitnELM Mrynr, 248. yi CONTENTS. ' Number 184. Page Art. XXIV.—Ionization of Different Gases by the Alpha Particles from Polonium and the Relative Amounts of Energy Required to Produce an Ion; by T. 8. Taytor 249 XXV.—Heat Generated by Radio-active Substances ; by W. DUANE 2. 502. 23 ie XXVI.—Contributions to the Geology of New Hampshire, IV. Geology of Tripyramid Mountain; by L: V. IPTRSSON ANG Vint. INFO er chai CL een ee 269 XXVII.—Note on a Method in Teaching Optical Mineralogy; by FW... McN Are Se ne em XXVIII.—New Paleozoic Insects from the Vicinity of Mazon Creek, Illinois ; by A. HanpirrscH---- .------- 297 XXIX.—Results of a Preliminary Study of the so-called Kenai Blora of Alaskac iby Aviioumcnemeee es. == anes 327 SCIENTIFIC INTELLIGENCE: Chemistry and Physics—Use of Calcium Carbide for the Determination of Moisture, Masson : Action of Water upon Phosphorus Pentoxide, BaLa- REFF, 351.—Fractional Crystallization of Argon, F. Fiscurer and VY. FROBOXKSE: Qualitative Chemical Analysis, BASKERVILLE and CURTMAN : Die Verwertung des Luftstickstofts, J. ZenNEcK, 332.—Allen’s Commercial Organic Analysis: Absorption Spectra of Solutions, H. C. Jonrs and W. W. Strone, 333. Geology and Mineralogy—Thirty-fourth Annual Report Department of . Geology and Natural Resources, Indiana, 333.—West Virginia Geological Survey, Vol. 5, Forestry Wood Industries, A. B. Brooxs : Geological Sur- vey of Tennessee; Atlas Phographique des Formes du Relief Terrestre, 334.—Illinois Oil Fields in 1910: Physical Notes on Meteor Crater, Ari- zona, 335.—Minéralogie de la France et de ses Colonies: Practical Min- eralcogy Simplified for Mining Students, Miners and Prospectors, J. P. Rowe: Calcites of New York, H. P. Wurttock: Minéraux des Pegmatites des Environs d’Antsirabé a Madagascar, L. Duparc, 337.—Production of Phosphate Rock in Florida during 1910: North Carolina Geological and Economic Survey, 338.—Note on the parietal crest of Centrosaurus apertus ’ and a proposed new name for Stereocephalus tutus, L. M. Lames, 339. Miscellaneous Scientific Intelligence—Carnegie Foundation for the Advance- ment of Teaching, 339.—Text-Book of General Bacteriology, EH. O. JoRDAN: Catalogue of the Lepidoptera Phalenz in the British Museum, G, F. . Hampson, 340. es Obituary—H. P. Bowpircu : J. H. van’t Horr: J. W. BrUaL, 340. CONTENTS. vil Number 185. Page Arr. XXX.—Melting Points of Minerals in the Light of Recent Investigations on the Gas Thermometer; by A. DLeeWaveandpits by SOSMAN 25202. seen 2S s'Le oss ae 2 341 XXXI.—Separation of Cerium by Potassium Permanganate ; ange imtebe bn OUST Se ete 2(2 7. cle oe can wien Se 350 XXXII.—New Paleozoic Insects from the Vicinity of Mazon Creek, lilimois:;by A. HANDIIRSCH 525)... 22-2 22.2254 358 XXXIII.—New Family of Reptiles from the Permian of INewnMexicaGiby S) Wee WiILLISTON, 2222 55).222 205508 378 XX XIV.—New Elasmobranchs from Solenhofen in the Car- negie Museum ; by C. R. Hastwan. (With Plates I-III) 399 XXXV.—Contributions to the Geology of New Hampshire, No. V ; Petrography of Tripyramid Mountain ; by L. V. IPATESSIORN regener cs cls Soe ee a | 405 XXXVII—Geologic and Petrographic Notes on the Region about Caicara, Venezuela; by T. A. BeNDRAT..------- 443 SCIENTIFIC INTELLIGENCE. Chemistry and Physics—Researches on Polonium, Mdme. Currie and A. DEBIERNE, 453.—Introduction to General Chemistry, J. T. Sropparp: Die Beziehungen zwischen Farbe und Konstitution bei organischen Verbin- dungen, H. Lay, 454.—Rays of Positive Electricity, J. J. Thomson, 455.— Focal Isolation of Long Heat-Waves, RuBprns and Woop, 456.—A Method of Calibrating Fine Capillary Tubes, T. R. Merron, 457.—An Intro- duction to Thermodynamics for Engineering Students, Jonn Mints, 468. Geology and Mineralogy—Denudation and Hrosion in the Southern Appa- lachian Region, L. C. Giunn, 458.—Preliminary Notes on the ‘‘ Chazy” Formation in the Vicinity of Ottawa, P. E. Raymonp, 459.—Die Fauna der Spiti-Schiefer des Himalaya, ihr geologisches Alter und ihre Weltstellung, VY. Unuic: Beitrage zur Geologie der Baren-Insel, Spitzbergens und des K6nig-Karl-Landes: Historical-Stratigraphical Review of the Silurian of Sweden, Jon. Cor. Mopere, 460.—Geologisch-petrographische Studien in der Patagonischen Cordillera, P. D. QuENSEL, 461.—Ueber einigige Japan- ische Vulkane, I. FRIEDLAENDER: Geological Survey of Ohio, 462.—Ele- ments of Geology, BLACKWELDER and BarRows: A Remarkable Crystal of Beryl, G. F. Kunz: The Mineralogy of Arizona, F. N. Guiip, 463.—Notes on a recent find of Zincite Crystals, A. H. Puriures, 464. Miscellaneous Scientific Intelligence—National Academy of Sciences, 465.— Bulletin of the Seismological Society of America, 466.—Commercial Geography, E. V. Ropinson, 467. Obituary—S. F. Emmons, 467: S. Catvin: J. M. Van BeMMELEN: Mrs. EK. H. Ricnarps: HE. E. Hows, 468. Vili _ CONTENTS. Number 186. ; . Page Art. XXXVIII.—Podokesaurus holyokensis, a New Dino- saur from the Triassic of the Connecticut Valley ; by M. Parson, (With Plate GV.) e222 22k 469 XXXIX.—Minerals Associated with Diamonds and Carbon- ados in the State of Bahia, Brazil ; by J. C. Branner.. 480 XL.—An Engelhardtia from the American Eocene; by E.. W.. BER By ee RE Re oo geet en 8 cha 491 XLI.—Use of Sodium Paratungstate in the Determination of Carbon Dioxide in Carbonates and Nitrogen Pent- oxide in Nitrates ; by F. A. Goocu and 8. B. Kuzirian 497 XLII.—Influence of Pressure on the Melting Points of Certain Metals; by J. Jonnsron and L. H. Apams ..-- 501 XLIII.—A New Occurrence of Pearceite; by F. B. Van Horniand Cx WicCoom a2 5. caer cee ae XLIV.—Mollugo verticillata L.; by T. Horm .--.---.---- 525 XLV.—Composition and Crystallization of Parisite and Occurrence in the Granite-Pegmatites at Quincy, Mass., - U.S. A,, etc., by C. Panacue and C, H. Warren ----- 538 XLVI.—Notes on the Absence of a Soil Bed at the Base of the Pennsylvanian of Southern Ohio ; by J. E. Hype __ 557 ' XLVII—A New Jolly Balance; by E. H. Kraus -----.--- 561 XLVIII.—Independence of the Coronas of the Thickness of the Woe Waiver, siloyg, @. gly AS seat eee eee 564 SCIENTIFIC INTELLIGENCE. Chemistry and Physics.—Radium Contents of some Uranium Minerals, MarckwaLp and RvussELL, 566.—Determination of Cane Sugar in the ~ presence of other Sugars, A. Jottes: Action of Sulphur Dioxide upon Ammonia, EPpHRam and ProrRowskt, 567.—Quantitative Chemical Analy- sis of Mixtures, H. FRIEDENTHAL: Metallic Coloring in Birds and Insects, A. A. MicHELson, 568.—Isolation of an Ion, a Precision Meas- urement of its Charge, and the correction of Stokes Law, R. A. MILLIKAN, 5970.—Homogeneous Rontgen Radiation from Vapors, J. C. CHAPMAN, 571.—Lehrbuch der Kristallphysik, W. Voter, 572.—Electrical Nature of Matter and Radioactivity, H. C. Jones, 573. Geology and Mineralogy.—Illinois State Geological Survey: New Zealand Geological Survey, 573.—Geological Survey of Western Australia: Geo- logical Survey of Canada, 574.—Mineral Production in the United States in 1909, 575.—Production of Gems and Precious Stones in 1909 : Production of Bauxite in 1909: Origin of the Thermal Waters in the Yellowstone National Park, 576.—Tables for the Determination of Minerals, H. H. Kraus and W. F. Hunt: Striiverite, R. C. WELxs, 577. Miscellaneous Scientific Intelligence.—Bulletin of the Bureau of Standards : Prehistoric Period in South Africa, J. P..Jonunson, 578.—Field Museum of Natural History : Astronomical and Astrophysical Society of America, 580.—Harvard College Observatory : Cincinnati Observatory: Princeton University Observatory: R. Comitato Talassografico Italiano, 581.—Con- gress of the Applications of Electricity : Congress of Applied Chemistry : Anthropological Society : University of Bologna, 582. Obituary.—SAMUEL H. ScuppER; M. Epovarp Dupont: THomMas RUPERT Jones ; J. Bosscwa, 582. InDEx TO VOLUME XXXII, 583. > 7 ‘f VOL. XXXI. JANUARY, 1911. Established by BENJAMIN SILLIMAN in 1818. HT AMERICAN JOURNAL OF SCIENCE. Epiror: EDWARD S. DANA. — \ ASSOCIATE EDITORS Proressors GEORGE L. GOODALE, JOHN TROWBRIDGE, W. G. FARLOW ann WM: M. DAVIS,: or CampBrince, Proressorns ADDISON EB. VERRILL, HORACE L. WELLS, L. V. PIRSSON anp H. E. GREGORY, or New Haven, Proressor HENRY S. WILLIAMS, or Irwaca, Proressorn JOSEPH S. AMES, oF Bartimmore, Mr. J. S. DILLER, or Wasuinerton. FOURTH SERIES VOL. XXXI—[WHOLE NUMBER, CLXXXIJ]_ No. 181—JANUARY, 1911.9" : NEW HAVEN, CONNECTICUT. dL Oalsk THE TUTTLE, MOREHOUSE & TAYLOR ©O., PRINTERS, 123 TEMPLE STREET. Published monthly. Six dollars per year, in advance. $6.40 to countries in the either by money orders, Postal Union ; $6.25 to Canada. Remittances should be made registered letters, or bank checks (preferably on New York banks). food B ol IMPORTANT TO MINERATLOGISTS. I have the pleasure to inform my numerous patrons that I have just secured:a collection of remarkable old finds from exhausted localities, which was collected by a well-known mineralogist. This collection represents some of the finest crystallized minerals from well-known American localities some of which are not at present procurable. This collection will be placed on sale by January 15th. Full details of special list sent on request. I can safely state that no dealer has ever had at one time as many Zincites crystallized as there are in my present stock, at prices which will tempt any: collector not having these remarkable rare crystals ; also a good stock of all the well known Franklin Furnace minerals. COPPER AND SILVER FROM LAKE SUPERIOR. I have just received a large lot of excellent crystallized Silver and also Copper, Calcite enclosing Copper, Psilomelane, etc., in large and small sizes at very reasonable prices. COLORADO MINERALS. Recent additions to my already large stock of Colorado Minerals include the following: Native Tellurium, Sylvanite, Calaverite, Gold Pseudomorph, Smoky Quartz Crystals, Amazonite Crystals, Topaz, Aquamarine, Calciovol- borthite, Crystallized Carnotite, Pyrite, Rhodochrosite, Fluorite. CALIFORNIA MINERALS. In addition to my already large stock of California Tourmalines, I have made special connections with one of the large mining concerns for all their specimens and I have just received as fine an assortment as ever was found, all ranges of quality and color, at reasonable prices. Tam making special efforts in my gem department to supply Gem collec- tors with rare stones in all sizes, shapes and colors. I have also added considerably to my stock of Reconstructed Gems and have the following to offer :. Rubies, sapphires, blue, yellow, pink and white ; in all sizes and shapes of the best quality. If you are interested in anything in Mineralogy send for my special lists or request me to send you a ship- ment for your selection, prepaid. Do not fail to send your name for my mailing list, mentioning AMERICAN JouRNAL OF SCIENCE. A. H. PETEREIT, 81—83 Fulton Street, New York City. 4 DS Oto) AMERICAN JOURNAL OF SCIENCE [FOURTH SERIES. ] Arr. I. — On Gravity Determinations at Sea;* by L. A. BaveEr. A etven mass changes its weight when transported over the Earth’s surface. How shall we determine these alterations ? It cannot be done with a balance; for a mass in one scale pan held in equilibrium by a set of weights in the other pan will remain in equilibrium all over the Earth, since the Earth’s gravitational force within our limits of measurement acts alike on all substances. The variations in gravity are composed of three parts: one called the ‘normal part,” varying simply with latitude; another the “anomaly,” a more or less irregularly distributed part, and the third part dependent upon altitude of station above sea-level. We might replace the beam of our balance by a pivoted magnetized steel bar or needle. Suppose we were to start out _ from Washington, where the magnetic dip is about 70°5°, with the beam made to le horizontal by suspending a suitable weight on its south end; we have thus balanced the earth’s gravitational force by the vertical component of the Earth’s magnetic force. Were we now to proceed with this balance to some distant point, the beam might no longer be truly hori- zontal but make instead an angle with the horizon, the magni- tude of which would depend upon the relation at that point between the earth’s magnetic and its gravitational force. Knowing the value of the vertical magnetic component, it is then possible to determine, within certain limits, from the inclination of the beam how much the suspended mass has i Presented before the Philosophical Society of Washington, November 5, 0. Am. Jour. Sci.—FourtH Series, Vou. XXXI, No. 181.—January, 1911. 2 L. A. Bauer—Gravity Determinations at Sea. altered its weight, or in turn find the change in gravitational force. We have in fact such an instrument as here described, viz., the dip circle used in determining the total magnetic force by Lloyd’ s method, one operation consisting of measuring the angle of dip of a suspended magnetic needle having a load on one end. Observations made thus on the ‘ Galilee” in the Pacific Ocean, covering a range in latitude from about 59° North to 42° South, showed that the variation in weight of the load over a large range in latitude might be taken into account. With such a dip circle, the latitude variation of gravity could just about be detected ; however, for smaller gravity variations the instrument would not be sensitive enough, both because of its construction and our inability to determine the magnetic force with sufficient precision. The method most commonly in use on land for determining the value of the linear acceleration of gravity, g, is by pendu- lum observations. But this method has not been found feasi- ble on board ship. If, then, it is desired to extend a gravity survey so as to include the oceans as well as the land, other experimental means than swinging pendulums must be devised. The method thus far used to disclose gravity anomalies at sea is that known as the “boiling point, mercurial barometer method.” The principle here is to measure the counterbalane- ing effect of the elastic pressure of a vapor on a column of mer- cur y, once by reading the height of the mercurial column, the counterbalancing vapor being the atmosphere ; next, by measur- ing the atmospheric pressure prevailing at the time by deter- mining the temperature ot the boiling point of pure water. Since the boiling point, other things being equal, depends solely upon the atmospheric pressure, it will not vary as we pass over the earth as long as the pressure is the same; how- ever, the height of the mercurial column, under the same con- ditions, changes with variationsin gravity. Hence the gray- ity anomaly is found by a direct comparison of the atmospheric pressure determined from the boilmg point with that read off on the mercurial barometer. Prior to the boiling-point method for measuring the prevailing atmospheric pressure inde- pendently of gravitational force, the use of the aneroid was proposed ; however, the latter instrument is found too variable and uncertain in its indications to possess the required sensi- tiveness. Guillaume in 1894 was led to suggest the use of boiling point thermometers in place of the aneroid, but Mohn* has the credit of having made the first practical use of the method, * Mohn, H.: Das Hypsometer als Luftdruckmesser und seine Anwendung zur Bestimmung der Schwerekorrektion. Christiania, Skr. Vid. Selsk. Math.-naturw. K1. I, 1899, No. 2 (1-69). L. A. Bauer—Gravity Determinations at Sea. 3 in determining between 1896 and 1898 the gravity correction to observed barometric heights at various stations of the Mete- orological Service in Norway. It was Mohn’s suecess which led Helmert and Hecker to consider this method for ocean gravity observations. Hecker, working under the direction of Helmert, has thus far made three expeditions, chiefly at the expense of the Inter- national Geodetic Association,—the first in 1901, in the Atlan- tic from Hamburg to Rio de Janeiro and return to Lisbon ; next in 1904, a eruise extending over the Indian and Pacific Oceans, and finally in 1909 in the Black Sea. All of his work was executed in a most painstaking manner and a very elabo- rate instrumental outfit was used; the observations were made only on good-sized steamers of 5000 tons and upward. In his Black Sea eruise, of which the results have just appeared, the Russian cruiser “ Pruth,” of 5500 registered tonnage, was put at his disposal; his instrumental outfit consisted of six boiling point apparatuses, each provided with a thermometer and two sets of photographically registering barometers, one set having five mercurial barometers and the other four, or nine specially constructed barometers in all. The thermometers were read with a telescope magnifying twenty times, so that to the observer 0°001° appeared of the length 0-4". The nine photo-baro- grams were independently read by two assistants and correc- tions for various sources of error were applied. Hecker also devised instruments for photographically recording the ship’s motions, with the aid of which further corrections were deter- mined. Finally, an elaborate adjustment by the method of least squares was made of the outstanding differences between the atmospheric pressure, A, derived from the boiling point work, and that, 6, resulting from the barometric readings referred to standard temperature and normal gravity for lati- tude 45°; there were thus determined further corrections, as explained later in this paper. The 1901 work was published in 1903, that of 1904 in 1908, and that of 1909 in 1910. In Hecker’s last publication are given, in addition to the Black Sea results, the revised results of the work done in 1901 and 1904, so that the previous publi- cations are superseded by this latest one. The revisions were made necessary by the correction pointed out by Baron Kétvos due to course and speed of vessel. Hecker’s work in the Black Sea was done partly for the very purpose of testing whether his methods were of sufficient accuracy to detect this theoretical correction; he reaches an affirmative conclusion and, accordingly, revises his previous reductions. He now also. excludes, in the least-square adjustments, observations in ports 4 L. A. Bauer—Gravity Determinations at Sea. made on vessel at anchor; his present results for Ag differ at times from the previously published ones by 0°15°™, Suggestions have been received from various sources that it would be highly desirable to include, if possible, gravity work on the “Carnegie.” At the request of President Woodward, I consulted in 1905 Professor Helmert, Director of the Geo- detic Institute at Potsdam, as to the possibility of attempting the boiling-point method on the “Galilee,” which had just been chartered for magnetic work in the Pacific Ocean. As the result of Hecker’s experiences on vessels exceeding the ton- nage of the “Galilee” by eight times and more, Helmert did not feel warranted in advising the undertaking of similar work on our vessel; he thought it best under all circumstances to await the conclusion of Hecker’s labors. No attempt was accordingly made on the ‘ Galilee.” However, on the “Carnegie” it was decided to include determinations of the temperature of the boiling point of water in the regular routine work aboard, the prime purpose being to obtain data for controlling the corrections of our ane- roids. The instrumental equipment was in accordance with this aim; it consisted merely of two boiling apparatuses of the pattern described and figured in the British Antarctic Manual of 1901, p. 94, two specially constructed thermometers by Green, of Brooklyn, N. Y., graduated into one-hundredths of a degree centigrade from 97°6° to 107-7°, the length of one degree being about 40 millimeters, and one Green mercurial marine barometer. In all 106 determinations of the boiling point were secured on the First Cruise, between Sept. 1909 and Feb. 1910, four of which had to be rejected because of manifest errors, leaving 102 values and representing 75 differ- ent points. While a few observations were made at the very beginning of the cruise, by Mr. J. P. Ault, the navigating officer, the work did not begin regularly until the vessel left Falmouth on November 6, 1909, but thereafter to Madeira, thence to parallel 20° North and return to Brooklyn via Ber- muda, the observations were made almost daily by Dr. C. C. Craft, and that too at times under very trying conditions of weather. A week’s series or more was obtained at each of the ports,—Brooklyn, Falmouth, Funchal (Madeira), and Hamilton (Bermuda). The two boiling-point thermometers were read visually, with the aid of a hand lens, to the nearest 0:001° by estimation of a tenth of a space 0°4"" long; and the mercurial barometer was read directly by vernier to 0-01 inch and by estimation to 0-005 inch or less. The pumping of the barometer, which is of the ordinary marine type, amounted at times under the severe conditions of sea encountered on the return trips to as much as LI. A. Bauer—Gravity Determinations at Sea. 5 5™™; several settings were made, and both the low and high readings were recorded. To reduce the pumping, Hecker had introduced a special capillary tube in about the middle part of his barometer, and, since his observations were made on large steamers, the pumping of his barometer was generally less than 075". A eareful scrutiny of our observations has encouraged me in the belief that it may be worth while to attempt also gravity work on the “ Carnegie,” which in her various cruises will have opportunity of getting data in regions not yet coy- ered, and will also at times cut across Hecker’s trips. Atten- tion is accordingly being given to the question of refining the instrumental appliances and simplifying the method of reduc- tions. As the best preparation it was thought well to review carefully Hecker’s work in order that full advantage might be taken of his experiences in this pioneer work, as also to determine what were the various sources of error and their rela- tive importance. (Cf. Bestimmung der Schwerkraft auf dem Schwarzen Meere, etc., Berlin, 1910.) Hecker’s Ocean Gravity Observations. Let us begin with the formule used, and the theoretical treatment applied to the observations. Let A=atmospheric pressure deduced from the temperature of the boiling point of pure water with the aid of tables, as for example Wiebe’s, given in the “ Landolt—Boérn- stein Tabellen” for 1905 ; £=the simultaneously observed atmospheric pressure with a mercurial barometer, reduced to standard temper- ature, to sea-level and to normal gravity for latitude, p=4o° ; thenis @= A— Binmms. mercury. (1) Were there no errors of whatever kind attaching to A or B, then 8 would be the gravity anomaly sought. To convert into ems. of the acceleration of gravity, g, we must multiply equation (1) by the approximate factor, 980/760 = 1:29, hence Ag = 1:29 (A — B) in ems. per second. (2) The reduction of & to normal gravity is made with the aid of Helmert’s formula of 1901* : ¢ = ( — 0002644 cos 26 + 0:000007 cos’ 26) B. (3) The coefficient of the second term was adopted by Helmert from the theoretical investigations of Wiechert and Darwin; * See note next page. 6 L. A. Bauer—Gravity Determinations at Sea, however, the first coefficient, 0:002644, he deduced empirically* from a least-square discussion of nearly one-fourth of the available pendulum observations, selecting undisturbed coast and inland stations. This first coetticient gives for the ellip- ticity of the earth, 1/2983; the value adopted in the Smithsonian Meteorological Tables, 1907, is 0°002662 as obtained by Professor Harkness in his work “The Solar Parallax and its Related Constants, Washington, 1901.” For B = 760, for example, the first corrective term for a point on the equator would be, —2-0095™™"s for Helmert’s formula and —2-0231°™S for the Smithsonian Tables; the second term for 6 = 760™™ and the equator amounts to +0:0053, so that the total correction, according to Helmert, would be —2°0042"™"*. For the poles, the corrections would be, +2°0095 and +2°0231™™s. In order, therefore, to detect by ocean observations the difference between the two formule, it would be necessary to secure an accuracy of about 0:01™ mercury or 0:015"™ acceleration or about 1/100,000 part of g. This matter is mentioned here since one of the conclusions drawn by Hecker from his ocean observations is that they accord with Helmert’s formula, But A and & are subject to various sources of error, partly due to instrumental causes and observational errors and partly due to motion of the vessel. Of the disturbances caused by the vessel there are two which may readily be disposed of. First that due to tbe possible attractive effect of the mass of the vessel, since this even for a 100,000 ton vessel would only be on the order of 1/1,000,000 of g, is negligible; second, that due to the course and speed of the vessel. Only the motion in longitude counts—for a vessel sailing east along a certain par allel the instruments aboard are being transpor ‘ted around the axis of rotation of the Earth faster than is a fixed point in the same parallel and the force of gravity aboard is accord- ingly diminished and the mercury in the barometer made to stand correspondingly higher than it would were the vessel not moving. For a vessel sailing west, the effect is reversed. So that if at a certain point on the earth gravity is measured aboard a moving vehicle, once when moving eastwardly and next moving westwardly at the same rate of speed, the values of g at the two times would differ by twice an error, the exact amount of which may be computed from the following formula : *Helmert, F. R. Der normale Theil der Schwerkraft im Meeresniveau. Sitzber. Akad. Wiss., Berlin, xiv, 1901. There is a misprint in Hecker’s publication of 1908; at top of. table, p. 226, Helmert’s coefficient is given as 0:00244 instead of 0°002644. ih al LI. A. Bauer—Gravity Determinations at Sea. 7 Let c’ = correction to an observed mercurial barometric height on account of speed and course of moving vessel, C= velocity of a point in the equator = 4:65 x 10° em. / see. y= “ * ship along a parallel of latitude. FR = earth’s mean radius — 6°38 X 10° em. then is Fa UOON AO rime. sane tec Ly Eau vind — 980 Bi s' 6=0'0146 cos’ ¢ inmm.mereury (4) — for vessel going east. + for vessel going west. To get the correction in ems. g, multiply the tabular quan- tities by 1:29. For a oe sailing east, for example, along the equator at. the rate of 7 degrees or 420 knots a day, the atmospheric pressure sheerved on board with a mercurial bar- ometer would have to be diminished by 0°:1™™. On the other hand, for the same vessel going west at the same speed, the barometer reading would have to be increased Oy Wee Lie the first case g would be too low by 0°15°" and im the second too high by the same amount, or the resulting error, if not taken into accouat between the two occasions, would be 0:30 or 1/3300 part of g—a respectable quantity. "On the aver age, Hecker’s corrections for the vessels on which he made his observations was about + 007°" in g; for the “ Carnegie ” the correction would usualiy be less than + 9°03. Since the cor- rection is a perfectly definite one and can readily be coniputed, it is worth while applying. But there are more troublesome sources of disturbance arising from a moving vessel not so readily disposed of as the preceding one—arising from the actual motions of the vessel, such as rolling and yawing, vibration due to machinery, and worst of all, pitching and accompanying vertical motions. Hecker, as above stated, undertook to determine these sources of error instrumentally with the aid of devices recording the ship’s motions. He then determined the reducing coefficients for each effect by a least-square adjustment of all the observa- tions made on any one cruise. The manner of mounting, as compared with those of the barometers, as well as an examina- tion of the results of Hecker’s laborious least-square adjust- ments, leads one to question the effectiveness of his devices for the elimination of the ship’s effects. Next are the errors due to purely instrumental causes, such as changes in the corrections of the thermometers with which the boiling point of pure water is determined, furthermore the relation of the zero of the boiling-point atmospheric pressure to that of the mereury barometer, and the variations in this relation, etc., ete. 8 L. A. Bauer—Gravity Determinations at Sea. Hecker’s final observation equation is of the following form : d B B+k,+a ae +bp+er+ds+e(t—t,) +h,=0. (5) I have substituted 8 for his quantity (Therm. —. Bar. — com- puted gravity reduction to 45° — correction due to speed and course of vessel); it is the same as the quantity in equation (1) after having had the correction ¢’ applied to B. k,, = relation of the two zeros above + constant part of other corrections. ad B ag bee correction to reduce the observations of boiling Me poimt and readings of barometer to same moment of time for any one set. bp = correction due to pumping, p, of barometer. cr = 6 “« © rolling of vessel. ds — ce oe 66 pitching oe 66 e(t—t,) = ue «“ “ time changes in instruments supposed to progress linearly. k, = constant which enters into the equation only for deep sea observations, say for depths beginning with about 2000 meters. There are thus in Hecker’s complete equation seven unknowns, kn, a, 6, c, d, e, and &,, which he determines by the method of least squares. Substituting next in each observation equation, of which there is one for each station, the derived values of the unknowns, a residual quantity, v, is obtained, supposed to be the gravity anomaly sought. To get a clear understanding as to the assumption implied in his formula, let us suppose first that all the corrective terms in (5) except k, and #,, by a suitable scheme of observation and in a calm sea, reduce to negligible quantities, then we have for a shallow water station, s, 8, = — = constant, (6) and for a deep water station, d, Ba= —k, — k, = constant, (7) or 8, — Bag =k, = constant. (8) The same result (8) is obtained if we suppose, in passing from s to d, the corrective terms for each station were the same in magnitude and sign. But the difference @s—d, under the conditions supposed, when multiplied by 1-29 (see equation 2) should be the difference in Ag at the two points, ’ which of course would not, in general, be a constant. In other L. A. Bauer —Gravity Determinations at Sea. 9 words f is composed of two distinct quantities, one, Ay, repre- senting the gravity anomaly and the other the various sources of error, 2, “and so we have: Ag mag. Ag Oe > 09 oe (9) Comparing this equation with Hecker’s (5), it must be evi- dent that his corrective terms include the effects of the very quantities—the Ag’s—to be determined. Since he applies the method of least squares to his equation, Hecker must asswme that during a cruise the local gravit y anomalies, 1.¢., the Ag’s, partake of the nature of accidental errors—that they either balance out in the long run or oscillate about a mean constant value, which enters into the constant terms of (5). But is not the proving whether such distributions of gravity anomalies exist, or do not exist, the very purpose of gravity surveys ¢ Furthermore, since Hecker adjusts each cruise by itself, then by the theory of least squares alone, the sum of his residuals or outstanding gravity anomalies must reduce to zero, or prac- tically so, because of the presence of the constaut terms in (5); hence his average computed § tor the cruise must be ¢heoreti- cally equal to his average observed 8, or in other words, the average gravity anomaly of @ whole cruise would be zero. It must be evident then that as Hecker derives his unknowns they are not true values but are affected by the gravity anomalies over the areas for which the adjustment is made. They might be different, for example, for a cruise from New York to Liverpool than for one from Hamburg to Rio de Janeiro, even - though all conditions remained precisely the same except that of difference in route followed. Manifestly then Hecker’s method of adjustment is open to grave objections and it is a question as to how much of his resulting conclusions may not already be contained in his fundamental assumptions. Let us hope that the variations in the gravity anomalies at sea about an average value will be found to be of asufficiently accidental nature to vitiate Hecker’s main conclusions! Strictly speaking, the values of the unknowns entering into equation (5) can only be derived from stations where “there exist accurate gravity observations from which the anomaly Ag can be derived. This means, however, restriction to shore and harbor observation, but these are the very observations which above all Hecker has been unable to reduce satisfactorily 10. LL. A. Bauer—Gravity Determinations at Sea. and hence either omits entirely in his final tables or brackets as doubtful. In the first place the coefficients, 6, c and d, can of course only be found from observation on a moving vessel. How poorly the Ag’s from the harbor observations agree, in general, with those resulting from shore pendulum observations when the former are computed with the values of the unknowns obtained from the observations on the moving vessel, is shown by a table which Hecker gives on p. 159 of his 1910 publica- tion. The differences amount at times to 1/6000 part of g. Hecker believes that the trouble arises chiefly from the fact that observations on a vessel moving and on one at rest are not comparable and, hence, require separate treatment, the difference arising chiefly from the dynamic conditions which enter in on the moving vessel. While he is undoubtedly in the main correct, still he does not appear to see that the un- knowns as he derives them are not strictly instrumental or ship constants, but depend, as has been shown above, upon the area (extent and geographic position) from which they are derived. In any case, beyond revealing the discrepancies, he does not make known any attempt at a satisfactory reduction of the harbor observations. This is doubly unfortunate, first, because the harbor observations ought to furnish the best criteria pos- sible of the absolute accuracy and possibilities of his method of observation, and secondly, since the connection of ocean results with land stations is correspondingly diminished in strength. Every series of observations made by Hecker on shore or in port has been investigated, and not a single case of satisfactory reduction or adjustment was found. On his first eruise in 1901, in the Atlantic Ocean, from Hamburg to Rio de Janeiro and return to Lisbon, he made shore boiling-point observations at Rio de Janeiro and Lisbon at precisely the same places where he swung his pendulums. There was thus afforded a fine opportunity to test his boiling-point method and the behavior of his instrumental appliances. But he makes no attempt at such a comparison. Instead, he merely adjusts the series of shore observations at Rio de Janeiro o, Aug. 24—-Sept. 11, 1901, by itself and similarly the series at Lisbon, Oct. 12-17, 1901, again by itself. His observation equation is the same as (5) above with the omission of the terms involving J, ¢, d, and &,, which apply only to observations at sea. While his adjustment improves the ¢ndividual day’s results at each of the two stations, it leaves unaltered the actual mean gravity anomaly observed at each station—in brief, he does not adjust Rio de Janeiro and Lisbon together, and the labor of his painstaking adjust- ments is practically. for naught. Hence, if we take the quantities, as derived from Hecker’s adjustments (or from the L. A. Bauer—Gravity Determinations at Sea. 11 direct observations), we may see what the extent of énstru- mental changes may be during even such a brief interval as six weeks, during which, thermometers are subjected to fre- quent and protr acted boiling. The mean Ag results for Rio de Janeiro—Lisbon derived from each of four barometers—two eye-reading ones and two photographically recording ones— differ from the pendulum value by —0-105 to +0: 200°", thus exhibiting a range of 0°3. Even the two visual barometers give results from shore observations differing by 0-1 and this in spite of Hecker’s laborious method of observation. The mean result here considered for any one barometer depended on 24 boiling point determinations and 8 barometric readings times the number of days, or for Rio de Janeiro, 360 B. Pts and 120 readings of each barometer and for Lisbon 216 B. Pts and 72 readings of each barometer ! Hecker made no shore observations by the B. P, method on any of his subsequent cruises, but he made a number in harbors on board vessels at anchor. These also exhibit most marked changes in but a few days, the effects of which if likewise experienced at sea, as must undoubtedly be the case, would exceed in impor tance the corrective terms in equation (5) due to motions of ship. Jn his Black Sea work, Hecker had repeated trouble with his thermometers so as to be obliged to discard some series entirely. The thermometers were made by Fuess of Steglitz of Jena borosilicate glass 59 III. Looking over Hecker’s scheme of observations, the suspicion is awakened that he ‘‘ boiled” too often and too protractedly —a fact he himself began to suspect in his later work. What accuracy was supposed to be gained by excessive observing was lost in resulting instability of his thermometers. he corrections jor Hecker’s thermom- eters were never re-determined after they had once been fur- nished by the German Pha ysihalische Reichsanstalt. Though some of the thermometers had been in use on the three cruises of 1901, 1904 and 1909, practically the same table of thermo- metric corrections is employed throughout. Three of them were provided with zero points but the zeros were never re-determined. Zhe corrections for the various barometers on a standard barometer for various barometric heights were never determined, or it so, they were not used, the observer supposing that all instrumental changes—both of thermometers and of barometers—would fully be taken account of by a con- stant term (#,, equation 5) and by a term, ¢ (¢—7,), progressing linearly with the elapsed time. Let it ‘be remembered that these two quantities #, and e were not derived from observa- tions at stations where Ag was known from pendulum work, but from the discussion of ocean observations for which a fictitious distribution of gravity anomalies had to be assumed in order that a least-square adjustment could be made. 12 L. A. Bauer—Gravity Determinations at Sea. The first point to be made, therefore, on the instrumental side is, that in order to secure desired accuracy in gravity determinations from boiling-point observations, it is essential that a method of observing be adopted which will protect, as nearly as possible, the instruments from changes of whatever kind, and next that the boiling-point thermometers be provided with zero points, the variations of which may be determined in the field with melting ice once a week or as often as may be found necessary. The next point is that the method of observations be such that they can be quickly reduced and that too in such a perfectly definite manner as to admit of no ques- tion with respect to the logical method of reduction to be employed. Hecker, as shown aboye, did not lay sufficient stress upon these vital points. It is believed that equally good, if not indeed superior results, can be obtained with less equipment than used by Hecker, using a simpler method of observation as well as of reduction. Hecker’s cumbersome adjustments at times appear to have caused much needless labor. See, for example, his Black Sea adjustments, where he has attempted to derive his many unknowns from an insufh- cient range of conditions. Another very important point introducing a source of error not considered by Hecker is with regard to the possible errors in the vapor tension tables used to convert boiling-point tem- peratures into corresponding atmospheric pressure. The latest of these tables are those of Wiebe’s given in Landolt-Born- stein’s “Physikaliseh-Chemische Tabellen” for 1905. The most recent observations appear to be those of Holborn and Henning. For the purpose of gravity work, it is essential to be able to obtain accurately the atmospheric pressure for a com- paratively limited range extending below and above 100° C.; the observations on which the tables are based were made at larger intervals and the interpolation is accordingly somewhat uncertain. It is quite possible that the atmospheric pressure as taken from the tables may be out by -05 to 0:1", which corresponds to 0-065 to 0°135™ in g. When dealing with only differential results, as we are in our case, the tabular errors are somewhat eliminated, though not wholly. Zhe problem of most accurate vapor-tension tables for water between 99° and 101° is here called to the attention of physicists. Hechers Gravity Results. From the explanatory statements on p. 150 of his 1910 pub- lication, it is seen that Hecker uses a different plane of refer- ence for the gravity anomalies, the Ag’s, over each ocean, and that the planes refer strictly only to the parts of the respective oceans traversed. No direct comparison can in consequence be made in passing from one ocean to another and il L. A. Bauer— Gravity Determinations at Sea. 18 even for the same ocean, eg., the Atlantic, it would not be possible to compare directly gravity anomalies between New York and Hamburg, with Hecker’s between Hamburg, Rio de Janeiro and Lisbon. He does not explain completely how he actually connected ocean results with pendulum stations; for example, how he. distributed the correction from one land station to another. Why he did not refer his Atlantic Ocean results likewise to his pendulum stations, e.g., at Rio de Janeiro and Lisbon, he does not say. All this confusion has come about because of Hecker’s method of adjustment, as already explained, whereby he discards the shore and port boiling-point observations ab inztio and gets his unknown coefficients from a least-square adjustment of ocean observations. Having done that, he finds that the port observations computed with these coefficients give results not only very discordant among themselves but also with the pendulum observations. He then has the difficult problem of connecting his ocean results with land pendulum stations by means of more or less discordant port and shallow water stations. Tables I and II were drawn up from the figures in Hecker’s 1910 publication ; a plus sign means that g at the place in ques- tion is greater than it would be did the local disturbing cause not exist, and a minus sign means, of course, the reverse. It must be recalled that the tabulated Ay’s are those as derived from Hecker’s adjustments ; if we may assume them correct, a mere glance shows at once that the disturbances in g are, in general, larger over the oceans than usually observed on land. Table I would show that the difference in Ag for two oceanic points may reach 0-4 and even 0°67. TaBLE I.—THE AVERAGE, THE MAXIMUM, AND THE MINIMUM VALUES OF GRAVITY DISTURBANCES AS SHOWN BY H&rCKER’S OCEAN OBSERVATIONS. (Revised figures 1910.) > |oA Ag Route = 168 A |AO|Average |Maxi’um| Mini’um) Range em em em em Hamburg—Rio de Janeiro-___- 1901} 47 | +0-048 | +0:172 | —0:095 | 0-267 Rio de Janeiro—Lisbon ___.__-_ 1901) 35 56 | +0°142 } —0-123 | 0°265 Spain—Suez—Colombo ----.--- 1904) 33 83 | +0°289 | —0°106 | 0°395 Colombo—Sydney -. .____....|1904/ 28 61 | +0°214 | —0-106 | 0°320 Sydney—Honolulu—San Fran- CISCO. =: eee a ee ELS 1904| 41 106 | +0°393 | —0°273 | 0°666 San Francisco—Honolulu—Yo- kohamla SS Sueeaee ese ee 1904) 33 54 | +0°310 | —0-067 | 0°377 Black Sea (Odessa—Batum) ____/1909) 15 31 | +0:079 | —0-052 | 0:131 Entire Work, 1901—1909___. 232 | +0:066 | +0°393 | —0-273 | 0-666 14 L. A. Bauer—Gravity Determinations at Sea. TasLe II.—Some Larce Gravity Distursances Saown By Hecker’s OBSERVATIONS, 1901-1909. mate riser figures 1910.) Cruise \Ye’r| Ag Lat. Long. | Depth Region | em m ° ( | | + 0°172|35° 02’ N) 11°56’ W)| 3600 |Deep sea. ge a ee a )ig01/4 1488 36 S| 34 58 W| 40 [Nea Pernambuco. ; | — 095/11 44 N) 26 59 W) 5600 |Deep sea. ( \— 12311 35 S) 86 49 W) 3200 |Deep 2D) AnD 100 km. or Rio de Janeiro— J 4901 + 123) 6 238 S} 38 20 W) 5000 of move from Bra- Lisbon } |+ °142| 2 15 N| 29 38 Ww! 2000 « \ zilian coast. L + 123) 1 04 N| 30 08 W) 2400 |Near St. Paul. {| + 289/48 34 N) 9 30 E) 200 |Med. Sea, N. of Corsica. | (+ °23512 34 N) 55 45 E) 3400 |Near Socotra. + °214/35 49 N|129 64 E) 5400 |Steep gradient. | + °346/33 49 S/151 54 HE} 200 Bt D f + °393/384 17 S172 07 HE) . 150 |Near N. point N. Zealand. So even i 24 Wig0dl ms pe6di8 20 S178 27? E) 2700 |Tonga Plateau. {| + 161/27 15 S177 40? E| 2700 iN | | — ‘196/23 12 S174 47 E) 8000 |Tonga Deep. | — 273/22 07 S|174 13 | 6500 ui — °245)17 09 S171 42 H 8500 re L + 26821 17 Nj157 50 KE) 20 |Roadstead of Honolulu. San Franciseco— ; 1904) + *310/21 18 N\157 387 E). 70 |Near Oahu. Yokohama + °304/21 17 N\158 17 | 1700 . | + ‘079/44 51 N| 82 46 E) 150 |Shallow water. BLISS ae i 1909)” -052/43 36 NI 35 50 El 2200 ‘Deep water. Table III gives a comparison between Hecker’s 1908 and 1910 values of Ag for certain characteristic points in the Pacific Ocean selected by Prof. J. F. Hayford in his paper before the meeting of the International Geodetic Association of 1909.* The last two columns are the differences between Hecker’s values of g and those computed by Hayford with the aid of his new method; I have myself added the 1910 figures. TasLeE III.—ComMpariIson OF SOME OcEAN GRAVITY ANOMALIES OBSERVED BY HECKER IN THE PACIFIC OCEAN WITH THOSE RESULTING FROM HAYFORD’S COMPUTATIONS. Depth | Hecker’s Ag | Ee Se iS Name of Station in Lat. Long. : Hayford Fe ec Niners 1908 | 1910 | 1908 | 1910 1/Between Honolulu and San cm cm em em Francisco, at sea -_-.___- 5100 |28° 10’ N|146° 35’ W|—0-001 —0-010) + 0.003) —0-006 2\Tonga Plateau, at sea-__---- 2700 |28 20 S\178 27 W)+ 215) + -264)+ -207/+ -256 3/Tonga Plateau, at sea______ 2700 27 15 S177 40 W/+ -135)+ +161)+ -124/+ +150 4\Tonga Deep, at sea ------_- 6500 |22 07 S174 13 W/— -271)— -273)— -181/— -183 5/Tonga Deep, at sea -. ___- 8500 17 09 S171 42 W\— -248/— -245/— -162)— -159 6|Near Hawaiian Islands, atsea) 4000 |22 50 N\160 23 W\+ 034)+ -062)+ -023)+ -051 7|Near Oahu, at sea _ ___---- 1700 21 17 Njl58 17 W/+ -278)+ -304)+ -203)+ -234 Mean with regard to sign ___ | +0020) +.0-038) +.0:031) + 0-049 Mean without regard to sign 0-168; 0:188| 0:129) 07148 *Hayford, J. F. The Effect of Topography and Isostatic Compensation upon the Intensity of Gravity. Cf. Report of the Int. Geod. Association for 1909, published in 1910, pp. 365-389. L. A. Bauer—Gravity Determinations at Sea. 15 In the first place it is seen that Hecker’s mean Ag for the seven points here considered is larger for the 1910 figures than for the 1908 ones—whether the mean is taken with or without regard to sign. Next, the differences on Hayford are in every instance larger quantitatively for the 1910 figures than for the original ones of 1908 except for No. 5. Furthermore the difference, Hecker-Hayford for station No. 2, viz., +0°207 for 1908 and +-256 for 1910, is greater than any residual thus far shown upon Hayford’s computed g’s. For 56 pendulum stations in the United States Hayford’s computed values differed from the observed ones, on the average, by less than 0:02, the maximum difference being 0°094, this occur- ring at Seattle, known to be locally disturbed. Here are the differences for some very disturbed pendulum stations : TasLeE IV.—Some VERY DistuRBED LAND STATIONS. Height Station above Latitude | Longitude! g,-g. sea level m | cm lslom@lmllhn = 22.0 a Seaesensogee 6 21°18’ N |157° 52’ W, +0:0838 Mauna Kea (voleano) -_-.---. 3981 19 49 N |155 29 W) + 184 Hachinohe (Japan)____---.--- 21 40 31 N |141 30 EH) + ‘111 St. Georges, Bermuda-------- 2 32 21 N | 64 40 W| + -019 Jamestown, St. Helena_---- -- 10 15 58 S| 5 44 W) + -059 Sorvaagen, Norway _____----- 19 67 54 N | 13 02 E) + 147 Kala-i-Chumb, Turkestan ---- 1345 Bem Ne 0m Gm Ep 052 Gornergrat, Switzerland _----- 3016 45 59 N| 7 46 EH + -050 St. Maurice, Switzerland ___-- 419 46 13 N! 7 00 E| + -004 It will be noted that only in the case of two very remark- able stations—the voleano Mauna Kea and Sorvaagen, Norway, Hayford’s computed g, differs from the observed g, by more than 0-11 and in both of these cases the differences are less than 0-2. But Hecker’s revised figures of 1910 give five out of seven residuals over 0-1 and two above 0°2. Whether Hay- ford’s method fails for such deep sea stations as here considered or whether we have thus afforded an indication of the absolute error of Hecker’s values, it is not for me to say. It is curious, however, that Hecker’s supposedly most correct values (those for 1910) accentuate the differences on Hayford. Other detailed examinations made have not revealed any superiority of the 1910 method of adjustment over the pre- vious one. The difficulty with some of the port observations, e.g., at San Francisco, was found to be chiefly due to énstru- mental changes (change in thermometer corrections). If the port observations are omitted, as Hecker desires, then the mean difference in Ay between his two computations without regard 16 LZ. A. Bauer—Gravity Determinations at Sea. to sign is 0:°022™. The individual differences occasionally amount to 015°". He finally s says: “All conclusions drawn in the previous publications remain unaltered.” These main conclusions are : “The acceleration of gravity over the oceans traversed is approximately normal and conforms with Helmert’s gravity formula of 1901. Pratt’s hypothesis of isostatic adjustment of the masses of the earth’s crust is thus, except for local anoma- lies, found to hold true generally. It can be regarded hence as proved that the lesser density of the water of the oceans is compensated for by the increased density of the masses below the ocean bottoms.” My contention is that this conclusion was already practically embodied in Hecker’s method of adjustment. The conclusion may be true, but it can not be considered as proved by his mode of attack. Since no attempt was made to test whether another formula for normal gravity might not still better con- form with the observations, ‘the statement at the close of the first sentence does not seem warranted. Observations on Hecker’s Ocean Gravity Work. 1. No wholly satisfactory measure of the absolute accuracy of the existing ocean gravity results can be secured by a mere perusal of the publications. If an independent examination is made and such checks applied as are possible, and when all sources of error are considered, it will not be surprising if it be found that many of the most recently published results are in error by an amount approximating to 01°, or about 1/10,000 part of g. In view of the pioneer nature of the work, it would have been desirable to have repeated observations, under dif- ferent conditions, over all regions previously traversed. 2. One of the chief sources of error is to be ascribed to in- constancy of the corrections of the boiling-point thermometers caused by their continued and protracted use; the error thus arising may at times transcend in importance all other ones, an error in the temperature of 0:01° C. corresponding to about 0°35 in g. Insuflicient attention was paid to purely instru- mental changes and corrections. Thus, for example, correc- tions for the boiling-point thermometers of the Atlantic Ocean work of 1901 were used practically unaltered throughout the subsequent cruises of 1904 and 1909—after having once been supplied by the Physikalische Reichsanstalt, the corrections were never again redetermined. No separate examination of the barometers by comparison with standard barometers appears ever to have been made. The belief that such purely instru- mental changes would be fully taken account of im the adjust- DOr val Bauer—Gravity Determinations at Sea. 17 ment is shown to be fallacious. A source of error also not considered is that due to possible imperfections of the vapor tension tables. 3. Insufticient evidence has been given to prove that, in the reduction of the observations, it is best to omit those made on board vessels at anchor. A method of adjustment which must assume practically what is to be proved, and which necessitates the rejection of data secured under supposedly the best conditions, weakening thereby the connecting link between the ocean results and the shore pendulum stations, can hardly be regarded as the best possible one. Instead some logical method of observation and of adjustment must be striven for, which will take advantage to the fullest possible extent of the shore and harbor results. 4, The problem of obtaining sufficiently reliable ocean grav- ity results still awaits solution. Method to be Tried on the “ Carnegie.” The method it is proposed to try on the “ Carnegie,” begin- ning, if possible, at Cape Town in about April of 1911, is practically the same as that employed in the magnetic work. At all ports visited there will be both shore and harbor observations, especially at those places where g has been observed with pendulums and where accordingly the anomaly Ag is known, thus permitting a logical determination of purely instrumental constants. Our provisional equation of condition for such stations will be of the followimg form, @ having the same significance as in equation (1) above : p— 9% —h + a(t-t,) + b(B-B,). (10) k=k,+ k’, = constant part (%,) of the relation between the zero otf the thermometer and the zero of the barometer plus the constant part k’, of the errors of the vapor tension tables. It is hoped also by zero point determinations of the thermom- eters and by comparisons of barometers with port stand- ards wherever there are such, to determine #, independently of &’, and thus gradually get some idea of the various errors. a (¢—t,) is to represent the change in instrumental constants with elapsed time from some mean epoch, ¢,; it may later be found necessary to introduce a quadratic term, a’ (t—2,)*, but it is believed that, with proper care of instruments and with sufficiently frequent zero determinations of the thermometers, this term may be avoided. b(L-—8,) is to take account of the variations not included in the time term, but dependent upon barometric height or upon Am. Jour. Sci.—Fourts Series, Vou. XX XI, No. 181.—January, 1911. 2 18 LZ. A. Bauer—Gravity Determinations at Sea. boiling point; we may possibly find that this term can be taken account of by the special observations for %, as mentioned above. Hecker’s term a = is to be eliminated partly by method of observation and partly by refinement of instrumental appliances. The three terms, bp +cr+ds of his equation (No. 5), supposed to represent the effects of the ship’s motions, we shall endeavor to make negligible as far as possible or reduce to one term, dp, partly by the manner and place of mounting and by construction of the barometers, and partly by the scheme of combination of the observations, so as to introduce varied conditions of motions of vessel. It is thus hoped to avoid any need of a laborious and time-consumizg adjustment of the ocean results, thereby enabling the observer to make as nearly a complete reduction of his observations aboard as may be possible, the determination of the effects from instrumental causes disclosed by the shore and harbor observations being left to the office computer. The various sources of instrumental error—thermometer and barometer—are at present being further examined. It is possible that the temperature of the boiling point will be deter- mined both with mercurial thermometers of special construe- tion and with electrical resistance thermometers. The chief difficulty now appears to be in the sufficient refinement of the barometric work. The hope is entertained, however, that the great importance of getting values of g within the accuracy demanded by geodosists—about 0:02 or 0:03°°—will lead some one to discover a method so superior as to eliminate the boiling point-barometer method altogether for ocean gravity work. R.S. Bassler—Deep Well at Wuverly, Ohio. 19 Art. Il.—The Stratigraphy of a Deep Well at Waverly, Ohio ;* by R. 8S. Bassiur. Some months ago, in the course of routine work at the National Museum, the writer had occasion to identify a num- ber of characteristic Eden fossils from a set of photographs sent for determination by Mr. Peru Hutt, of Waverly, Ohio. The occurrence of this fauna at Waverly, in a region of Missis- sipian strata more than 60 miles from the nearest outcrops of the Eden formation, led to a correspondence with Mr. Hutt, in which it was learned that the originals of the photographs had been obtained from a deep well drilled for oil at that place. Mr. Hutt had carefully saved enough samples from all of the material resulting from the boring to prepare two very detailed logs, which he was kind enough to forward for study. He is to be commended for his zeal in the matter, for, without his eare, the following determinations, which are believed to be of some interest, concerning the underground stratigraphy could not have been made with any degree of accuracy. The notes resulting from the study of the two series of sam- ples were discussed with Dr. E. O. Ulrich at the time, and then set aside for future reference. Later the subject was mentioned by Dr. Ulrich to Professor Schuchert, who, in turn, deemed the section of sufficient importance to request a short article upon it for this Journal. The well was drilled to a depth of 3,320 feet. The upper 1,100 feet were cased, so that the samples from this portion were little mixed and afforded an accurate idea of the various formations penetrated. The lower 2,220 feet, however, were left open, and the samples from this part required more careful study. Still, this latter portion was not difficult to decipher, since the predominating foreign material in the lowest samples was the blue limestone and shale of the Cincinnati group, which had fallen from above, and which, on account of litho- logic characters totally different from those of the white mag- nesian limestone and sandstone of the lower formations, could easily be eliminated. Instead of giving a detailed description of each of the many samples, the results of the study are arranged below in the form of a geologic section. Drilling commenced at a point 100 feet below the top of the sandstone quarries of the town, and the first 35 feet of the well passed through the lower por- tion of this sandstone. Then in descending order came the black Ohio shale, the limestones and sandstones of Devonian and Silurian ages, a good representation of the various Cincin- * Published by permission of the Secretary of the Smithsonian Institution. 20 R.S. Bassler— Deep Well at Waverly, Ohio. ‘natian formations, a fair thickness of Trenton, Lowville, and Stones River, typical Saint Peter sandstone, and, finally, about 300 feet of rocks assigned to the Canadian. The various thick- nesses given must be considered as only approximate, mainly because they were calculated from two distinct logs. For some reason the samples had been arranged in two sets, one measuring from the top to a depth of 2,020 feet, and the other from the bottom to a height of 2.200 feet. The two overlapping por- tions were correlated with little difficulty, because this part of the section was the most fossiliferous. The base of the well is of particular interest and will be discussed later. Geologic Section at Waverly, Ohio. Thickness in feet. Depth. Mississippian: (b) Fine grained, drab ‘‘ Waverly ” sand- stone exposed in hills of town above mouthviok well: p:2 44; ae te as ee same OO) Similar standstone forming lower part of Waverly, series yess pease ee 35 0- 35 (a) Bituminous, fissile, black Ohio shale-. 450 35— 485 Devonian and Silurian: Mainly white, fine-grained sandstone with traces of white limestone. This material is so ground up by the drill that the limestone which it may have contained in some quantity has been mainly pulverized and washed away. The sandy material is evidently mostly from the Obio Silurian formations. At the base of this portion are the red and brown calcareous sandstones of Clinton ager? 227 22a ae eee eres 415 485— 900 Cincinnatian: (c) Blue shale with a few fragments of blue limestone. Fossils scarce, small portions of only Dalmanella jugosa being seen, but the strata are evidently of Richmond and Maysville age, with probably the Upper Eden shales rep- ; ReseNbe eee. oe oe 1065 900-1965 (b) Blue shales containing rather numer- ous Middle Eden fossils. ‘The species identified are: Rafinesquina alternata (Eden variety), Plectambonites seri- ceus, Dalmanella multisecta, Trematis millepunctata, Pholidops cincinnati- ensis, Callopora sigillarioides, Clima- cograptus typicalis, Byssonychia vera, Protowarthia cancellatu, Ceratopsis R.S. Bassler—Deep Well at Waverly, Ohio. 21 Thickness in feet. Depth. chaumbersi, Bythocypris cylindrica, Trinucleus concentricus, Calymene cal- licephala, Proetus? spurlocki, and Nereidavus varians. From this point to the bottom of the well, blue shale fragments holding this fauna were en- countered, but, eliminating them, the remaining formations were, for the main part, clearly represented in the SACOLOS arse ban Sear = ye ee 55 1965-2020 (a) Unfossiliferous blue and greenish shale associated with the blue shale holding the overlying Middle Eden fauna. This portion probably represents the Lower Eden and Utica divisions... -- 80 2020-2100 Mohawkian: Trenton formation. Blue clay and shale, with a few frag- ments of blue limestone. Zygospira recurvirostris, Trinucleus concentricus, and a species each of Khinidictya and Callopora, known elsewhere from the Lower Trenton, were noted. At the base of this formation a small amount of glauconitic grains were present in the*sanmples ey wae keen heh oe 125 2100-2225 Lowville and Stones River formations. Each of these formations is probably represented, but their lithology is so similar that no distinction could be made in the samples, which consisted mainly of white, clayey, and dove un- fossiliferous limestone with blue argil- laceous limestone at the bottom __-.- 600 2225-2825 St. Peter sandstone. White, saccharoidal sandstone ---- -- 175 2825-3000 Canadian. This portion naturally contained the greatest mixture of materials, but after excluding all the rock formations of the overlying beds, a few fragments of white dolomitic limestone remained. These are quite similar to the Cana- dian rocks of the Appalachians, and for that reason and the absence of chert, which is more characteristic of the Ozarkian, as well as the stratigraphic position, the correlation was made as above. .At the very base of the well a few small fragments of an igneous rock were detectedinn - 98222 2225... 320 3000-3320 92 Rk. 8. Bassler—Deep Well at Waverly, Ohio. This section presents no new facts regarding the rocks younger than the Trenton, for between Waverly and Cincin- nati, about 80 miles west, the same strata have been studied along numerous surface outcrops. The earlier Mohawkian formations are not exposed until central Kentucky, over 100 miles distant, is reached, while the Saint Peter sandstone and Canadian limestone are not known at all by surface outcrops in the Ohio valley. The section of strata penetrated by a deep well at Oxford, Ohio, is of interest in this connection. A detailed description of the log of this well was given by Joseph F. James in vol- ume X of the Journal of the Cincinnati Society of Natural History, but for present purposes only the general section determined by him is of interest. Arranged in the same form as the one given aboye, this section, with the formations as identified by James, but with the correlations of the present day inserted in brackets by the present writer, is as follows : Geologic Section, Deep Well at Oxford, Ohio. Thickness in feet. Cincinnati group: Blue limestone and shale [Richmond and Maysville]-- 360 Blue shale |[Maysvallevand ident] S25 ee = oe ae oer 380 Dark limestone! | @remt om) = eS cae eon seen ae Trenton group: White limestone with magnesia [Lowville and Stones River | 2 oes eee e)e ep ere Ce ae ee BO os Calciferous sandrock: White, arenaceous limestone [Saint Peter] ._--.------- 40 Unfortunately this well did not go deep enough to show the strata underlying the Saint Peter sandstone, nor did certain deep wells bored at Cincinnati pass beyond this formation. These Cincinnati wells showed the same stratigraphy and essen- tially the same thickness as in the Oxford well, so that the latter can be taken as typical for the region of. the Cincinnati axis. Comparing the Oxford and Waverly sections, the follow- ing conclusions may be drawn: (1) From observations on both sides of the Cincinnati axis, the Maysville and Richmond divisions of the Cincinnatian do not vary enough in thickness to suggest marked decrease of deposition across the apex of the axis. The Utica is seldom more than a few feet thick at Cincinnati. In northern Ohio it has become greatly thickened, as shown in the gas wells; it has likewise attained a considerable thickness in the Appala- chians. The increased thickness of the Cincinnatian as a whole a R. 8. Bassler—Deep Well at Waverly, Ohio. 23 in the Waverly well is thus probably due to the presence of greater deposits of Utica shale. (2) The same eastward increase in thickness may be stated for the Trenton rocks with less doubt. At Cincinnati the lower 50 feet of the Trenton are exposed with the thin Utica shale resting upon its eroded surface. Proceeding southeast along the Ohio River, this thickness increases to over 100 feet, in a distance of 30 miles, by the addition of higher beds of the formation. The occurrence of 125 feet of Trenton strata at Waverly, 80 miles east, is therefore in line with the idea that the Trenton and the Utica are alike in having a minimum thickness along the Cincinnati axis. These same facts, among others, have convinced several of the students of Cincinnati geology that this axis did not pass along a northeast-southwest line 25 or 30 miles east of Cincinnati, as commonly believed, but close to the city itself. Allowing for a certain amount of error in determination, the Stones River—Lowville sequence is practically the same in cach section. At any rate, both the Stones River and Low- ville are among the most widespread Ordovician formations, extending from Oanada to Alabama, and from New York to the central Mississippi Valley. Both the Oxford and Waverly wells are interesting, therefore, in indicating the presence of both formations in the northern part of the Ohio Valley, where they have no surface outcrops. The presence of such a typical fauna of Middle Eden species, a hundred or more miles from the shores of the sea of the time, is evidence for the shallowness of the early Paleozoic continen- tal seas. Deep wells elsewhere have furnished abundant proof of this same fact, and the one at Waverly simply furnished an additional well-established example. This Eden fauna is well known in New York, in the Cincinnati uplift,and in the Appa- lachians, and not being pelagic, it could not have had such a eeeereipe had deep seas intervened between these several shore ines. The occurrence of glauconitic material in the samples from the base of the Trenton is likewise noteworthy in indicating the unconformity between this formation and the underlying Lowville strata. Detailed studies of the early Paleozoic rocks have shown glauconite to be a common ingredient of the basal sediments of several overlapping formations. Perhaps the most interesting fact brought out by this well is the presence of a few fragments of igneous rock at its very bottom. The importance of this occurrence was suggested at once by Doctor Ulrich, for the area about Waverly is on the northward extension of the uplift which he has named the Car- ter axis. The igneous nature of these fragments was verified 24 RR. S. Bassler—Deep Well at Waverly, Ohio. by Dr. George P. Merrill, who determined them as peridotites which had become altered into serpentine. That the presence of this igneous rock is to be considered as indicating that the well passed through the base of the Paleozoic cannot be posi- tively determined from the facts at hand, as the material may possibly have been derived as a bowlder from above. However the facts so far as known are highly significant, and it may be tentatively suggested that the Canadian rocks rest upon pre- Cambrian. Whether this area was a part of an axis of uplift, such as the Carter axis, or was included in a broad southern extension of the Laurentian shield during Cambrian and Ozark- ian time, cannot be decided with the present evidence. U.S. National Museum, Washington, D. C. eX, ee ee eee ee ee Foote and Bradley—Solid Solution in Minerals. 25 Arr. IL.—On Solid Solution in Minerals with Special Reference to Nephelite; by H. W. Foorm and W. M. BRADLEY. Ir is a fact well known to mineralogists that there are certain minerals to which no satisfactory chemical formule can be assigned which agree with the results of analysis. The reason for this in many eases, particularly where the mineral is rare and little investigated, is probably that the material is impure, containing included foreign matter, or else the analysis is incorrect. There appear to be cases, however, where the material has been so carefully selected that foreign matter could not be present except in traces, and where analyses have been made with the greatest care and still the formula cannot be definitely assigned. A case of this kind is that of the mineral nephelite, to which the formule NaAISiO, and Na,A1,Si,O,, besides others more complicated have been given. An examination of several good analyses of this mineral will show that the analytical data do not support any one formula, but that there are considerable variations from it which are greater than can be accounted for by the ordinary errors of analysis. In general the composition of a mineral as obtained in analysis varies from the composition of the ideal pure com- pound for two reasons, aside from errors of analysis. Either there is (a) isomorphous replacement of one element or radical by another, or (>) there are mechanical impurities present. Where there is merely isomorphous replacement, the formula of the pure compound can be derived from the analysis by the ordinary methods of caleulation, which need not be considered here. The presence of mechanical impurities can usually be determined by other means, for instance, by the use of heavy solutions or by microscopic examination. We wish to call attention to another influence which must probably be taken into account in cases like that of nephelite. It appears to us necessary to assume that in certain cases a substance on crystal- lyzing forms a solid homogeneous solution with foreign matter which cannot be assumed to be isomorphous with any constit- uent, and which is not to be regarded as a mechanical mixture. It can be compared to the solution of salt in water, in which the salt takes on the appearance and form of the water without taking any part in the formula of the water. A case of this kind in minerals would not be a mechanical: admixture of the foreign substance, comparable to the suspension of a solid in water, but would form a homogeneous mass with the rest of 26 Foote and Bradley—Solid Solution in Minerals. the mineral comparable to the salt solution. If such an impurity were present in appreciable amount, it is obvious that the formula of the pure compound could not be caleulated correctly from the analysis. This type of solid solution must be clearly distinguished from isomorphous replacement, which is also commonly considered as solid solution. In the latter case, the formula of the compound can be derived directly from the analysis, as previously mentioned. Before considering the application of these statements to nephelite, we wish to mention a simple case of solid solution which is known in artificial crystals. It has been shown by Roozeboom* that when ammonium chloride crystallizes from a solution containing ferric chloride, the crystals deposited are colored and may contain as much as seven per cent of ferric chloride. Here there can be no question of isomorphous replacement, and on the other hand the ferric chloride is not mechanically enclosed by the ammonium chloride. The latter point is proved partly by the fact that the crystals appear pertectly homogeneous, and it is proved much more definitely by the fact that the solubility of such erystals varies with their composition. If a mechanical mixture were present, the solu- bility would not vary with the composition of the mixed crystals, but there would be a definite solubility at a given tempera- ture independent of the composition. The colored crystals are to be regarded as one homogeneous phase in which the ferric chloride is held in solid solution by the ammonium chloride. Similar occurrences have been noted in artificial minerals with a good deal of probability. Day and Shepherdt have observed an artificial calcium metasilicate crystallizing with tridymite which differs slightly in optical properties from the pure silicate. The variation appears to be due to the presence of silica taken up in solid solution by the metasilicate. The same metasilicate is also capable apparently of absorbing a considerable amount of the orthosilicate and still appear homogeneous. Again, Shepherd and Rankin{ have shown that artificial corundum may take up a limited amount of sillimanite (or silica) in solid solution and also a small quantity of calcium oxide. We believe such cases also exist in certain minerals such as nephelite. Several years ago, the late Prof. S. L. Penfield suggested to one of us (Bradley) that the reason for the variation in the composition of nephelite might be due to the presence of mechanical impurities and that if material of undoubted purity * Zeitschr. f. phys. Chem., x, 145, 1892. + This Journal (4), xxii, 265, 1906. {This Journal (4), xxviii, 293, 1909. Foote and Bradley—Solid Solution in Minerals. 27 could be obtained so far as mechanical admixture was con- cerned an analysis would show the correct formula of the mineral. A sample of nephelite from Eikaholmen, Norway, was chosen for analysis and freed from other minerals by use of acetylene tetrabromide. The sample used in analysis floated when the specific gravity of the liquid was 2°638 and sank when it was lowered to 2°632, so that variation in the density of the mineral was not more than -006. The resulting nephel- ite contained a minute amount of albite which was insoluble * in hydrochloric acid, but the quantity was so small thatit could be neglected. The material obtained was, we believe, as pure as it is possible to obtain nephelite by mechanical means, since observations under the microscope showed the sample to be of excellent quality and practically homogeneous. Two complete analyses and two other partial ones were made on this material with the greatest care (by Bradley). Only brief mention seems necessary of the methods employed im the chemical analysis. Silica was determined in the usual way, after dissolving the mineral in hydrochloric acid and by testing its purity traces of alumina were recovered. Alumina was precipitated as hydroxide and this was dissolved, repre- cipitated and weighed in the usual manner. The small per- centage of iron was determined volumetrically with potassium permanganate. A Smith’s fusion was made for the alkalies. The results in detail with the ratios obtained are given below. Taste I, Analyses of Nephelite (Bradley). Nephelite from Hikaholmen, Norway. perc 2 3 4 Average Ratio SiO, 44:59 44:31 44:37 44°59 44-46 -736 = 2:93 BOR asso yn on 02 es ola Wari t 185 Wie 824 anes ier O: 96 “96 ‘96 sons 96 :006 ay KEOe (562) 5:62 . 155% 15:59. 5-61. -060 BIG) Sy NO SiG Onn 1631 Gia W606, 16-325 263 tae” H,O emt cae ap asa cae 101:43 100-60 100-29 100:84 The above analyses were not published at the time they were made, as the formula derived from them was complex, and the results could not be regarded as establishing the formula. * Taken from anal. I. 28 Foote and Bradley—Solid Solution in Minerals. The composition of nephelite was further investigated the following year by Morozewicz,* who also gives an excellent review of the literature on the subject. The author gives analyses of six different nephelites which were apparently made with the greatest care on carefully purified material. By his method of analysis, he was able to free his material from even a trace of albite. The results and the ratios derived are given below. Taste II. Analyses of Nephelite (Morozewicz). i II Tit Nephelite (Elezolite) from Mariupol. Nephelite (Elolite) from Mariupol. Porphyritic Crystals. Coarse and granular. Ratio Ratio SiO, occas oes RSTO Mere Sem 43-46 TO. 61. ane 10 { 22 Oa oe 07 f Ae EAR O ae oes 33°12 ‘ AON eet ee 32°82 ; Re es bas) (200 Maney a ors} 100 CAO ye ieeee 0°49 CaQ a2 ee 0°31 EO yee 5°69 0799) 1) AS O Pee ee 5°55 0°99 NaI O's eee 15°91 Nia: Oe see 16°12 EE OS 0°74 AU Ress Meare 0°89 100°18 99:97 IV ; Vv Nephelite (Eleolite) Mariupol. Nephelite (Elzeolite) from Miask. Reddish Crystals. Ratio Ratio SiOmes ee 43°55 SlOve sg. eee 42°71 : 9) sap . PHO, sone st 0 t ar OP ab 0-04 } gue INCOM eat 32°96 Poa Ni OhGuu at a: 33°83 He Oe 0°66 t 100 ss tel) hy. Mena G A Bee CaO Sere ee 0°25 } CaOiee ser 0°32 KP Ope oi 6°09 100 TE OL se 5°86 1:00 Manan bee 16-00 J NaiOne ns 16-46 YO ee ee ae 0°33 EO Peeires 0°18 MgO Be ae trace Impurities -. 0°06 99°86 99°86 * Bull. Acad. Sciences Cracovie, 958, 1907. Foote and Bradley—Solid Solution in Mineratls. 29 VI VII Nephelite from Vesuvius, Nephelite from Vesuvius. Different specimen from VI, Ratio Ratio Ste aes 49°53 SiOn so) ae 4334 Ona eae 0-01 ; Serle TON - Neen at aie ALO Cee ae 33:92 AO yee 33°75 Pome We ple? pots ray. 1200 CaQ pas Sr. 1b9'7) CaQ ie tees 2°20 | is pe 0:07 | MeOus. 5. 0:24 | 9 5 iFe(O ue ela! eps @enn lc ik Omen cl. Aevinig 08 Na,O meee Sate 15°12 | INEEOR See: 15°66 J EI Oye eepre ses 0°13 HORS ee 30 0528 Impurities... 0°24 100711 100'26 We consider the seven analyses given above to be the best which have been made on nephelite. A considerable number of other analyses have been made, however, and we give below a summary of the ratios obtained from the analyses given in Dana’s Mineralogy, page 425. The numbers are the same as in the Mineralogy. Tasxe III. Ratios obtained from Analyses of Nephelite given in Dana’s Mineralogy. SiO, Al,O; etc. Na,O etc. No. 1. OM 1:00 1:00 4 Oe 9°95 (45 “99 OF BOUT o6 1:00 4, 2-29 ss *95 5. 2°20 sf “91 6. 2°24 a 1°04 7. 2°18 ‘6 1°02 8. 2°24 ce 101 9, 2°31 ‘s 97 10. 2°60 “ 1°16 11. 2°24. << 1°05 UR, 2°06 oY °94. 13. 9°14 6G *93 14, 2°29 < “97 15, 2°19 Se “98 The summary of the ratios obtained in the seven analyses first given is as follows: 30 Foote and Bradley—Solid Solution in Minerals. Taste IV. Summary of Ratios from Analyses by Bradley and Morozewicz. No. SiO, Al.O; ete. Na.O ete. It 2°93 1:00 0:98 TT: Zroili ef 0°99 1606 2°21 ee 0°99 Ve 2°21 fs 1:00 ais Pa as 1:00 VI. 2-11 cs 1:02 VIL. 2°15 re 1:03 - In this table the ratio of Na,O: Al,O, is as nearly 1:1 as could be desired. There can be no question that soda and alumina are present in this proportion. The ratio for silica varies from 2°11 to 2°23, and this variation is greater than can be accounted for either froin errors of analysis, or from the presence of impurities. For imstance, analysis No. I contains more than two per cent excess of silica if the ratios were to be the same as in No. VI. There is no case known, we believe, where silica can be considered as replacing isomorphously either alumina or soda, and if it did in this case, the ratio between these two would not be simple. The same general conclusion as regards composition may be drawn from the ratios derived from older analyses given in Table III, though many of the analyses are probably not as good as the more recent ones. Morozewicz* has shown that the nephelites may be con- sidered as consisting of two series of compounds, a normal series and a basic one. The normal series should be repre- sented by the formula K,Na,Al,4.Si,43Oin410, in Which n=8, 9, 10, and 11, and the basic series by the formula K,Na,,Al,,O,,. By this series of variable formule, the variation in composition can be expressed. This method of representing the com- position is open to the serious objection that a chemical com- pound, so far as we know, does not vary in type. Isomorphous replacement, for instance, varies the composition, but the type of compound remains the same. If nephelite be considered a solid solution, the case becomes very different. A solution may be defined as a homogeneous mixture of substances which cannot be separated by mechani- cal means and whose composition varies continuously within certain limits. This definition distinguishes a solution from a suspension on the one hand and from a chemical compound on the other. It characterizes a solution of a salt in water, and a solid solution of ferric chloride in ammonium chloride and we * Loc. cit. —— Foote and Bradley—Solid Solution in Minerals. 31 can see no reason why nephelite should not be treated in the same class. This method of considering the composition of nephelite has the advantage of being much more simple than using a series of complicated formule, and it appears to us to agree with the facts. It need hardly be said that a chemical formula could be assigned to any solution but a different one would have to be used for each change in concentration of the solution, just as Moroscewiez uses a different formula for each nephelite. From what has been said, we think it fair to consider that nephelite as it occurs in nature is not a pure compound but a solid solution analogous to the solid solution of ferric chloride in ammonium chloride. It then becomes of interest to con- sider the probable formula of the pure compound which forms the basis of nephelite. This appears to be the orthosilicate NaAIlSiO,. This formula is supported in two ways: (1) Nephe- lite has the same crystalline form as eucryptite LiA1ISiO, and kaliophilite K AISiO, which are in the same group, making it very probable that the type of formula is the same in all three eases. (2) Artificial nephelites have been prepared by Doel- ter* which have the same general characteristics as natural nephelite and vary in composition from the formula NaAISiO, to compounds containing potash and an excess of silica corre- sponding to the mineral. Perhaps the point should be emphasized that nothing what- ever is known about the actual condition of the dissolved silica, whether it is present as dissolved albite or silica or leucite or in any other form, just as very little is known about the condition of dissolved substances in liquids as to whether they are combined with the solvent. It is certain, however, that the dissolved silica does not have the properties of either ordinary quartz or albite, since it is soluble in hydrochloric acid. In the same way, the properties of a dissolved salt are entirely different from the properties of the solid. The excess of silica which can be taken up by nephelite to form a saturated solution can apparently be determined from the data given by Morozewicz and ourselves. Where albite is found intimately mixed with nephelite it is evident that the nephelite must be saturated with silica and the excess of the latter has formed albite. In this case, therefore, the nephelite should have a constant ratio of silica to alumina, and these nephelites should contain the maximum amount of silica that ean be taken up. The influence of temperature in determining the composition of the saturated solution can apparently be neglected. In our * Zeitschr, f. Kryst., ix, 321, 1884. 32 Foote and Bradley—Solid Solution in Minerals. own specimen, albite was associated with the nephelite, and Morozewicz states that albite was present in the specimens con- taining the nephelites of analyses II and III and microcline- microperthite, which would have a similar effect, in analysis IV. The ratio for silica in these four cases is 2°23, 2°21, 2°21 and 2°21, which is as nearly constant as could be desired. In analysis V, where the ratio for silica is only 2°12, the mineral is stated to be exceptionally pure, with biotite crystals on the outside. In VI or VII, where the ratios are 2°11 and 2-15, sanidine was present which might have the effect of albite, tending to raise the ratio to the saturation point, but in just these two cases (from Vesuvius) the nephelite appears to be a later growth on the sanidine and not intimately mixed with it. In these cases, then, where albite or its equivalent was not formed with nephelite, the ratio of silica to alumina shows that the nephelite has not taken up the maximum amount of silica. The most basic rock containing nephelite with which we are acquainted is an iolite described by Hackman.* This rock contains essentially pyroxene and nephelite with smaller amounts of titanite, apatite and ivaarite. There is no albite, quartz or feldspar present. The nephelite in this rock had the following composition and ratios: Ratio si0, BOO 2 ps wee ee 43°98 O13 Al,O, a AY Aa 34°93 1°00 CaO pea laa ties ee 0.36 INaiOR: Sacre ee 16 mf 0°94 K,O Eres Das eae Sh Meta 3°83 99°86 Here, again, the silica is below what we may call the “saturation ratio” of 2°21. It would be of considerable interest if nephelites could be found which closely approximated the formula NaAISiO,,. From what has been said above, such an occurrence could only be expected where crystallization had taken place from a magma so deficient in silica that albite did not form. In conclusion, the authors consider that the arguments advanced in the present article may be applicable to other minerals. Work has already been begun on the mineral pyrrhotite with the hope that similar deductions may be applied to this mineral. Chemical and Mineralogical Laboratories of the Sheffield Scientific School of Yale University. New Haven, Conn., October, 1910. * Bull. de la Commis. Geol. de Finlande, 1900, p. 9. Watson and Powell— Age of Virginia Piedmont Slates. 33 Arr. [V.—Fossil Evidence of the Age of the Virginia Piedmont Slates; by Tuomas L. Warson and 8. L. Pow tu. CoNTENTS: Introduction. Virginia Piedmont Province. Slate areas of the crystalline region. Quantico slate belt. Fossils. Arvonia slate belt. Introduction. Recent detailed field study of the slate areas in the crystal- line (Piedmont) region of Virginia by the State Geological Survey has resulted in much important information bearing on the lithologice characters, structural and age relations of the rocks, and on the sulphide ore-bodies (veins) associated with the slates of the northeastern belt. Of especial interest are: (1) discovery of fossils in the easternmost one of the slate areas ; (2) recognition of voleano-sedimentary beds intimately associ- ated with the slates in several of the areas; and (3) evidence of - the age relations of a part at least of the sulphide veins in the northeastern portion of the crystalline region, which hitherto have been assumed to be pre-Cambrian. The present paper treats only of the discovery of fossils in the Quantico slate belt, with a brief statement of its strati- graphic position, and of that of the other slate belts in the Vir- ginia crystalline region. Discussion of the age relations of the sulphide veins and of the volcano-sedimentary beds associated with the slates will be treated in another paper, now in pre- paration. Virginia Piedmont Province. The Virginia Piedmont province (crystalline area) lies be- tween the Coastal Plain and the Appalachian Mountains. It extends from the Blue Ridge eastward to the western margin of the Coastal Plain, and it widens southward (map, fig. 1). Its width increases from 40 miles in the northern portion along the Potomac River to nearly 175 miles along the Virginia— Carolina boundary. The rocks of this region are the oldest in the state, and, excepting the areas of Triassic rocks, they are all erystalline. They comprise both igneous and sedimentary masses, in many places so altered from metamorphism, chiefly pressure and recrystallization, that their original character is indistinguishable. The region is made up of a complex of schists, gneisses, and granites, with which are associated some slates, quartzites and Am. JOUR. Sct.—FourtH SERIES, VOL, XX XI, No. 181.—January, 1911. >) Aa ee “a S ~ aS = Sn ae aoa a Se STS “ol -= S y i eN fyatnoa” oss 5 iG oe i S er, ae S | & \ r=

S ~ I's V, Age of Cc SS | ajeos azewixosddy SS aS ‘ “ (S09 puoypag pue yssayWy)-41aq UspmoUs > aN INP VSS : (09 ajewaqyy) “3184 3uoWs| NVINEWWS aN ‘ ; p . A QR BIpUENA TVG 7 Neue Lode, f (sop SERED) pue Jainbney) -}aq UopUaIe\\ nN é (seg e1uerjAsq}0ds = f Z pue “paojyers “wWeIj|!AA e2Ul4d) HEGoruenh B | i. oquo Ss : /. ; (sop euueAn|4 pue Wweysulyong) “}ieq eluoniy ~ QN3931 > > 18 -118°52 The values of 27, remembering that a centimeter scale was used, are again surprisingly good. The shift is computed by the above equation. It may be eliminated in the mean of the two methods. The lens Z’ may be more easily and firmly fixed than Z. 7. Collimator Method.—The objection to the above single- lens methods is the fact that the whole spectrum is not in sharp focus at once. Their advantage is the simplicity of the means employed. If a lens at LZ’ and at Z are used together, the former as a collimator (achromatic) and with a focal dis- tanee of about 50°, and the latter (focal distance to be large, say 150°") as the objective of a telescope, all the above diffi- culties disappear and the magnification may be made even excessively large. The whole spectrum is brilliantly in focus at once and the corrections for the shift of lines due to the plates of the grating vanish. Both methods for stationary and rotating gratings give identical results. The adjustments are easy and certain, for with sunlight (or lamplight in the dark) the image of the slit may be reflected back from the plate of the ovating on the plane of the slit itself, while at the same time the transmitted image may be equally sharply adjusted on the focal plane of the eye-piece. It is, therefore, merely necessary to place the plane of spectra horizontal. Clearly a’ and @” are all infinite. 92 OC. Barus and M. Barus—Plane Grating Similar to In this method the slide Sand D are clamped at the focal distance apart, so that flame, ete., slit, collimator lens and grat- ing move together. The grating may or may not be revoluble with the lens Z on the axis a. 8. Data for the Collimator Method.—The following data chosen at random may be discussed. The results were obtained at different times and under different conditions. The grating nominally contained about 15,050 lines per inch. The efticient rod length ab was R=169'4™. Hence if 1/C=15,050~x 3937 X 338°8, the wave-length A = C-2x. Grating Lines Qa! 2a cs Stationary es eee DD. 118°30 118'19 Rotate eso .. -eae D, 118-08 118719 Stationary) suse e ese D, 118:27 118716 Rotate eles ae D 118-05 118°16 Rowland’s value of D, is 58°92X10-°™; the mean of the two values of 2x just stated will give 58° 8710-2, The dif- ference may be due either to the assumed grating space, or to the value of # inserted, neither of which were reliable abso- lutely to much within 1 per cent. Curious enough an apparent shift effect remains in the values of 2x for stationary and rotating grating, as if the colli- mation were imperfect. The reason for this is not clear, though it must in any case be eliminated in the mean result. Possibly the friction involved in the simultaneous motion of three slides is not negligible and may leave the system under slight strain equivalent to a small lateral shift of the slit. 9. Discussion.—The chief discrepancy is the difference of values for 2v in the single lens system (for D,, 118-7 and 118'5°", respectively) as compared with a double lens system (for D,, 118-2°") amounting to *2 to ‘4 per cent. For any given method this difference is consistently maintained. It does not, therefore, seem to be mere chance. The detailed investigation, which must be omitted here,* made it clear that the effect of focusing is without influence on the diffraction angle and much within the limits of observation. It is, there- fore, probable that the residual discrepancy in the three methods is referable to a lateral motion of the slit itself due to insufficient symmetry of the slides AA and BL in the above adjustment. This agrees, moreover, with the residual shift observed in the case of parallel rays in § 8. 10. Reflecting Grating—The adjustment of the plane grat- ing if cut on specular metal is nearly identical to the above, except that the collimator is fixed as a whole in front of the grating, either to the slide carrying the standard of the grating, B, or else quite in front of the cross slide AA, fig. 5 above, * See Proc, Am. Phil. Soc., 1. ¢. Rowland’s Method for the Concave Grating. nos so as to give clearance for the to and fro motion of the rail, /2. This admits of measurement of w on both sides of the slit, so that 2v, the distance apart of the two symmetrical positions for a given spectrum line, is again observed. 11. Rowland’s Concave Grating.—For the case of the con- cave grating, the accurate adjustment for symmetrical meas- urement on both sides of the slit is not feasible, because the slit and eye-piece would have to pass through each other. It is possible, however, to find conjugate foci at different distances » from the grating in the normal position, which approximately answer the purposes of measurement. Rowland’s equation (cosz/p — 1/ &) cost + (cos 6/p’ — 1/p') cos 6 = 0 where p and p’ are the conjugate focal distances for angles of incidence and deviation ¢ and @, may for 0 = 0 be written 1 lpm aligeecs a CIO ~ Th CIs Say FI where p, is the normal distance of the eyepiece, so that 1 1 2 If in figure 6, the slit S is put at p’ > & from the grating G (normal position), the image is at Hat the end of p, from G, where p, < H p. But this excess need not be so large as to inter- fere with adequately sharp focusing. The following table gives an example, in which the difference of p, and p,’ in the normal position is even over 1 foot, an excessive amount, as the distance necessary for clearance need not be more than a few inches. The grating has 14436 lines to the inch and a radius about R = 191. TABLE II. Conjugate foci of the concave grating. R=191°™, 14436 lines to inch, 5683 lines to em. D= ‘000,176. po=166™,. p=198™, p—po=32™. 1/p.—1/R= 000788. 6=0, sin i=A/D. Fraunhofer ae p Po cos t Diff. (p/cos i)? Lines a em. em. 0° 166°0 166:0 0 27500 B 99° 59! 5 165°3 165°3 0) 27500 C 91° 54! 10 163°2 163°5 c=) 93} 27400 D 19° 34’ 15 159°6 160°3 — "7 27300 EH 17° 26° ! 20 154°7 156°0 —1°3 97100 Ji 16° 02’ 25 148°5 150°4 —1°9 26800 G 14° 10’ 30 140:9 143'7 —2°8 26500 35 132°2 136°0 --3°8 26000 40 122°3 12771 —4'8 25500 94 CO. Barus and M. Barus—Plane Grating Similar to The greater part of the visible spectrum is thus contained between 7 = 15° and ¢ = 20°. It follows that the excess of Po CoS 7 — p lies between 7 and 13"". Hence the eye-piece may be placed at a mean position corresponding to 10™™" and give very good definition of the whole spectrum without refocusing, as I found by actual trial. Within 1™ the focus is sharp enough for most practical purposes. If the distances po and p,’ are selected so that eye-piece and slit just clear each other the definition is quite sharp. The diffraction equation is not modified and if 2a corre- sponds to the positions + ¢ and — ¢ for the same spectrum line, De (D/f) Phos It is, therefore, not necessary to touch the eye-piece, and this is contributory to accuracy. Fic. 6. If Rowland’s equation is differentiated relatively to p and p’ —dp = (—* / dp, , where the dp,’/p,’ factor is constant. Po COS 7% Hence — dp varies as (p/cos 2)’, given in the table. If further- more a comparison is made between dp, and dp this equation reduces to a/dp,/dp = |(R—p,)(i—cos 7) |/F# cos z which becomes unity either for 7=0 or for p, = (Rowland’s case). 12. Summary.—By using two slides symmetrically normal to each other and observing on both sides of the point of ————— Rowlands Method for the Concave Grating. 95 interference, it is shown that many of the errors are elimi- nated by the symmetrical adjustments in question. The slide carrying the grating may be provided with a focusing lens in front or again behind it, if the means are at hand for actuat- ing the slit which is not shar ply in focus on the plane of the eye-piece carried by a second slide throughout the spectrum at a given time. It is thus best to use both lenses conjointly, the latter as a collimator and the former as an objective of the telescope in connection with the eye-piece. It is shown that a centimeter scale parallel to the eye-piece slide with a vernier reading to millimeters is sufficient to measure the wave lengths of light to few Angstrom units, while the wave lengths are throughout strictly proportional to the displacements along the scale. The errorsof the three available methods and their counterparts are discussed in detail. The method is applica- ble both to the transparent and the reflecting grating. It is furthermore shown that in case of Rowland’s concave grating observation may be made symmetrically on both sides of the slit, or for reasonable clearance of slit and eye-piece passing across each other, although one conjugate focal dis- tance is now not quite the projection of the other. Brown University, Providence, R. I. % HH. Z. Kip—Determination of the Hardness of Minerals. Arr. XI.— Determination of the Hardness of Minerals, IT ; by H. Z. Kier. In the issue of this Journal for July, 1907, an article by the writer on the subject of mineral hardness appeared, whose threefold object was outlined as follows: 1. To invite general acceptance of a single definition of hardness. 2. To establish theoretically in conformity with the definition the best method of investigation. 38. To put this method in practice by means of suitable apparatus and adequate mathematical calculation. Inasmuch as it is my present purpose to act as my own critic as well as to publish the results obtained in carrying out the investigations indicated above under 3, it will be found excuse- able, perhaps, if I depart from the general practice of contribu- tors to the extent of speaking in the first person instead of the third.* In regard to the formula for determining hardness, estab- lished in my previous paper, H=~/a’+y’, I may say that no mineralogist or physicist who has favored me with his opinion has taken exception to this equation. Indeed so long as the generally accepted definition of hardness prevails (resistance to abrasion) this is, and ean be, the only adequate formula. If in what follows I appear to view my own results with some scepticism, I wish it to be understood that this is not the result of a lack of faith in the method employed, but merely an acknowledgment of the difficulty of dealing accurately with molecular forces by mechanical means, such means, at least, as I have had at my disposal. The apparatus employed was described in its general prin- ciples in my previous paper. As actually constructed it dif- fered from the description given in two points only. —_—__. Arr. XIX.—The Transmission of Light through Transparent Inactive Crystal Plates, with Special Reference to Observa- tions in Convergent Polarized Light; by Frep. EvcEnr Wricu'. Introduction. Tue problem of the refraction and reflection of light on inactive, transparent crystal plates has long attracted the atten- tion of physicists and crystallographers, and has proved a fruitful field of investigation from the standpoints both of theory and of applied physical optics. The general problem was first successfully attacked in 1835 by F. Neumann’ in Ger- many and by J. MacCullagh’ in Ireland, Neumann using strictly analytic methods; MacCullagh, on the other hand, inclining rather to geometric methods and attaining thereby greater simplicity in his treatment of the whole. Both Neu- mann and MacOullagh showed keen mathematical insight and judgment in overcoming the inherent difficulties of this prob- lem; their work, moreover, was remarkably thorough and comprehensive, and has served as the foundation on which all subsequent investigations have been based. Their general conclusions have remained intact and valid to the present day, even though their methods of calculation have been superseded by simpler and more effective methods and their fundamental assumptions have been modified to some extent and expressed in terms more nearly in accord with modern views on the nature of light. , 1 Theoretische Untersuchungen der Gesetze, nach welchen das Licht an der Grenze zweier vollkommen durchsichtiger Medien reflektiert und gebro- chen wird, Berliner Akad. Abh. 1835, Math. Abt. p. 1-160; also Poggen- dorf’s Annalen. xlii, 1-37, 1837. ? Phil. Mag. (3), viii, 103, 18385; x, 42,1837; On the Laws of Crystalline Reflexion and Refraction, Trans. Roy. Irish Acad. xviii, p. 31, 1887; Col- lected Works, 1880. Am. JouR. Sci.—FourtH SrrRizs, Vout. XX XI, No. 183.—Marcz, 1911. 12 158) FE. Wright—Transmission of Light through Besides the papers by Neumann and MacOullagh, the most important contributions to this subject have been made by D. Brewster’ (1819), A. Seebeck’ (1831), A. Cauchy® (1836), C. G. Stokes* (1852), J. Grailich® (1855), A. Cornu’ (1867), G. Kireh- hoff’ (1876), H. A. Lorentz* (1877), F. Kohlrausch’ (1878), R. T. Glazebrook” (1882), Th. Liebisch” (1885), J. Danker" (1885), J. Conroy” (1886), C. Spurge™ (1886), J. Norrenberg”’ (1888), Lord Rayleigh” (1888), P. Drude’ (1889), C. Pulfrich* (1890), A. Potier’® (1891), W. Voigt” (1896), C. Viola” (1899), A. Ost- hoff” (1905), P. Kaemmerer* (1905), F. Pockels™ (1906). In all these investigations, interest has centered in the reflected rather than in the refracted waves. The phenoniena, however, resulting from the transmission of light through crystal plates are of great importance in practical microscopic diagnosis and merit detailed consideration, from the standpoint both of general theory and of observation. The present inves- tigation was undertaken primarily to determine the influence 1Phil, Trans., 1819, p. 145. BS Pogg. Ann., xxi, 290, 1831; xxii, 126, 1831; xxxviii, 276, 1836; x1, 462, 1837. 3Compt. Rend., ii, 364, 1836. 4On the composition and resolution of streams of polarized light, etc., Cambridge Trans. ix, 399; Phil. Mag. (4), ii, 316, 1852. ®'Wien Sitzungsber. (II), xi, 817, 1853; xii, 280, 1854; xv, 311, 1855; xix, 226, 1856; Denkschr. Math. Nat. K1 ix, 57, 1805; xi, 41, 1856; Pogg, Ann. xcviii, 205, 1856. ® Recherches sur la reflexion cristalline. Thése fac. Science. Paris, 1867; Ann, Chim. Phys. (4), xi, 1867. 1 Uber die reflexion u. Brechung des Lichtes an der Grenze krystalliner Mittel., Abh. Berliner Akad., 1876. 8Uber die Theorie der Reflexion u. Refraction d. Lichtes, Schlémilch’s Zeitschr. xxii, 1, 1877. 9Wied. Ann., iv, 1, 1878. 10On the Refraction of plane polarized Light at the Surface of a uniaxial Crystal. Phil. Trans. clxxiii, 595, 1882. 1Uber Totalreflexion an doppeltbrechenden Krystalien. Neues Jahrb. i, 245, 1885 ; ii, 47, 1886 ; Lehrb. d. Physik. Kryst., 1891. 12 Neues Jahrbuch, Beil. Bd. iv. 241, 1885. 13 Proc. Roy. Soc., xl, 173, 1886. 4 Proc. Roy. Soc., xli, 468, 1886; xlii, 242, 1887. 15 Uber Totalreflexion an doppeltbrechenden Krystallen. Wied. Ann., xxxiv, 843, 1888. : 16 Phil, Mag. (5), xxvi, 241, 1888. “Wied. Ann., xxxvi, 532, 865, 1889; xxxviii, 265, 1889. Physik. d. Aethers, 1894. Lehrbuch d. Optik (2d edition), 1906. 18 Das Totalreflectometer, 1890. 19 Sur la principe du Retour des rayons et de la Reflexion cristalline, Journ. de Phys. (2), x, 349, 1891. *0Compend. d. theoret. Phys. ii, 622, 1896. 1 Zeitschr. Kryst., xxxi, 40, 1899; xxxvi, 245, 1902. 22 Uber die Reflexion u. Brechung des Lichtes an Zwillingsebenen volkom- men durchsichtiger inaktiver, einachsiger Krystalle. Neues Jahrb., Beil. Bd. xx, 1, 1905. 23 Uber die Reflexion und Brechung des Lichtes an inactiver, durchsichti- gen Krystallplatten, Erster Teil. Neues Jahrb., Beil. Bd. xx, 159, 1905. 24 Lehrbuch d. Kristalloptik, 1906. Transparent Inactive Crystal Plates. 159 of certain factors which underlie the methods for the measure- ment of the optic axial angles, especially the method of Pro- fessor Becke’ and the writer’s modification*® of the same.° These methods are based on the degree of curvature of the dark hyperbolas or zero isogyres of the interference figure and depend, therefore, on the polarization directions of waves transmitted along ‘different paths. In microscopic work, the influence of the boundary surfaces, not only of the crystal plate, but also of the intervening glass plates, on these waves enters the problem and tends to render it more complicated. In the following pages the general mathematical treatment of the problem of ‘ight transmission through transparent inactive erystal plates is given in. Part 1 and ‘several important and apparently new yelations are deduced which simplify the presentation materially. In Part 2 results of calculation are checked by series of observations with apparatus specially designed for the purpose. The results of the investigation show that the methods pro- posed by Professor Becke and by the writer are approximate methods only; both furnish results of about the same order of accuracy, the one advantage of the writer’s method being that of slightly greater simplicity. They show, furthermore, that a theoretically correct method is not attainable because of many factors, each of only slight intluence, which enter the problem and complicate the relations seriously. Part 1.’— Theoretical. The Boundary Conditions.’ Light waves, in passing through a erystal plate, encounter peculiar conditions, both on entering the plate and emerging from it. At the limiting surfaces of the plate, the crystalline material ends abruptly and the system of forces which result from the erystal strueture are suddenly cut off from further action. On emerging from the plate the light waves pass from the influence of these forces to that of an entirely different 1 Tschermak’s Mitteil., xxiv, 35, 1905; xxviii, 290, 1909. This Journal (4), xxiv, 332-338, 1907; Tschermak’s Mitteil., xxvii, 293, 1908. *In the course of this investigation the writer has corresponded frequently with Professor Becke and is indebted to him for several suggestions and for his open consideration of the points in question. “In the preparation of this section the following books and papers have been consulted especially: Drude, Lehrbuch d. Optik; also Drude in Win- kelmann’s Handbuch d. Physik; Liebisch, Lehrbuch d. Kristalloptik ; Pockels’s Lehrbuch d. Kristalloptik ; and P. Kaemmerer, Uber die Reflexion u. Brechung des Lichtes an inaktiver, durchsichtigen Kristallplatten, Neues Jahrb., Beil. Bd. xx, 159, 1905. 5 The subject of boundary conditions is thoroughly treated by P. Drude in Winkelmann’s Handbuch der Physik, vi, 1169, 1906; also in Drude’s Physik d. Aethers, 511, 1894. 160 FL E. Wright—Transmission of Light through system, but this passage from the one set of conditions to the second, although very rapid, is a continuous process, since, physically speaking, there are no discontinuities in nature. On the one side of the surface the light waves are entirely within the influence of the crystal forces; on the other, within that of the second medium; at the boundary surface, the transition from the one sphere of influence to the second is accomplished. There the two sets of forces meet and the result is a continuous passage of the one set to the second, so far as their influence on external forces is concerned. Whatever theory or hypothe- sis of light is adopted to explain the phenomena, this contin- uity must be taken into account. In the electromagnetic theory of light a “ boundary’ surface between two substances of dielectric constants e, and e, must be considered an inhomo- geneous surface in which the dielectric constant passes contin- uously though very rapidly from the value e, to e, in the direction of the normal to the surface.” The general equations of the electromagnetic theory are valid even in this film: 4r , Ow NA _ ou Ow 47, Ow dou Gn” Foy! Gee? oom be) del ice) a co Gy * 10a. soy! 2 ee Ok ios ate ~ meno i umnro ne In these equations of Maxwell, u, v, and w are the compo- nents after the a’, y’, 2’ axes of the magnetic force (the 2’ axis being normal to the surface and the ~’ axis in the plane of incidence); X, Y, Z the components of the electric force; Jx's Jy Juy the components of the electric density in electro- static units ; Ss, ., Sy, S,, the components of the magnetic current, and ¢ a constant, expressing the ratio between the electrostatic and electromagnetic units. The components of the electric and magnetic currents, 7., Jy, J, and s,., sy, 8, are finite quanti- ties. The right hand side of the equations with differential quotients must therefore also be finite, even when the thickness of the film approaches 0, and the components of the electric and magnetic forces parallel with the boundary surface must _ be continuous on passage through the boundary surface. This condition is realized mathematically by stating that on either side of an infinitely thin film the two forces are equal. (@),. = @),6°(%), =(@), (C2 =o, 1), = eee) These conditions are perfectly general and must always be fulfilled at boundary surfaces. 'P. Drude in Winkelmann’s Handbuch der Physik, vi, 1169-1170, 1906. Transparent Inactive Crystal Plates. 161 Boundary Conditions Applied to Transparent Inactive Crystal Plates. A erystal is distinguished electromagnetically from an iso- tropic body by the variation of its specific inductive capacity with the direction. If «€,, €,, «,, be the three principal dielec- tric constants of a crystal and mw the magnetic permeability, = |, as is practically the case in all known dielectrics, then the general differential equations, referred to any codrdinate sys- tem, for the electromagnetic field in a crystal, are: Dea a0) es 20S e en ot + €, ot ar 13 a) ay Bev | Ox era ot) _ 9 Jw (4) = —— Sih de = = _ —= ec 1 ot ap or ot 4 €o3 Ot dz! Ox! Y] Mv OX en Ov OZ Oo Ou Cc («: at E50 ot + €,, =] Ty Ghee Vo Lau _9¥_ 9% 180 _ 9% 9X 1dw_ dX OY aero cnon mon) O20. ue OF Oy’ 1 Oz! In these equations, €),—=€xn ('). If the magnetic force be taken as light vector, the compo- nents X, Y, Z, of the electric force can be eliminated from (4) and (5) by differentiating equations (5) with respect to ¢: Ohh 6 (ey 9 ear 1 . A ee 9 @); Goin Oe NOt) MeOu Not) 6 Ot Onl ob) m2) \ an) 2 Idi 2 0 /oX do /dY (6) e Ot a) cal ae IX IY OZ and substituting the values from equations (4) which Ot? dt? dt are linear functions of these quantities. If, for abbreviation, the right hand side of the equations (4) be made equal respect- : OX JY OZ : ively to &, 7, & then Dt? on? oe Can be expressed as linear functions of &, », €, thus: ‘A simple proof of this relation is given in Drude, Lehrbuch der Optik, 294, 1906. 1620 FL EB. Wright—Transmission of Light through Oxia dl on = Cc (4,,€+4,,9+4,,0) Vesa Ot = Cc (4,,€ + 49 + Aa6) (7) Amel Ot a Cc (45, + G9 +4,,) in which the so-called polarization constants @,,...,a,,..., @,,. + . 5 aresimple determinate functions of ¢,...,6,.. €,--., and ¢’ and for which the relations a, = dy, hold true just as €,, == €,. These equations indicate that a function of the second degree is possible whose partial differential quotients with respect to & 7», ¢, are equal, respectively, to aX oY 97 ot? dt? Ot” 2cG = 4,6 +4,,4' +4,,0 + 20,6 + 2a,,CE+2a,,€) = constant. (8) From energy considerations it is evident that this equation (8) must represent an ellipsoid; and if, in it, the constant be 2¢G=1, the equation is then that of a triaxial ellipsoid referred to a coérdinate system of any position. This ellipsoid is the “index ellipsoid” of MacCullagh, or “ellipsoid of elasticity ” of Kirchhoff, or the “indicatrix”’ of Fletcher. The codrdi- nate axes can be brought to coincide with the principal ellipsoidal axes by use of the usual transformation equations : This function is ap + Bp", +p’, = 0. 2,2 22 2,2 ae OP POG, Vegan =a 2,42 PLP) taza =: OT OTS Ci 9 =e 2 r 2 a : 42 a a le OOP +e 9,1; ap, Ws (9) wT Dp, ar br, DP, 3 CT, Dp, = 4,, WPT, + OG. + OPI, = % in which 7, Py Doo Yrs Yoo Yon And 7,, 7,, 7,, ave the direction cosines between the new coordinate axes #, y, 2, and the w’, v’, 2’ of the old system respectively. Referred to the principal ellipsoidal axes equation (9) becomes a*&’+6'n’+c*O* = 1, in which a, 6, c, are the principal light velocities of the erystal. The symmetry axes of this index ellipsoid are the reciprocals of a, 6, ¢, or directly the principal refractive indices of the crystal. In geometric problems of reflection and refraction, the index ellipsoid and the index surface derived from it are specially useful. Transparent Inactive Crystal Plates. 163 On substituting the values of Bae ae Ze of equation (7) in (6) we obtain a system of partial differential equations : ae = = (Qe Gant 2.56) — 5 (14,8 + 529 + Ua06) oe 3s = (Gygé + Gag + 4.6) — S (4,,€+@,.9+4,,6) (10) ce = iu (a,,€+4,.n+4,€) — ee (4,,E + U9 + 4,06) which are free from the components of the electric force and of the electric current. 5: Equations (4) and (5) are of general validity and obtain therefore, even at the boundary surface of a crystal plate." alt is apparent from the last equation of (5) that, as (X),= (X),, (Y),=(Y), at the boundary, (3 i) = in or (w),=(w), for periodic vibrations. The boundary conditions for a crystal plate may therefore be written : (0, = Hy =v (= oe (FE) = (Ge), (ar), (5), of which only four are independent. The last equation of the set may accordingly be discarded. The fourth equation a) = (Fe) can be expanded by means of (7) and (4) and becomes for the general case of two adjoining crystal plates : Pee) io aa, (2 Se tu 50) — 5p +4,,(55— 5a T ys aay) = , fow — ov’ Bay (Oi an os fou why a le a oa) + a aCe Ox! ) +a da a (12) These boundary equations, together with equation (8), are of general validity for transparent, inactive plates, and form the basis on which all detailed work rests. The partial differential equations (10) representing the move- ment of the magnetic vector are satisfied by the components u, v, w of the vibration of a plane polarized, advancing wave of constant amplitude. This vibration is defined by the usual equations of the general form : 164 FL EL Wright—Transmission of Light through U / / “= Al cos aC Siw) T q if / ' v = Am cos all (« = ee (13) T q t J / w = An cos p(e—-S ts) T q in which A is the amplitude of the vibration or magnetic force; 7, m, n, the direction cosines of the line of vibration 7, which in this case is that of the magnetic light vector and also the polarization direction; T, the period of vibration; A, #, v the direction cosines of the normal N of the wave propagated with the velocity g. To simplify these equations, let the boundary surface of the plate be the a’, y’ plane, fig. 1; the AE plane of incidence, the «’ 2’ plane; the angle between the normal of the advancing wave N, and the 2’ axis, 7, the positive direction of z’ being on the crystal side of the boundary surface; let also the polarization azimuth yw be the angle between the plane of incidence (the «’z’ plane) and the plane of polarization (the angle Kz, fig. 1) counting from the 2’ axis in the direction of the +7’ axis and passing beyond this axis if necessary. In this case, = cos 7 | oS _ == sin i t= —cosrcosy, m=siny, n= sin? cos y. and the equations (13) reduce to Transparent Inactive Crystal Plates. 165 a x’ sin 7 +2! cos } u= — A cosy cosr cos — —- i qY , 2 a'sin 7 + 2’ cos7r Qs A sin y cos a(t — (14) i q : Qn a’ sin r + 2’ cosr 2o= Acosysinr cos —(t — T q For 2’ = 0 it is evident from the boundary conditions, (wv), =(w),, (v), = (v), of equations (11) and (14), that for all pos- sible reflected or refracted waves at the limiting surface, T, the period of vibration (color) remains constant (T, = T,); also oF By BS = fs , which is the sine law of wave normals; while # =O signifies that all wave normals lie in the plane of incidence. By means of equations (14), the general differential equations (10) can be solved and the fundamental formulas obtained for the refraction, reflection, and polarization of light waves in erystals. Thus from equations (4) and (14) ow ow é Syme: sin yw cos 7 Ou Ow ee acme ping. ha COSY dv ou , : = gi Dy! = K sin wy sin 7. W herein 9 i D) i I SI “4 I a eS es sin sp (¢— CaS eas o*) T.g T qd From these expressions, we find: p ; ) Oz! (2,, 5 a Ayo) a Aes é) = | a cos 7 (—a@,, Sin ¥ Cos 7 —a,, cos y +4a,, sin y sin 7’) ) Ox’ (a,,€ oF a, ate Ass é) = Cae : ; A ee sin 7 (— @,, sin y cos 7 — @,, Cos + Q,, Sin W sin 7) } Oz! (4, g — a). SUE a,, é) = C ; : ; ae cos r (— a@,, sin y cosr — @,, COS Y + @,, Sin y sin 7) 166 FF. EB. Wright—Transmission of Light through 4) ° Ox" (4, 3 ai; 4 tr Ms é) == —; sin 7 (— @,, sin y cos 7 — a,, cos y + @,, sin y sin 7) Wherein Cay all ie Qa (+ a! sin 7 + 2’ cos 7 T de q similarly from (14) Ou >A atari Cosh tan Ns 3 ov ‘ ; oF = C sin y ow é cava ae C cos wy sin 7 Substituting the values from these last two sets of equations in (10), we obtain the two equations 1 4 “ 4 —cos Ye Phe sin y cos 7—a@,, cos W+a,, Sin ysin”) and ; sin r 4 : F | sin y= Fe (—a,, sin Y cos r—a,, cos Y+@,, Sin y sin 7) — | cos 7 6 ; (—a,, sin Wy cosr—a@,, cosy+a,, sin y sin 7) 2 which, on rearrangement, become (a) cos y (¢’—a,,)=sin wp (a,, cos r—a,, sin 7) (15) (0) cosy (a,, cos r—a,, sin 7) = sin y (¢°—a,, sin’r7—a,, cos’r+ 2a,, sin 7 cos 7) By division of 15 (a) by 15 (6) an expression results which is free from y; (7’—4a,,)(¢°—4,, cos’*”r—a,, sin’*r +2a,, sin r cos 7)= (16) (@,, COS 7—a,, Sin 7)” and which reduces to’ [4, 1 a 20,19 ui a (4s, —k'/ig’r | [Go ar (Go —k*)tg’r| — (16a) ; (@,,— A, .9 ry(1 +t9*r) if & be substituted for the constant value = which, by reason of the sine law of refraction, is equal, for all possible waves, to es where g, is the velocity of light in the isotropic medium enveloping the crystal and 7 the angle of incidence. By means of this standard formula, which can be derived in different . 1G. Kirchhoff, Uber die Reflexion und Brechung an der’ Grenze Krist. Medien, Berliner Akad. Abh., 1876.—Th. Liebisch, Neues Jahrbuch, II, 191, 1885. Transparent Inactive Crystal Plates. 167 ways, the angle of refraction or reflection of any light waves in the erystal can be calculated. From 15 (a), the following equation is readily derived : t/a gw= q 22 @,, COS T—A,, SIN 7 which may be written (°—a,,)tg’r—4a,, (@y.— Ayal 1) 4/1 tg" age r (17) wherein & = 2 as in (16). From (17) the azimuth of the plane of polarization can be determined, provided 7 be known. Equation (16) is biquadratic and indicates that in a er ystal there are four possible waves, of equal significance,—-two reflected and two refracted waves,—which must be taken into account in the general boundary conditions for the erystal. The general equations (11), (12) and (14) for the magnetic light vector on passage through the boundary between two inactive, transparent crystal plates may therefore be written :' gv 4 4 > A, cos 7, cos y, = > A’, cos 7’, cos wy’, (from (w),=(u),) 4 > A, sin y, = > A’, sin W’, (from (¢),=(),) (18) 4 = A, sin 7, cos y, = > A’, sin 7’, cos y’, (from (w),=(w),) eH Me »# Men 4 A, sin 7, > (sin ¥. (@,, COS 7,—G,, SIN 7) +a,, COS Y,) = ATS Sim en . —*—__*(sin y',(#’,, cos7,—@’,, sin 7",) +.a',, cos ’x) (tm (=) =(3),) In the last equation of this set both sides of the equation sin TAS sin 7’, have been multiplied by the equality —— ee k k In the first three equations, the factors of the amplitudes, A,..A,, are the direction cosines, J, m, nm, of the line of vibration 7 with the axes 2’, y’, 2’; if the factors of the amplitudes in the fourth equation be indicated by p, the equations can be written in the abbreviated form,” 1G. Kirchhoff, Ges. Abhandlungen,. 367-370, 1882. 2A. Potier, Journ, Phys. (2), x, 300, 1891. P. Kaemmerer, N. J., Beil. Bd. xx, 174, 190 1 4 > 1 168) FE. Wright—Transmission of Light through S4.,= 5 4107 1 1 4 4 Ayn, = ¥ Alm! > erga > ok (18a) 4 4 SA, = > Ain, at 1 4 4 LAAs A’ p's 1 1 In case the crystal plate is surrounded by an isotropic medium, these general equations become simpler; the index ellipsoid for the isotropic medium is a sphere and its coeffi- cients are @’,,—= @',,= @,,.—=q, anda’,,=a,,=a',,=0. For the passage of light from the isotropic medium to the crystal plate, there are, in general, one incident wave (I), one reflected wave (Rt) and two refracted waves, W,, W, (fig. 2); the boundary equations are, then, Fig. 2. I R Wu; ~W, D, cos 0, cos r, + D, cos 0, cos 7, = (E cos e—R cos p) cos 7 D, sin 0, +D, sin 0, = E sine+Rsinp D, cos 0, sin7,+D, cos 0, sin 7, = (E cos «+R cos p) sin 7 sin 7,,. : iD), 7 *[sin 0, (a,, cos 7,—a@,, sin7,)+4,, cos 0,]+ 1 D SU 5 ; 3 ; . 7p [sin 0,(a,, cos 7,—a@,, sin 7,) +a,, cos 0,|= (19) 2 (E sin e—R sin p) sin @ cos 7 wherein, for the incident wave (1), the reflected wave (R), the faster refracted wave W, and the slower refracted wave W.,, respectively, E, R, D,, D,, are the amplitudes; «¢, p, 6,, 6,, the polarization azimuths; g,, g,, @,, d,, the normal velocities of the wave; 7, 7—17, 7,, 7,, the angles of the wave normals with Transparent Inactive Crystal Plates. 169 z'. In these equations 2, E, ¢, g, of the incident wave are known ; also by calculation (equations (16) and (17)), 7,, 7, and 05:05 of the refracted waves; and 7, g, of the reflected wave; unknowns are R, p of the reflected wave and DD, ot the refracted waves W, Woe In abbreviated form, corresponding to (18a), these equations may be written :' D/, +DJ, = (E cos e—R cos p) cos 7 Dm, +D,m, = E sin e+R cos p Dy», +D,n, = (E cos e+ Ros p) sin z Dp, +D,p, = (E sin e—R sin p) siné cost (19) At the second boundary surface where the two refracted waves emerge from the crystal plate into the isotropic medium, two sets of boundary equations obtain, one for each refracted wave, W,and W,. At this surface, there are for each incident wave, W, and W.,, two reflected waves and one refracted wave as indicated in fig. 2. For the refracted wave W, the boundry conditions reduce to D, cos 6, cos 7,+R’, cos p’, cos 7’, +R”, cos p", cos 7”, = IDINCos) 0) cos) 7 ‘ PW Ie 1 View ee Ti whe. D, sin 6, +R’, sin p’, +R’, sin p”, ee : D’ sin 0’. Heaps aes ! if yak ,! Tse WS a 7 ae D, cos 8, sin7, +R’, cos p’, sin 7’, +R’, cos p sine = D’, cos 8’, sin ¢. (20) SUD PAT os : D, 7 | sin 8, (4,, cos 7,-a,, Sin 7,) +4,, Cos 3, | + 1 R’, sin 7! 1 =o ’ ‘ / = ! 2 I ese | sin p, (@,, cos? —a@,, sin r,)+-a,, Cosi p,) + 1 R”, sin 7” 1 1 is etd : a7 Picante te apn [sin p’, (a,, cos 7”,—-a,, sin 7”,) +a,, cos r’,) = 1 D’, sin 0’, cos @ sin 7. wherein, for the incident wave W, the faster reflected wave W.,, the slower reflected wave W,,, and the refracted wave WwW"; respectively, D,, R’, R”, D’, aie the amplitudes; 6,, p’,, p”, 0; the polarization azimuths ; PT Pte a the inclination of the wave normals with 2’; and I Gs Fy Yo the wave normal velocities. Similarly, for the slower refracted wave W, the boundary equations are D, cos 6, cos 7, +R’, cos p’, cos 7’, + R”, cos p”, cos 7” De cos ‘yy cos 7 D, sin 6, +I’, sin p’, +R", sin p’, Ett eu) D', sin 6’, 1P,. Kaemmerer, Neues Jahrb., Beil. Bd. xx, 176, 1905. 170 EL EL Wright—Transmission of Light through D, cos 6, sin 7, +R’, cos p’, sin 2, +R”, cos p, sin 7”) = (21) cos p, sin 7”, = IDE veos! a sin 2 Sankey ple ; . De (sin 8, (4,, COS 7,—a,, Sin 7,)+4a,, cos oy) + Vs ; R 5 sin ae J, Sey 1 as 1 f , —2___—+*| sin p, (@,, cos 7’,—@,, Sin 7”) + @,, COBip’, | 4. Yu p sibel : ” ay wv beet) R’, oR sin p”,(a,, cos 7”,—a,,sin 7”,)+ a,, Cos p”,. | = 2 D’, sin 0’, cos @ sin @. - / These equations (19), (20), (21), agree with the fundamental equations of Neuman, MacCullagh and Kirchhoff derived from the mechanical theory of light. They are, however, exceed- ingly complicated, and in their solution certain auxiliary geo- metric and analytic relations are used which simplify and facilitate the practical calculations considerably. The most important of these aids are the index surface introduced by MacCullagh ('), Potier’s (7) generalization of the Neumann- MacCullagh relation, and the conception of the uniradial azi- muth as given by MacOullagh(*) and Neumann.(‘) The Newman-Mae Cullagh-Potier relation. The index surface, whose radii vectors are proportional to the reciprocal wave normal velocities or directly to the refrac- tive indices for the direction of propagation, is best adapted to Fig. 3. present graphically the relations between refracted and re- flected waves. Itis derived from the index ellipsoid in the same manner that the ray surface is derived from Fresnel’s ellip- soid. The index surface (1) of a crystal is the reciprocal of its ray surface (2), just as the index ellipsoid (I) is the reciprocal 1J. MacCullagh, Trans. Roy. Irish Acad., xvii, 252, 1833. * Jour. Phys. (2), x, 349, 1891. 3 Trans. Roy. Irish Acad., xviii, 31, 1837. 4 Berliner Akad. Abh., Math. Abt. 144, 1835; Pogg. Ann., xlii, 9, 1837. Transparent Inactive Crystal Plates. 171 of Fresnel’s ellipsoid (Ff). Each point S of the ray surface (2) defines a ray direction; the normal OQ to the tangent plane SQ through § of the ray surface is then the radius vector of the normal tu the wave producing the ray, 8. (Fig. 3.) The extension of this wave normal vector to its reciprocal length ON determines a point N on the index surface. The two points N and § are said to be corresponding points, and the plane NOS is normal to the polarization direction. Similarly, a point p of the index ellipsoid (I) is the corre- sponding point P of Fresnel’s ellipsoid (F), (fig. 4), if its radius vector coincides in direction with the normal op’ to the tan- gent plane Pp’ through P and is equal in length to the recip- rocal of the normal Op’=g. The radius vector, OP repre- sents a ray of velocity s (fig. 4) while the radius vector Op = Lis g the reciprocal of the corresponding wave normal velocity g. The normal to the plane Pop is then the polar- rey, ization direction. By obtaining the codr- a tk dinates of such corresponding points on the two ellipsoids (F) and (1), Potier discovered a simple relation between the expressions l,m, n, p, of equations (18a) which has proved of great value in the solution of B problems of reflection and refraction. For the sake of simplicity, let the equa- tion of Fresnel’s ellipsoid and the index ellipsoid be referred to the principal axes; Fresnel’s ellipsoid is then represented by Fa eto; the index ellipsoid by (22) MWe w+ yy +e 2=1 (23) If the codrdinates of a point P on Fresnel’s ellipsoid be 2°, y°,, 2°, then the equation of the tangent planes through P is: 5 (ok ile or WOE ey) oe (x x 1 Gas ats y yY Aa)ye, + (2 & (Fn )e = The equations of the normal to this plane are az Yy @ 2) = e = eo (24) By definition, the point p is common both to the normal and the index surface and its codrdinates, 2,, y,, 2,, are readily found by means of (22), (23), and (24), to be: 1720 EF. EB. Wright—Transmission of Light through 2 =O, © Oa ]xe, oy]y®, ; d2 2°, Similarly it can be shown that Fee) AS) Bll, OL ay ful (25a) (ae) aie ees This reciprocal relation of polar quadries obtains for all posi- tions of the codrdinate system, as can be readily proved by adopting general coérdinates in the above equations.’ oe 4 i : In case two pairs of corresponding points be taken, p, P, and p, P,, then, in general codrdinates, 2 gi! aie, p(s) f= (5 PO Out! a, ? D3 dy! y', FS aioe Ba ,/0 = f=(— = (2 27 Oa’ com > Y= ahh ? ona ai The general form of 21=a,,v"+a,,y" + ,,2” +2a,,y'2' + 2a,,2'x' +2a,,x'y'=1 is that of a homogeneous equation of the second degree, from which it follows that ! ol , ol Tr ol Tg oelt . t ! ib pot x, Ox’ ae Oy’ tia PE gt gy Y gH yg? 24 + A +Yy'2',) ar a, (2,2", +2 @ ,) ar A, (2',Y'o ae wy’) ol aol TREN Sos =0(57) pee ay),* ee x) i tat PET ae eRe SAW cee Bell's 8K) fF aati r lO Ce YY -+e 2 ee FY DY , +22 He (26) Fig, 5. Y! Pp S ao 3 ! ».¢ K 1MacCullagh, Geometrical Propositions applied to the wave theory of light, Trans. Roy. Irish Acad., xvi, pt. 2, 67; also xvii (1833); Coll. Works, p. 20-22, 1880. Transparent Inactive Crystal Plates. 173 To apply this relation, discovered by Potier, between two pairs of corresponding points on the two ellipsoids (F) and (1), it is necessary to obtain the codrdinates of the points. In the stereographie projection (fig. 5), let y and P be the projections of the two corresponding points p and P, Ky’ the wave front, x'z’ the plane of incidence, N the wave normal, 7 the angle of inclination of N with 2’, the azimuth of the plane of polari- zation, and s the angle pOP (O being the center of the sphere). The coordinates of p are then: '—O Iyosd 1 G 7) a Bee sin w, cos 7, 1 1 DOCS BY Coe (27) 1 ' Ets ol ee : z,=Op cos pz a sin y, sin 7, 1 and the codrdinates of P are oe ees c Sal a feo x'° =OP cos Pa' =OP (sin s, sin 7,—Ccos s, cos 7, sin y,) y,=OP cos Py’=—OP cos s, cos y, (28) z' =OP cos Pz’ =OP(sin s, cos 7,+¢os s, sin 7, sin y,) Op" qs But from fig. 4, OP= and equations (28) ean cos POp' cos s, be written: x’ =9,(t7 s, sin 7, — cos 7, sin y,) Y 49. COSY (28a) 2 =q,(tg7s, cos 7,+sin 7, sin y,) In fig. 5, P is situated between N and p; in case P lies beyond p, tg s, in (28a) changes sign and becomes negative. If the sign + be placed before tg s,, therefore, all possible relations are taken into account; for each particular case, the proper sign must be determined. On substituting these codrdinate values (27) (28a) in the Potier relation (26) for two sets of corresponding points p,, P., P., P, of the waves W,,W, whose normals lie in the plane of incidence az’, we obtain, q : : : : ~ | sin y, cos 7, sin y, cos 7,—sin y, cos 7, tg s,sin7, + 1 cos , cos Y, +sin y, sin”, tg s, cos 7, +Sin y, sin”, sin W, sin 7, ] q, q, [ sin y, cos7, Sin y, cosr,—sin y, cos 7, tg s, sin 7, + cos Y, cos Y,+sin y, sin”, tgs, cos7r,+sin y, sin 7, sin y, sin 7, | This equation may be rearranged to read (7, —¢’,)[sin YW, sin YW, COS (7,—7,) + cos W, Cos v|— (29) (g 1 tg $, Sin Word 2 tg s, Sin ,) sin (7,—7,)=0 Am. Jour. pets uEre SERIES, VoL. XX XI, No. 183.—Marcg, 1911. 1 174 FE. Wright—Transmission of Light through This general relation of MacCullagh-Neumann-Potier exists between any two of the four possible waves, W,, W,, W,, W,, within a erystal plate for each of which the sine law g, sin 7, q, Sin 7, qg, Sin 7, SS Tc ae hg Sentara rh. POT) Sete Mens Ses Nei) sin é is valid. These values introduced into (29) give after division by sin (7,—7,) the equation sin (7,+7,)[sin y, sin y, cos (7,—7,)+cos y, cos y,]— (29a) sin’ 7, tg s, sin y,—sin’ 7, tg s, sin y,=0 Six different equations of this general form are possible, as six combinations of two can be obtained from the four different waves. Equation (29a) can be simplified by substituting for ty s, an expression containing g,, 7,, ¥,, and three constants, @,,, d,.) Us.) of the index ellipsoid. In fig. 4, the coordinates of P are rope BIN oI ah al (e (equation (25@)) and the length d1\? BNE Ol\2 Oz Accordingly Op' q PO '— a —— u cos J) =COS S, OP Vy nie a O1L\? (30) V (5.),+ (5;),+ (a) The direction cosines of the radius vector OP are propor- tional to the codrdinates of P and therefore from (29) and (29a) ol Cos baaae ee (=). _tg s, sin r,—cos 7, sin y, COS Py) Tn ole ae —cos tp, (33), from which (5 ) =(3,) —cas yy, Oy), \dx/, tg s, sin r,—cos 7, sin y, Similarly (81) OLN | tg s, cos7r,+sin 7, sin y, & a \e ,tg Ss, sin 7,—cos7 sin y, On substituting these values from (31) in (30) we find Lee. q% Se I Le /1+tg’s, (5) Vitis’ 8, “/, tg 8, sin 7,—cos 7, sin y, COs S$, = Transparent Inactive Crystal Plates. 175 or tg s, sin 7, = ~(5) +cos 7, sin y, AM), ol / / But (Fe) Oa Cb Yb A which from (27) il : : ; = @ ( —d,, cosr,siny,—d,, cos ¥, + a@,, sm 7, sin ) 1 7 1 Accordingly 2 S te oe 2 Fes (97, —4@,,)cos ”, sin ¥,—a,, cosy, +a,, sin 7, sin y, (*) (32) 1 q°, sin 7, On eliminating tg s from (29a) by means of (32) we find sin (r,+7,)[sin y, sin y, cos (7,—7,) + cos y, cos y,]+ sin 7, sin Wy, q sin 7, sin y, q, which on rearrangement becomes [a,,—7',) cosr, sin y,—4,, sin 7, sin W,+a,, cos y,]+ (a,,— q,)° cos 7, sin W,—4,, Sin 7, Sin Y,+a,, cos y,]=0 1This expression was first derived by P. Kaemmerer (Neues Jahrb., Beil. Bd. xx, 206, 1905), though by a method different from the above. 2From the equations (80) and (81) the following relations can also be derived: —cos P= It (5) = 1 —@i2 COS 7; SIN Y;—Aeo COS J, + G13 SiN 7 Sin a) qi \dy/» 2 —a or tg y= Osim 2aa Be (82a) G12 COS 71 —Ao3 SIN 1} an equation identical with (17). — % mee ae Vixntg s, 0z ; tg si cos7; +sin 7; sin yp, Also —@13 COS 7; SiN W;— 3 COS 1 +(As3—Q"1) SiN 71 Sin W, (32b) @*1 Cos 1, an expression for fg s, which is apparently novel. On equating (82) and (320) we find a 12 COS 7; —Ao3 SiN 1; tg i= 5 a - D) @?1— M11 COS’; —G33 Sin? 7; + 2ai3 SiN 7; COS 11 ) From (82a) and (32c), we have (q?1 —@22)(q?1 — G11 COS? 71] — M35 Sin? 7; +2 a3 SiN 1; COS T;) =(die COS 7; —Ag3 Sin 71)? an expression free from wp and identical with (16). . or tg = 176 FL EB. Wright—Transmission of Light through cos 7, cosy, sin 7, cos w,+ cos 7, cos , sin 7, cos y, + sin 7, sin W, j [sin y,(a,, cos 7,—a,, sin 7,)+a@,, cos ,]+ (33) q, sin 7, sin ; f ea a [sin w, (@,, CoS 7,—@,, Sin 7,)+a,, cos w,] = 0 To i an equation, which, like 18¢, can be written in the abbreviated form Ln tln,+p,m,+p,m, = 0 (33a) This equation is the general Potier relation’ and is applicable to any two of the four possible waves within the crystal. Uniradial azimuths. At the boundary surface of a crystal plate with an envelop- ing isotropic medium, an incident plane polarized, monochro- matic light wave furnishes, in general, one reflected wave in the isotropic medium and two refracted waves, W,, W,, within the erystal. The directions and azimuths of the two refracted waves are definitely fixed by equations (16) and (17) and a rotation of the plane of polarization of the incident wave can produce a change in the amplitudes only of the two refracted waves. Fora certain value of the azimuth, the amplitude of either W, or W, becomes zero, and but one refracted wave is transmitted. Such azimuths, e,, of the plane of polarization of the incident wave, for which only one refracted wave results, are called uniradial azimuths, and were first inves- tigated by MacOullagh* and Neumann.’ For D, = 0 in equa- tion (19) we find (a) (E cos e,—R cos p) cosz ==) (6) Esin e, +R sin p == DN simso), (c) (E cos «, +R cos p) sin 7. = D) cos 0, sin 7, |. (34) (d) (E sin e,—R sin p) sin? cos? = D, sinv « 1 1 = a eee [sin 0,(@,, cos 7, —a,, sin 7,)+4a,, cos 0,] L The last equation of this set can be readily reduced by means of (32) to the form (d’) (EK sin e,—R sin p) sin 7 cos7 = D, sin*’r,(cot 7, sin 0,—¢g s,) By multiplying the first of these equations with sin ?, the third with cos 7 and adding; also the second with cos 7 sin 2, and adding to the fourth, we obtain 1 Potier, Jour. de Phys. (2), x, 352, 1891. 2Trans. Roy. Irish Acad., xviii, 31, 1837. Collected Works, p. 110, 1880. 2 Berliner Akad. Abh., Math. Abt. 144, 1835; Poge. Ann., xlii, 9, 1837. Transparent Inactive Crystal Plates. diving . (a) E cos «,(2 cos ¢ sin 7) = D, cos 0, sin (¢+7,) (6) Esin «(2 cos¢sinz)= D, (sin 0, sin 7 cos 7+sin’ 7 (cot 7, sin 0, —¢g s,) On division of (0) by (a) 4 sin 0,(sin 7 cos7+sin r, cos 7,)--sin’ 7, tg s, cos 0, sin(¢+7,) ge= sin’ 7, tg s, or tg «, = tg 6, cos (t—7,) - ane an Ol 1 1 (35a) a sin 7, tg S, cos 0, sin(#+7,) Similarly, tg «, = tg 0, cos(t—r,) — By means of these formulas the uniradial azimuths for the refracted waves W, and W, can be calculated. At the second boundary surface of the crystal plate, the refracted wave W, produces two reflected waves W,,, W,, and one emergent wave W’.. (fig. 2). To caleulate the azimuth of the plane of polari- zation of this emergent wave, the relations of -Potier are important. The general boundary conditions for this surface and wave W, are defined by equations (20), which, after the manner of (18a) can be written in the abbreviated form: DL © BP, = RB’. =D’ cos 0! cosé Dim,+ Rm’ + Rm’ = De sin o”, Dan, + Rin’, + Rn’, = D’, cos 0", sin? (20a) Dp, + Rp’, + R’ p’, = D’, sin 0’, sin 7 cos 7 On multiplying the first of these equations by n,, the second by p,, the third by Z,, the fourth by m,, and adding, we find D,(n,J,+m,p,+1,n,+m,p,) +R’ (nv +pm',+1,n',+m,p',) + R! (fn, +m, p",+4,m"_ +m,p",) = D’,(n, cos 4’, cos i+p, sin 0’, +0, cos 0’, sin ¢+m, sin 0’ sin 7 cos @) In this equation the coefiicients of D, R’, R’, are=0 by virtue of the Potier relation (83a), and as the amplitude D’ is not in general zero, the equation reduces to n, CoS 7+/, sin @ Pp, +m, Sin 7 Cosz On replacing 7,, Z,,9,, m, in this expression by their respective tg Os = values from (19), we obtain lg W =_— cos 0, sin 7, cos 7+cos 0, cos 7, sin 7 (36) Ringe 5 : ‘ vote ; P [sin 0,(@,, cos7,—a,, sin 7,) + @,, cos 0,] + sind, sin @ cos ¢ 2 178 EB. Wright—Transmission of Light through This Bapree eH can be simplified, as was (84), by the intro- duction of tg s, cos 0, sin(? +7,) sin* 7,(cot 7, sin 0,—¢g s,)+ sin 0, sin Z cos 7 tg O” =— cos 0, sin(¢+7,) sin 0, sin(¢+7,) cos (¢—7,)— sin’ ”, tg s, (36a) or tg 0" =— On comparing this relation with (350), it is evident that 1 Te g 1 tg €, oe = 6 aE 90° (37) also Of Nemo Oe This apparently new and important relation greatly simpli- fies the labor of calculating the azimuths of the planes of polarization of waves transmitted through a crystal plate. In form it is similar to the relation deduced by Potier,’ that in case a wave W, within a crystal plate emerges into an isotropic medium, the emergent wave is polarized at right angles to that wave which, entering the crystal plate in the opposite direc- tion, produces the so-called “Hilfswelle W’” of W’. The above relation states that the azimuth of the plane of polariza- tion of the emergent wave W’, from W,, is at right angles to the uniradial azimuth e, of the wave W,. To calculate the azimuths of the emergent waves W’,, W’,, it is only necessary, therefore, to calculate the uniradial azimuths e,, e, of the inci- dent waves which produce the refracted waves W, and W,,. Uniaxial Crystals. In the preceding pages the formulas for the transmission of light through crystal plates have been developed for the most general case, that of biaxial crystals. When applied to uni- axial plates, these formulas become somewhat simpler, and deserve brief consideration as they will be used in the observa- tional part of this paper. The equation of the index surface for uniaxial crystals referred to general codrdinates is [0° (ae? +4" +2’) —1][@,,2" an U e +, ze 14-20, WV 2 ate 2a, ge! +2a,,0'y' —1]=0. If, as usual, the plane of incidence be the «’ 2’ plane, (y’=0), and the zg’ axis the normal to the plate, the positive direction of z' being within the crystal plate, this equation can be written 1 Journ. de Phys. (2), x, 354, 1891. 2H. HE, Neumann, Berliner Akad. Abh., Math. Abt., 1835. Transparent Inactive Crystal Plates. 179 [o°(a’? +2") —1][a, ,v" +a,,2° + 2a,,2'x’—1]=0. (38) In this formula w’ and 2’ can, by virtue of the sine law, be readily expressed in polar coérdinates; if 7 be the angle of incidence and 7 the angle of refraction and m, the refractive index of the isotropic medium, then v =n, sin 2, and sin @ : Zo== aa . With these values (38) reduces to WL [o’n,” sin? ¢ (1 +ég’r) —tg’r][(@,,2. sin® ¢—1)tg’r + 2a,,2,' Sin” ttgr+a,,n, sin? i]=0, (38a) From the first half of this equation (88a) sin 7,=7, O Sin 2@. To evaluate the coefficients of tg 7 in the second half of the equation (38a), let the plane @’ y’ in fig. 6 be the boundary surface of the crystal plate; a’ 2’, the plane of incidence, the positive direction of 2’ being on the crystal side of the bound- ary surface; «, y, 2, the principal ellipsoidal axes of the erys- tal; @ the polar angle zz’, and », the azimuth of the principal plane 22’; let also the angles of inclination of the wave normals, Q., Q., with the 2’ axis be Q,2’=7, Q.2’=7.; with the 2 axis, be Q.2=¢,, Q.2=¢.. In this case the direction cosines of wv’, y', 2 with x, y, 2, are respectively P, = — Cos 8 cos w Pp, = Sino P, = sin 6 cos w 7, = — cos sin w 7, = — Cos w q7, = sin 0 sin w ri sin a) ii, == GOS (uy also, inequations|(9)i@: = er. nbs )6' ce = 07, 180 Ff. EL Wright—Transmission of Light through Substituting these values in equation (9), we obtain the usual equations a,, =e +{o°—e’) sin’ 6 cos’ w A,,=e +(0°—e’*) sin’ 6 sin® w a@,,=€ sin* 6 + 0° cos’ 6 (39) ,,=(0°—e*) cos @ sin @ sin w d,,=(0°—e*) cos 6 sin 6 cos w a,,=(0*—e’) sin’ @ cos w sin and the coefficients of tg 7 in the second half of (88) become @,,%) Sin* i—1=n,° sin’ 7 [e’+ (0’—e’) sin? 6 cos’ w]—1 2a,,m0° sin® 4=2n," sin’ 7 (o"—e") cos sin Ocosw (40) U,,%) Sin® t=n, sin* 7 (e’ sin” 6+ 0? cos? 6). < From equations (88a) and 40), 7, for the extraordinary wave can be calculated. If n,=1, as is practically the case when the crystal plate is surrounded by air, equation (38a) can be written in the following form, which is logarithmically con- venient : sin 7,=0 sin 7; (41) uf 1 Bcos wo + Y B’ cos’ o + c( ; sat A cos? w — ¢) sin? ¢ tg — 1 e — A cos?’ wo — —,. sin® 7 wherein A = (e’—0o’) sin’ 0 B = (e’—o’) sin @ cos 6 C = e’ sin’ 6 + 0’ cos’ 6. To find the azimuths 6,, 6, of the planes of polarization of the refracted waves W,, W,, fig. 6.is again useful. In the spherical triangle Q, 2 2’, the relation obtains sin 7, cot 6 — cos 7, cos w cot 0, = a (42a) while in the spherical triangle Q, 2 2’ iq bee cos 7, COS w — sin r, cot 6 (428) sin w The uniradial azimuths e, and e, are calculated from equa- tions (35a). For the ordinary wave the wave normal and ray direction coincide and the angle s,—0. Accordingly tg €,==tg 9, cos (t—1,). (43) In the analogous expression for #9 €,, sin’? 7, tg 8, cos 0, sin (¢+7,) ee) tq 8, occurs but can be expressed in terms of known angles.’ 1 MacCullagh, Coll. Works, 1880. Trans. Roy. Irish Acad., xvii, 1833. Pockels, F., Lehrbuch der Kristalloptik, 194-195, 1906. tg «, = tg 9, cos (t—r.) — Transparent Inactive Crystal Plates. 181 In uniaxial crystals the usual formula for tg s, is 2 2 is = a cos ¢, sin ¢, (a) where ¢, is the angle Q,.z of fig. 6. From the spherical tri- angle Q, 2 2’, fig. 6, cos 7, sin 0, — cot w cos 0, | (2) cot ¢, = sin 7, cos 0, also cot ¢. = — (cot 6 cos 7, + CoS w sin ”,). (0’) sin w Furthermore, in the principal section through Q, and z Je = 0 cos’ d + e’ sin’ ¢, from which expression, we find (c) e—o’ ., ; O° 4 sin’ 7”, sin (7,—7,) sin (7,+7,) - Sino, == fo = ws : Oe $e Yo sin’? 7, sin’ 7, By means of the three equations (a), (0), (c), equation (44) becomes sin (7,—7,) sin (7, +7.) (cot o—cos r, tg 0.) sin (¢+7.) sin 7, (44a) tg «, = tg 9, cos (i—r,) + or from equations (a), (0’), (¢) tg «. = tg 9, cos (i—r,.) + sin (7,—7,) sin (7,.+7,)(cot 8 cos r,+ Cos w sin 7.) sin (¢+7,) SIN w (44d) Having thus found e, and e¢,, the azimuths 6’, and 6’, are readily obtained by the use of equations (37). 0’, =€.+90° 0’, =e+90- (45) e Isotropic Plates. In the case of isotropic plates, the index ellipsoid reduces to a sphere and the constants of the equation become a,, = ,, = a. — Y Ay5 = A, aa a, = 0 Equations (16), (85a), (86a), then reduce (if the surrounding medium be air g,=1) sin @ VW (0) tg «, = tg 0, cos (t — 1) (a) Sin 7) 19 7sin? = (0) tg 0, = 00h: cos (t—7) (46) 182. Fk. EL Wright—Transmission of Light through In these last two equations, (4) and (¢), the azimuth 6, may assume any value, since the structure of isotropic substances does not prescribe definite planes of polarization for transmitted waves, as do anisotropic substances; but if 6, be once given, 6, is then = 6, + 90° and the last equation, (¢), may be written tg 6, ig €, ey cos (t¢—7) cos (i— 7)? (46c’) From this formula, the angle 8’ can be caleulated, provided e,, «@ and 7 be given. The difference (e,—8’,) is.then the amount of rotation which the plane of polarization of incident, monochromatic light suffers on transmission through the isotropic plate. In case the light wave passes through several plates, cemented together as in a thin section mount where n, is the refractive index of the object glass, , that of the Canada balsam, and , that of the cover slip, an incident wave 2 ; similarly, becomes 7, in the object glass, where sin 7, =~ nN, 7, and 7, are the angles of refraction in m, and n, and can be calculated by the general sine formula above. From formula (460) and (46c) it is evident that the total rotation of the plane of polarization of a transmitted wave under these conditions is cot 0’ = cot € cos(t—r,) cos(7,—7,) cos(7,—7r,) cos(¢—7,) Summary.—In the foregoing pages the formulas have been developed which are especially useful in a consideration of the phenomena observed on mounted crystal plates in convergent polarized light. In this discussion, the effects of the plates on reflected light waves have not been treated in detail, nor has a study been made of the relative amplitudes of the reflected and refracted waves; attention has been directed rather to the effects of transparent, inactive plates on the planes of polariza- tion of transmitted light. In the calculation of these effects, four steps are necessary; (1) if. the angle of incidence 7 of the entering light wave be given, the angles of refraction 7,, 7, of the two transmitted waves are found by means of formula (16) in the case of biaxial plates, or by (41) for uniaxial plates, or by (46a) for the single transmitted wave in isotropic plates. (2) The azimuths of the planes of polarization of the two refracted waves are then found by use of equations (17) for biaxial plates, or (42) for uniaxial plates. In the case of iso- tropic plates the plane of polarization of transmitted waves may have any azimuth, so far as such azimuths are dependent on the structure of the material. (3) Having given the angle of Transparent Inactive Crystal Plates. 183 refraction and the azimuth of the plane of polarization of a trausmitted wave, the azimuth of the plane of polarization of the incident wave which produced it is obtained by use of equa- tions (35) for biaxial plates, (48) or (44) for uniaxial plates, and (460) for isotropic plates. (4) To find the azimuth of the plane of polarization of the emergent wave, provided that of the incident wave which produced the refracted wave be known, equation (37) is useful. This last equation, which is apparently new, states that the azimuth of the plane of polarization of an emergent wave 0’,, resulting from the refracted wave W.,, is 90° from the uniradial azimuth of the incident wave e, which ~ produced the refracted wave W,. A detailed discussion of the above formulas is deemed unnecessary in this summary, as they are in large measure standard and the effects of the different factors will appear more clearly in the discussion of the data of observation. In the development of these formulas, no account has been taken of the effects of surface films on the rotation of the planes of polarization of transmitted waves. These films have been shown by P. Drude’ and others to be occasionally of great influence, especially on plates which have been highly polished, while on freshly cleaved plates they are practically absent. The observations listed in the following pages were made largely on cleavage plates, which, however, were usually ex- posed for a month or more before the observations were finished and may have accordingly suffered some deterioration. Parr 2. Observations. Apparatus.—All observations recorded below were made in sodium light, the crystal plate being mounted on a universal stage on the new model petrographic microscope recently described by the writer.’ To insure accuracy, the microscope was carefully adjusted and its adjustment tested at intervals in the course of the measurements. The nicol prisms were of the square end Glan-Thompson type and were crossed by pointing the microscope, from which all lenses had been removed, directly toward the sun whose rays are parallel and so intense that a rotation of less than 1’ of are from the posi- tion of exact crossing of the nicols is readily discerned. By means of the iris diaphragms the sun’s rays were sent through the microscope centrally so that no rotatory effect of the nicol surfaces on the planes of polarization of the transmitted waves was possible. The ordinary type of nicol prism with oblique 1 Wied. Ann., xxxvi, 532, 865, 1888; xxxviii, 265, 1889; xliii, 146, 1891. ? This Journal (4), xxix, 407-414, 1910. P 184 OE. EL Wright— Transmission of Light through ends was first used, but was soon discarded because of the effect of its oblique surfaces on the plane of polarization of the transmitted waves. For exact work in extinction angles, the ordinary type of nicol prism is much inferior to the Glan- Thompson type with square ends. Having thus crossed the nicols accurately, the crosshairs of the ocular were adjusted by using the ocular and Bertrand lens as a microscope and focus- ing on a mounted anhydrite cleavage plate through which par- allel sun’s rays were passed centrally. Here again the sun’s rays are so intense that the position of total extinction of the anhydrite plate can readily be fixed within 1’ of are. By means of the anhydrite plate which extinguishes parallel with its cleavage edges, the principal sections of the nicols and the ocular crosshairs were brought to coincidence. The universal stage was then attached to the microscope stage and its hori- zontal axis of rotation brought to coincidence with the hori- zontal crosshair of the ocular, by use of the lines engraved on the glass disk of the universal stage. This glass disk, together with its supporting ring, was then remoyed and in its place a second ring of precisely the same dimensions substituted, on which a strip of thin glass plate was cemented, and to which in turn one corner of the crystal plate was cemented, the glass plate serving merely as a support for the erystal plate whose major part was left free and exposed on both sides to air. The surface of the crystal plate was then brought to approxi- mate parallelism with the horizontal circle of the universal stage ; it was adjusted to exact parallelism by viewing, through a mounted telescope, the image of a distant light source as reflected from the surface of the crystal plate. The horizontal circle H, of the universal stage’ was then rotated and the erys- tal plate tilted and turned by means of the horizontal circle H, and the vertical circle V, until the reflected image remained stationary on rotation of H,. The circles on the universal stage could be read to 5’ by means of the vernier, while on the microscope stage the vernier intervals were 3’. In neither case, however, were the lines on the circles and verniers suffi- ciently fine to insure greater accuracy in reading than the 3’ or 5’ intervals on the vernier. In actual work each position of total extinction was determined 10 times and the average taken. On sharp extinctions it was found that the different settings were usually within 10’ of the average. In the earliest preparations measured, the positions of total extinction were determined by use of the bi-quartz wedge plate,’ but the fact that, in making the observations with this plate, it was necessary to use the objective and ocular, the glass surfaces of which in turn influence the plane of polar- 1See page 186 below. ? This Journal (4), xxvi, 391, 1908. Transparent Inactive Crystal Plates. 185 ization of transmitted waves, was sufficient reason to discard it. In the final arrangement adopted, no glass surfaces intervened between the nicols and the erystal plate. An enlarged image of the plate was obtained by means of a weakly magnifying microscope consisting of the Bertrand lens and ocular, above the upper nicol. Intense sodium light flame.—To increase the intensity of the sodium light flame, and with it the accuracy of the observa- tions, an arrangement was adopted which in practice has proved entirely satisfactory. A 25° platinum crucible was filled with a mixture of sodium chloride and sodium carbonate and heated over a Bunsen burner, a special mounting of thick platinum wire having been made for the erucible as indicated in the diagram. ), in which the axial bars are drawn for the two positions of the extinguishing plane of the upper nicol (—10° and +15° ) as indicated by the dotted lines. The 1 This Journal (4), xxix, 423, 1910. Transparent Inactive Crystal Plates. 205 observations were made by using the cross grating ocular as shown in the microphotograph fig. 15. The observed codrdi- nated values were reduced to their angular equivalents by use of the apertometer and these in turn reduced to the corre- sponding erystal angles by means of the sine formula and the refractive index 8. The use of the refractive index 8 for all direction introduces an error, but experience has shown that this error is not great and in general may be disregarded. Points were located as accurately as possible along each axial bar and then plotted in projection (indicated by small circles, figs. 14a, 146). Although the axial bars were not perfectly sharp they were well detined and the points were taken along the central line of the bar, the position of each point being determinable to within about 1°, or less for certain positions. In fig. 14a, the results which were obtained from an unmounted cleavage plate are represented ; in fig. 140, the interference figure is that from the same plate mounted in Canada balsam between cover glass and object glass. In each of these figures, the positions of the line of vibration were determined graphically, both by the method of Professor Becke (indicated by small crosses) and by that of the writer (indicated by small circles). A comparison of the relative positions of these small circles and crosses relative to the dot- ted line which represents the position of extmguishing plane of the upper nicol shows that in a few instances the points as determined by Professor Becke’s method are slightly more accurate than the equivalent points of the writer’s method; in the majority of instances, however, the small circles are more nearly correct than the small crosses. As a general] rule, it may be stated that the order of accuracy of the two methods is about the same, the writer’s method having the single advantage of greater simplicity. A critical comparison of the results of observation on mounted flakes with those on unmounted flakes show clearly the effect of rotation by the glass surface, causing the axial bars and axes to shift slightly, so that the direct reading of the optic axial angle is not quite the same in the two cases. The difference is not great, but it is noticeable, and is sufficient to make it advisable to use unmounted plates wherever possible, in optic axial measurements, if results of the highest accuracy are desired. Ordinarily, however, this precaution is unnecessary, since such accuracy is not required. A rotation of the crossed nicols through 90° also generally produces a slight shift of the axial bars from mounted plates, as indicated by fig. 16, which is a direct record to scale of the observed phenomena. In each case the points along the cen- tral line of the axial bar were plotted. The position of this Am. JOUR. Sch SE oree SERIES, VoL. XX XI, No. 1838.—Marcgu, 1911. 1 206 =F. EL Wright—Transmission of Light through central line for an angle of rotation of 15° of the crossed nicols is indicated by the curve I, fig. 16; it sposition for an angle of rotation of 105° is shown by curve II. These two curves do not coincide, and although such measurements cannot be made very accurately, they show that a rotation of the crossed nicols Fie. 16. causes a slight shift of the axial bars of the interference figure of a mounted crystal plate. The amount of shifting rarely exceeds several degrees and is usually less, but it is often sufii- cient to be perceptible and shows the importance of referring the data, when plotting, to the correct position of the extin- guishing plane of the upper nicol. It is, therefore, not imma- terial which one of the principal nicol sections be chosen. If the observations themselves were of a higher order of accu- racy, this fact would be a serious objection to Professor Becke’s method. Anhydrite.—A series of observations (figs. 17a, 6) on a cleay- age plate of anhydrite, unmounted (17a) and mounted (176), cor- roborates the conclusions stated in the last paragraph. The degree of accuracy of the two methods in question is about the same here as in muscovite. A rotation of the crossed nicols through 90° also produced a slight shift of the axial bars on mounted plates, as in muscovite, and it is important, therefore, that the plotting be done with reference to the correct prin- cipal nicol section. Transparent Inactive Crystal Plates. 207 Fie. 17a. Fie. 170. 208 EE. Wright—Transmission of Light through A device to aid in the graphical solution of optical problems involving the use of the stereographic projection. In the measurement of optic axial angles in convergent polar- ized light,’ and also in all measurements by means of the uni- versal stage methods, the stereographie projection plat of Prof. Wulff has proved a useful and necessary adjunct, the angular values of observation being plotted directly on thin transparent paper placed above the plat and held in the center by means of a needle. This needle, however, is not entirely satisfactory, since it does not hold its place rigidly and tends thereby to injure the stereographic plat below. To overcome this dithculty the writer has constructed the device of fig 18 Fic. 18. C D (one-eighth actual size). A heavy brass bar fits into two end blocks of brass, A and B; at its center a small hollow brass rod, C, containing a needle backed by a spring is introduced. By this device the needle is rigidly supported in a vertical posi- tion, and as the distance between the end blocks A and B is 44, there is more then sufficient space available for the pro- jection plat and the overlying drawing. The writer has used this device for several years and has found it satistactory and a time saver.” Summary. Minerals are determined under the microscope by the effects they produce on transmitted light waves. Plane polarized light waves are ordinarily used and examinations are made 1 For determining the Mallard constant of the microsope, whichis required in the measurement of optic axial angles by means of the microscope, Dr. J. S. Flett of London uses a Zeiss Abbe apertometer. His method is simple and accurate and is superior to any method yet suggested. He introduces the micrometer scale in the ocular as usual and then determines the divi- sions covered by the different angles of the apertometer. Since with this device any angle can be instantly set off, an objective can be calibrated rapidly for all possible angles within the field of vision, and an empirical, correct table prepared which is independent of the Mallard formula, thus obviating all errors due to a lack of correction in the objective lenses. * Recently, Prof. Nikitin has had constructed a graduated porcelain hemi- sphere (made by R. Fuess, Steglitz, Berlin, Germany) which the writer has found very satisfactory in optic axial angle projections and slightly more accurate than the projection plats. chiefly because of its lack of distortion toward the margin and consequent acute angled intersections of great circles. This hemisphere has the advantage of serving as amodel in the study of optical phenomena and is a useful piece of apparatus for the petrological laboratory. Transparent Inactive Crystal Plates. 209 partly with and partly without the aid of the upper nicol (ana- lyzer). In anisotropic erystals the planes of polarization of light waves, transmitted along a given direction within the erystal, are prescribed by the crystal structure. On entering or emerging from a crystal plate, plane polarized light waves transmitted obliquely usually suffer a slight rotation of the azimuth of their plane of polarization. The amount of this rotation is rarely more than a few degrees. In practical micro- scope work but little attention has been given to this phenomenon, but in accurate work it is a factor which must be considered. In the foregoing pages the attempt has been made in Part 1 to present, in terms of the electromagnetic theory of light, the general mathematical treatment of the transmission ot light waves through a transparent inactive crystal plate, spe- cial attention being given to the rotatory effects of the boun- dary surfaces of the crystal plate on the plane of polarization of a transmitted wave. This problem was first solved in 1835 by J. MacCullagh and also by F. E. Neumann; since their time a number of investigators have made important contribu- tions to its solution. Interest, however, has centered chiefly in the reflexion of light waves by erystal surfaces and no con- nected presentation of the mathematics covermg the phenom- ena of refraction in crystal plates appears to have been made. This has been essayed in Part 1. The greater part of the ground covered therein is familiar, but several of the formu- las derived appear to be new, notably (326) and (87). Of these (87) is important and states that the uniradial azimuths of the plane of polarization of the emergent waves W’, and W’, are 90° from the uniradial azimuths of the entering waves whieh, on refraction, produce the waves W, and W,. In other words, the positions of extinction on emergence for either one of the two possible refracted waves, W, or W,, resulting from a single plane polarized light wave, incident at the surface of a crystal plate, are precisely 90° apart. The positions of extinction for the two waves do not, however, coincide and there is in general, therefore, no position of total extinction for waves transmitted obliquely through a crystal plate. Both theory and the observations of Part 2 show that as a general rule, a uniradial, plane polarized light wave, after transmission through a bare crystal plate (preferably a cleav- age plate so that the disturbing effects of surface films caused by polishing are not serious), is still plane polarized, but its plane of polarization has suffered a slight rotation depending on the direction of transmission, and if examined under crossed nicols does not appear perfectly dark in consequence. In thin crystal plates the two refracted waves W, and W, overlap to a 210 Ff. EB. Wright—Transmission of Light through large extent and there is no position of total extinction for the tilted crystal plate even if the upper nicol be rotated alone. In general it may be stated that from an incident plane polar- ized wave two refracted waves are formed, which on emerg- ence from the plate are each still plane polarized, but their planes of polarization are not precisely 90° apart. The result- ant light as observed through the analyzer is consequently elliptically polarized and there is no possible position of total extinction of the plate, but rather a region of minimum illumi- nation which may extend over several degrees. These relations have an important bearing on methods based on the determination of the positions of extinction of obliquely transmitted waves, and preclude at once a high order of accu- racy in the measurements. If the observed crystal plates are mounted in Canada balsam, the rotatory influence of the sur- faces of the glass and Canada balsam mount enter the problem and tend to complicate the phenomena still further. The measurements of Part 2 show: (1) That a tilted glass plate may rotate the plane polarization of a transmitted plane polarized light wave several degrees, and that the amount of rotation increases with the angle of tilting; (2) that the observed uniradial azimuths of tilted cleavage plates of calcite agree closely with the calculated values; (8) that for the central areas of tilted plates of calcite, nephelite, muscovite, and ara- gonite, there are no positions of total extinction. It settings be made at the apparently darkest positions of the plate during the rotation of the microscope stage, these positions are often several degrees from 90° apart, and if the observed azimuths of the plane of polarization be taken as the azimuth of the refracted waves within the crystal, errors of several degrees are easily possible. (4) An obliquely transmitted wave will be extinguished provided its direction of vibration after emerg- ence is contained in the extinguishing plane of the analyzer. The direction of vibration of an observed dark point on the axial bar of an interference figure is therefore the line of intersection of the extinguishing plane of the upper nicol with the polar plane of the given point. This construction, sug- gested by the writer, does not take into consideration the rotatory effects of the surfaces of crystal plate and glass mount, and is accordingly only an approximate method. Prof. Becke has suggested another method, which is, in effect, to find the inter- section of the polar plane with the great circle in stereo- graphic projection, which is tangent to a line parallel with the principal section of one of the nicols. The points obtained by Prof. Becke’s method are slightly different from those obtained by the writer’s method, but not sufficiently different to affect the degree of approximation obtainable by such methods. In Transparent Inactive Crystal Plates. 211 principle, however, the two methods are fundamentally differ- ent, and a detailed discussion, together with a series of meas- urements on interference figures of muscovite and anhydrate, indicate the general validity of the principle on which the method proposed by the writer is based; in this method the rotatory effects of all boundary surfaces are disregarded and for this reason the results obtained by its use are only approxi- mately correct. Several devices are described which have been found ser- viceable in connection with this work: (1) An apparatus for securing an intense and constant sodium light. (2) A simple and accurate method for adjusting the petrographic micro- scope. (8) A device to aid in the work with the stereo- graphic projection plat. Geophysical Laboratory, Carnegie Institution of Washington, Washington, D. C., November, 1910. 212 Gooch and Boynton—Estimation of Barium Art. XX.—The Separation and Estimation of Barium Associated with Calcium and Magnesium, by the Action of Acetyl Chloride in Acetone Upon the Mixed Chlorides ; by F. A. Goocs and O. N. Boynton. [Contributions from the Kent Chemical Laboratory of Yale Univ.—cexviii.] In former papers from this laboratory* it has been shown that certain chlorides may be quantitatively precipitated for purposes of analysis by treating their water solutions with’ aqueous or gaseous hydrochloric acid and ether. The present paper is an account of procedure for the pre- cipitation of barium chloride from water solution and its sepa- ration from calcium and magnesium by the use of acetyl chloride to decompose the water of the solution according to the reac- tion CH,COCl + H,O = CH,COOH + HCl, inconvenient vio- lence of reaction being moderated by the addition of acetone which mixes in all pr oportions with both acetyl chloride and water and by itself exerts no appreciable solvent action upon barium chloride. When a mixture of acetone and acetyl chloride, preferably 4:1, is added slowly to a very concentrated solution of barium chloride in water, the water is attacked at once, hydrogen chloride is liber ated, and precipitation begins immediately. Gi the temperature is kepi down during the process by immers- ing in cool running water the vessel in which reaction takes place, no more than a mere trace of barium can be detected by sulphuric acid in the residue left after evaporating the liquid separated from the precipitate by filtration through asbestos. When, however, the temperature is allowed to rise, in conse- quence of the heat liberated in the reaction, an appreciable amount of barium may be found by sulphuric acid in the filtrate. Below are given the data of experiments in which the residue obtained (a) by treating a solution of barium chloride in 1 of water with 30°? of a 4:1 acetone- -acety| chloride mixture and collecting the precipitate upon asbestos in a perforated crucible, washing with acetone and with ether, was weighed after drying in air, then (0) treated on the asbes- tos for ten minutes with 15-20 of acetyl chloride, washed with acetone and with ether, dried in the air and weighed, then (¢) digested for ten minutes with 20-25°* of 2:1 ace- tone-acetyl chloride mixture, washed with acetone and with ether, dried in the air and weighed, and then (d@) heated in the air-bath, or to low redness, and weighed. * Mar, this Journal [8], xliii, 521; Havens, this Journal [4], ii, 416; iv, 111; vi, 45; vi, 396. Associated with Calcium and Magnesium. 213 Experiment I PLE ae II —_——_-+7~ —— Weight Loss Weight Loss grm. grm. grm. grm. BaC],.2H,0O taken ..-- -_-- OghOM 2M ee OULOOO a eee (a) Residue after precipitation, washing, and drying in air, 071008 0:0004 0:0996 0:0004 (b) Residue after treatment with acetyl chloride, washing, andl dyamereim a, 2-1 071006 0:0002 90:0996 0:0000 (c) Residue after treatment with acetone-acetyl chloride mix- ture, washing, and drying Maas SO las 0:0985 0:0021 0°0981 0-0015 (d) Residue after heating, BaCl, 0:0846* .... 00839 © .--- BaCl,.2H,O corresponding to ba@ le Townde easy: = sans OOO 35) oe Ser COLO9SD ein. cae. ‘Loss of BaCl,.2H,O due to solubility and dehydration ---. 00027 ---.- 0°0019 Loss of BaCl,.2H,O due to solubility, calculated from BaCl,.2H,O taken and BaCl, OUMKC ee a tes eee eee TS ES ff OLOOLINE |. era =z 0'0015 Wosspby, dehivdiratione jae =)t ae. 010008) "2-3-1" 020004 * Heated to low redness. + Heated to 135° for 114 hrs. From these results it appears (a) that when the acetone- acetyl chloride mixture (4:1) acts upon the cooled concen- trated water solution of barium chloride the precipitate is the hydrous chloride, BaCl,.2H,O, only the water in excess of that needed to form the hy drous salt being immediately attacked ; (6) that acetyl chloride by itself pr oduces only slight dehydra- tion of the salt without marked solubility ; and ‘(e) that pro- longed action of the acetone-acetyl chloride mixture (2:1) results in appreciable dehydration and considerably increased solubility of the salt. By fnrther experimentation it was shown that when the acetone-acetyl chloride mixture is added without cooling to the water solution of barium chloride the heat of reaction favors dehydration of the hydrous salt, and the anhydrous salt may go into solution to the amount of several milligrams in 10° of the precipitating mixture. Upon filtering the mixture and treating the filtrate with acetone, with acetyl chloride, or with the acetone-acetyl chloride mix- ture the dissolved anhydrous salt is not thrown out of solution, 214 Gooch and Boynton— Estimation of Barium but the addition of a drop of water is sufficient to induce immediate precipitation in the form of the hydrous salt. Incidentally it is interesting to note that when water acts upon the colorless mixture of acetone and acetyl chloride the solution becomes yellow, and then reddish, and develops a distinetly fruity odor, condensation taking place between the acetone and acetyl chloride. The boiling points of the col- lected filtrates from a series of barium chloride precipitations after standing about a week ranged from 50:5° to 250°, and left a resinous residue at that temperature. From the results of the experiments described, it may be inferred that the best conditions for the quantitative precipita- tion of barium chloride by the acetone-acetyl chloride mixture should be found in the use of minimum amounts of water, the preservation of ordinarily low temperature, a liberal propor- tion of acetone, and not too prolonged digestion of the precipi- tate in the excess of the precipitant. These conditions have been complied with in the quantitative tests. Barium chloride was prepared for the work by precipitating it with strong hydrochloric acid from a water solution of the presumably pure salt, recrystallizing twice from water, and drying in the air. On gentle ignition the salt lost water cor- responding to the ideal composition of the hydrous chloride, - BaCl,.2H,O. In each test a portion of this salt was weighed out into a small beaker and dissolved in 1°™° of water. The beaker was cooled by immersion in a water-bath preferably supplied with running water at a temperature of about 15°. To the cooled solution, constantly shaken, the acetone-acety] chloride mixture was added from a dropping funnel at the rate of five drops to the second. Other data of the experi- ments with barium chloride are given in Table I. The pre- cipitate was filtered off upon asbestos in a perforated crucible, dried, or ignited, and weighed as the anhydrous chloride, BaCl.,,. From these results it appears that the best of the conditions studied for the handling of 0-1 grm. of hydrous barium chlo- ride are the solution of the salt in 1°° of water, treatment with 30° of the 4:1 mixture of acetone and acetyl chloride, washing with acetone, and drying in the air-bath at 135° or at low redness. ' The application of these conditions to the separation of barium from moderate amounts of calcium and magnesium proves to be easily feasible. When acetone is added to the concentrated solution of calcium chloride or magnesium chlo- ride in water two liquid layers are formed, the acetone above and the aqueous layer below; but the addition of a few drops of acetyl chloride renders the liquids miscible while further addition causes no precipitation. When the 4:1 mixture of Associated with Calcium and Magnesium. 215 TABLE J. The Estimation of Barium. Amount of mixture and BaCly Water to composition by volume taken as BaCl, dissolve ———_—_—— tH BaCl,.2H,0 found Error BaCl,.2H.O To precipitate To wash grm. grm. grm. 1-0°0859 0:0859 —0-0000t ems Get eal OGL 2-0:0861 0:0854 —0-0007t ES RE Oe TOE 3-0°0861 0°0858 —0:0003f ss aE oil HOSP et 4-0°0862 0:°0854 —0-0008* ee Ge! Weil LOPE Bei 5-0°0857 0:0854 —0-0003* sf ecm? 931 UIE Weil 6-0°0858 0:0860 +0:°0002* ve (HOE BT, i Oto abe 7-0°0860 0:0859 —0-:0001* © § Ge! Beil SO Za 8-0°0853 0°0850 —0:0003* WY Ger acetone 9-0°0854 0:0848 —0:0006* fs Gone 2 il y 10-0°0852 0°0851 —0-:0001* “s 6cem® 221 ss 11-0°0857 0:°0856 —0-0001t oo Ge Beil a 12-0:0852 0°0845 —0-0007+ os Goat) Beil ce 13-0°0855 0°0852 —0-0003+ oe Gore o a Hy 14-0°0862 0:0862 —0-0000t He SOc es 15-0°0868 0:0868 —0-0000t ee BOY Zep ss * Ignited at low redness. + Dried at 185° for 144 hrs. acetone and acetyl chloride is added at the rate of five drops in the second to the solution containing no more than 0°5 grm. of the calcium and magnesium salts, barium chloride is pre- cipitated and calcium chloride aud magnesium chloride are dissolved ; but when the soluble chloride is present in the pro- portion of 1-0 erm. to 0'1 grm. of the barium chloride, the rate of addition of the precipitating mixture should not be greater than two drops in the second at the start in order to avoid inclusion of the soluble salt in the insoluble barium salt. Even in such cases the mixture may be added at the rate of tive drops in the second, after the greater part of the barium is down. Tables II and III contain the data of experiments upon the separation of barium from calcium and magnesium. The results obtained in the separation of 0-1 grm. of the barium salt from 0°5 grin. of calcium and magnesium salts are excellent. The separation of barium from strontium proves not to be so simple. When the 4:1 mixture of acetone and acetyl chloride is added to the concentrated water solution of 0-1 grm. of strontium chloride a partial precipitation takes place. When the precipitate thus produced was filtered: off, washed with ace- tone and with ether, and dried in air, it lost SET amounting to 19°93 per cent and 20-00 per cent of its weight on heating 216 Gooch and Boynton—LKstimation of Bariwm TABLE II, The Separation of Barium from Calcium. » Ba), Water used Amount of takenas CaCl,.2H.O BaCl, to dissolve mixture BaCly.2H;0 taken found Error salts (4:1) used grm. erm. grm grm. em?, em?, 1-0-0859 01000 0:0859 0°0000* hs ae 30 2—0:0867 071040 0°0867 0:0000* 1 S30) 3—0°0868 0°1022 0:0868 0:0000* il 30 4-0:°0865 0°1020 00865 0°0000* 1 30 5-0°'0868 01017 0°0869 +0:0001* 1 30 6—0°0864 071016 0:0861 —0:0003* 1 30 7—-0°0866° 0°3025 0°0867 +0:0001*. 14 30 8—0:0859 075025 0:0859 0:0000* 2 30 9-0°0860 1:0020 0:0878 +0:0018* 3 30 10—0°0859 10020 0:0855 — 00004 2 30 11-0°0864 1°00385 0:0867 + 0°0003F 2 30 * The precipitant was added at first at the rate of five drops in the second. + The precipitant was added at the rate of two drops in the second at the outset and later of five drops in the second. TaBLeE Lil. The Separation of Barium from Magnesium. BaCl, Water used Amount of taken as MgCl..6H.0 BaCl, to dissolve mixture BaCl..2H.,0 taken found Error salts (4:1) used grm. germ. grm. germ, cm’. em’, 1—0°0858 071000 0°0857 —0-:0001* u 30 2—0°0869 0'1025 0:0870 + 0°0001* i 30 3-0°0858 0°1025 0°0858 0:0000* 1 30 4—0°0862 0°1010 0:08638 =+0°0001* h 30 5-0'0858 071006 ~=—- 00860 +0:0002* 1 30 6—0'0860 0°1020 0°0859 —(:0001* i 30 7—0:0860 0'1010 = 0:0862 “+ 0:0002* 1 30 8—0'0865 0°3010 0°0867 +0°0002* 14 30 9-0°0864 0°5000 =0°0867 +0:0003* 2 30 10—0°0868 1:0015 = 0°0878 +0°0010* 3 30 11-—0°0853 10010 0:0854 . +0:0001t 3 30 * The precipitant was added at the rate of five drops in the second, +The precipitant was added at first at the rate of two drops in the second and later of five drops in the second. to 185°. Obviously the salt was essentially SrCl,.2H,O, which should contain theoretically 18°51 per cent of water. This precipitate of SrCl,.2H,O when treated with a mixture of ace- tone and acetyl chloride containing a larger proportion of the latter, goes into solution and is again partially precipitated upon increasing the proportion of «cetone, essentially as SrCl,.2H,O. Associated with Caleium and Magnesium. 217 When a mixture richer in acetyl chloride, the 2:1 mixture of acetone-acetyl chloride, is added to the concentrated water solu- tion of strontium chloride, the precipitate first formed is slowly redissolved in a sufficient excess of the mixture and again partially precipitated upon the addition of more acetone. This second precipitate of SrCl,.2H,O is not completely soluble, however, when the proportion of acetyl chloride is again increased, but will dissolve upon the addition of an acetone- acetyl chloride mixture to which a few drops of water have been previously added and which is, therefore, charged with hydrogen chloride. So it appears that the solubility of the strontium chloride, SrCl,.2H,O, depends to a very large extent upon the concentration of hydrogen chloride in the mixture. In attempting the separation of barium from strontium, therefore, it was the 2:1 mixture of acetone with acetyl choride which, on account of its higher power as a solvent for SrCl,.2H,O, was added to the concentrated water solution of * the barium chloride and strontium chloride, though this mix- ture has been shown to be somewhat less favorable to the com- plete precipitation of barium chloride than the (4:1) mixture containing the larger proportion of acetone, and the addition was made at the rate not greater than two drops in the second or 30 in ten minutes. ‘The precipitate, filtered upon asbestos and washed with acetone, was dried at 135°. The data of these experiments are recorded in the table. The separation of barium from strontium by the process TABLE IV. The Separation of Barium from Strontium. BaCl, Water used Amount of taken as SrCl, BaCl. to dissolve mixture BaCl,.2H,0 taken found Error salt (2:1) used grm. grm. grm. grm. em?, em?, 1-0°0867 0°0385 0:0923 + 0°0056 1 30 2—0'0866 0°0304 0°0857 —0°0009 1 30 3—0:0860 0°0320 0°0861 +0°0001 1 30 4—0°0856 0:0315 00840 —0°0016 1 30 5—0°0856 0°0307 0°'0848 — 0°0008 1 30 6-0'0859 0°0304 0:0839 —0°0020 ] 30 7—0:0862 0:0307 0:0859 —0°0003 1 30 8—0°0857 0:03815 0'1160 + 0:0303 0°5 30 9-0°0857 0:0317 01058 +0:0201 0°5 30 10-0°0853 0°0315 0'1083 + 0°0230 0°5 30 11-0:'0869 0°6305 0'1043 +0°0174 0°5 30 12—0°0863 0°0307 0:0906 + 0:0043 0°5 30 13—0°0859 0'0808 0:0849 —0:0010 0°5 30 14-0:°0869 0°0108 0:0870 +0:0001 1 30 15—0°0865 0°0110 0°0858 —0:0007 1 30 16—0°0853 0°0115 0°08538 + 0°0000 1 30 17-0:0861 0:0109 0:0870 + 0:0009 1 30 218 Gooch and Boynton—Estimation of Barium. described is obviously only approximate, some barium chloride going into solution in the 2:1 mixture of acetone and acetyl chloride, while the solubility of the strontium chloride turns upon the amount of water originally present—that is, upon the development of hydrogen chloride: It appears, therefore, that the method which rests upon the action of a 4:1 mixture of acetone and acetyl chloride upon the concentrated solution of the chlorides affords easy and exact means for the separation and estimation of barium associated with calcium and magnesium. It is not recommended for the separation of barium from strontium. : — ee = = Arr. XXI.—A Feldspar Aggregate Occurring in Nelson Co., Virginia; by Witt1am M. Tuornton, Jr. Near Rose’s Mill in Nelson Co., Virginia, the General - Electric Company recently carried on some mining operations with the view of obtaining rutile, where it occurs as a rock- forming mineral in the unique rock type ‘‘nelsonite.”* To give some idea of the nature of this peculiar rock, an analysis of the rutile phase is here inserted : Analysis of Nelsonite.—General Electric Company’s mine? 14 miles northwest of Rose’s Mill. Essential minerals :—Rutile (TiO,) and apatite (Ca,FP,O,,). Accessory minerals :—Ilmen- ite (FeTiO,), pyrite (FeS,), and yuartz (Si0,). SION SOLER ene ho feta bene eee 0°67 per cent Bei O) rae SR NGhe oe wets serene tae EOI ee HeOr.=_ a es wie aD ee lke La ee aes ema) () fy ss MeO re aye oo 2 eae ae eee eae OmlaeD es CaO See ee eee eee He “ HO! (ab 110° Cpe ee ae Oe EVOy(above 110 21@:) Gee sens ae 0-1] a DiO ie eo), aie ge 6967“ Pe io be Oe au L Seed eee ote RON ee Aa ee (Cleese Ben ies ate Ss ENS : AE eimai age em (10) Fe SS eee ie Ur Se ne Ree eee 0°34 ce 101°21 Dp (orejfsy OM ae veered Hayy 8 0°39 SUD EMIT AO 11 —— an ye 100°82 At this place the narrow, well-defined dike of nelsonite intersects a metamorphosed pegmatite. The pegmatite is *Name proposed by T. L. Watson, Director Virginia Geological Survey. Seve Mineral Resources of Virginia, 1907, p. 300. W. MW. Thornton, Jr—A Feldspar Aggregate. 219 composed essentially of feldspar and blue quartz and in places much hornblende. The accessory minerals ilmenite and pyrite and apatite are also present. The feldspar is by far the domi- nant portion of the rock; and, since it presents some unusual features, it was thought by the author that a study of its com- position would prove of some interest. Of course the natural procedure would be to isolate mechanically the feldspathic por- tion and to analyze the most homogeneous material obtainable. But three analyses of the pegmatite were required for geolog- ical purposes; and since time was lacking in which to make a fourth of the feldspar alone, it was decided to employ the analysis of the extreme acidic phase for calculating the compo- sition of the feldspar. The color of the feldspar is light bluish gray. Under a magnifier of twenty diameters it appears decidedly transparent and glassy. The texture is one of very close crystalline aggregation. Specific gravity = 2°68. From all outward appearances one would suppose it to be a definite species; but the analysis and portioning of the molecules to form the respec- tive feldspars in the accompanying table shows it to be a mixture of orthoclase and plagioclase, and that the plagioclase is made up of albite and anorthite in the ratio of 10 to 7. This is also confirmed by microscopic examination of thin sections. * Pegmatite (feldspathic facies) near Rose’s Mill, Nelson Co., Virginia. Per Mo. Rel. no. cent wt. mos. Rel. no. feldspar mos. SLO Smee se ean ae ae 59°92 = 60 = 0:9987 2K AISi;0, = 0°08117 Als Oy eee Sea ee a eal 24°23 +102 = 0°2375 2NaAl1Sis0, = 0:081 Hie) © seeks Sytner aa 0:29 CaAl.Si,0, = +1137 LSE Os SR Da ee aN 0°24 [CasP20,; = -0006] Mie Oia cee a IS Soars as 0:23 Cia Os Sra eh ais as ek 6°47 = 56= 01155 .-. NaAISi;0; : CaAl.Si.0, = Nias OE ee ae a le Fe 5°03 + 62 = 0:081 Ql == LON 7 NEG Os Sa TS eS Ne a 2°93 + 94 = 0:03117 HeOlat T07C:) 222255 0-08 H.O (above 110° C.)_.-_- 0°28 (CLO) Sts ae lars eee mee trace AUTO Pps Gee eR Sis aeetee ea ae 0:22 To O) plete ie ye leet eos 0-09 +142 = 0:0006 SC eee Pees Me eo trace IVErn Ocha ely apn see Re trace 100°01 Pegmatite: = .*. orthoclase — 17°337 per cent, plagioclase (Ab;,An;) 74:057 per cent. Composition of feldspar, aggregate : orthoclase 18°96 per cent ; plagioclase (Ab,.An7) 81:04 per cent = 100°00. * Bull. 430-D, U. S. Geological Survey, p. 57. ‘‘The Virginia Rutile Deposits,” by T. L. Watson and S. Taber. 220 =W. M. Thornton, Jr.—A Feldspar Aggregate. After combining all the potash molecules with alumina and silica to form orthoclase and likewise all the soda molecules to form albite and all the lime molecules (except enough to satisfy the phosphoric anhydride to form apatite) to form anorthite, there is a little alumina in excess. This can be accounted for by assuming the hornblende to contain some alumina, which is probably the case, or by errors in the determinations. In general the methods of analysis employed were those in use by chemists of the U. S. Geological Survey. In the determination of the alkalis the purest reagents of Dr. T. Schuchardt were used, and the filtrate from the first lime precipitation was concentrated in a silver basin; glass and porcelain vessels were avoided where any error might arise from their use. Virginia Geological Survey, University of Virginia, January 16, 1911. O. F. Cook—History of the Coconut Palm in America. 221 Arr. XXII.—Wistory of the Coconut Palm in America ; by O. F. Coox.* Many scientific text-books and works of reference support the popular idea that the coconut palm is specially adapted to tropical seacoasts and is confined to maritime regions. No other example of special adaptations of plants to their environ- ments has had longer currency or more confident belief. Nevertheless, it seems that the botanical romance of the coco- nut, protected by its thick husk and floated from island to island in advance of human habitation, must go the way of many other pleasing traditions. What natural agencies have been supposed to do for the coconut is now to be recognized as the work of primitive man. The truth proves again to be stranger than the fiction. The coconut exists in the lowland tropics only as a product of cultivation. It does not plant or maintain or distribute itself on tropical seacoasts, and would entirely disappear from maritime localities if human care were withdrawn. The habits of the palm from the botanical standpoint, its signifi- cance in human history, and even its agricultural possibilities are misunderstood unless we are able to lay aside the maritime traditions. An outline of the evidence for the American origin of the coconut palm and of its distribution by human agencies has been published in a previous number of the Contributions.+ The present study carries the subject further in two principal directions. It brings additional facts to show that the coco- nut palm was already widely distributed in the New World before the arrival of the Europeans, and that it is not naturally a maritime or humid tropical species, but a native of drier and more temperate platean regions in South America. A com- parison of the habits of germination of the coconut with those of other related American palms shows other and very differ- ent uses for the characters that have been looked upon as special adaptations for maritime dissemination. The huge seed with its immense store of food materials and its thick fibrous husk make it possible for the coconut to propagate itself im the relatively dry interior localities where it appears to have originated. The inability of the palm to withstand shade explains why it has been unable to establish * Extracted from Contributions from the U. S. National Herbarium, vol. xiv, pt. 2, pp. 271-342, 1910. The portions here given are from the intro- duction and summary. + The Origin and Distribution of the Cocoa Palm, Cont. Nat. Herb., vol. vii, pp. 257-298, (1901.) Am. Jour. Sci.—FourtH SERIES, Vor. XX XI, No. 183.—Marcga, 1911. 16 222 O. F. Cook— History of the Coconut Palm in America. itself as a wild plant on any tropical seacoast. The application of these facts to cultural problems shows that the possibilities of an extratropical extension of the coconut palm are not to be realized on seacoasts, but in interior desert regions where larger amounts of heat and sunlight are to be obtained. Though the biological evidence of the American origin of the coconut palm appears complete and adequate, recent years have brought to light several additional facts which may be of use to those whose training and habits of thought lead them to attach great weight to the historical arguments of De Candolle and other writers who believed in the Old World origin of this palm and its dissemination by the sea. The reader is impressed by De Candolle’s references to many old and rare books, and will naturally remain loth to believe that so eminent an authority could have come to an erroneous conclusion, unless all the foundations of his opinions are carefully reex- amined. It is important to trace and clear away any mistakes or false deductions which obscure the early history of cultivated plants. Misconceptions regarding the origin and dissemination of any important economic species tend to distort human history as well as to mislead botanical and agricultural investigation. It is only when we view the past with the right perspective that we gain correct ideas of the factors which control our present interests and our future progress. Civilization itself is based on cultivated plants, and history may be written with as much propriety from the agricultural standpoint as from the mili- tary, political, or commercial. SuMMARY OF RESULTS. The history of the coconut palm has relation to several different kinds of scientific questions, so that the facts require to be summarized from several different standpoints. Botanical Conclusions. All the palms that are related to the coconut, comprising about 20 genera and 200 species, are natives of America, with the possible exception of a single species, the West African oil palm. All the species of the genus Cocos and of the closely allied genera are natives of South America. The species of Cocos that are most related to the coconut are natives of the interior valleys and plateaus of the Andes, where the coconut also thrives, remote from the sea. Comparison of the structure of the fruit and the method of germination of the coconut with those of the related palms indicates a high degree of specialization, but not for purposes of maritime distribution. The unusually large, heavy seed and O. F. Cook—LHistory of the Coconut Palm in America. 223 the thick, fibrous husk are to be considered as adaptations for protecting the embryo, assisting in germination, and establish- ing the young plants in the dry climates of interior localities, the only conditions where this palm could be expected to maintain its existence in a wild state. The habits of the coconut palm afford no indication that its original habitat was on the seacoast, and none of its closer relatives have maritime habits or maritime distribution. The coconut palm does not appear to be able to maintain itself under littoral conditions without the assistance of man. Though carried by man to all of the warmer parts of the earth, it has not been able to establish itself as a wild plant on any tropical coast, but. is always crowded out by other vegetation after human care is withdrawn. Wafer’s circumstantial account of the existence of large numbers of coconut palms on the Cocos Islands, 300 miles west of Panama, in 1685, taken together with their almost complete disappearance at the present day, affords a striking illustration of the dependence of the coconut upon human assistance, not only for distribution, but for its continued exist- ence on oceanic islands. The dissemination of the coco palm along the tropical coasts is to be ascribed to the agency of primitive man, as with the sweet potato, banana, and other domesticated plants which were widely distributed in prehistoric times. The theory that it has been disseminated by ocean currents is gratuitous, unproved, and improbable. The development of distinct varieties of the coconut has not been confined to the Polynesian and Malayan islands. Dis- tinct varieties are also to be found in isolated localities in America, such as the Soconusco region of Mexico and the island of Porto Rico. The existence of many and diverse varieties in the Malay region does not indicate that the species is native there, but the opposite, since the proximity of the wild stock of a species is likely to hinder the appearance and preservation of mutations among its cultivated representatives. The relative uniformity of the coconuts of America is in accord with the probability of an origin in this hemisphere. The discovery of distinct varie- ties in isolated localities in America accords with the proba- bility that the Malayan varieties have arisen, like other culti- vated varieties, through segregation and mutation rather than by gradual evolution and natural selection. Historical Conclusions. At the time of the discovery of America the coconut was not confined to the Pacific side of the Isthmus of Panama, as 224 O. F. Cook—Tistory of the Coconut Palm in America. De Candolle believed, but was already widely distributed along the Atlantic side of the American tropics. Early records show its presence in Cuba, Porto Rico, Brazil, and Colombia at dates so early as to preclude the idea of introduction by the Spaniards, The statement of Pickering, frequently quoted in works of reference, to the effect that coconuts were reported by Columbus on the coast of Central America during his fourth voyage, proves to be erroneous. On the other hand, there appears to be a definite reference to the coconut in Cuba in the journal of the first voyage of Columbus. De Candolle’s inference from Acosta’s report of coconuts in Porto Rico at the end of the sixteenth century, that they had recently been introduced by the Spaniards, proves to have no warrant in history and is directly opposed by the more extended reference to the coconut in Porto Rico by the Duke of Cum- berland’s chaplain, who visited the island only a few years after Acosta. De Candoile’s use of the testimony of Piso and Marcerave to support the idea of the introduction of the coconut into Brazil by Europeans is also unwarranted, since those writers only indicated that the plant was cultivated. An earlier and more explicit record, unknown to De Candolle, gives an account of the coconut as one of the native products of Brazil. The journal of Cieza de Leon, who accompanied the first Spanish expedition to the interior of Colombia, indicates the presence of the coconut palm in localities where it still con- tinues to exist, as shown by the accounts of Velasco, Humboldt, and more recent travelers, down to the present decade. Ethnological Conclusions. The American origin of the coconut palm and the strict limitation of its status in maritime tropics to that of a culti- vated plant are facts of ethnological significance. The wide distribution of the coconut in prehistoric times is evidence of the antiquity of agriculture in America and of very early communication across the Pacific. The American origin of the coconut palm, along with its inability to maintain itself on tropical seacoasts without human assistance, compels us to believe that its trans-Pacific distribu- tion was the work of primitive man. The dependency of the Pacific islanders upon the coconut may be taken to show that these islands could not have been occupied without the pre- vious domestication and dissemination of the coconut. In view of the fact that several other palms of unquestioned American origin have been domesticated by aborigines of the American tropics, no ethnological objection can be raised to O. F. Cook—History of the Coconut Palm in America. 225 the idea that the coconut palm was originally domesticated in ancient America. The name “coco” does not appear to have been applied to the “Indian nut” till after the discovery of America and is to be considered as a word derived from the natives of the West Indies. Other native names for the coconut are found among primitive tribes of Costa Rica, as well as in Brazil. The presence of large numbers of coconuts on Cocos Island in the time of Wafer (1685) and their subsequent disappearance should be considered as evidence that the island was formerly inhabited, or at least regularly visited, by the maritime natives of the adjacent mainland. The fact that the coconut is largely restricted to islands and tropical countries of low elevation explains its importance among the preéminently maritime people of the Old World tropics and its relatively slight importance among the nonmari- time natives of the lowland tropics of America. The evidence of the prehistoric dissemination of the coconut and other American cultivated plants across the Pacific Ocean is such as to warrant a careful consideration of other indica- tions that agricultural civilization developed originally in America and was distributed to the shores of the Pacitie and Indian oceans by a primitive people with agricultural and maritime habits, like those of the Polynesians and Malays. The existence of a distinct tribe of frizzle-haired people near the Isthmus of Panama at the time of the discovery does not rest alone on Peter Martyr’s casual mention of the finding of negroes, but is supported by Oviedo’s contemporary history written directly from the testimony of Balboa and other mem- bers of his expedition, just after their return to Darien. The facts are not to be explained reasonably by assuming a chance arrival of African negroes, but indicate that prehistoric com- munication across the Pacific continued after the frizzle-haired Melanesian race had spread westward in the Pacitic. Such communication would account for the existence of the banana plant in America previous to the arrival of the Span- jards, as well as for the Old World distribution of the coconut palm and other cultivated plants of American origin. The banana plant is as evidently a native of the eastern continent as the coconut palm of the western. Evidence of these facts appears very definite and concrete from the biological stand- point, and is worthy of careful consideration by ethnologists. Agricultural Conclusions. The coconut is confined to seacoasts only in the humid low- lands of the Tropics; in dry regions it is not restricted to coasts, but thrives in many districts remote from the sea. The 226 O. F. Cook—History of the Coconut Palm in America. fact that it received scientific study only as a maritime plant should not longer obscure the fact that it is also adapted to interior localities with saline soils. The cultural problems of the coconut palm should be investigated quite apart from the idea of maritime habits and distribution. The possibility of raising coconuts in frost-free localities out- side the Tropics is not to be tested along the seacoast, but in interior districts where larger amounts of sunlight and heat are available, as in the valleys of southern California and Arizona. The coconut, like many other plants, is not tolerant of shade nor of long-continued cool and cloudy weather. Other species of Cocos that are less exacting in their requirements of sun- light and heat have been found to do well along the California coast. The possibility of introducing coconut palms into southern California is not disproved by the absence of these palms from Egypt and Palestine. Though the climatic conditions are probably favorable, it does not appear that any adequate effort has been made to introduce the palms in those countries. The ability of the coconut to thrive on seacoasts shows that its requirements of heat are not as great as those of the date palm. Though probably less hardy than the date palm, it is not impossible that the coconut may be able to exist in frost- free localities that have not enough heat for the ripening of dates. The possibility of introducing the coconut palm into southern California and Arizona can not be fairly tested by the planting of the maritime varieties. The chances of success will be very much greater with the varieties that are adapted to the dry interior localities of the temperate plateaus of the Andes. Loomis—New Mink from the Shelt Heaps of Maine. 227 Arr. XXTIL—A New Mink from the Shell Heaps of Maine ; by F. B. Loomis. Dorine the summer of 1909 the Amherst Biological Expe- dition collecting in the shell heaps along the Maine coast, opened the heap on the east side of Flagg Island in Casco Bay, near South Harpswell. This heap is distinct from any of the others in several features, but especially in having large num- bers of mink bones in it, the mink being, however, larger than any species now living in New England and markedly different from any that are known. It is as large as the largest species from Alaska.* In the course of the week spent in the Flage Island heap no less than 45 individuals were found in which there were 10 upper and 34 lower jaws of males, and 2 upper and 11 lower jaws of females. Beside these 3 lower jaws of the same species were found in the heap on Sawyers Island near Boothbay, 2 in the Seward Island heap in Frenchman’s bay, and one in the Winter Harbor heap. The other localities worked did not offer any of this mink; so that it would appear that Flagg Island was more or less overrun with these minks during the shell heap period, while they occurred also in small numbers along the coast to the east and north. The exact time when they lived is difficult to estimate, but the heaps contain nothing of European origin, so they were accumulated before 1627 and are probably as much older as it took to build them up, perhaps 200 to 400 years more. The mink is not confined to any one level on Flage Island but occurred all through the heap; so that it is to be thought of as having lived on the Maine coast for some hundreds of years. None of the skeletons were found associated, nor were any of the skulls perfect. In every case the mink had served as food for the aboriginal campers, so that the carcass had been pulled to pieces and the bones thrown away in various direc- tions. Every skull has the brain case broken and lost, the brain having apparently been used for food. The facial por- tion of each skull is, however, pretty much intact, indicating that the meat was simply picked off it. Many of the lower jaws are marked with tool scratches (see fig. 2) apparently made while removing the meat from the bones. This form is much larger than any of the living New England species, being all of 25 per cent larger than Lutreola (weson) lutreocephalus* Harlan, the large brown mink, and *N. American Fauna, Dept. Agriculture, No. 19, 1900, p. 42. + Bangs, North American Minks, Boston Soc. Nat. Hist., Proceedings, vol. xxvii, 1897, p. 1-6. ‘ 228 Loomis—New Mink from the Shell Heaps of Maine. equal in size to the Alaskan Z. vison ingens Osgood. This new form may be described as follows: Fie. 1. Fic. 1. A, the upper dental series of the type specimen, nat. size. B, the skull seen from above, nat. size. Lutreola vison antiquus sp. nov. The type is a male skull lacking the brain case together with an associated right lower jaw, numbered M 1401 in the Fie. 2. Fie. 2. 6 lower jaw, nat. size; a is scratch made by aboriginal tool in removing the flesh. ¢? lower jaw of female, showing relative size. Amherst Collections: and also a right lower jaw of a female numbered M 1402, both from Flage Island shell heap. Loomis — New Mink from the Shell Heaps of Maine. 229 The form is large and heavily built, the skull with a low sagittal crest and short wide postorbital processes. The frontal region is slightly arched between the orbits. The teeth are those typical of the genus but rather stouter and heavier than usual. The inner tubercle of the upper carnassial is single and rather small. With this larger and typical form occur numerous individuals about 20 per cent smaller, but otherwise with the same char- acteristics, which I take to be the females, as there is in the family usually about this difference in size between the sexes. The following measurements give the data for comparisons: CG uppermental’ series). 22. 8222 2 es 99) 1 /2mm fe) T3 75 Oi A na he eR LE Warts ee SON Bs 95mm & lower dental series...-._-..--...-. -. 30™™ Q “ “cc CA Awe ao ene Oe mmM eG uuppen Catnassiple:t of paar tek es 8 1/47™ fe) “e ce ey gcie swe MUTE AS RIOR aS eR aey ee! 7 3/4m™ 2 width between the orbits.._.____._-- _ 20mm For other measurements see figures, which are drawn to scale. ‘Amherst, Mass. 230 Scientific Intelligence. SCIENTIFIC INTELLIGENCE, I. Cnemisrry Ann Purysics. 1. Mesothoriwn.—The existence of this radio-active element as one of the products of thorium was established in 1907 by Hahn, who showed that its half decomposition period was about 54 years, and that it was separated from thorium in the commercial extraction of the latter from its ores. Later, Hahn was able to find this substance in the residues from the extraction just men- tioned, and he showed that it is not directly transformed into radiothorium, but that there is an intermediate product with a half-period of 6-9 hours, which he called mesothorium II. Hahn has recently been able to concentrate this last substance to such an extent that the radio-activity of the product is several times greater than that of pure radium salts. W. Marcxwatp bas now made some interesting observations in regard to these substances, for it appears that nothing has been published concerning the chemical properties of mesothorium I. He had occasion to examine a “radium preparation” which had been manufactured from the residues of uranium and thorium ores. This preparation, consisting chiefly of barium chloride, gave a y-ray radiation corresponding to more than 1 per cent of radium chloride, but the radium emanation obtained from it corresponded to only about 0°2 per cent of radium. A further study of the product showed that about 80 per cent of the y-radi- ation came from mesothorium II, for when an aqueous solution of the salt was treated with a trace of ferric chloride and made ammoniacal the mesothorium II was precipitated with the ferric hydroxide. The precipitate gave a strong y-radiation, while the barium chloride recovered from the filtrate by evaporation had lost almost the whole of this radiation. However, while the pre- cipitate lost its activity at the rate of one-half in about 6 hours, the salt regained the greater part of its activity within a day. The precipitate by ammonia must have contained also the radiothorium produced by the mesothorium, and in fact this was found to be the case, as the precipitate gave the a-rays of the thorium emanation after the disappearance of the mesothorium II. Mesothorium is evidently entirely analogous to radium in its chemical behavior, for Marckwald has been unable to find any means of separating the two. This is interesting in connection with the fact that no chemical method is known for separating the four elements, thorium, radiothorium, ionium and uranium X, and it appears that radium and mesothorium form a similar group, which possibly may contain other members also. It is evident that radium preparations are liable to be contaminated with meso- thorium, and since radium has a period of existence about 300 times as long as the other, this contamination is of much import- Chemistry and Physies. 231 ance. There is a possibility, not only of accidental contamina- tion, but of wilful adulteration. The best means for testing radium preparations for mesothorium is to remove the radium emanation either by heating or by solution and evaporation ; then the presence of y-rays after a few hours shows the presence of mesothorium. The proportion of y-rays before and after this treatment gives an indication of the amounts of the two radio- active substances present. — Berichte, xliii, 3420. H. L. W. 2. The Combustion of Hydrocarbons.—In a recent lecture before the British Association, W. A. Bonz has given a review of the present knowledge of gaseous combustion, much of which is due to his own important researches. The opinion which formerly prevailed among chemists that in combustion the hydro- gen of hydrocarbons is first attacked by oxygen with the forma- tion of steam is incorrect. It has been known for a long time that when ethylene and acetylene are exploded with equal vol- umes of oxygen, carbon monoxide and hydrogen are practically | the only products, as follows : C,H, +0, = 2C0+2H, C,H, +0, =2CO+ H Bone has shown that the oxidation of the hydrocarbons at com- paratively low temperatures proceeds by addition of oxygen to the molecule and the successive formation of hydroxyl prod- ucts—alcohols, aldehydes, formic acid and finally carbonic acid. It appears that combustion at higher temperatures goes on in the same way, and it has been found that oxygen has a much greater affinity for the hydrocarbons than for hydrogen and car- bon monoxide. For example, when detonating gas is exploded with acetylene in the proportion O,H,+2H,+0O,, there is abso- lutely no separation of carbon nor formation of steam, and practi- cally the same thing holds good in the case of a mixture of ethylene, hydrogen and oxygen corresponding to C,H, +H,+0.,. In the presence of a hydrocarbon, carbon monoxide is attacked by oxygen even less readily than is hydrogen. These observations have an important bearing on the chemistry of flames. Hitherto hydrogen has been considered as one of the most combustible of gases, but in reality it is very much less so than the hydrocarbons. It is probably not so much th eoriginal hydrocarbon as its hydrox- ylated molecule which decomposes in ordinary flames, and experi- mental evidence does not warrant the view, so often encountered in scientific literature, that hydrocarbons are resolved into their elements prior to being burnt.— Chem. News, cii, 809. Hu. L. w. 3. Supposed Chemical Distinction between Orthoclase and Microcline.—Two ov three years ago the view was advanced by Barbier that orthoclase differs from microcline in the fact that the former contains traces of lithium and rubidium, while these elements are not found in the latter. It appears, however, that Ramage had previously found these alkali metals in a microcline from Dalkley in Ireland, and that Vernadsky, somewhat later, 2 232 Scientific Intelligence. found both rubidium and cesium in the bluish-green microcline from Miask in the IImen mountains. In order to examine the matter more thoroughly, Virnapsky and Revoutsky have now examined a number of samples of microcline spectroscopically, and have found the following elements present: Miask, Iussian, K, Na, Rb, Li. Arendal, Norway, Ba, K, Na, Rb. Pike’s Peak, Colorado, K, Na, Rb, Li, Cs. Hunttila, Finland, K, Na, Ba, Rb, Li. Lojo, Finland, K, Na, Ca, Li (Rb?). These results were obtained by heating fragments of the mineral directly before the gas-oxygen blowpipe and observing~ the spectrum of the flame. The results show that Barbier’s view is evidently incorrect.— Comptes Rendus, cli, 1372. H. L. W. 4, Preparation of Argon.—G. CLauDE has found a convenient way to prepare large quantities of argon. As a starting point he uses the oxygen produced by the liquefaction of air, which may be obtained 95 per cent pure, and with argon as its prin- cipal impurity, amounting to about 3 per cent. This oxygen is, therefore, about three times richer in argon than ordinary air. The oxygen is absorbed by hot copper, and the nitrogen by hot magnesium, only a small amount of the latter being required. _A tube of hot copper oxide serves finally to oxidize any hydro- gen that may have been formed from moisture present in the materials employed.— Comptes Rendus, cli, 752. H. L. W. 5. Die Stellung der neueren Physik zur mechanischen Natur- anschauung ; von Dr. Max Piano. Pp. 33. Leipzig, 1910 (8. Hirzel).—This pamphlet contains a lecture given before the eighty-second meeting of the German Scientists and Physicians, held in Koénigsberg, last September. It deals with the Theory of Relativity and with the philosophical views to which it leads in the minds of many German mathematicians and physicists. To the Anglo-Saxon mind, these views appear to touch the limits of philosophical idealism. Ether and matter, in fact all substance, is apparently discarded, and the physical universe consists of a vacuum mitigated by the presence of the Principle of Least Action, and Maxwell’s Equations; the “building stones,” of which the physical world is constructed, are no longer material particles but the so-called universal constants: the velocity of light, the charge and mass of the electron, the “elementare Wir- kungsquantum, ” etc. H. A. B. 6. History of the Cavendish Laboratory, 1871-1910. Pp. x, 842. London, 1910 (Longmans, Green & Co.).—This volume has been prepared in commemoration of the twenty-fifth anniver- sary of Sir J. J. Thomson’s election to the Cavendish Professor- ship of Physics in the University of Cambridge. It is the work of several different authors, each period of the history of the laboratory being treated by one who was intimately connected with it during the time in question. Messrs. Fitzpatrick and Chemistry and Physics. 233 Whetham deal with the building of the laboratory ; Professor Schuster, with the Clerk Maxwell period ; Mr. Glazebrook, with the years during which Lord Rayleigh was professor ; and Sir J. J. Thomson, himself, gives a general survey of the past twenty- five years. In chapters V to VIII the activities of the labora- tory during these twenty-five years, are discussed in greater detail. Professor Newall gives an account of the researches conducted between 1885 and 1894 : Professor Rutherford recounts the memorable achievements which marked the years 1895-1898 : C. T. R. Wilson deals with the period from 1899 to 1902, and N. R. Campbell completes the record to 1909. Chapter IX by Professor Wilberforce of the University of Liverpool deals with the development of the teaching of physics in Cambridge, a development which has had great influence upon physics teach- ing throughout the world. The volume closes with a list of memoirs containing accounts of research performed in the Caven- dish Laboratory, and a list of those who have worked there. The volume cannot fail to be of great interest to all students of physies ; it is a valuable contribution to the recent history of the science, and a more appropriate way of celebrating Sir J. J. Thomson’s twenty-fifth anniversary could scarcely have been found. H. A. B. 7. The Principles and Methods of Geometrical Optics, espe- cially as applied to the theory of Optical Instruments . by Jamus P. C. Sourmany. Pp. xxiii, 626, with 169 figures. New York, 1910 (The Macmillan Company).—This is a notable book which surpasses all others in the English language treating of the same subjects. The very great number of propositions in geo- metrical optics are presented clearly, in a carefully studied notation, which is, except in a few cases where other consider- ations are of greater weight, consistent and lucid. The diagrams are sufficient in number and very clear, with the too rare quality of good taste in respect to all the details which determine the character of such illustrations. Most excellent features of the book are its bibliography and historical notes, which are very complete. The only striking omission observed is that of the admirably convenient—perhaps the most convenient of all—col- lection of formulas for rigid computation of the constants of a system of centered lenses by P. A. Hansen. These features make the volume invaluable to one who seeks a knowledge of what has been accomplished in this field during the three cen- turies in which the problems of geometrical optics have been continuously increasing in importance. When we come, however, to consider the utility of the methods and equations deduced in the text to the designer of optical apparatus, we must give more restricted praise, since they empha- size what is relatively unimportant in practice and thrust more or less into the background those features which are essential. Excellent examples in support of this assertion are afforded by the only numerical calculations in the book, namely, the calcula- 234 Scientific Intelligence. tions of some of the optical constants of a 12-inch Taylor tele- scope objective. The geometrical data are given to three-figure accuracy and are susceptible, perhaps, to a maximum of ten times this precision. The calculations of focal lengths are carried out to seven-figure accuracy, that is to some 1000 times the precision warranted by the data; it is three or four hundred thousand times as great precision as is deemed important by the designer himself, if we infer that he expected to make the ratio of focal length to aperture the standard for telescopes of this size. Never- theless, this incommensurate labor of calculations is necessary in order to deduce a sufficient value, to two figures only, of the spherical aberration. The other case is equally striking. The spherical aberration of the same objective is calculated to a two-figure precision, which is all that is of practical significance, by a tedious computation with seven-figure logarithms by application of Seidel’s analysis and with the disappointing error of one hundred per cent. These considerations are enough to show that there is some radical defect in a method which demands such efforts for such meager returns. Probably there is no hope of material improvement until the mathematician informs himself thoroughly as to the relative importance of the magnitudes which enter his analysis and then deals with the physical realities of wave surfaces and refracting surfaces instead of the unnecessary fictions of rays and of radii of lens surfaces. Cc. 8. H. 8. Chemische Krystallographie; von P.Grotu. Dritter Teil. Aliphatische und Hydroaromatische Kohlenstoff-verbindungen. Pp. iv, 804, mit 648. Text figuren. Leipzig, 1910 (Wilhelm Engelmann).—This monumental work on Chemical Crystallog- raphy begun in 1906 has now reached its third part, or as esti- mated, three-quarters of the whole. It is devoted to the aliphatic and hydroaromatic hydrocarbons. ‘The whole makes a work of 800 pages, with perhaps 1200 or more individual compounds, whose crystallographic and optical constants are given with great thoroughness. In a large number of cases the crystal form is illustrated by figures. The congratulations of those immediately interested are due to the veteran author for his success in carry- ing through a work of such magnitude and importance. II. Gerortoay anp Naturat History. 1. United States Geological Survey, Thirty-first Annual Re- port (1909-1910) of the Director, GrorcEe O. Suite. Pp. 131 with two plates. Washington, 1911.—This report contains a statement of the work done by the various divisions of the Survey during the fiscal year ending June 30, 1910. The pro- gress in land classification consisted in, first, the preparation of withdrawals covering power sites and coal, oil, gas, and phos- phate lands ; second, the classification of withdrawn lands and Geology and Natural History. 235 restoration of such as were found to be not underlaid by valuable deposits. The work involves also the performance of various executive and advisory functions connected with the classification and valuation of the public lands. The mining and technologic work of the Survey, which in the last few years has assumed large importance, was transferred on July 1, 1910, to the newly established Bureau of Mines (see v. xxx, pp. 292, 419). Thus another child of the Geological Survey, having grown to adult proportions and demonstrated its useful- ness, has been launched on an independent career. No small part of the great value which the Survey has been to the nation con- sists in the foresight, efficiency, and high scientific grade with which new branches of government work have been developed under the care of that organization. The increase in the corre- spondence of the Geological Survey and in the distribution of its publications is a measure of the increasing appreciation by the people of the work which is done. The correspondence increased more than 20 per cent over that of the previous year and the total number of reports and maps distributed has increased more than 13 per cent. The publications of the Survey during the year measure a part of its returns for the money expended. They consisted of 4 professional papers, 47 bulletins, 18 water supply papers, one volume on mineral resources, 6 geologic folios and 94 topographic maps. J. B. 2. Publications of the U. 8S. Geological Survey.—Recent publications of the U. 8. Geological Survey are noted in the fol- lowing list (continued from vol. xxx, p. 417). The thirty-first Annual Report of the Director is noticed above. Torocraruic ATLAS.—Seventy-three sheets. Fouro, No. 174. Johnstown Folio, Pennsylvania; by W. C. PuatEen. Pp. 15, with 1 columnar section, and 3 maps. Sacramento Folio, California; by W. LinpGren. Pp. 3, 4 maps. ProFreEssionAL Parser, No. 72. Denudation and Erosion in the Southern Appalachian Basin ; by Leontpas C. Gurenn. Pp. 137, 21 plates, 1 figure. Buxtietins.—No. 430. Contributions to Economic Geology (Short Papers and Preliminary Reports, in part earlier issued as separates.) 1909. Part I. Metals and Nonmetals except Fuels. C. W. Hayrs and Watpremar LinpereEn, geologists in charge. Pp. 653, 14 plates, 75 figures. No. 431—A. Advance Chapter from Contributions to Eco- nomic Geology, 1909. Petroleum and Natural Gas; by A. G. Lronarp, H. E. Grecory, C. W. Wasnpurne, and RosBertr ANDERSON. Pp. 83, 3 plates, 1 figure. No. 433. Geology and Mineral Resources of the Solomon and Casadepaga Quadrangles, Seward Peninsula, Alaska ; by Puivie S. Smiru. Pp. 234, 16 plates, 26 figures. No. 436. The Fauna of the Phosphate Beds of the Park City Formation in Idaho, Wyoming, and Utah; by Groner H. Girry. Pp. 82, 7 plates. 236 Scientific Intelligence. No. 440. Results of Triangulation and Primary Traverse for the years 1906, 1907, and 1908. R. B. Marsuarn, chief geog- rapher. Pp. 688, 1 plate. No. 441. Results of Spirit Leveling in Alabama, Georgia, North Carolina, South Carolina, and Tennessee, 1896 to 1909, inclusive. R. B. Marsuatt, chief geographer. Work done in cooperation with the State of Alabama during 1899 to 1905, inclusive ; with the State of North Carolina during 1896 and from 1902 to 1909, inclusive. Pp. 145. No. 442. Mineral Resources of Alaska. Report on Progress of Investigations in 1909; by Atrrep H. Brooke and others. Pp. 426, 7 plates, 8 figures. No. 470-A. Advance Chapter from Contributions to Eco- nomic Geology, 1910. Phosphates in Montana; by Hoyt S. GatE. Pp. 9, 2 figures. Warer Suppty Parers.—No. 254. The Underground Waters of North-Central Indiana ; by SrepHen R. Capps, with a chap- ter on The Chemical Character of the Water, by R. B. Doux. Pp. 279, 7 plates, 12 figures. Nos. 262, 264.—Surface Water Supply of the United States, 1909 [prepared under the direction of M. O. Lerentron]. No. 262. Part II, South Atlantic and Eastern Gulf of Mexico; by M. R. Hatt and R. H. Botster. Pp. 150, 5 plates. No. 264. Part IV, St. Lawrence River Basin ; by C. C. Covert, A. H. Horton and R. H. BotstErr. Pp. 130, 5 plates. 3. Bureau of Mines, Joseru A. Hormes, Director.—Four additional bulletins have been recently issued ; these are as fol- lows: Bulletin 2, North Dakota Lignite as a Fuel for power- plant Boilers ; by D. T. Ranpati and Henry Kreisincrer. Pp. 42, 1 plate, 7 figures. Bulletin 3, The Coke Industry of the United States as related tothe Foundry; by Richarp MoLtpENKE, Pp. 32. Bulletin 4, Features of Producer-Gas Power—Plant Development in Europe; by R. H. Fernarp. Pp. 27, 4 plates, 7 figures. Bulletin 5, Washing and Coking Tests of Coal, at the Fuel-Testing Plant, Denver, Colorado, July 1, 1908, to June 30, 1909 ; by A. W. Berpen, G. R. Detamarsr, J. W. Grovus, and K. M. Way. Pp. 62, 1 figure. Miners’ Circulars Nos. 1 and 2 have also just been issued ; they are the first of a series to be written in plain, non-technical language for the benefit of the miner. They contain the names of the per- missible explosives tested by the bureau at its Pittsburg station up to November 15, 1910, and gives precautions as to their use. 4. Florida State Geological Survey. Third Annual Report, 1909-1910, EK. H. Srtiarps, State Geologist. Pp. 397 with numerous plates and figures. Tallahassee, Fla., 1910.—This report, like the preceding of the series, is of high scientific as well as practical value. ‘The scientific interest lies in the unique character of the geologic province of Florida as com- pared with other portions of the United States and the way in which the subjects have been treated. The value to the citizens Geology and Natural History. 237 of Florida lies in the information which it contains on the mineral and water resources of the State. The volume contains, besides the administrative report and index, the following papers: A Preliminary Paper on the Florida Phosphate Deposits, by E. H. Sellards ; Some Florida Lakes and Lake Basins, by E. H. Sel- lards ; The Artesian Water Supply of Eastern Florida, by E. H. Sellards and Herman Gunter; A Preliminary Report on the Florida Peat Deposits, by Roland M. Harper. J.B: 5. The Badland Formations of the Black Hills Region ; by Crropnas C. O?Harra. 144 pp., 50 pls., 20 figs. South Dakota School of Mines, Bulletin No. 9, Department of Geology. Rapid City, South Dakota, November, 1910.—The “ badlands” of South Dakota form one of the most interesting physiographic sub- provinces in the world, and taken in connection with the Black Hills, forms a type area which in many respects is unique. While the structure and stratigraphy are not complicated, yet the details are so important as to fully justify the prominent place given to this area in the literature. Heretofore students have had to search through widely scattered technical reports in order to obtain information regarding the origin and topographic develop- ment of the badlands, as well as of the large and interesting col- lection of fossils. Thanks to Professor O’Harra, we now have a single volume accessible to students and amateurs wishing to become acquainted with this country,—a volume which does not require advanced scientific training to understand. From an educational standpoint the publication is, therefore, abundantly justified in spite of the absence of essential facts and interpreta- tions new to science. 6. West Virginia Geological Survey, I. C. Wuirr, State Geologist. Bulletin 2. Pp. 358. Morgantown, 1911.—Follow- ing Bulletin No. 1, which gives a bibliography of the state, the West Virginia Survey has now issued a volume containing tables of levels and distances, and also coal and coke analyses. The levels are compiled from records of the State and Federal Sur- veys, and have been supplemented by data collected by the vari- ous West Virginia railways. The analytical tables contain results of tests of coal made from all the economic horizons of importance,—namely, the Pottsville, Kanawha, Allegheny, Cone- maugh, Monongahela, and Dunkard series. The estimated coal production for 1910 is 65,000,000 short tons. H. E. G. 7. New Zealand Geological Survey, J. M. Bell, Director. Bulletin No. 9 (New Series), The Geology of the Whatatutu Subdivision, Raukumara Division, Poverty Bay ; by Jamus Henry Apams. 1910. Pp. ili and 46, 3 ills., 5 maps.—In age the rocks included within this subdivision are upper Miocene and consist of shales, argillites, sandstones, with coarse sandstones and conglomerates in the upper portion. Pumiceous deposits, possibly of Pliocene age, also occur. An interesting problem is presented by the fact that many of the igneous pebbles in the conglomerate are unlike any rocks thus far discovered in the Am. Jour. Sci.—FourtTH SERIES, VoL. XX XI, No. 183.—Marcu, 1911. 17 238 Scientific Intelligence. north island of New Zealand. The fossils collected in this area have been studied by Professor Marshall of Otago University. He finds that out of a total of forty-four species of mollusca recent species number twelve, and the conclusion is reached that the strata are. Upper Miocene in age rather than partly Cretace- ous, aS previously assumed. In the Whatatutu area the terraces developed along the streams at 200 and 400 feet give a clue as to the amount of elevation since the end of Miocene time. A much dissected coastal plain at an elevation of 3000 feet is indicated by the structure and attitude of Tutamoe ridge. Owing to the economic importance of this field the structure has been studied in detail. It is found that the rocks have been folded into broad anticlines and that accompanying the folds are faults of slight dis- location. The existence of fourteen oil seeps attracted attention to the Waitangi hill as early as 1874, but later developments have not led to discoveries of oil or gas in quantities sufficiently large to be of commercial importance. Bulletin No, 10 (New Series), The Geology of the Thames Subdivisions, Hauraki, Auckland ; by Coun FRAsER. 1910. Pp. ii and 129, 9 ills., 19 maps and sections.—The Thames section in the northern island of New Zealand was brought into promi- nence by the discovery of gold in 1865. By 1871 the production had reached about $5,940,000, and this largely from one bonanza. The production at the present time is below the $500,000 mark, and the main hope is in exploiting lower levels. The oldest rocks of the area, the Tokatea Hill series, consist of argillites and graywackes of pre-Jurassic age. The Manaia Hill series (Jurassic) overlies uncomformably the older terranes. A long interval, dur- ing which folding and faulting occurred and a submarine topog- raphy was developed, elapsed between the Jurassic and the Eocene. Three periods of volcanic eruptions are revealed by an examina- tion of the Tertiary strata. The upper Eocene and Miocene volcanics consist of andesitic and dacitic tuffs, breccias, con- glomerates and lavas, while the Pliocene eruptions were rhyolitic in nature. Large and small folds have been observed in the dis- trict, and have been found to be of direct economic importance. The great Moanataiari fault with a down-thrust of 595 feet is represented topographically by a partially dissected fault scarp. Pages 50-115 of this report are devoted to detailed descriptions of existing mines and mining areas. ; H. E. G. 8. Geological Survey of Western Australia, Bulletin No. 33, Geological Investigations in parts of the Gascoyne, Ashbur- ton and West Pilbara Goldfields ; by A. Gipp Marrianp. 1909. Pp. 77, 13 maps and 65 figures.—The area covered by this report is the extreme western portion of Australia including the coast line from Port Hedland to the mouth of the Wooramel River. The geological sketch-map shows the Gascoyne beds (Carbonifer- ous) well developed in the lower Gascoyne River ; the Bangemall beds (Nullagine ?) from Frederick River to Mount Flora; and the Ashburton beds (age undetermined), chiefly in the neighbor- hood of Ashburton and Hardey Rivers. There are also small Geology and Natural History. 239 areas of pre-Carboniferous granite and gneiss. The relations and character of the formations near the coast line are revealed by a 3011 foot well at Carnarvon, which shows 1211 feet of Mesozoic, 1650 feet of Carboniferons. ‘The Carboniferous section in the Arthur River valley consists of (1) grits and fine conglomerate, (2) fossiliferous limestone, (3) limestone conglomerate, (4) gla- cial bowlder bed, (5) sandy and flaggy limestones. This glacial bowlder bed is well exposed at a number of localities and in the Wyndham River valley contains Spirifera, Productus, Polyzoa, and Aviculopecten. The Carboniferous series as a whole rests upon metamorphosed sedimentary rocks of unknown age. Owing to the economic importance of this part of West Australia some- what detailed studies of structural relations were made. In the vicinity of Bangemall slates, limestones, quartzites and diabase are arranged in a denuded anticlinal fold and are intersected by numerous quartz reefs. Mount Augustus, one of the most con- spicuous scenic features of West Australia, was found to be a sharp monoclinal fold of schist and conglomerate. Both normal and thrust faulting are revealed at Coorabooka Gap, and the position of this gap as well as the arrangement of the drainage lines suggests interesting physiographic studies. In the Uaroo copper district of Ashburton the rocks are sedimentaries of unknown age and have undergone deformation since mineraliza- tion. A chapter on petrography by J. Allan Thomas contains a dis- cussion of dolomite and cherts, pyroxenites and amphibolites, together with conclusions regarding magmatic sequence. Sixty- nine slides are described in detail accompanied by a list of eighteen analyses. While this bulletin is chiefly devoted to economic studies, it adds considerable to the meager information regarding the ceology of this interesting country. Bulletin 38, The Irwin River Coalfield ; by W. D. Came- BELL, 1910. Pp. 101, 7 plates and 53 figures.—The pre-Car- boniferous rocks of the Irwin River district are gneisses and granites “traversed by dikes of diabase, basalt, and norite, and lodes of lead and copper in addition to quartz veins.” This crystalline complex was greatly eroded before the deposition of extensive beds of quartz conglomerates and submarine tuffs of pre-Carboniferous age. The Carboniferous rocks are fossiliferous and consist of clays, shales, sandstones and limestones. One important stratum is the glacial bewlder bed which has been recognized at several localities in Western Australia. In the area under discussion this bowlder bed is found within the Carbonifer- ous and not at its base. Jurassic strata, 300 feet in thickness and containing lignite, rest upon the denuded surface of the Car- boniferous. ‘Tertiary limestones and sandstones seems to overlie unconformably the Jurassic rocks of the Hutt River district. H. E. G. 9. Paleontological Contributions to the Geology of Western Australia. Geol. Surv. Western Australia, Bull. 36, Pt. III, pp. 133 and 12 pls. 1910.—A series of eight papers is here 240 Scientific Intelligence. included, as follows :—(1) Hinde on isolated sponge spicules that are “newer than the Cretaceous”; (2) Arber on some Juras- sic plants; (3) Etheridge on 19 Oolitic invertebrates ; (4) Glau- bert on a fossil cave marsupial, Sthenwrus occidentalis, (5) on a list of West Australian pre-Tertiary fossils known to the end of 1908, (6) on Paleozoic fossil plants, (7) on peel fossils, and (8), on Cretaceous chalk and fossils. c. 8. 10. Report of the Vermont State Geologist for 1909-1910 ; by Grorer H. Perkins. Pp. xii, 361, pls. 71. 1910 [Jan. 1911]. —The volume opens with an account of the History and Condi- tions of the State Cabinet, by the State Geologist. The granites of the state are described by T. Nelson Dale, in an ar ticle which is practically a reprint of a bulletin of the U. 8. Geological Sur- vey. C. H. Hitchcock has a chapter on the Surfacial Geology of the Champlain Basin and Percy E. Raymond brings together all that is known about the trilobites of the Chazy formation in Vermont. The latter comprise 36 species, all of which are illus- trated. Professor Perkins describes the geology of the Burling- ton Quadrangle and Professor Seely has a preliminary report on the Geology of Addison County. Asbestos in Vermont is treated by C. H. Richardson and the mineral resources by the State Geologist. oS} 11. A Contribution to the Geologic History of the Floridian Plateau ; by Tuomas Waytanp Vaucuan. Carnegie Institution of Washington, Publication 133, pp. 99-185, 15 pls., 6 text figs. :910.—This well written and very interesting work should be studied by all stratigraphers and geologists because here we have worked out with care the present conditions of deposition and geologic work now going on in the Floridian region as a basis toward a proper interpretation of the Tertiary history of the peninsula. The author first describes the topography of the Floridian Plateau and then goes into considerable detail in regard to the marine bottom deposits forming in the bays and sounds behind the keys. Limestones are here being made by precipita- tion from the sea water as amorphous calcium carbonate and are apparently not of detrital origin. It is a soft ooze into which a rod can be forced down ten feet or more ; in fact, the depth of this soft material has not been determined. Vaughan then discusses the transporting agents (currents and winds) of the Florida coast and their effects. The smaller half of the work treats of the geologic history of the Floridian Pla- teau. The history is worked out in some detail and the book is abundantly illustrated by maps, one of which presents the geolo- gic formations of the state. There are also many photogravures of vegetation, sea shores, and geological deposits. ©. Ss. 12. Recent Discoveries Bearing on the Antiquity of Man in Europe ; by Gnorck Grant MacCurpy. Smithsonian Report for 1909, pages 531-583, pls. 1-18. 1910.—The author brings together here the accounts of the many wonderful discoveries that have been made in the past ten years bearing upon the Geology and Natural History. 241 antiquity of man in Europe. It seems that the oldest undisputed flint implements of man go back to the Upper Miocene and the oldest bone, a jaw (Homo heidelbergensis), has been found near the base of the Quaternary. Man, as man, has lived, therefore, in western Europe at least throughout the entire Glacial period, developing into Homo primigenius, a stocky robust type, of low stature, and with relatively short arms and legs much as in the Eskimo. In the Upper Quaternary, or at least 30,000 years ago, there came into western Europe, probably from the East, a more intellectual race of men, the Aurignacians, and it is these people who sculptured, engraved, and frescoed the walls of the caverns and their tools and ornaments. Their descendants, the Magda- lenians, introduced the rudiments of writing, and this seemingly was more than 10,000 years ago, if we may judge by the time standards accepted by geologists for the duration of time since the last glacial climate. The negroid people also passed into western Europe, possibly by way of Gibraltar, probably soon after the arrival of the Aurignacians. It is also becoming more and more certain that man did not originate out of any of the existing ape stocks, but rather that the human stock is as old as any of the tailless primates. According to Professor Klaatsch, Homo primigenius is more closely related to the gorilla of Africa, while Homo aurignacensis bas closer affinities with the chimpanzee of Asia. All of these stocks had their origin in the far distant past, certainly not less than one million years ago. C.us! 13. On the Fossil Faunas of St. Helen’s Breccias ; by HENRY S. Wittrams. Trans. Royal Soc. Canada, III, pp. 205-246, pls. 1-4, 1910.—The Devonian faunas of St. Helen’s island near Mon- treal have long perplexed students of fossils as to the exact age of these fossil horizons when compared with similar formations in New York. Professor Adams of McGill University had quar- ried out from three isolated limestone masses underlying the agglomerates of the island about three-fourths of a ton of mate- rial which Williams has here subjected to a detailed study. The author, therefore, has had far greater advantages than any other paleontologist studying these early Devonian biotas. The older fauna of about 30 species is of Helderbergian age and apparently of about Becraft time. Williams names it the Gypidula pseudogaleata fauna. The age is clearly seen in the following species: Schizophoria multistriata, Stropheodonta planulata, Gypidula pseudogaleata, Camaroteechia ventricosa, Spirifer concinnus, and Meristella princeps. The faunal rela- tions are clearly with New York and nothing exactly like it is known farther northeast in the Gaspé region. The younger fauna is from another isolated limestone mass, and is the one furnishing much new information. Williams calls it the Spirifer arenosus fauna and ascribes to it 25 species. The more striking forms are: 1 Dalmanella subcarinata, 2 Hodevo- naria hudsonicus gaspensis, 3 Chonetes striatissimus (near C. can- 242 Scientific Intelligence. adensis), 4 Rhynchonellu eminens, 5 Eatonia peculiaris, 6 &. cf. whitfieldi, 7 Spirifer arenosus, 8 S. gaspensis, 9 S. montrealensis n. sp. (look almost like genuine S. granulosus), 10 S. cumber- landice, 1\ 8S. pennatus helene, 12 Metaplasia pyxidata, and 13 Cyrtina rostrata. To these must be added 14 Chonostrophia montrealensis described by Schuchert but not seen by Williams. In the light of our American Devonian assemblages we see here a very much mixed fauna. Numbers 1 and 4 are Helderbergian forms, while 9 and 11 are decided early Hamilton reminders. The remainder of the fauna suggests later Oriskanian. In the Oriskany of Cumberland, Md., the reviewer has also collected shells of the type of 8S. montrealensis, but these are by no means so near the Hamilton S. granulosus as seemingly are the~St. Helen’s specimens. Further, the reviewer, while collecting on the island in 1900, noted that the two species 9 and 11 occurred together, but he then saw no other forms in the “flat block of limestone” in the agglomerate. For this reason he held their age to be Onondaga. It was Mr. Ardley who directed him to these fossils and associations and Williams’ S. arenosus fauna of 25 species has since been quarried out of this same block. The majority of the fauna is undoubtedly Oriskanian and yet the aspect is more recent than any fauna of this series in the Appa- lachian—New York area. Williams clearly recognizes that the St. Helen’s Spirifer are- nosus fauna is unique and believes it to be somewhat younger than any Oriskanian fauna of New York but older than the Onondaga. Faunas of the same age as that of St. Helen’s but of another basin, linking more directly with the European Coblen- zian, he holds are those of Nictaux, Nova Scotia, York river, Gaspé, and Moose river, Maine, in which “ are seen traces of the Hamiltonian magnafauna.” He further holds that the Onondaga fauna came in along the western side of the Cincin- nati axis, finally spreading to the St. Lawrence valley and there met and mixed with the northern Atlantic fauna “ on the Ameri- can border at the time of the departure of the Oriskanian ele- ment rather than at the opening of the Hamilton epoch. This interpretation is in harmony with the mingling of these same two magnafaunas [lower Devonian and Hamilton] in the lower Devonian (Coblenzian) of Europe.” The reviewer agrees with Williams that the Oriskanian faunas of the maritime “province of eastern Canada are considerably Coblenzian in faunal aspect and that the Hamilton aspect appears earlier in this European assemblage, but he still believes that the York river fauna near the base of the Gaspé sandstone as described by Clarke (1908) is considerably younger than the Spirifer arenosus fauna, for the reason that the latter assemblage at Gaspé occurs at a very much lower horizon, in fact, at the base of the Grande Gréve limestone. C. SCHUCHERT. 14. Paleontologia Universalis, ser. JI, fasc. II, 46 sheets, July 26, 1910.—In this new part of the Palzontologia Univer- Geology and Natural History. 243 salis there are redescribed and well figured 20 Lamarckian species of mollusks and corals described by him between 1801-1819. The studies were made by Dollfus, Boussac, Pervinquiére, Cossmann, Lemoine, and Germain. CS: 15. Hine Botanische Tropenreise, Indo-Malayische Vegetations- bilder und Reiseskizzen ; by Prof. Dr.G. Hasmrianpr. Second edition. Pp. viii, 296; 12 plates and 48 text-figures. Leipzig, 1910 (Wilhelm Engelmann).—The first edition of Professor Haber- landt’s book appeared in 1893 and soon became widely and favora- bly known on account of its graphic and satisfactory descriptions of various types of tropical vegetation. The work is based on the autbor’s personal observations, most of which were made during a visit to the famous botanical garden at Buitenzorg in Java. Among the many interesting chapters those dealing with tropical trees, tropical leaves, vines, epiphytes, and mangroves should per- haps be especially mentioned, although several of the others treat subjects of equal importance. The twelve plates in the second edition are all new; nine are made from photographs, while the three others, in color, are reproduced from water-color sketches by the author. A. W. E. 16. Plant Anatomy, from the Standpoint of the Development of the Tissues, and Handbook of Micro-technic; by Wittiam Cuase Srevens, Professor of Botany in the University of Kansas. Second edition. Pp. xv, 379, with 152 text-figures. Philadelphia, 1910 (P. Blakiston’s Son & Co.).—The first edition of this excellent work appeared in 1907, and was reviewed in this Journal for April, 1908 (xxv, 363). The most important new matter in the second edition is the, chapter on reproduction, which includes discussions of the following topics : the reduction of chromosomes, the behavior of hybrids interpreted according to Mendel’s Laws, the bearers of hereditary characters, and the theory of pangeneic exchange. A. W. E. 17. A Text-Book of Botany and Pharmacognosy ; by Henry KrarmMeEr, Ph.D., Professor of Botany and Pharmacognosy in the Philadelphia College of Pharmacy. Fourth edition. Pp. viii, 888, with 344 figures, mostly in the text. Philadelphia and Lon- don, 1910 (J. B. Lippincott Company, price $5.00 net).—The first edition of the present text-book appeared in 1902 and contained 384 pages ; the second edition, of 1907, had already been enlarged to 840 pages; while the third edition, of 1908, numbered 850 pages. The rapid succession of new editions proves conclusively that there is a strong demand for a work of this character by students of pharmacognosy and that the book in question is well fitted to their needs. In the first part, entitled ‘‘ Botany,” the morphology and classification of plants are clearly treated, with special reference to medicinal plants. In the second part, “ Phar- macognosy,” detailed descriptions of important drugs are given, their minute structure being fully illustrated by figures. The third and fourth parts are much shorter than the others. The third deals with “Reagents and Microtechnic,” and the fourth, 244 Scientific Intelligence. which is new to this edition, discusses “ Micro-Analysis.” The eighteen figures illustrating the fourth part are reproduced from microphotographs of erystals, A.W. EB. 18. Biology: general and medical; by Joseru McFaruann, M.D. Pp. 440, with 160 illustrations, Philadelphia and Lon- don, 1910 (W. B. Saunders Company).—This book differs widely from most of the other elementary text-books in biology, which have recently appeared, in subordinating the morphological almost entirely to the physiological aspects of the subject. It is essentially a treatise on general physiology, with such descrip- tions of the anatomical structures as are absolutely necessary for the understanding of the processes concerned. For elementary courses in colleges and universities where large numbers of-stu- dents elect biology as a general culture study, and where the laboratory work is necessarily confined mainly to the morpholog- ical side of the subject, the book forms an admirable supplement to the laboratory and lecture portions of the course. The immediate adoption of this book by some of our largest universities shows the need that has been felt for a work of this kind. There are, however, certain defects which appear when the book is subjected to the test of the classroom. Numerous instances of statements that are misleading or actually erroneous are brought to light, and complaint is made that an unnecessarily formidable array of technical medical terms is introduced. The general excellence of the plan of treatment, however, more than compensates for such emendations as the experienced teacher is ~ required to make in the classroom. The properties of living matter, cells, and their arrangement in different groups of organisms, reproduction, ontogenesis, con- formity to type, divergence, structural and blood relationships, parasitism, infection and immunity, mutilation and regeneration, grafting, senescence, decadence and death, indicate the subjects of the principal chapters into which the book is divided. W. Rs,.0. Ill. Miscernanrous Sorentiric InreLLicENoE. 1. Carnegie Institution of Washington. Year-Book, No. 9. 1910. Pp. xvi, 258, 5 plates. Washington, January, 1911.— Especial interest is connected with the appearance of the ninth Year-Book of the Carnegie Institution because of the recent gift © by Mr. Carnegie of an additional $10,000,000 to the Institution, making its total fund equal to $25,000,000. This addition to its resources is particularly opportune at this time, since in the pres- ent volume Dr. Woodward calls attention to the serious effect of increase of prices as limiting the future income available for promoting research. The Institution was organized in 1902 and since that time the magnitude and importance of the work it has accomplished are truly remarkable. The total amount of money Miscellaneous Intelligence. 245 expended up to date is $4,590,000, of which a little more than one-half has been applied directly to the prosecution of research, and about one-third is represented in land, buildings, and other permanent forms; about 8 per cent has been used for expenses of administration and somewhat less for publications. Twelve hundred individuals have contributed towards the researches and publications undertaken by it. The volumes already published are 167 in number, and aggregate more than 40,000 printed pages. Twenty-five additional volumes are now in press : further, some 1200 shorter papers have been contributed to current scientific periodicals by those working under the Carnegie foundation. Of particular importance in the work of the past year is the occupation of the new administration building, which was dedi- cated in December, 1909, and has proved in all respects a thor- oughly satisfactory and dignified permanent home for the Institution. During 1910, also, the non-magnetic ship Carnegie completed its first voyage of 8,000 miles with important results, and a second cruise, planned to last three years, was begun on June 29th : at present the vessel is off the coast of Brazil. As is now generally known, there are ten departments, to the sup- port of which the income of the Institution is chiefly devoted, the total sum appropriated towards them amounting to $450,000. A considerable number of minor grants have been made in addi- tion, although these are few as compared with the situation earlier in the history of the Institution. For these last, the ageregate amount allotted was about $70,000. In the opening pages of the present volume, Dr. Woodward gives a very interest- ing résumé of the investigations of the present year, particularly in connection with the ten lines of work already alluded to. This same subject is discussed in detail on pages 53-204 by the Directors of the different departments. It is impossible here to go into details in regard to these special lines. Some of the most interesting concern the work of the Geophysical Labora- tory, under Dr. A. L, Day; the Department of Marine Biology at Tortugas, Florida, under Dr. A. G. Mayer; and the Solar Observatory at Mt. Wilson, California, now represented by W.S. Adams, Acting Director during the absence of Professor Hale. Dr. Bauer also gives a summary of the work accomplished in ter- restrial magnetism, with a chart showing the projected cruise of the “ Carnegie ” alluded to above. The volume closes with brief statements, thirty-four in number, as to the results accomplished in the various lines of investigation represented by the minor grants. 2. Publications of the Carnegie Institution.—Recent publica- tions of the Carnegie Institution are noted in the following list (continued from vol. xxx, 295): No. 74. The Vulgate Version of the Arthurian Romances, edited from manuscripts in the British Museum; by H. Oskar Sommer. Volume III. Le Livre de Lancelot del Lac. Part I. Pp. 430. 246 Scientific Intelligence. No. 88. Dynamic Meteorology and Hydrography; by V. BserKNES and different collaborators. Part I, Statics by V. Bserknes and J. W. Sanpstrém. Pp. 146 and appendixes, 31 figures. No. 119. Determination of the Solar Parallax, from photo- graphs of Eros made with the Crossley Reflector of the Lick Observatory, University of California ; by Cuartes D. Perrine, Harotp K. Patmer, Freperic C. Moorr, Apretarwe M. Hose. Pp:98.- See ip. 153. No. 120. The Symmetric Function Tables of the Fifteenthic including an Historical Summary of Symmetric Functions as relat- ing to Symmetric Function Tables; by Froyp Fiskr Decker, Pp. 16, 5 large tables. No. 127. Superheated Steam in Locomotive Service; by WiturAm F. M. Goss. Pp. 144, 1 plate, 108 figures. i No. 130. A Study of the Absorption Spectra of Solutions of Certain Salts of Potassium, Cobalt, Nickel, Copper, Chromium, Erbium, Proseodymium, Neodymium, and Uranium as affected by Chemical Agents and by Temperature ; by Harry C. Jones and W. W. Srrone. Pp. ix, 159, 98 plates. No. 132. Department of Marine Biology, ALrrep G. Mayer, Director. Papers from the Tortugas Laboratory. Volume III, pp- 1-152, 17 plates, 38 figures. Contains twelve papers by dif- ferent authors. No. 132. Department of Marine Biology, ALFrep G. Mayer, Director. Papers from the Tortugas Laboratory. Volume IV, pp. 1-186, 43 plates, 17 figures. Contains three papers by Henry S. Pratt, Epwin Linton and T. W. Vaueuan. Pp. 185. No. 185. Researches upon the Atomic Weights of Cadmium, Manganese, Bromine, Lead, Arsenic, Iodine, Silver, Chromium, and Phosphorus ; by Grecory Pav Baxter, in collaboration with M. A. Hines, H. L. Frevert, et al. No. 136. Metabolism in Diabetes Mellitus ; by Francis G. Benepict and Erxriotr P. Josuin. Pp. vi, 234. No. 141. The Water Balance of Succulent Plants; by D. T. Macpoueat and E. 8. Spatpine. Pp. iii, 77, 8 plates. 3. Annual Report of the Board of Regents of the Smith- sonian Institution, showing the operations, expenditures and condition of the Institution for the year ending June 30, 1909. Pp. x, 751, 73 plates, 1 map. Washington, 1910.—The Annual Volume of the Smithsonian Institution for 1909 opens with the Report of the Secretary, Dr. Walcott, issued in advance about a vear since, and at that time noticed in this Journal (vol. xxix, 196). It also contains, in the General Appendix (pp. 119-751), the usual series of well-selected papers devoted to a wide range of subjects, many of them republished from foreign journals. These papers are all more or less popular in method of presentation, so that they appeal to the intelligent public, which finds here a remarkable résumé of recent scientific progress not to be found in so con- venient a form elsewhere. Among them may be mentioned one on Radio-telegraphy by J. A. Fleming ; another by Marchis on Miscellaneous Intelligence. 247 the production of Low Temperatures and Refrigeration ; on the return of Halley’s Comet, by W. W.Campbell ; on the British Antarctic Expedition of 1909, by Lieut. Shackleton; on the Antiquity of Man in Europe, by G. G. MacCurdy ; on Panama and its People, by Eleanor Y. Bell; on the Natural Resistance to Disease, by Simon Flexner. The following are recent Bulletins issued by the Bureau of Ethnology of the Smithsonian Institution: No. 30. Handbook of American Indians North of Mexico ; edited by Freprrick Wess Hopner. In two parts. Part 2, N-Z. Pp. iv, 1221. See vol. xxiv, p. 91. No. 37. Antiquities of Central and Southeastern Missouri, by GERARD Fowxer. (Report on Explorations made in 1906-07 under the auspices of the Archzological Institute of America.) Pp. vii, 116, 19 plates, 20 figures. No. 45. Chippewa Music; by Frances Densmore. Pp. xix, 216, 12 plates, 8 figures. No. 49. List of Publications of the Bureau of American Ethnology with Index to Authors and Titles. Pp. 32. 4. Publications of the Allegheny Observatory of the Univer- sity of Pittsburgh ; edited by Frank ScuLesincEer.—The follow- ing have recently been issued : Vol. I, No. 23. The Orbits of the Spectroscopic Components of v Andromede; by Frank C. Jorpan. Pp. 191-201. Also title page and contents of volume I. Vol. If. No.1. A Description of the Mellon Spectrograph ; by Frank Scuiesincer. Pp. 1-12, two figs. No. 2. On the Relative Motions in 61 Cygni and similar Stars; by Frank ScHLESINGER and Dinsmore Atrer. Pp. 13-16. No. 3. The Orbits of the Spectroscopic Components of « Herculis ; by Ropurr H. Baxer. Pp. 17-23. No, 4. The Orbit of I H. Cassiopeiz ; by Roxserr H. Baker. Pp. 25-28. No. 5. The Orbit of 30 H. Urs Majoris; by Rosrerr H. Baker. No. 6. The Orbits of the Spectroscopic Components of 57 Cygni; by Roprerr H. Baker. No. 7. Further Observations of 9 Aquile ; by Roperr H. Baker. No. 8. The Orbit of 7 Andromede ; by Frank C. Jorpan. No. 9. The Hclipsing Variable « Herculis; by Frank ScuLesincEeR and Rosrerr H. Baker. Pp. 51-62. No. 10. The Spectrum and Orbit of o Persei ; by Frank C. Jorpan. Pp. 63-71. 5. Bref och Skrifvelser af och till Carl'von Linné. Part 1V. Pp. iv, 365. Stockholm, 1910.—The fourth part of the corre- spondence of Linnzeus (see vol. xxix, 200), published under the auspices of the Upsala University, contains a very interesting series of letters to and from Abraham Bick, dating from 1741 to 1755. 6. Seismological Society of America.—The Seismological society, of which Prof. J. C. Branner is president, has decided to issue a Bulletin, the first number of which is about to be issued ; it will be sent to all members. The dues of the society are $2.00 per year ; life membership, $25.00. 7. Das Electrokardiogramm des gesunden und Kranken Men- - schen ; von Prof. Dr. Frrmprica Kraus und Prof. Dr. Grore 248 Scientific Intelligence. Nicotar Pp, xxii, 322. Leipzig, 1910 (Veit & Co.),—This volume represents the first more elaborate attempt at a syste- matic presentation of the scientific basis and latest technic of the electrocardiographic method applied to the study of heart functions in animals and man. The authors have particularly emphasized the possibilities of the electrocardiogram as an aid in clinical diagnosis, and have furnished a review of the rapidly growing literature on the subject. Such pioneer work deserves commendatory mention and will assist many physiologists and clinicians in orienting themselves in the newer methods of research. L. B. M. 8. Plane Trigonometry ; by Epwarv R. Rosstrys, Senior Mathematical Master in the William Penn Charter School. Pp. 166. New York, 1910 (The American Book Company).—The only valid excuse for an addition to the multitude of text-books in ‘Trigonometry is that it be written by a teacher of Trigonome- try in order to minimize his labor of teaching by giving to his pupils his own methods in print instead of by dictation. Mr. Robbins has in this way recorded the economies that he has secured by 15 years’ experience. He seems to have systematized the work in elementary plane Trigonometry a little better than any of his predecessors, and thereby in so much diminished the labor of thinking for his pupils. w. B. 9, Shop Problems in Mathematies ; by W. E. BRECKENRIDGE, S. F. Mersereav, and ©. F, Moorn. Pp. 280. Boston (Ginn & Co.).—This volume is designed for students in trade schools. It discusses materials and machines for both wood and metal work, and outlines construction work of various kinds. Mathematics, including trigonometry, is also reviewed with particular refer- ence to usefulness in shop practice. A large number of problems drawn from practical work are given. D. A. K. 10. Ostwald’s Klassiker der Hxakten Wissenschaften. Leipzig, 1910 (Wilhelm Engelmann).—Recent volumes in this important series are the following: Nr. 176. Mikroskopische Unter- suchungen uber die Ubereinstimmung in der Struktur und dem Wachstume der Tiere und Pflanzen, von Ta. Scuwann. Heraus- gegeben von F. Htnsever. Pp. 242. Nr. 177. Untersuchungen tiber Gegenstiinde der hdéheren Geodisie ; von Cart Friepricu Gauss. Herausgegeben von J. Friscuaur. Pp, 111. Nr. 178. Physikalisch-chemische Abhandlungen; M. W. Lomonossows, 1741-1752. Aus dem Lateinischen und Russischen mit Anmerkungen herausgegeben von B. N. MenscnuTKin und Max Sreter. Pp. 61. OBITUARY. Sir Francis Garon, the veteran English explorer and con- tributor to many departments of science, died on January 17 at the age of seventy-nine years. His most important writings were those on Heredity, but his activities extended into a remarkable number of different fields involving the application of quantita- tive methods to science. Dr. M. Wiruerm Meyer, the German astronomer, died two months since, at Meran, at the age of fifty-eight years. VOL. XXXL ? APRIL, 1911. Established by BENJAMIN SILLIMAN in 1818. THE AMERICAN JOURNAL OF SCIENCE. Epitor: EDWARD S. DANA. ASSOCIATE EDITORS Proressors GEORGE L. GOODALE, JOHN TROWBRIDGE, W. G. FARLOW ann WM. M. DAVIS, or Camsringe, Proressors ADDISON E. VERRILL, HORACE L. WELLS, L. V. PIRSSON anp H. E. GREGORY, or New Haven, Proressork HENRY S. WILLIAMS, or Irwaca, Prorressorn JOSEPH S. AMES, or Bautimore, Si Mr. J. S. DILLER, or Wasuineron. RUREAU GF neem ant EF THNO! OGY Ak Rail e] ; Vd he ! VOL. XXXI—[W HOLE NUMBER, CLXXXI.] No. 184—APRIL, 1911. “ansonian Ins P / os ity b NEW HAVEN, CONNECTICUT. APR 29191] | ore National Muse ; a ‘J THE TUTTLE, MOREHOUSE & TAYLOR O©O., PRINTERS, 123 TEMPLE STREET. Published monthly, Six dollars per year, in advance. $6.40 to countries in the Postal Union ; $6.25 to Canada. Remittances should be made either by money orders, ___—scregistered letters, or bank checks (preferably on New York banks). as Of Especial Interest to Mineralogists. HIDDENITE FROM NORTH CAROLINA. It has been some years since this rare gem mineral was procurable at the mineral dealers; through very fortunate circumstances I procured a large lot of these crystals at a remarkably low price; they range in size from 14 to 84 in. of very good color and quality. No doubt many collectors will be glad to have the opportunity to procure a representative of this variety of Spodumene, with a deep emerald green color; they range in price from 50 cents to $2.00. KUNZITE FROM CALIFORNIA. I also received a large lot of Kunzite crystals, showing remarkable nat- ural etchings; I am. now in a position to furnish a series of these etched crystals which should be in every collection; they range in color from white to deep lilac; former lots of this beautiful gem crystal were beyond ~ the average price for regular collectors; the present prices are far below any material of this quality ever offered before; crystals range in size from 34 in. to 2144 in. long, from 75 cents to $5.00; will send a series of these crystals for selection to anyone. OTHER CALIFORNIA MINERALS. Can safely state that my stock of California minerals is the largest in the country, considering their quality; the following will suggest a few of the additions to my present stock; Stibiotantalite, which is at present extremely rare, both loose xls. and in matrix, prices ranging from $2.50- $15.00. Pink Beryl, good crystals, fair color, from $2.00-$12.00; Tourmalines, all colors, loose and in matrix, from $2.00-$25.00. Awaruite, a new lot of these interesting metallic pebbles, from the Smith River; their appearance is something like Platinum nuggets; price from 25 cents to $1.50; from the above locality I have also a fine lot of black and red Obsidian and brown: polished from 214 in. to 344 inches, prices from $1.50-$2.00. In addition to the above I also received quite a number of other minerals too numerous to mention, from this state. COLORADO. Recent shipments have brought a large lot of Amazonstone, in groups and loose crystals, single and twins, some of which are remarkable; also a number of the celebrated Cripple Creek Tellurides, such as Tellurium, Calaverite, Sylvanite, Gold Pseudo after Calaverite ; Calciovolborthite crystallized, Carnotite, Amethystsin parallel growths, Topaz, Smoky Quartz, Pyrites, Rhodochrosite quartz, with Fluorite and other well known minerals at remarkably low figures. I shall be pleased to send anyone on request an assortment, prepaid, for selection, and guarantee satisfaction. A. H. PETEREIT, 81—83 Fulton Street, New York City. Phone Beekman 1856. THE AMERICAN JOURNAL OF SCIENCE [FOURTH SERIES.] Art. XXIV.—On the Ionization of Different Gases by the Alpha Particles from Polonium and the Relatwe Amounts of Energy Required to Produce an lon; by T. 8S. Taytor. Introduction. Iy previous papers,* the writer has shown that the air-equiv- alents + of metal foils decrease with the speed of the alpha par- ticles entering the foils. For sheets of different metals of equal air-equivalents, the rates of decrease are approximately proportional to the square roots of the respective atomic weights. On the contrary, the air-equivalents of hydrogen sheets increase while the hydrogen-equivalents of air sheets decrease with the speed of the entering alpha particles, and at such a rate as to be in agreement with the square-root law ob- served for the decrease of the air-equivalents of the metal sheets. A comparison of the Bragg ionization curves, obtained in atmospheres of air and hydrogen, when the pressure of the air ~ was so reduced that the range of the alpha particles from polo- nium was the same in air as it was in hydrogen at atmospheric pressure, showed differences which are sufficient to account for the variations in the air-equivalents of the hydrogen sheets with the speed of the alpha particles. These differences be- tween the Bragg ionization curves in air and hydrogen sug- gested that some such differences might be found between the ionization curves obtained in other gases, and it was for the purpose of making a detailed comparison of the ionization * This Journal, vol. xxvi, pp. 169-179, Sept. 1908; ibid., vol. xxviii, pp. 307-372, Oct. 1909. Phil. Mag., vol. xviii, p. 604, Oct. 1909. + By air-equivalent is meant the amount by which the range of the alpha particle is cut down by its passage through the foil. Am. Jour. Sct.—FourtH SERIES, VoL, XXXI, No. 184.—Aprin, 1911. 18 250 Taylor--- Ionization of Different Gases by the curves obtained in different gases that the present experiments were begun. Continuation of Experiments. The apparatus used was the same as had been used in the previous experiments.* The sheet iron case, enclosing the apparatus proper, was replaced by a solid iron case which could be readily exhausted. Polonium was used as the source of rays and was placed in a brass cylinder of such dimensions that the rays emerging from the cylinder fell well within the limits of the ionization chamber for all available distances of the source of rays from the ionization chamber. In the determination of the ionization curve in any gas, the vessel enclosing the apparatus was first evacuated and then the gas admitted very slowly till the pressure it exerted was such that the range of the alpha particles was exactly 1171 centi- Fie. 1. weer" 0 / e 3 4 5 6 7 8 9 Jo Ww Fic. 1. The ordinates are the deflections in millimeters of the electrome- ter needle per second. The abscissas are the distances in centimeters of the polonium from the ionization chamber. Curves I, II, and III were obtained when the maximum range of the alpha particle was exactly 11°1 centime- ters in hydrogen, air, and methy] iodide, respectively. meters, which was the maximum range available with the apparatus. The Bragg ionization curve was then obtained in the usual manner by observing the deflection of the needle of the Dolezalek electrometer in scale divisions per second for * Loc. cit. Alpha Particles from Polonium. 251 various distances of the source of rays from the ionization chamber. In this manner, the Bragg ionization curves were obtained in the gases and vapors given in Table I. The curves Fie. 2. 0 / % 3 4 5 6 7 8 9 10 Fic. 2. The ordinates are the deflections in millimeters of the electrome- ter needle per second. The abscissas are the distances in centimeters of the polonium from the ionization chamber. Curves I, II, and III were obtained when the maximum range of the alpha particle was exactly 11:1 centimeters in methane, ethyl chloride, and carbon disulphide, respectively. in figures 1 and 2 and the dotted ones in figure 3 represent the ionization curves obtained in the above manner in the gases as indicated below the figures, respectively. The dotted por- tion of each curve in figures 1 and 2 is assumed to be the form it would take were it possible to move the polonium entirely up to the ionization chamber. At any rate, such assumed por- tions of the curves can differ but little from the actual curves. It is to be noted, that the ionization curves shown in figures 1 and 2 are plotted differently from the regular Bragg ioni- zation curve in that the values of ionization are taken as ordi- nates and distances of the source of rays from the chamber as abscissas, instead of vice versa as is usually done. Although the curves in figures 1,2, and 3 represent some differences from one another in regard to the relative amounts of ionization for corresponding distances of the source of rays from the ionization chamber, all of them are of the same general form. From a re-determination of the velocity of the alpha particle at different points in its path, and the assumption that 252 Taylor—Ionization of Different Gases by the the ionization produced at any point in the path of the particle is proportional to the energy consumed, Geiger * has shown that the ionization Z at any point in the path is given by the rela- tion ee ae (r—a) 4% where ¢c and 7 are constants and « is the distance from the source of rays. By comparing this theoretical ionization eurve with the experimental curve obtained in hydrogen for a pencil of rays, Geiger found the two to agree very closely: This theoretical curve has been compared with the experi- mental curves obtained in each of the gases and vapors given in Table I and a very close agreement between theoretical and experimental curves was found for each gas. To make this comparison, it was necessary to determine the constants 7 and ¢ for each gas. For the value of 7, Geiger used the aver- age range of the alpha particles in the pencil of rays. Since the maximum range of the alpha particles in the cone of rays used in the present experiments was always 11'1 centimeters, the average range of the alpha particles in this cone of rays emerging from the cylinder containing the polonium was slightly less than 11:1 centimeters. Consequently 10°8 centi- meters were taken as the value of the average range of the alpha particle, that is, 10-8 centimeters are supposed to repre- sent the average distance the alpha particles traveled in each gas before losing their power of producing ions. In order to determine c for any one gas, the ionization (ordinate of the ionization curve figures 1, 2, and 3) and the corresponding dis- tance x of the source of rays from the ionization chamber (abscissa of curve) were substituted in the equation zal c ~ (10°8—2x) 4 and the equation solved for c. Separate values of ¢ were thus obtained for various distances of the source of rays from the ionization chamber between «—0O and 9°5 centimeters, and the mean value of these separate determinations found for each gas. The mean values of ¢ as found in the above man- ner for all the gases and vapors used are recorded in column 2, Table I. * Proc. Royal Society, Series A, vol. Ixxxili, No. A 565, p. 505. Alpha Particles from Polonium. Fie. 3. 4 r¢) (0) / 2 3 4 5 Gly Meera tin Se, hy Ot Fie. 3. The full line curves I, II, and III are the theoretical ionization curves for nitrogen, sulphur dioxide, and ether, respectively, as obtained by substituting the corresponding values of ¢ given in column 2, Table I, in the equation = as “ where r = 10°8. r—x The dotted curves I, II, and III are the experimental ionization curves for nitrogen, sulphur dioxide, and ether, respectively, and are plotted similarly to the curves in figures 1 and 2 The full line curves, I, II, and III in figure 3 represent the theoretical curves for nitrogen, sulphur dioxide, and ether, respectively, as obtained by using the values of ¢ as recorded in column 2, Table I, for the respective gases. The dotted curves are the corresponding experimental curves and, as can be seen, agree very well with the theoretical curves. The agreement “petween the theoretical and the experimental eurves for the other gases was equally as good as it was for those given in fioure 3. In some cases the agreement was much closer. This agreement between theoretical and experi- mental curves confirms the assumption that the energy assumed is proportional to the ionization produced The ionization at any point of the path of the particle being given by the relation G a (r—a) % ‘ 253 254 Taylor—TLonization of Different Gases by the the total area under this theoretical curve is a measure of the total ionization produced by the alpha particle in the gas. If A, represents the area under the theoretical curve, then A ye dx = ve ea (r—a)4 0) ro) = 3/2 c(r)% = 7°33 ¢ (7 being equal to 10°8 centimeters). Hence c is 3/22 of the area under the theoretical curve when the average range of TABLE JI, ic ¢ der expe. anetee Relative Gaston or area under) rimental under ex-|2atio of the total ioni-) energy re- Vapor theoretical | curve as perimen- zation in the gasto | quired to eurve divid- measured | al METS that in air. produce an ed by 7°33. jwith plan- fray ion. | imeter. | ; Taylor. Bragg. Air 11:24 980 | 87 ms on 1:00 ic 10°00 966 9€ 0:99 1-00 101 CH,I 14°73 1301 88 1°38 1°33 0°75 CH, 12°65 1156 91 118 0°85 C,H,Cl 14:05 1251 |) 89 1°29 1°32 0-77 CS, 15°60 B50) Me Sa 1°38 1:37 0°78 Air 14°64 1249 85 Pe: ae 1°00 N, 1381) 1) 0206 87 0:96 0:96 1-04 CO, 15°01 1262 84 1-01 1:08 0°99 O, 16°72 | 1415 | 85 1:13 1:09 0°88 C,H,,0 19-42 | 1702 88 1°36 1°33 0-74 Air SO 82 89 ae ac 1:00 so, 15°30 1223 80 1'03 aa 0:97 HCl 17°70 1530 86 1:29 ae 0:77 HBr 18732 po a 27 83 1:29 ae 0-77 Air 13°36 | 1190 89 Ae, aS 1:00 HI 17°68 | 1535 87 1°29 Se 0-77 the alpha particle is 10°8 centimeters in any gas whatever. The values of ¢ recorded in column 2 of Table I are then 3/22 of the area under the theoretical ionization curves in the respective gases. The areas under the ionization curves being proportional to the energies consumed in the production of ions in the respec- tive gases, the value of ¢ in any one gas depends upon the total ionization produced in the gas, and consequently upon the energy required to produce an ion in the gas. Then the ratio of the area under the experimental curve to ¢ should be a constant. By dividing the areas under the experimental Alpha Particles from Poloniwm. 255 curves as measured with a planimeter and recorded in column 3, Table I, by the values of ¢ for the corresponding gases, the values recorded in column 4 were obtained and, as can be seen, are approximately constant. The areas under the ionization curves being the measures of the relative ionizations produced in the gases, the ratios of the total ionization produced in the gases to that produced in air were determined by finding the ratio under each curve to the area under the corresponding comparison air curve. After the determination of the ionization curve in each gas, the ioni- zation curve was always obtained in air to be used as a basis of comparison. The ratios of the ionizations produced in the different gases to that produced in air are recorded in column 5 of Table I. Bragg,* by a less direct process, determined the ratio of the total ionizations in gases to that in air and his values are recorded in column 6. There is a fairly good agreement between the values as found by Bragg and those found by a _ more direct process of measurement of the area enclosed by the axes of references and the ionization curve for each gas. Since the energy of the alpha particle is entirely consumed before it ceases to produce ions, the energy required to pro- duce an ion in any given substance will vary inversely as the ratio of the total ionization in the substance to the total ioniza- tion in air if the energy required to produce an ion in air is always taken as the basis of comparison. The values of col- umn 5 of the table are the ratios of the total ionizations pro- duced in the gases as compared with the total ionization produced in air. Consequently the reciprocals of these ratios are the relative amounts of energy required to produce an ion in the substance as compared with the energy required to pro- duce anion in air. The values recorded in column 7 are these reciprocals of the values in column 5, and hence are the rela- tive amounts of energy required to produce an ion in the gases as compared with that required to produce an ion in air. These values indicate a considerable variation of the energy required to produce an ion. The heavier and more complex molecules are apparently more readily ionized than the lighter and less complex ones. This is probably due to the electrons in the heavier and more complex molecules being in a less stable arrangement than they are in the lighter and less com- plex molecules and hence more readily drawn out. In conclusion, I wish to express my thanks to Professor Bumstead for his valuable suggestions in connection with the work and for loaning me the apparatus. I am also indebted to Professor Boltwood for furnishing me the preparation of polonium. * Bragg, Phil. Mag., vol. xiii, pp. 8383-357, March, 1907. 256 Taylor—Ionization of Gases by the Alpha Particles. Results. 1. The ionization curve obtained in various gases and vapors with polonium as the source of rays is of the general form e (72) where J is the ionization; ¢ is a constant for any one gas depending upon the total ionization produced, and conse- quently upon the energy required to produce an ion in the given gas; 7 is the average range of the alpha particles in the cone of rays ; and z is the distance from the source of rays. 2. The agreement between the theoretical and the experi- mental curves confirms the assumption made in previous papers by the writer* and by Geiger,t that the ionization produced by the alpha particle is proportional to the energy consumed. 3. The values of the ratio of the total ionization produced by the alpha particle in different gases to the total ionization produced in air as found by Bragg have been confirmed by a more direct process. 4. The energy of the alpha particle consumed in the pro- duction of an ion depends upon the nature of the molecule ionized. It apparently requires less energy to produce an ion in the gases or vapors which have heavy or relatively complex molecules than it does in those gases of lighter or less complex molecules. i= y% Laboratory of Physics, University of Illinois, Urbana, Illinois, January 28, 1911. * Loe cit. + Loe cit. Duane— Heat Generated by Radio-active Substances. 257 Art. XXV.—On the Heat Generated by Radio-active Sub- stances ; by Witt1am Duane. Since the discovery of radio-activity questions relating to the source and the transformations of the energy involved in the processes have been considered of prime importance. Early in the history of the subject Curie and Laborde* dis- covered that radium generates heat continually, and also that the heat effect increases as the emanation accumulates. A little later Rutherford and Barnest found that the emanation and the first few products of radium that form its induced activity produce their shares of heat, and more recently still Pegram and Webb¢ have succeeded in detecting a small heat effect in a large mass (about four kilograms) of thorium oxide. The ordinary methods of measuring heat (an ice calorimeter for instance) are sufficiently sensitive to detect and measure the heat generated by the quantities of radium, its emanation and its induced activity now at our disposal. I have made recently a number of experiments on the heat effects of other radio-active substances, and in these I have had to use special methods. At first I employed a modification of the differen- tial air calorimeter devised by Rutherford and Barnes (1. ¢.), but this was not sensitive enough and I then constructed a new instrument which is considerably more sensitive than the differential air calorimeter. The method is based on the rapid increase in the vapor tension of a very volatile liquid when the temperature rises. A and A’ (fig. 1) represent two glass vessels, which are joined by the capillary tube B. The vessels are half filled with the volatile liquid, and almost all the air is pumped out by means of a water aspirator through the tube C, which is then sealed off. A small bubble formed out of the residual air left in the vessels is inserted in the tube B, and the displacement of this bubble is observed by means of a reading telescope or by projection with a lamp, lens and seale. I usually employ the latter method, and the displace- ment of the image on the scale is about eight times that of the bubble in the tube. It is not difficult to place a bubble of any desired length in the tube B. It is sufficient to turn the apparatus upside down, and let the liquid run out of the tube. Then on replacing the apparatus right side up one finds the tube more or less completely filled with air. The bubble is * Comptes rendus, cxxxvi, p. 673, 1903. + Nature, Oct. 29, 1903; Phil. Mag., Feb., 1904. tScience, 1904; Le Radium, 1908. 258 Duane—Heat Generated by Radio-active Substances. usually much too long, and to reduce its length all that is necessary is to tilt the apparatus up a little so as to cause a current of the liquid to pass through the tube. This current pushes the bubble down into the portion of tube below and at the side, which is larger than the capillary portion. The bubble remains in this portion of the tube, and the current of the liquid passing it carries along the air little by little, thus reducing the bubble’s volume. On repeating this process, | Fie. 1. SN Se causing the current to flow first in one direction and then in the other, one can reduce the bubble to any desired length. After the bubble has been replaced in the tube, and the apparatus has been prepared for the experiment, the bubble remains in the horizontal part of the tube. It never descends into the large portion, no matter how much the temperature of the room may vary: but it slowly disappears. The air in the bubble dissolves in the liquid more or less rapidly accord- ing to the nature of the liquid, the pressure of the air and the dimensions of the apparatus. In my experiments it is neces- sary to renew the bubble once in two or three weeks: and this is a process requiring about five minutes time. The form and dimensions of the capillary tube B have been carefully studied. The length of the horizontal part is 44™ Duane—Heat Generated by Radio-active Substances. 259 and the internal diameter is a little more than -5"™. The internal diameter of the two large parts is about 8™", and the yertical parts joining the horizontal with the larger parts should not have a larger diameter than the horizontal capillary part. It is easier to control the movement of the bubble, while placing it in the tube and reducing its size, if the capil- lary tube is not joined to the ends of the larger parts, but to the tops as indicated in the figure. The volume of each vessel is about 50°. The interior of the vessels and of the tube must be cleaned most carefully. The least dirt or grease stops the bubble, and in the experiments it is well to choose the part of the tube where the bubble moves most freely. If a source generating heat is introduced into the tube D, the vapor tension is increased and the liquid pushes the bub- ble toward the vessel A’. The instrument is very sensitive. In my experiment I found that 1:5 10-* gram-calorie of heat displaced the image of the bubble 1™ on the scale. This sensitiveness is due to the rapid increase of the vapor tension with the temperature. Among the liquids I have tried, ether seems to be the best. Ether cleans and wets the surface of the glass well, it has very little viscosity and its vapor tension increases rapidly with the temperature, about 17™™ of mercury per degree centigrade at ordinary tempera- tures. Ethel chloride works well also but is much less easily manipulated. The sensitiveness of the instrument varies a great deal with the quantity of air in the vessels. If there is very little air the displacement of the liquid does not change the pressure of the gas much (saturated vapor tensions depending only on the tem- perature), and an increase of pressure in A due to a slight pro- duction of heat is opposed only by a change of level of the liquid in A and A’. As ether is a light liquid, this change of level opposes only a slight force to the displacement of the bubble. For great sensitiveness, therefore, one must remove almost all the air from the vessels, leaving only enough to form the bubble. The sensitiveness depends also upon the ratio of the eross-section of the capillary tube to the surface of the liquid in the vessels. A decrease in the cross-section increases the sensitiveness. I have found, however, that (if the liquid is ether) a tube of less than °5™™ internal diameter does not work well on account of the capillary forces. Further, the displacement of the image of the bubble is increased by the lens (or reading telescope). It is not desir- able, however, to multiply the displacement more than eight or ten times, as the loss in sharpness of image counterbalances the advantage of increased displacement. 260 Duane—Heat Generated by Radio-active Substances. In actual practice the protection of the instrument against outside thermal disturbances is just as important as great sen- sitiveness. In my earlier experiments Damibedden the two vessels in a block, E (fig. 1), of lead (weighing 25 kilograms). The vessels were held in place by a layer of paraffine, which filled the space between them and the lead at the bottom. At the top this space was filled with cotton wool. Two metal rods, F (normal to the plane of the figure) support the block of lead inside a brass box, G. These rods serve as axles about which the lead can be turned, and thus the bubble of air shifted to any desired position in the capillary tube. The box G was completely enveloped in cotton wool contained in a second box of zine (not represented in the figure). This system of good conducting metal screens separated by spaces filled with non-conducting material furnished excellent pro- tection against thermal disturbances, but was not sufficient where the greatest sensitiveness was required. The whole apparatus, therefore, was placed in an electrical thermostat similar to the one described some years ago in this Journal.* In my later experiments I have replaced the cotton wool with eider down, and I have added two large blocks of lead on top of the box G. These blocks equalize the variations of temperature coming from above. They are placed one beside the other, leaving just enough space between them for the tubes by which the substances to be examined are lowered into the calorimeter. With these modifications I have found it unnecessary to set the thermostat going, except on those days when the temperature of the room undergoes wide fluctuations. Very often the heat due to the radio-active processes is pro- duced in relatively large masses of matter. In these cases it is necessary to leave the substance to be examined for a long time in the upper part of the tube by which it enters the calor- imeter, in order to be sure that its temperature is as nearly equal to that of the calorimeter as possible. This part of the tube should lie between the two large blocks of lead, and should be of metal to facilitate the equalization of temperatures. If the generation of heat by the source is relatively large, an appreciable quantity of it may be conducted down the column of air into the calorimeter. In order to avoid this a small quantity of eider down fastened to the end of a very fine glass rod may be inserted into the tube just above the calorimeter. In making an experiment the eider down is removed, the substance to be examined lowered into the calorimeter and the eider down quickly replaced. Any one of several methods may be used in measuring the heat generated by the source. On lowering the source into the calorimeter one can wait until a sort of thermal equilib- * Duane and Lory, this Journal, 1900. Duane—Heat Generated by Radio-active Substances. 261 rium is reached, when the heat conducted away from the calorimeter equals that given it per second by the source, and observe the maximum displacement of the bubble of air. This method works well provided the instrument is not arranged for very great sensitiveness. If, however, the apparatus is very sensitive it is better to take the velocity of the bubble as a measure of the heat generated per second. Although the instrument is well pro- tected against thermal disturbances from the outside, yet the bubble does not stay in the same place. The zero of the instrument is not fixed. Nevertheless, if the apparatus has remained undisturbed for a long time, and the temperature throughout has become as nearly equalized as possible, the natural drift of the bubble is slow and regular, and the change in its velocity due to the heat from the source, when it is lowered into the calorimeter, can be measured with consider- able precision. A third method is to compensate the effect of the heat generated in the tube D by generating a known quantity of heat in the corresponding tube D’ (figure 1). The best method, however, is to compensate the heat effect by absorbing the heat in the tube D itself as fast as it is generated. This can be done by means of a current of electricity flowing across the junction of two metals. Peltier discovered that if the current passes in one direction heat is generated, and if in the opposite direction heat is absorbed at the junction. In my earlier experiments I inserted a thermo-couple P of iron and nickel wires into the tube D, and I determined the current that absorbed the heat as fast as it was generated, by varying the strength of the current until the velocity of the bubble was the same as its natural drift. In the later experi- ments I have replaced the simple thermo-couple by a metal tube. The walls of the tube are 1™™ thick, and its external diameter is just enough less than the diameter of the tube D to allow of its being inserted easily into the latter. The length of the metal tube is about 4°, so that the entire tube lies inside the calorimeter. Half of the tube is of iron and the other half of nickel, the two surfaces between the two metals being vertical and parallel to the axis of the tube. An iron wire is soldered to the outer edge of the iron half of the tube and a nickel wire to that of the nickel half, so that a current of electricity descending by the iron wire into the iron half of the tube can pass across the joints into the nickel half and ascend by the nickel wire. With this arrangement, when a source of heat is lowered into the middle of the iron-nickel tube, it is surrounded by a good conductor of heat, and the distribution and compensation of the heat takes place easily 262 Duane—Heat Generated by Radio-active Substances. and quickly. Thus the thermal equilibrium of the apparatus is not disturbed much by the heat generated by the source. This method is capable of considerable precision and can be used, without changing the apparatus, to measure heat effects varying from ‘001 gram-calorie to 2 gram-calories per hour. Larger heat effects could be measured by increasing the thickness of the iron-nickel tube and iron and nickel wires so as to decrease their electrical resistance and the heat gen- erated in them according to Joule’s law. The iron-nickel tube has been carefully standardized by inserting a small coil of manganine wire of known resistance into the tube, by heating this with a known electric current, and by determining the current in the tube that would exactly absorb the heat produced. The electric currents were produced by small storage batteries, and their intensities were varied by changing the resistances in plug resistance boxes contained in the circuits. The resistances in these boxes, as well as the other resistances in the cirenits, were carefully measured by a standard Wheat- stone’s bridge. The electric currents were measured by com- paring the electromotive forces of the storage cells with that of a standard Weston cell by the potentiometer method, and by dividing these electromotive forces by the total resistance in the circuits. The following table contains the data obtained in standard- izing the iron-nickel tube. The resistance of the small coil inserted into the calorimeter was 9°20 ohms, and that of the lead wires attached to it was negligible. The electromotive force of the standard Weston cell was 1,018 volts, and that of the two cells forming the storage battery 4,153 volts. TABLE 1. Total resistance Heat produced Current iniron- Heat absorbed in heating cir- in heating cir- nickel tube. per hour per cuit. cuit calories . Ampere ampere Ohms per hour —_—_- Observed Corrected 480 "593 0716 0716 § 29 550 ‘450 *0540 °0542 8°30 650 "324 “0392 "0393 8°25 910 165 “0200 0197 8°37 The compensations were not always exact, and a small cor- rection was made in the values of the current in the iron- nickel tube. This correction was determined by observing the velocity of the bubble of air. The heat effect in the iron-nickel tube is due to two causes. Firstly, the heat generated or absorbed at the junctions of the metals according to the direction of the current (Peltier effect), and, secondly, the heat generated according to Joule’s law, Duane— Heat Generated by Radio-active Substances. 268 which is proportional to the square of the current and to the resistance. ‘The fifth column in the table contains the heat absorbed per hour and per ampere by the tube, and it appears that this quantity is independent of the current in the tube. This means that the absorption of heat is proportional to the intensity of the cooling current, i. e., the resistance of the tube is so small that the heat generated according to Joule’s law is inappreciable, if the cooling is no larger than -6 calorie per hour. The mean value of the heat absorbed (or generated) per hour and per ampere in the iron-nickel tube is 83 gram- calories. I found 8-2 calories for the couple used in the earlier experiments. It is interesting to note that the electromotive force in the Fig. 2. Centimeters Minutes surface between the iron and the nickel must be about -00055 volt to produce this effect. In order to determine the sensitiveness of the instrument I sent a very small current through the iron-nickel tube, and observed the change in the velocity of the bubble due to the heat absorbed or generated. The curves in figure 2 represent the displacements of the image of the bubble. The lines ad and ed represent the bubble’s natural drift. The abscissas of the points 6 are the instants at which the electric current commenced to flow through the tube, and the abscissas of the points ¢ are those at which the current was broken. For the first curve the direction of the current was such as to generate heat, and for the second to absorb it. It appears that the dis- placement of the bubble due to the current was about the same in the two cases but in opposite directions. This confirms Peltier’s law, and indicates that the Joule effect is negligible. 264 Duane—Heat Generated by Radio-active Substances. The strength of the electric current was ‘00019 ampere, and the heat generated or absorbed *0016 calorie per hour. In ten minutes ‘00027 calorie was generated or absorbed and _ this quantity of heat displaced the image of the bubble about 1-6". It follows that one millimeter displacement of the image corresponds with :00017 calorie of heat absorbed or generated. In some other experiments I have found that the displace- ment of the bubble is proportional to the quantity of heat absorbed or generated, provzded that the absorption or genera- tion is not too rapid. These results are used in estimating the small correction that must be applied, if the value of the current that exactly absorbs the heat generated by a source has been determined only approximately. Centimeters Minutes I have measured the heat generated by radiothorium and by polonium. These experiments were described in two notes presented to the Paris Academy of Sciences on June 1 and June 21, 1909. Since the first experiments I have measured the heat gen- erated by the polonium several times to see if the heat effect decreased with the time according to the law of decay of polonium. Half a given quantity of polonium disappears in about 142 days. The curves in figure 3 represent the displacement of the bubble in these experiments, the dates being for curve 1 the 4th of May; for curve 2 the 4th of June, and for curve 3 the 25th of June. In each case the lines ab and fg represent the natural drift of the bubble. The abscissas of the points 6 are the instants at which the polonium was lowered into the Duane—Heat Generated by Radio-active Substances. 265 calorimeter, and the abscissas of the points / are the instants at which the polonium was raised again. The numbers written near the lines cd and de are the electric currents in amperes which were flowing through the iron-nickel thermo-couple dur- ing the corresponding intervals of time. Between the points ) and eI was searching for the proper value of the current to counterbalance the heat effect, and between the points e and 7 I was reducing the current to zero. The last experiment was not as good as the others, because the natural drift of the bubble was large and changed a little dur- ing the experiment. I compared the ionization due to the polonium when spread out in a thin layer on a disk of platinum (4 ¢) with that due to a thin layer of radium. The results of the experiments appear in Table 2. TABLE 2. Heat Weight of effect RaBr~ that of this Date of Current that Heat pro- Ionization hassame quantity experiment compensated duced cal- current due activityas of radium 1909 heat effect. orie per topolonium polonium calorie Ampere hour g. per hour May 4 00143 012 Mes} Se Or" 75 °0110 June 4 “00110 ‘009 119+ 107 66 0095 June 25 ‘00100 008 "99 X10)" “Bil 0084 It is evident that the generation of heat and the ionization current due to the polonium decrease with the time. The ionization current decreases at a rate indicating decay to half value in 136 days, which is very close to the value previously found by other experimenters. The heat effect decays a trifle faster than this, but the differences are not greater than one would expect considering the magnitude of the quantities of heat evolved. It follows from this that the heat effect was certainly due to the polonium. . On account of the difficulty of obtaining saturation in measuring the ionization of the radium and of the polonium, the method of comparing the activities of the two substances must be regarded as approximate only. Remembering this. it appears that the heat generated by the polonium is ver close to that generated by the quantity of radium that would produce the same ionization as the polonium. I have made «a number of experiments on phosphorescent salts to see if they generate heat when in the phosphorescent state. Every time I examined such a salt one or two hours after it had been withdrawn from the light of the sun (or of an ultra-violet ray are lamp), I found a small but measurable generation of heat. Twenty-four hours later not the slightest effect could be detected in the majority of cases, but a few times I observed a small generation of heat and on withdraw- Am. JOUR. Poteau Series, Vou. XX XI, No. 184.—Apnrin, 1911. ell 266 Duane—LHeat Generated by Radio-active Substances. ing the salt from the calorimeter found that the phosphorescent light had disappeared. It is impossible to affirm, therefore, that the heat effect is directly related to the emission of vzscble phosphorescent light. It may be due to the emission of visi- ble and invisible rays together, or it may be caused by some reaction of a secondary nature. These researches, however, have suggested to me the follow- ing question: if a quantity of radium is mixed with a phos- phorescent salt, causing it to phosphoresce brilliantly, does the mixture generate the same quantity of heat as the radium would generate alone? There appear to be three possibilities : (a) the ener gy of the rays is absorbed (at least in part) in pro- ducing chemical reactions in the phosphorescent salt. In this case the heat effect of the mixture should be less than that of the radium alone, at least at first. (0) the radium rays acting on the atoms and molecules of the salt liberate a part of their chemical or subatomic energy. In this case the heat produced by the mixture should be larger than that produced by the radium alone. (¢) the energy of the radium rays is rapidly transformed (in part) into the energy of the phosphorescent light without producing other reactions, and in this case, if all the light is absorbed in the vessel containing the mixture, the heat produced should be the same as that due to the radium alone. In order to investigate this question I made the following experiments. On December 3d, 1909, a certain quantity of a salt containing finely pulverized radium chloride and barium chloride was divided into two parts. One part, A, weighed "0314 gram and was sealed into a small glass tube. The other part, B, weighing :0206 gram, was thoroughly mixed with ‘267 gram of phosphorescent zine sulphide, and then sealed in a second glass tube similar to the first. Several times during the five weeks following the sealing of the tubes I measured the heat effects of each of them, and I also compared the intensity of the y-radiation emitted by them with that due to a standard tube containing 26°5 grams of radium chloride. The following table (3) contains the results of these experiments : TABLE 3. Date of Quantity of RaCl-? Production of Ratio of that produces the heat calories A experiment same y-rays per hour Tay = a => = Si => A B A B December 7 1°66 1:07 Be is pir 1°55 December 7 se: deen sales 2 +098 5S December 21 DAVY 1°57 nae eM tinh 153 December 21 ere eee 199 129 1:54 January 5-- 2°39 1°56 oe Be 1°53 January 6.. Bese sits = 201 “2 1°58 Duane—Heat Generated by Radio-active Substances. 267 The y-radiation and the generation of heat increased between the 7th and the 21st of December, but after that no increase was perceptible. This was due to the accumulation of the emanation and the induced activity, which after three weeks attained approximately their saturation values. The sixth column contains the ratios between the tubes A and B. It appears that these ratios are the same no matter what the date of the experiment was and no matter whether the y-rays or the heat effect was measured. It follows that the presence of the phosphorescent salt does not appreciably change the rate of generation of heat by the radium. The following facts may be noticed in passing: The phos- phorescence of the mixture has become less imtense than at first but is still brilliant. The hght has also changed its color, becoming more orange. In order to investigate the heat effect in the case where the phosphorescence is produced by the 8 and ¥ rays in willemite and in platinum-barium cyanide, I arranged the following experiment: A long, fine tube is inserted into tube B of the calorimeter and around the end of this tube is packed the phosphorescent salt. A very small glass capsule hermetically sealed containing radium can be lowered down the fine tube to the center of the salt. The walls of the tube and capsule are so thin that under these conditions the salt phosphoresces brilliantly. I compared twice the y-rays from the capsule with those from the standard and found that the quantity of radium in the capsule corresponded with 1:91 and 1:92 mg. of RaOl,. The heat effects observed on lowering the radium to the center of the phosphorescent salt were the following : TABLE 4, Heat calorie Salt used per hour 67 gr Platinum-barium cyanide -_-------- 170 DONG) Ore Aan Nieves 5 5 ey oes ee wee ca Nie No phosphorescent salt -_....--.--- Styl No phosphorescent salt _--- .-- ee MAlG It is evident that the generation of heat is the same whether the phosphorescent salt is present or not. It follows from these two series of experiments that there is no appreciable absorption of energy in producing chemical reactions, and that the rays do not liberate an appreciable amount of chemical _ subatomic energy. These results are interesting from the point of view of the amount of energy necessary to effect the organs of sight. 268 Dwane—Heat Generated by Radio-active Substances. In the first series of experiments the phosphorescence was produced for the most part by the a-rays of the radium. We know that each a-particle that strikes the phosphorescent zine sulphide produces enough light to affect the eye, and it follows from the experiments described above that the energy of this light is no larger than the energy of the a-particle. The smallest velocity of an a-particle that has been measured and at the same time detected by its scintillation is 510° —— , and the kinetic energy of the a-particle at this velocity is 8107 erg. . ; : 6 1 : . This energy is about that required to raise G0 of a mille- 1 : sas gram 3559 of a millimeter. The energy necessary to pro- duce the sensation of sight is less than the above quantity, since only a part of the total light energy enters the eye, and since probably the whole energy of the a-particle is not trans- formed into luminous energy. The heat generated by one gram of pure radium can be calculated from the data of Table 3. It is for tube A 110 and for tube B 108 calories per hour. The difference between these two numbers is not greater than the errors of experiment. The heat effect of one gram of radium calculated from the data of Table 4 is 117 calories per hour, a value considerably larger than the preceding values. This difference cannot be explained by errors of experiment. It is probably due to the - fact that the radium employed in the second series of experi- ments is several years older than that employed in the first series, and contains, therefore, more of the disintegration products of the radium, especially polonium, which generate heat. I have made a number of attempts to measure the heat pro- duced by the rays from radium at a distance from their source. In the first experiments a thermopile, a bolometer, and a radiometer were tried, but none of these instruments gave satisfactory results. A modified form of differential gas ther- mometer gave positive indications of a heating effect, but the only instrument that proved satisfactory was the differential calorimeter described in the present paper. I hope shortly to publish some details of these experiments, but will state here simply that the problem is somewhat different from that of measuring the energy of ordinary radiations (at least as far as the penetrating radium rays is concerned), because a relatively large amount of matter is required to stop these penetrating rays, and the heat is generated throughout the mass of this matter. Pirsson and Rice—Geology of Tripyramid Mountain. 269 Arr. XX VI.— Contributions to the Geology of New Hamp- shire, IV. Geology of Tripyramid Mountain; by L. V- Prrsson and Wm. Norru Ricr.* Introductory.—Tripyramid Mountain is in the southern part of the White Mountains in New Hampshire. The point formed by the intersection of 44° N. and 71° 30’ W. is about two miles northwest of its northwest lower slope. It is entirely within the township of Waterville and a little east of its cen- ter. Surrounded by other mountains, Osceola, Kancamagus, Passaconaway, Whiteface, Sandwich Dome and Tecumseh, peaks which rise 3-6 miles distant, it is much concealed, and there are not many places where it can be observed in its full proportions from below. ‘The retired character of its situation is much enhanced by the wild and heavily wooded nature of the region in which it stands, the only habitations in the upper valley of Mad River, which drains the township, being a sum- mer resort hotel and a few scattered farm houses. The east- ern slopes of the mountain are drained by headwater branches of Swift River, whose upper basin is a similarly wild and heavily for ested region. Consequently its summit is not easily accessible and is little visited by tourists, especially as the view is largely circumscribed by the neighboring peaks and obscured by the thick growth of spruce serub covering it. The best point to reach it from is the hotel at Waterville, which is 12 miles from the railway at Campton. A walk of about four miles, partly on trails through the forest and partly a scramble up rough and overgrown mountain brook beds, brings one to the lower slopes and the slides described beyond. Topography.—Tripyramid Mountain is a roughly oval mass which rises about 2000 feet above the floors of the valleys about it. It is crested by three peaks with saddles between, called the North, Mid, and South Pyramids, to which it owes its name. Its appearance from the west is seen in the accompanying view, which we owe to the kindness of Mr. A. L. Goodrich. It was taken looking across the mead- ows above the old lumber dam at a place on Slide Brook called Swazeytown, below the junction with it of Cascade Brook. On the north the mountain is connected with the peaks of Kancamagus by a high ridge with an intervening point upon *Some years since one of us (L. V. P.) went to Waterville, N. H., to see the occurrence of the ‘‘ossipite”” mentioned in this paper and to collect material. While engaged in this hefound W.N. R. had also studied the Tri- pyramid rocks in the field. We joined forces and the present paper is the result. It is proposed to follow this with one dealing with the petrology of the rocks. 270 = Pirsson and Rice— Geology of Tripyramid Mountain. it, sometimes called Fourth Pyramid. On the south it descends to a bench, or roughly level area, known as Flat Mountain, whose elevation is about 2500 feet above sea level and which in tur again descends into the valley of Cold River. North Pyramid is about 4200 feet above the sea, the other pyramids are a little lower. The mass, as thus defined, is over two miles long, by about one and a half broad; the Fie. 1. Fie. 1. View of Tripyramid Mountain. Looking east from above the Swazeytown Dam. The North Slide and Ravine of Avalanches are seen to the left, the South Slide on the right. distance between the North and South Pyramids about a mile along the crest. The details of the topography are shown on the accompanying map, which has been compiled from various sources, the approximate expression of the topography of the older map of the Hitchcock Geological State Survey being cor- rected in details by later maps of parts of the area made by the Yale School of Forestry under the direction of Mr. Henry Gannett and Prof. Il. H. Chapman, by Mr. C. W. Blood, and one of trails and stream courses by Mr. A. L. Goodrich, and to Pirsson and Rice—Geology of Tripyramid Mountain. 271 these gentlemen we desire to express our indebtedness for the use of this material. The western two-thirds of the map has been made chiefly from these sources, the eastern third is from the older state map. We have added a few corrections of our own. Nearly everywhere the mountain, and also for the most part the surrounding area, are covered with a dense forest growth. On the slopes of the mountain and on its top this is composed of a thicket of small spruce trees which rise through a floor mat composed of intermingled dead and fallen tree trunks, more or Jess decayed, accumulations of spruce needles, shrubs and moss, into which one often sinks to the waist, and through which progress is extremely difficult. On the lower slopes the kinds of vegetation are somewhat different but the character of the thicket remains the same and the mantle is often swampy in addition. Were it not for the slides and the chan- nels of the streams running from them, the underlying rocks, except on the summit, would be completely concealed by this vegetable growth and deposit, which on a rainy day, to one immersed in it, calls to mind Darwin’s description of Terra del Fuego. The Slides.—The most interesting features of the mountain are what are locally known as the “slides.” These are two tre- mendous landslides, or avalanches, which have occurred, one on its north, the other on its south slope. The North Slide has left a bare face of underlying rock extending from the narrow Ravine of Avalanches, which separates the mountain from the next peak to the north, upward for half a mile along the slope and with over a thousand feet of elevation, with an average angle to the horizontal of 30°. Starting at a point not far below the summit, it gradually widens until at its base nearly the whole north face of the mountain, up into the head of the Ravine of Avalanches, is exposed. The naked rock surface left by this, which is about as steep and smooth as one can comfort- ably climb upon, is interrupted here and there by piles and trains of rock débris and lines of small trees and shrubs grow- ing in crevices. The most conspicuous lanes of rock face exposed are separated from several minor similar ones east of them on the north slope and these from each other, by long strips of soil and forest. The exposed rock of these smaller eastern lanes appears quite weathered. Minor slides have also occurred from the opposite slope of the neighboring elevation into the Ravine of Avalanches, which appears to be well named; see fig 1. A view of the chief double lane of sliding of 1885 is seen in fig. 3, taken from the opposite mountain side by Prof. E. L. Rice. Small drainages pass down these lanes and empty into Avalanche Brook, which heads below. “ay MOZMOT, “S@]IJ JO e[vog 779, BS = oa “o1qqey eee 01449 NM eS +4, = <1 {=\— Ag any zy 7 D Ve Yer |e ji “AJLMIOLA puUv ULeyUNOW; pimerddiary, Fo ‘slaqaq pur FMP TRIOLTH “opTURIH) 2a od eos Hes S| sft pai es so a= os lon! Oreos Ge Dn OF sya} S28 O58 °S on S SE°2 Seler aS) = a's ac LY —oetn 2 RE, St) = FS — oot ee: YH ee ow R nl 13 (aS Be° > Og See hy a= ano > cs 2 as EH = iS) A 8 As & = Oo aa Pirsson and Rice—Geology of Tripyramid Mountain. 273 Fie. 3. Fie. 3. View of the North Slide from the south slope of Fourth Pyramid. 274 = =Pirsson and Rice—Geology of Tripyramid Mountain. small amount of débris spread down the ravine below, com- pared with the extent of surface denuded. The condition which has caused the slides appears to be the steep and smooth rock surfaces on which the accumulating lay- ers of largely organic deposits rested. When these became heavy enough, at a time when they were saturated with water after long and torrential rains, which also lubricated the under- lying rock surface, they broke away and slid down. The steep- ness and smoothness of the bed rock is occasioned by a certain sheeting which it possesses and which is discussed later. Judg- ing from the conditions, and from what has occurred, it seems possible that other slides may occur in the future. The South Slide is in essential respects, as to size, height, ete., quite com- parable to the North one, only in this case the thickness of the débris of soil, rocks, ete., which moved, was apparently greater. Thus the underlying rock i is exposed only at the upper part of the slide, the earth mantles the middle part and increases in thickness as one descends, while the glen below in which the Slide Brook heads is choked with the accumulated material that moved down into it. This lower part has the hummocky surface characteristic of landslides and is furrowed by shallow ravines which the drainage has cut into it. Many large blocks of rock, some of them 8-10 feet long, are exposed in these ravines ; most of them are of syenite from the mountain above, but others are of black trap, porphyritic granite, dark gabbro, ete., and are evidently transported glacial erratics. This avalanche occurred on Oct. 4th, 1869, and a second one on Aug. 18th, 1885, as a sequence to the terrible downpour which also caused the largest North Slide. It can be just seen in fig. 1 to the right as a white line showing through the dark forest. An admirable account of these slides and the causes which produced them has been given by Mr. A. A. Butler.* Listory.—The first mention of the geology of Tripyramid Mountain that we have been able to tind is in a description of the South Slide by Prof. G. H. Perkins+ written shortly after its occurrence. The mountain is called by him Passa- conaway ; there appears at that time to have been some. con- fusion in regard to the use of this name, and later it became fixed to the mountain east of Tripyr amid, which now bears it. In his description of the slide he states that the upper part of the mountain is composed of a gray syenite. As this term was then used it bore reference to the fact that the rock contained hornblende, it did not mean that it was free from quartz, or nearly so. The name, however, proves correct as * Appalachia, vol. iv, No. 3, p. 177, 1886. + This Journal (2), vol. xlix, p. 158, 1870. -7 Pirsson and Rice—Geology of Tripyramid Mountain. 275 used in the former, or in the later, petrographic sense, as will be shown later. He speaks also of the presence of trap dikes in it from an inch or two up to a foot in thickness. He does not mention their color, so it is uncertain whether this refers to aplitic or lamprophyrie dikes, or to both, but the use of the word trap suggests the latter. He speaks also of extensive layers of black hornblendic rock a mile below the slide on the stream ; this evidently refers to the gabbro mentioned later. The next account bearing on Tripyramid is found in Hitch- cock’s Geology of New Hampshire.* In this, references to the mountain, to its rocks and geology, are made in a number of places, and in vol. II, p. 211 and following, a general de- scription of its geology is given. As weshall have occasion in s sy 6>¢e and their direc- tions are labeled respectively A, Band ©. This ellipsoid gives the wave shell, or ray surface, by the following construction : if through the center perpendicular to a given ray direction a plane is passed, the section is an ellipse, and if on the ray direction, distances be laid off proportional to the semi-axes of the ellipse the locus of all points so determined is the wave shell. Furthermore, the vibration directions belonging to the given ray lie in planes determined by the ray and the axes of the elliptical section. If one may judge by the text books, the ellipsoid, whether that of Fresnel or this of McCullagh, is used in the non- mathematical presentations of this subject to obtain the wave shell, or rather its three principal sections, and is then imme- diately abandoned. The device which occurred to me some years since, and which I have found useful in obtaining results with my students, is to carry the use of the ellipsoid into a considerable number of the applications of the theory to the properties of crystals. The “section cut from the ellipsoid” McNair—A Method in Teaching Optical Mineralogy. 293 is used at once in accounting for the phenomena exhibited by thin crystal plates between crossed nicols. It is applied next to the simple cases of superposition of plates of even thick- ness, then to those cases in which the plates are not of even thickness. By its use the average student gains some facility in predicting the phenomena which will be observed in a given case, such, say, as that of an even plate over one shaped like a convex lens. Afterward the method is applied to crys- tal plates in convergent plane polarized light and perhaps the most conspicuous example of its usefulness lies in its applica- tion to the distinction between positive and negative crystals in the convergent beam of plane polarized light. Fic. 1. Let fig. 1 represent a central section through the converg- ent cone which traverses a plate cut perpendicular to the acute bisectrix. If the crystal is positive, this acute bisectrix is C, the direction of minimum elasticity. If for each ray direction the ellipsoid is constructed, and the section perpen- dicular to the ray is cut, the elliptical sections will appear in this figure only as traces marked E,, E,, ete. If now a dia- grammatic plan be made on the assumption that the above section coincides with the axial plane, and, for brevity, with that plane in “the 45° position”, and if the ellipses be dia- grammed on this in their full area, that is, if each be tipped into the plane of the diagram before being drawn, we shall obtain something like fig. 2. On this figure lines represent- ing the position of the hyperbola have been drawn to enable it to be referred to the biaxial interference picture, and the 294 McNair—Method in. Teaching Optical Mineralogy. ellipses have been numbered to correspond with those shown in fig. 1. The ellipse for the central ray of the cone being drawn perpendicular to the C direction has for its semi-axes a and 6. If now we consider the series of ellipses belonging to rays in the A-O, or axial, plane, represented in the section above, it will be seen that as we leave the central ray, passing outward to rays making a continually increasing angle with the central one, the axis perpendicular to the A—O plane remains constant Fic. 2. and of the value 2b, while the axis in the axial plane decreases from its maximum, 2a, toward the value 2c. Furthermore, when the ray has the direction of the ray axis of the crystal the ellipse obtained for it is a circle. Beyond this ray the axis in the A-C plane is the minor axis of the ellipse. Thus it appears that the ellipses on one side lie with the major axes at right angles to the major axes of those on the other side, the McNair—A Method in Teaching Optical Mineralogy. 295 change occurring at the ray axis, which is close to the nose of the hyperbola of the picture. Proceeding now from the center of the figure at right angles to the axial plane, the ellipses undergo no such change, the major axis now remains constant at 2a while the minor axis erows smaller, reducing from 2b toward 2c. The major and minor axes of these ellipses represent the speeds of the two disturbances vibrating respectively in the planes of these axes. If we call the difference of these speeds A, as is convenient, it is easily seen that A near the nose of the hyperbola is zero, and that from this spot it increases toward the center of the figure, while at right angles to the trace of the axial plane it increases away from the center. Outside of the hyperbola and away from the center it increases in numerical value but is negative. Suppose now that the quartz wedge is pushed over the plate, thin edge first, with its vibration directions as indicated by the ellipse drawn in its corner. It will readily be seen that this ellipse lies parallel to that for the central ray of the cone, and to all others inside of the hyperbola. The faster ray in the plate will also be the faster in the wedge, and the slower in the plate will be slower in the wedge. The effects of plate and wedge will be added. Now the location of the color bands depends directly on the value of A, hence for a given position of the wedge a particular band will be found where the combined A of wedge and plate equals the A of the plate alone at the original location of the band before the super- position of the quartz wedge. As the wedge is advanced the combined A at any point of the plate is increased continuously, while any particular value of the combined A necessarily shifts toward that spot in the plate where A is least. Therefore, as the wedge is pushed over, the color fringes will travel from the center toward the nose of the hyperbola. Likewise, on the line at right angles to the axial plane they must travel in toward the center. Outside of the hyperbola, however, the ellipses are in crossed position, and the slower ray in the plate is now the faster in the wedge. S>~ =e Sn eee ee ee Ne eee ie AR eee eS a a SIN gee rae TOS NN egress 17 a Boo \ Shee weleee Y SCY SSS e < Bi a ey ~ Se Fic. 3. Syntonoptera schuchertt. no further, the anterior giving origin to 3 main twigs, the fore- most of these widely diverging and sending forth 4 S-shaped twigs directed backward. Each of the 2 other main twigs forms 3 short apical veinlets. The 2 long-vaulted anal veins, directed like the branches of the cubitus to the posterior mar- gin of the wing, send out 3 or 5 partly simple, partly forked S-shaped branches. Of a third anal vein there is but a frag- ment to be seen. By the peculiar distribution of the veins and by the parallel position of the last radial with the first medial branch, of the last medial with the first cubital branch and by the bipartition of the medial vein, the surface of the wing seems to be divided into 4 triangular areas almost equal in value. It is of great interest to see the manner in which the convex and concave veins alternate in this wing. We find here a proof that the position of the veins above or below merely follows mechanical rules and has nothing to do with the origin of the veins. The costa is convex, the subcosta concave; the Vicinity of Mazon Creek, Lllinois. 301 radius convex, its sector concave ; of the sector branches, the first and the third are concave, the second convex; the first main fork of the media is convex, the second concave and their branches again alternate like those of the sector ; the anterior branch of the cubitus is convex, the posterior concave, the twigs like those of the media; first anal vein convex, second concave. The spaces between all the main branches and veins are traversed by cross veins, sometimes forming a diffuse net- work. Many of the twigs or branches of the longitudinal veins have the appearance of intercalary veins like those of Plectoptera and Odonata, but this is only a superficial resemblance caused by their strongly S-shaped curve Holotype in : Peabody Museum, Yale University, Cat. No. 19. Family uncertain. AMOUSUS, new genus. Amousus mazonus, new species. Fig. 4. The basal third, 50™™ long, of probably a posterior wing. There are to be seen the nearly straight costa, the moderately remote subcosta and the radius, whose sector evidently rose in Wie. 4. Fie. 4. Amousus mazonus. the second third of the wing. The medialis splits near the base, the anterior branch running for a short distance close together with the radius, the posterior braneh soon splitting into 2 twigs. The cubitus is divided immediately after its origin, its 2 forked branches running steeply to the posterior 302 Handlirsch—New Paleozoic Insects from the margin. Farther backward follow 6 simple short anal veins, vaulted in the characteristic manner. Interstices filled with irregular network resembling that of the Dictyoneurida, to which family this fossil possibly may belong. Holotype in Peabody Museum, Yale University, Cat. No. 20. Direxopvs, new genus. Diexodus debilis, new species. Fig. 5. The specimen consists of 21" of the basal half of a left (? posterior) wing. Costal area moderately broad, pointed. Costa and subcosta nearly straight, radius close to and parallel Fie. 5. Fic. 5. Diexodus debilis (negative). x 2°8. with the subcosta, sector rising at about a quarter of the length of the wing. The media sends forth its first branch somewhat distally of the origin of the sector and splits soon after into 2 branches. Cubitus forked very near the base, the anterior main branch splitting into 2 branches, the first of which again splits into 3 twigs running in a long vanlt to the posterior margin, the second forming only a long fork. The second main branch remains undivided. Of the 5 anal veins to be seen, some may have a common stem. Cross veins widely spread. This species possibly may belong to the Lithomantide. Holotype in Peabody Museum, Yale University, Cat. No. 21. ScEPASMA, new genus. Scepasma gigas, new species. Fig. 6. A fragment, 60™" long, of the anal region of a posterior wing which had probably a length of 180™" and evidently was very broad. There are preserved: a strong vein, split into 3 Vicinity of Mazon Creek, Illinois. 303 branches, probably the cubitus ; 2 long forked anal veins strongly curved downward ; and 10 simple anal veins, joined together by delicate and irregular cross veins, forming a diffuse network. Though this anal area by the somewhat fan-like arrangement of the numerous anal veins recalls the anal areas of more highly specialized orders, there is no support for the opinion that it eould be folded. I am not able to say to which category this giant species may belong, but it is certainly a Paleeodiectyopteron. _ Holotype in Peabody Museum, Yale University, Cat. No. 22. Fig. 6. Fig. 6. Scepasma gigas. AMETRETUS, new genus. Ametretus levis, new species. Fig. 7. A fragment of 37™™ from the base of a ? posterior wing measuring probably about 150™". Costa slightly vaulted, par- allel with and rather remote from the subcosta, which gives off oblique simple or forked veinlets in the costal space. The simple basal pieces of the radius and of the medialis are fol- lowed by the eubitus, which is curved downward, splitting into 3 branches. Of the 8 anal veins, which I can distinguish, the first, fourth and sixth are forked, the fifth trichotomous, the others simple. I cannot see cross veins. 304 Handlirsch—New Paleozoie Insects from the This wing seems to belong to one of the more highly special- ized Paleodictyoptera, having a narrower base and no network between the veins. Holotype in Peabody Museum, Yale University, Cat. No. 23. “= Fie. 7. Ametretus levis. Order PROTORTHOPTERA Handlirsch. Family SPANIODERIDA Handlirsch. SpaNIoDERA Handlirsch. I have based this genus on a single species. Having now before me 7 other forms, it is possible to make a more exact description of the generic characters. Anterior wing generally with rounded apex. Costa marginal ; subcosta shortened, uniting with the radius; radins simple, not split at its apex; sector rising very near the base but never cleft before the middle of the wing, forming only a small number of branches; medial vein cleft near the base into 2 main branches, remaining either simple or splitting into a few twigs. ‘The medialis is always free and never meets with the sector. Cubitus long, more or less S-shaped, with 3-6 simple or forked branches, bent downward to the posterior margin. Vicinity of Mazon Creek, L[llinois. 305 Anal area limited, comparatively slender and filled by a moder- ate number of sinuate veins. Cross veins oblique in the costal area, stretched and diffuse in the other areas. Posterior wing similar to the anterior, but provided with a large folded anal lobe. Prothorax elongated. Head apparently with progna- thous mouth parts. Legs homonomous, moderately long. Fie. 8. ok FEN CAN fo of : Q SS EFA Tes} : | LLL Pore Sie Fie. 8. Spaniodera longicollis. x1°'8. Spaniodera longicollis, new species. Fig. 8. Prothorax very slender, 10™ long. Anterior wing 34™™. Very similar to S. ambulans Handlirsch, from which it may be distinguished by the longer neck and by some details in the wing venation. Apex of the wings broadly rounded ; space between the radius and its sector moderately narrow; sector cleft into 8 branches; media forked very near the base, its 306 = Handlirsch—New Paleozoic Insects from the anterior branch forming a long fork with, in turn, forked branches in the right wing, a simple shorter fork in the left ; posterior main branch of the media simple; cubitus provided with 1 forked and 4 simple branches in the right, with 6 simple branches in the left wing. Anal area containing 6 veins. Holotype in Peabody Museum, Yale University, Cat. No. 24. Spaniodera lata, new species. Fig. 9. Anterior wing 30" long and comparatively broad. Sector very far removed from the radius, split into only 2 branches. Medial vein branched near the base; in the right wing the Fic. 9. a5 — (rer aaa] She DLE ! eras y hip | | A, ae | I | Sai GS-2 Fic. 9. Spaniodera lata. x 2°6. anterior, in the left wing the posterior of these two main branches is forked. Cubitus giving rise to 4 branches, the first of which is forked in the left, the second in the right wing. Costal and subcostal area filled by numerous oblique cross veins ; the other cross veins far removed. Anal area showing about 7 veius of a peculiar specialization. Holotype in Peabody Museum, Yale University, Cat. No. 25. Vicinity of Mazon Creek, Lllinois. 307 Spaniodera elatior, new species. Fig. 10. Prothorax slender, measuring 8™. Anterior wings 38". Subcosta strongly shortened, scarcely reaching the outer half ; gry renes Ns of the wing. Cross veins in the costal and subcostal area much Fie. 10. Wie. 10. Spaniodera elatior. x2. less numerous. Sector far removed from the radius, split into 3 twigs. Media cleft near the base into an anterior forked and a posterior simple.branch. COubitus with 4 branches, the third of which is forked in the right wing. Anal area filled by 7-8 simple veins. Holotype in Peabody Museum, Yale University, Cat. No. 26. 308 Handlirsch—New Paleozoic Insects from the Spaniodera schucherti, new species. Fig. 11. Anterior wing quite elliptical, not so distinctly enlarged in the apical half as in the other species, 30" long. Sector twice forked. Medial vein cleft at a greater distance from the base, much nearer the middle of the wing, its twigs bent distinctly downward to the posterior margin of the wing, the anterior Hre. 11. Fie. 12. Rs Fic. 11. Spaniodera schucherti. x 2. Fie. 12. S. acutipennis. x 2°d. one split into 3 twigs, the posterior simple. Cubitus somewhat shortened, with but 4 simple branches. Avnal area less slender, provided with 7 simple veins. Holotype in Peabody Museum, Yale University, Cat. No. 27. Spaniodera acutipennis, new species. Fig. 12. Length of the anterior wing 33™". Avex distinctly pointed. Sector twice forked. Media cleft near the’ base, its anterior branch with a terminal fork, its posterior branch simple. Cubi- tus shortened, with but 3 branches. Only a few anal veins. Holotype in Peabody Museum, Yale University, Cat. No. 28. Vicinity of Mazon Creek, Illinois. 309 Spaniodera parvula, new species. Fig. 13. The smallest of all known species; the wings are only 22™" in length. Prothorax 5™™. Sector forming only a simple fork. Anterior branch of the media trichotomous, posterior branch simple. Cubitus long, with 5 branches. Posterior leg very slender and comparatively long. Holotype in Peabody Museum, Yale Univer sity, Cat. No. 29. Fie. 18. Fie. 14. ai Fie. 138. Spaniodera parvula. x3. Fie, 14. S. angusta. x1°9. Spaniodera angusta, new species. Fig. 14. Of a very slender shape. Length of the anterior wing 38™", of the pronotum 7™™. Radial sector split into 3 or 4 twigs. Media cleft very near the base. Holotype in Peabody Museum, Yale University, Cat. No. 30. Note-——Some of the species described by Scudder and Melander may possibly belong to the genus Spaniodera. This 310 = Handlirsch—New Paleozoic Insects from the would be the case with Propteticus infernus Scudder, if the drawing of the subcosta should prove to be wrong. The same thing is true of Camptophlebia clarinervis Melander and Paracheliphlebia extensa Melander, perhaps also of Miamia bronsoni and Dieconeurites rigidus Scudder. All these forms need careful reéxamination. DigrconruRA Scudder. Dieconeura mazona, new species. Fig. 15. | Body conspicuously long and slender. The 3 segments of the thorax nearly equal in length; prothorax narrow, about Fie. 15. Fic. 15. Dieconewra mazona. x 2°5. one and a half times as long as broad, 3°5™™ long. The head seems to have been short, with strongly vaulted lateral eyes. Abdomen narrower than the thorax, the segments becoming shorter from the first to the fourth, then increasing to the eighth and ninth. Apex of abdomen extending beyond the wings. . Vicinity of Mazon Oreek, Lltinors. 311 Anterior wing slender, 25" in length, its apical border obliquely rounded. Subcosta shortened, united at its apex with the radius and with the costa. Radius reaching to the apex of the wing, not split; its sector rising at one-fourth of the wing length, giving off 3 simple and arched branches. Media cleft about in the middle of the wing, its anterior main branch uniting for a short distance with the sector, then remov- ing again and forming a terminal fork. The posterior main branch of the media simple. Cubitus strongly S-shaped, with 3 or 74 simple branches, bent to the posterior margin. Anal area long. In the posterior wing the sector has 4 simple branches. The media is a long fork, quite independent of the sector. Holotype in Peabody Museum, Yale University, Cat. No. 31. Family SCHUCHERTIELLIDAL, new family. This family is based on a rather incomplete anterior wing, undoubtedly belonging to the Protorthoptera, but differing essentially from all the other types of this order. Protokol- laria alone affords a comparison. SCHUCHERTIELLA, new genus. Schuchertiella gracilis, new species. Fig. 16. A right anterior wing, reaching about 30™ in length, only 19™™ of which is preserved. The form of the wing seems to Fie. 16. Schuchertiella gracilis (negative of right wing). x3. have been an elliptical one, with slightly arched anterior mar- gin. Costal area moderately broad, the subcosta very well developed, united with the marginal costa by oblique cross veins and certainly not very much shortened. Radius nearly parallel to the subcosta and no farther removed than the sub- costa from the costa. Radial sector rising at about a quarter of the length, splitting about the middle of the wing, probably 312 Handlirsch—New Paleozoic Insects from the into a few branches only. Medial vein cleft somewhat after the origin of the sector into 2 main branches, which probably were forked again in the apical half of the wing. The media is followed by a long vein, slightly curved down toward the posterior margin and forked somewhat before the middle of the wing. This vein may be either the whole cubitus or the anterior branch of it and is attached to the media by an oblique bridge which we must suppose to be a proximal branch of the media, uniting for but a short space with the cubitus. This interpretation being correct, it would tend to reclaim the fol- lowing richly ramitied vein for the cubitus, for it is not to be assumed that the cubitus could be reduced to a simple vein. In the other case it would be possible to designate the above mentioned richly ramified vein as the first anal vein. As un- doubted anal veins we must acknowledge the 3 oblique veins provided with terminal forks, which close the venation at the posterior angle of the wing. I see no cross veins but many small folds, which induces me to suppose that this wing was a delicate and membranous one. Holotype in Peabody Museum, Yale University, Cat. No. 32. Family GERARIDA Handlirsch. The Yale collection contains several new forms evidently belonging to this family. The examination of these leads me to make some corrections in my first description. The prothorax is not, as I supposed, short and stout, but on the contrary forms a long neck, which I previously had held for a part of the head. The new material shows that this fam- ily has no close relationship to the Cidischiide, approaching rather to the Spanioderide, from which it principally differs by the much more expanded radial sector and the much more reduced cubitus. The subcosta shortened, but discharging into the costa, not into the radius. Costa marginal. GERARUS Scudder. Prothorax with a broad base, either provided with tubercles or smooth, but in every case produced into a long, neck-like part bearing the head. Gerarus latus, new species. Fig. 17. Anterior wing elliptical, nearly three times longer than broad, measuring 44™™. Subcosta not conspicuously shortened, with numerous oblique cross veins, and equally far removed from the costa and from the radius, the latter nearly reaching the apex of the wing. Sector rising very near the base, moving Vicinity of Mazon Creek, Illinois. 313 far away from the radius and sending obliquely toward the apical border 6 partly simple, partly forked branches, the first of which rises quite in the basal half of the wing. Medial vein independent, not meeting with the sector and branched in a singular manner: beginning from the base, we first find 2 Fic. 17. Fie. 18. | \ N\\ A Fie. 17. Gerarus latus. x1°8. Fie. 18. G. collaris. x 1:4. short branches directed backward and not reaching the mar- gin; the stem then bends obliquely to the posterior margin, emitting obliquely forward 4 branches which retain the primi- tive direction of the media. Cubitus comparatively little expanded, scarcely reaching the apical half of the wing and giving rise to but 3 short branches which are directed back- ward. The long and pointed anal area may have been pro- vided with-about 5 veins. Traces of the long neck are to be seen. Holotype in Peabody Museum, Yale University, Cat. No. 33. Am. JouR. Sct.—Fourts Serims, Vou. XXXI, No. 184.—Aprin, 1911. 22 314 Handlirsch—New Paleozoie Insects from the Gerarus colluris, new species. Fig. 18. This form seems to be very nearly related to G@. longus Handlirsch and may possibly be but a variety of this species. The anterior wings have a length of 52™™. Subeosta strongly shortened, extending only a little distance in the apical half of the wing and provided with numerous oblique veinlets very regularly arranged. Radial sector differing somewhat in both fore wings; with 5 branches in the right, the first of which forms a short apical fork, the second and third trichotomous, the fourth and fifth dichotomous; and with 6 branches in the left wing, the third of these being trichotomous, the fourth and fifth simple, the sixth forked. Media cleft into a long fork, the branches of which fork again. COubitus strongly reduced, probably with not more than 1 or 2 branches. Sec- tor, media and cubitus very close together near the base. Prothorax bottle-shaped, without tubercles, more than twice as long as broad at the base. Holotype in Peabody Museum, Yale University, Cat. No. 34. Gerarus (?) reductus, new species. Figs. 19, 20. As only fragments of the positive and negative of the right anterior wing and of the left anterior and posterior wing are preserved, I cannot be positive on the point of the generic Fie, 19. Fic. 19. Gerarus reductus (negative of two left wings). x 2°2. determination of this fossil. The total length of the wing may have surpassed 40™, the breadth about 12™™. Costal margin gently curved, subcosta shortened, cross veins of the costal area forming a network near the base, elsewhere far remote, simple and oblique. Radial sector rising very near the base, not very far removed from the radius and provided with about 6 very regular simple branches directed obliquely backward and outward. The media is cleft near the origin of the sector into 2 main branches, the anterior of these forming a large fork, the posterior cleft no farther. There are to be seen but two branches of the cubitus, the anterior of which is Fie. 20. Fie. 20. Gerarus reductus (negative of right fore wing), x22. united to the media by an S-shaped twig. ‘The cross veins are straight between subcosta, radius and sector, waved and some- times ramified in the medial and cubital region. Holotype in Peabody Museum, Yale University, Cat. No. 35. Gerarus longicollis, new species. Fig. 21. Prothorax 14™" in length, provided in the basal part with 2 lateral tubercles and with another situated in the middle of the hind margin; the neck-like slender anterior portion of the prothorax conspicuously thinand long. Head evidently broader than the neck. Anterior wing 386" long. Costal area broad, subcosta extending about two-thirds of the wing. Radius conspicuously far removed from the costa in the apical third. Sector rising at a quarter of the length, first diverg- ing from and then converging with the radius; of its 4 branches the first is forked. Media cleft somewhat below the origin of the sector, its posterior main branch simple, the anterior forked into 2 twigs, the anterior of these uniting for a short distance with the radial sector. Cubitus free, giving rise to probably but 2 branches directed backward. The posterior wing being very fragmentary, it is difficult to distinguish media and cubitus, but it is possible to see by the different arrangement of these veins that there was present a large anal lobe. Holotype in Peabody Museum, Yale University, Cat. No. 36. GERARULUS, new genus. Evidently related to Gerarus but distinguished by the strongly expanded apical half of the anterior wing and by the radius being branched before the end. Costa marginal ; sub- costa shortened, united with the costa. The space between the sector and the radius conspicuously enlarged, sector with 316 Handlirsch—New Paleozoic Insects from the numerous branches, media cleft into 2 ramified main branches. Cubitus of the anterior wing cleft near the base into 2 forked main branches, the anterior of which unites for a short dis- Fie. 21. Fie. 21. Gerarus longicollis. x18. tance with the posterior branch of the media. Anal area of the anterior wing not distinctly limited, with but few partly ramified veins. Posterior wing with an anal fan. Gerarulus radialis, new species. Fig. 22. Length of the anterior wing about 30™™ (24"™ only being preserved). Costal area comparatively narrow, extending about two-thirds of the wing and filled by simple and very regularly distributed oblique cross veins. tadius very near the subcosta, provided in the apical third with branches Vicinity of Mazon Creek, Lllinois. 317 directed toward the costa. Sector rising at about a quarter of the length, widely diverging and giving off about 9 very regular branches, the first of which is forked. Media cleft into 2 main branches, the anterior of which closely approaches the sector. Each of these main branches splits into 3 twigs. Cubitus cleft immediately after the origin, its long anterior branch touching for a moment the media and then splitting Fie. 22. Fie. 22. Gerarulus radialis. x2°2. into 2 twigs, the posterior branch forming a long fork. First anal vein sinuated, with 2 twigs directed ‘backward. In addi- tion, there are ? 3’short and simple anal veins to be seen. In the posterior wing the radius is ramified in a similar man- ner to that in the front wing. ‘The sector rises very near the base, does not diverge so far and gives off its branches much nearer the base. Media cleft very near the base into 2 rami- fied branches. Cubitus with a simple anterior branch approach- ing the media and with a waved posterior branch splitting into a few twigs. Anal veins simple and straight. Holotype in Peabody Museum, Yale University, Cat. No. 37. ANEPITEDIUS, new genus. The following species, which I consider the type of ies new genus, is likewise very closely related to Gerarus and shows a slender pear-shaped prothorax, forming a long neck. The long wings, with a strongly shortened subcosta, meeting with the costa. The radius in its apical portion sends forth some ramified S-shaped branches, directed forward. The sector rises near the base and appar ently gives off but a few branches. 318 Handlirsch—New Paleozoic Insects Srom the The media likewise forms only few branches and unites in the anterior wing with the sector by a short cross vein but remains quite independent in the posterior wing. Anal area of the posterior wing enlarged and evidently fan-like. Fie. 23. Fic. 28. Anepitedius giraffa (negative). x 2°2. Anepitedius giraffa, new species. Figs. 23-25. Prothorax 14" long, somewhat more than twice as long as broad at the base, produced into a neck-like form, traversed by a ridge and evidently depressed on the sides of the basal half. Vicinity of Mazon Creek, Illinois. 319 Wings long and slender, measuring about 40™, 33™" of which is preserved. By the superposition of all the four wings, the venation becomes very confused and difficult to decipher. Subcosta scarcely extending below the middle of the wing. Costal area comparatively narrow, traversed by about 8-10 regular oblique cross veins. Radius at the same distance from the subcosta as the latter is from the costa and giving rise to 2 S-shaped forked branches, directed toward the Fie. 24. Fic. 24. Anepitedius giraffa (anterior wing). x 2°2. costal margin. Sector rising near the base and separated from the radius by a broad interstice, traversed in the anterior wing by oblique, in the posterior by straight cross veins. The sec- tor seems to have had but few branches. Media cleft in the anterior wing not far above the middle, its anterior branch Fie. 25. Fic. 25. Anepitedius giraffa (posterior wing). x2"2. advancing in a blunt angle until near the sector, uniting with it by a short and strong cross vein; the posterior branch is forked. In the posterior wing the bifurcation of the media seems to take place quite near the base, the anterior branch being independent of the sector and forking again near the end, the posterior apparently remaining undivided. Oubitus forming in the anterior wing a quite free vein cleft into 2 long branches. Of the anal veins I cannot see more than 3, which are almost parallel. Holotype in Peabody Museum, Yale University, Cat. No. 38. 320 Handlirsch—New Paleozoic Insects from the Family APITHANID AL, new family. A very singular slender insect, differing from all other groups of the Protorthoptera, induces me to establish a new family. The pear-shaped prothorax is of a moderate length; the meso- thorax somewhat enlarged, bearing the comparatively broad anterior wings, which show a singularly specialized venation and could be laid backward over the abdomen. The metatho- rax conspicuously shorter. Abdomen slender, its first segment being the shortest of all, nearly twice as broad as long, the second about quadrate, the following to the seventh much longer than broad. The end of the abdomen and the posterior wing not being preserved, I am not able to say anything regarding the appendages and the anal lobe. The (? anterior) legs were short and delicate. The venation of the anterior wing is a characteristic one: the costa marginal; subcosta shortened, united with the costa; radius with a short sector, rising outward from the middle of the wing and forming a simple fork. Media independent, cleft near the middle of the wing into 2 branches. Cubitus divided near the base into 2 main branches, the anterior strongly waved and making a short terminal fork, the posterior trichotomous. Anal area slender, reaching half the length of the wing and containing few veins. Cross veins simple and oblique in the costal and subcostal area, elsewhere evidently very rare. APITHANUS, new genus. Apithanus jocularis, new species. Figs. 26, 27. Prothorax 5™™ long, one and a half times longer than broad. Mesothorax nearly quadrate. Metathorax one and a half times broader than long. First segment twice as broad as long, shorter than the metathorax; second segment quadrate; the third to the seventh each longer than broad. The middle line of the abdomen showing tubercles on the base and apex of each segment. The femora of probably the fore legs short and delicate, doubtless not exceeding the length of the protho- rax. Anterior wings 30™", three and a fourth times longer than broad, dilated in the apical half. Apical margin obliquely rounded. Costa marginal, slightly waved. Subcosta extend- ing nearly two-thirds the length of the wing, uniting with the costa and giving off about 7 oblique veinlets. Radius very close to the subcosta in the basal third, farther outward more remote, the interstice being traversed by 5 oblique cross veins. Sector not rising before the middle of the radius, from which it is separated by a broad interstice, and forming only one broad fork. Media quite free, cleft somewhat before the Vicinity of Mazon Creek, Litinois. 321 middle into 2 branches converging toward the margin, the anterior of these branches being attached to the posterior branch of the sector by a long, oblique cross vein. The cubitus run- ning nearly parallel with the posterior branch and with the stem of the media, splitting quite near the base into an ante- rior branch forming a short apical fork and into a posterior branch cleft into 8 long branches which descend to the poste- rior margin. Anal area long and lancet-shaped, showing dis- Fic. 26. Fie. 26. Apithanus jocularis. x 2. tinctly only 2 veins, but probably there were 2 more which are ° not preserved. Holotype in Peabody Museum, Yale University, Cat. No. 39. Family NARKEMID At, new family. Anterior wings laid backward over the abdomen, broad, elliptical, with a broad costal area; much shortened subcosta uniting with the radius and giving rise like the free end of the radius to very regular oblique veinlets reaching the costa and being themselves united by cross veins. Sector rising at about a quarter of the wing length, giving off a small number of branches which are directed toward the apical border. Media independent of the radius, simple and only provided 322 Handlirsch—New Paleozoic Insects from the with a short apical fork. Cubitus very well developed, cleft near the base, the anterior main branch splitting by repeated forking into a greater uumber of twigs, one of which reaches the media, the others going parallel with the branches of the Fig. 27. Fig. 27. Apithanus jocularis. x 2°9. sector and with the media to the margin; the posterior main branch splits into a bundle of divergent twigs, falling obliquely to the posterior margin. Anal area broadly lancet-shaped, separated from the wing by an S-shaped waved vein. All the interstices between the main veins are bridged over by verti- cally arranged simple cross veins. It is a curious fact that original color patterns, consisting of a number of broad and very regular transverse bands, have resisted destruction through petrifaction. Similar patterns have been found in some fossils of Commentry, as in Cnemi- dolestes and Protophasma. Unfortunately the venation of the Cnemidolestide not being sufficiently known, it is at pres- ent impossible to suppose that any relation exists between the two groups because of their similar color patterns. NARKEMA, new genus. Narkema teniatum, new species. Fig. 28. Front wing elliptical, scarcely two and a half times as long as broad, about 43™" in length, 37™™ of which is preserved. Costal area very broad, subcosta extending but little beyond the middle of the wing, uniting with the radius. Radius not far removed from the subcosta, not reaching quite to the tip of the wing and sending forth, like the subcosta, a num- ber of very regular oblique and straight veinlets, directed for- ward and bridged over by delicate cross veins. Vicinity of Mazon Creek, Illinois. 328 The sector rises at a quarter of the length, runs parallel to and not far from the radius and sends off 2 simple and 1 forked branch, directed toward the margin. The media passes near the radius and sector without meeting with them and forms a great and wide bow, only forking near the end. A great deal of the wing is occupied by the cubitus, cleft near the base, the anterior main branch splitting into 2 twigs, the posterior end of which is not divided and attains the posterior margin, while the anterior approaches the media, sending a small branch to this vein, and then regaining the original direc- Fig. 28. Z FEEL Lat Fic. 28. Narkema teniatum. x 2. tion splits into 8 twigs, the first and third of these forming apical forks. ‘The posterior main branch splits into 3 divergent twigs, the hindmost of which again splits into 3 twigs. The first anal vein being gently S-shaped, allows us to suppose that. there was a lancet-shaped anal area, provided with a small number of veins. Oross veins have been present in a rather great number; they are vertically situated on the main veins. Both wings are equally decorated by 7 transverse bands ex- tending across the whole wing without regard to the venation. Holotype in Peabody Museum, Yale University, Cat. No. 40. Family CACURGID, new family. In this family, which I propose provisionally, are placed forms regarding which I cannot with certainty discern whether they are Protorthoptera or Protoblattoidea. They all agree in the sector rising at a greater distance from the base and in the presence of a short oblique vein, going from the stem of the media to the cubitus and forming a sort of basal cell. This vein evidently is a branch of the media, which unites with the cubi- tus for the entire distance or for a certain space only. The 324 Handlirsch—New Paleozoic Insects from the cubitus being separated from the anal area by a comparatively large space occupies a great part of the wing. Costa marginal. Cross veins irregular, oblique and often ramified, the ram- ifications of the main veins never being very rich. I consider as the type of this family a new genus, Cacurgus, which is the best preserved of all. Besides this another new genus, Splomastax, is placed here, and of the previously described forms, Palwomustax Handlirsch from the Westpha- lian of Belgium, Avchimastax Handlirsch from Fayetteville, Arkansas, and Archwologus Handlirsch from Mazon Oreek. These latter I had described as genera of doubtful position. It is not impossible that Awiologus thoracicus Handlirsch may also belong in this family. CacurGus, new genus. Cacurgus spilopterus, new species. Figs. 29, 30. Both anterior wings are preserved, partly crossed over the abdomen in a quite natural position, so that there can be no Fic. 29. Fic. 29. Cacuwrgus spilopterus (right fore wing). x 1°29. doubt that they belong to the same individual. Nevertheless the venation seems to be rather different in both wings. The form of the wing probably has been an elliptical one, two and a half times longer than broad, measuring about 75™™ in length. Costa marginal. Costal area broad, with numerous oblique branches from the subcosta, bridged over by cross veins. Subcosta attaining three-quarters the length of the wing and uniting with the costa. Radial sector rising some- what before the middle and splitting into but 3 branches. The radius itself gives off 2 strongly ramified branches, Vicinity of Mazon Creek, Illinois. 325 directed toward the costal margin. The media in its basal part bends closely toward the radius and is cleft before attain- ing the end of the first quarter of its length. Its anterior main branch continues in a normal way and splits in the apical half; the posterior has the aspect of an oblique cross vein uniting the media with the cubitus. At a certain distance from this point of fusion we see that in the right wing a new separation of the two veins takes place, the branch of the media forming a large terminal fork. In the left wing there is but a temporary separation of the fused veins, which then unite again and remain so to the end. Nevertheless it is pos- sible, but not determinable, that the last fork of the cubitus may take rise from this second medial branch. In this case we would be forced to assume a strong reduction of the cubitus in the left wing in comparison to the right, where this vein Fic. 30. Fie. 80. Cacwrgus spilopterus (left fore wing). x 1°25. with its 8 or 9 partly forked branches occupies the whole hind margin. There being only 8 branches in the left wing, their number would be reduced to 6 by counting the apical fork as belonging to the medial vein. The anal area is broadly lancet-shaped, limited by an S-shaped vein and seems to have been filled by only a few veins of a similar curve. Jn all the broader interstices the cross veins are twisted into a moderately regular and polygo- nal network. Irregularly spread about there are to be found in both wings shallow, more or less regular, round bodies of different sizes, situated in darker patches of the wing mem- brane and showing a small groove in the middle. These structures give the impression of membrane thickenings. I probably should have regarded them as extraneous and not pertaining to the insect if they had been observed only in this 326 Tlandlirsch —New Paleozoic Insects. ease, but as they are also present in the wings of the related form Spilomastax, I think they will prove to be organie strue- tures of these insects. Dr. Reis has found similar but smaller structures on the wings of one of the Protodonata of Triassic age. Tubercles on the membrane have also been found in one of the Palsodictyoptera. Holotype in Peabody Museum, Yale University, Cat. No. 41. SPILOMA STAX, new genus. Spilomastax oligoneurus, new species. Fig. 31. Two fragments evidently appertaining to the left and right anterior wing of a comparatively small insect, whose wings can be estimated at 30-35". Subcosta far removed from the marginal costa and provided with regular oblique branches. Radius giving off its sector at about the middle. Media send- ing the typical short and oblique branch to the eubitus at about a quarter of its length and split into a large fork before attain- ing the middle of the wing. COubitus cleft immediately after its Fie. 31. Spilomastax oligoneurus. x2. origin, the anterior branch waved, sending 4 branches obliquely outward and backward, the first 2 of them not reaching the margin; but the posterior main branch running straight and undivided to the end. First anal vein gently curved, second swung forward toward the first. Cross veins irregular, here and there ramified but not twisted into a polygonal network. Both wings show numerous quite irregularly distributed shallow round bodies similar to those in Cacurgus. Holotype in Peabody Museum, Yale University, Cat. No. 42. [To be continued. ] A. Hollick—Kenwi Flora of Alaska. 327 Arr. XX1IX.— Results of a Preliminary Study of the so-called Kenai Flora of Alaska; by Arraur HoxxtoK.* Introduction. Durie a recent preliminary study of a series of collections of fossil plants from Alaska Peninsula, especially from the vicin- ity of Herendeen Bay and Chignik Bay, tentatively assumed, for the most part, to represent the flora of deposits to which the name Kenai formation has been applied, a number of facts were brought to light which are of considerable biological interest and which may prove, after critical analysis, to be of vaiue from the standpoint of stratigraphy. The name Kenai was originally applied to a series of beds exposed in southeastern Alaska, particularly on the shores of Kachemak Bay, Kenai Peninsula,t which at that time were regarded as Miocene or Oligocene in age but are now generally recognized as Kocene. The results attained from the recent studies indicate that, if the use of the name is to be restricted to the beds of the type locality and their equivalents elsewhere, there is also a series of beds, more or less closely associated with them stratigraphically, which may or may not be included in the formation. The ultimate inclusion or exclusion of these latter, either in whole or in part, cannot be determined, how- ever, until all of the paleobotanical evidence has been carefully weighed and compared with whatever stratigraphic observa- tions may be available. For example, some of the collections contain only Tertiary species. In others, often from the same localities, there is a preponderance of Tertiary species and a minority of Cretaceous. In others the majority are Cretaceous, with certain genera which are identical with those of the Jurassic. The several collections appear to merge into each other without any abrupt break in the paleobotanical sequence, and further careful investigation will be necessary before attempting to draw any stratigraphic line intended to indicate a differentiation of the flora into one undoubtedly of Tertiary and another of Creta- ceous age; but the facts may, at least, be described and their biological significance discussed. Description of the Flora. The flora represented in the collections thus far studied, if regarded as a unit, is unique so far as North America is con- cerned. There is none other described from either the United * Published with the permission of the Director of the U. S. Geological Survey. + Dall and Harris, Bull. U. S. Geol. Surv., No. 84 (Correlation Papers, Neocene), p. 233, 1892. 328 A. Hollick—Kenai Flora of Alaska. States or Canada with which it may be satisfactorily correlated. It contains many species which are identical with those of cer- tain well-defined Teitiary horizons in the United States and, if these species were the only ones represented, the Tertiary age of the beds in which they occur would not be questioned : such species, for example, as Zaxodium distichum miocenum Heer, Populus arctica Heer, P. Richardsoni Heer, Corylus McQuarrii (Forbes) Heer, Carpinus grandis Ung., Betula Brongniartii Heer, Paliurus Colombi Heer, ete. Associated with these, however, not only in collections from a single locality but frequently in the same pieces of matrix, are species which elsewhere occur in strata of recognized Ore- taceous age, such as Adiantum formosum Heer, Sequoia rig- ida Heer, Sagenopteris elliptica Font., ete., and others which, if found by themselves, would almost certainly be considered as closely allied to certain Jurassic species, such as Pterophyl- lum concinnum Heer, Anomozamites Schmidtit Heer, Vils- sonia comtula Heer, etc.! When first examined ,it was thought that this association of undoubted Tertiary angiosperm species with apparently Mesozoic types of gymnosperms was impossible, and that col- lections from different geologic horizons must have somehow become mixed; but the fact that some of these diverse floral elements were often included in the same piece of matrix proved that they must have been synchronous. Discussion of the Flora. Unless all of our previous knowledge and experience in rela- tion to the beginning and subsequent evolution of the angio- sperms is at fault, it is evident that the presence of highly developed angiosperm species in any flora at once precludes the possibility of regarding it as Jurassic inage. Furthermore, the fact that certain of the angiosperms in the flora under con- sideration are undoubted Tertiary species makes it imperative to regard the apparent identity of certain of the associated cycads with Jurassic species as untenable and to regard such apparent identity as due to superficial resemblances only. The theory that a specific type could persist throughout such a great length of time as that implied could hardly be accepted on evidence based entirely upon such inconclusive factors. The genera in which the species belong, however, are unques- tionably identical with Jurassic generic types; but it is more logical to assume that such types could have continued into Tertiary times than to imagine that highly developed angio- sperms could have been in existence in the Jurassic period. While searching through paleobotanical literature for any possible description of, or reference to, a flora similar to ours, A. Hollick—Kenai Flora of Alaska. 329 three papers were noted which have a bearing on the subject. Saporta* describes the discovery of a single species of a cycad, Zamites epibius Sap. (2. ¢. p. 322), in the middle Tertiary of Provence, France, associated with many of the same angio- sperm genera as those represented in the Alaskan flora. ‘He describes and figures it again in his “Wtudes sur la Végétation du Sud-Est de la France a PEpoque Tertiare, Part III eit compares it with Z. formosus Heer from the Jurassic of Switzerland, and says (J. ¢ p. 11):“....Zamites epibius, despite its analogy with the Jurassic genus, must necessarily be specitically distinct, especially when one realizes the vast time interval which separates them.” The interest and impor- tance which Saporta attached to this single specimen may be inferred from the lengthy discussion which he gives to it in each paper. lt is possibie that a small fragment in one of our collections, evidently a Zamites, may belong to Saporta’s spe- cles. The third paper is by Heer,t in which he describes and illustrates a Tertiary flora almost identical with ours, from the island of Saghalien, in northeastern Asia, where the same asso- ciation of angiosperms and cycads occur and in regard to which he remarks (/. ¢. p. 9): “The most striking....is the family of the Cycads..... There are two species, which dif- fer widely from all living ones, but which show a striking and unmistakable identity with Jurassic and Rhaetic forms.’ Two species of Wilssonia are described and figured (1. sero- tina and WV. pygmea). The former species is undoubtedly represented in our collections by a number of specimens, and it is interesting to note that one of his figures (J. ¢. fig. 1, pl. IT) depicts this species associated with a leaf of Populus arctica Heer on one and the same piece of matrix: an association of species which is duplicated in several of the fragments of matrix in the Alaskan collections. Significance of the Facts. To those who are familiar with the factors which influence the distribution of our living flora, the presence of cycads, representing a tropical type of vegetation, associated with spe- cies of Populus, Corylus, Carpinus, Betula, Juglans, ete., in far northern latitudes, will appear mcongruous; but the fact that such an association existed in those regions in the Terti- ary period cannot be questioned. It is evident, of course, that cycads must have continued their existence somewhere through- out both Tertiary and Quaternary times, otherwise they would * Bull. Soe. Géol. France, Ser. II, xxi, 314-328, pl. 5, f. 1-3, 1864. + Annales Sci, Nat., Ser. V (Bot.), viii, pl. 1, £. 1, 1867 t Fl. Foss. Arct., v (Mioc. Fl. Sachalin), 1878. Am. Jour. Sct.—Fourts Srmrigs, Vou. XX XI, No. 183-—Aprin, 1911. 23 330 A. Hollick—Kenai Flora of Alaska. not be represented in our living flora. Nevertheless, so far as the paleontologic record indicates, they apparently disappeared from the southern and central parts of the North American continent in the Tertiary period and were almost exterminated in similar European latitudes at the same time, but continued to exist in northwestern America and northeastern Asia, until their descendants were in part exterminated and the remainder driven southward by the advancing cold of the Quaternary period to where they are now growing. Their present range of distribution in the New World is between northern Mexico and Bolivia, which affords an approxi- mate indication of the possible extremes of climatic conditions which might have prevailed in Alaska at the time when this flora was growing there. The climate could not have been colder than that of northern Mexico or southern California at the present time, if the cycads are to be regarded as adequate climatic indicators; nor could it have been much warmer, if the associated angiosperm genera are to be regarded in the same light. The logical inference is, therefore, that the climate which was synchronous with this Alaskan flora was about the same as that of southern California and Florida at the present day. We may also be warranted, apparently, in assuming that, at the time when this flora flourished, either the climate of the northern Pacific region was warmer than that which prevailed in the mid-continental areas farther to the south, or else that their meteorological conditions were not identical, thus giving rise to floral differences similar to those which prevail at the present day in the coastal and interior regions of the West. Finally, the identity of the Tertiary floras of northwestern America and northeastern Asia is confirmatory evidence of a former land connection between the two continents in recent geologic times, which is so strongly indicated by the well recog- nized physiographic and topographic features.* * See ‘‘The Probable Tertiary Land Connection between Asia and North America,” Adolph Knopf. Univ. Calif. Pub., Bull. Dept. Geol., v, 413- 420, 1910. Chemistry and Physics. 331 SCIENTIFIC INTELLIGENCE. I. Cwemisrry And Puysics. 1. Use of Calcium Carbide for the Determination of Moisture. —This application of calcium carbide has been discussed by Masson. Difficulties occur in the ordinary methods of determin- ing moisture from the fact that other vapors besides water may be given off by the substance, or that the substance is oxidized and gains weight when heated in the air to expel its moisture. Calcium carbide has the advantage that water is the only ordi- nary substance that will react with it, and the acetylene pro- duced by the reaction can be readily measured. Hygroscopic organic substances are mixed with an excess of finely powdered carbide, when the reaction takes place quickly without artificial heating, and the gas evolved may be measured or its pressure determined by means of a manometer. The method has been applied for the determination of moisture in wool, explosives, petroleum, etc. The author has studied the application of the method to the determination of water of crystallization in salts, and finds that salts may be roughly divided into four classes in respect to their behavior with calcium carbide. In the first are those, such as sodium carbonate and sodium sulphate, which react at once and completely. In the second class are those that react rapidly and completely only on heating, such as barium chloride. The third class, which includes such salts as copper sulphate and the alums, react, either in the cold or on heating, in such a way as to lose only a part of their water, leaving a hydrated residue which belongs to the fourth class, characterized by being quite stable towards carbide even at 170° C. Itisa noteworthy fact that ammonium salts do not lose ammonia, nor do crystalline acids react as such, when these are heated with the carbide and calcium hydroxide present as a result of the acety- lene production. This is to be attributed to the complete absence of free water. It is interesting to notice that various observers have found that 18™ of water corresponds to 10°5° of acetylene at standard conditions, while the theoretical volume is, of course, 11:2°°.— Chem. News, ciii, 37. H. L. W. 2. Action of Water upon Phosphorus Pentoxide-—Various statements are found in the literature concerning the products of this reaction, that is, whether metaphosphoric acid, HPO,, the pyro- acid, H,P,O,, the ortho- acid, H,PO,, or mixtures of these are first formed. BatarerF has now investigated this matter by allowing pure sublimed P,O, to deliquesce very slowly, and more rapidly in more or less moist air, and also by throwing it directly into water. In every case where the product was examined imme- diately after deliquescence, only metaphosphoric acid, HPO,, could be detected. In about 18 hours some of these products 332 Scientific Intelligence. were found to be completely converted into H,PO,, and the remarkable fact was noticed that HPO,, produced from P.O, as described, becomes hydrated to H,PO, much more rapidly than the same compound which has been produced by the dehydra- tion of H,PO, by heating.—Zeitschr. anorgan. Chem., \xix, 215. H. L. W. 3. The Fractional Crystallization of Argon.—According to its position in the periodic system of the elements, argon should have an atomic weight of less than that of potassium, 39°10, instead of 39°88, the atomic weight now attributed to it. FRANZ Fiscurr and Vicror FRopoxrs® have attempted to fractionate it by freezing the liquid. Since argon has been purified previously by fr actional distillation, they reasoned that it might form a Gon- stant boiling mixture w ith some unknown i impur ity, as is the case, for instance, with hydrochloric acid and water. However, the results of fractional crystallization were negative, so that no evi- dence was found that the atomic weight is not exceptional, like that of tellurium, in the periodic system. — Berichte, xliv, 92. H, L. W. 4, Qualitative Chemical Analysis ; by BASKERVILLE and Curt- MAN. 8vo, pp. 200. New York, 1910 (The Macmillan Company). —Nearly every teacher of qualitative analysis writes a book on the subject, because there are many variations in the course of analysis and also many ways of presenting the subject. The book under consideration appears to be a very good one. There is a suitable amount of description and theoretical matter, and equations are very fully supplied. The analytical processes, which are given in tabular form, with numerous notes, are in general well selected. An attempt is made to have the student distinguish qualitative reactions in a somewhat quantitative way by comparing his tests with tests of measured solutions of known strength. The authors say in regard to this plan, ‘“‘ The value of such training cannot be overestimated. Our students rarely find any difficulty in differentiating between a trace and a significant amount.” The last sentence is somewhat disappointing when compared with the preceding one, for as much might be expected of the student even without the special method of comparison. H. L. W. 5. Die Verwertung des Luftstickstoffs ; von Prof. Dr. J. ZENNECK. 8vo, pp. 29. Leipzig, 1911 (Verlag von S. Hirzel).— This pamphlet gives a recent lecture on the production of air- saltpeter by the aid of the electric arc. The principles of the method and a description of the apparatus are described in a very interesting way, with the aid of many excellent illustrations. The development of this new industry in connection with Nor- wegian water-power is of the greatest significance for the ad- vancement of agriculture and other industries. It will be recalled that attempts to produce nitric acid commercially from the air were made in the United States anumber of years ago, apparently without complete success. H. L. W. — ‘Geology and Mineralogy. - 333 6. Allen’s Commercial Organic Analysis. Volume IV. 8vo, pp. 466. Philadelphia, 1911 (P. Blakiston’s Son & Co.).—This, the fourth volume, of the entirely rewritten fourth edition of “ Allen,” treats of Resins, India Rubber, Rubber Substitutes and Gutta-Percha, Hydrocarbons of Essential Oils, and Volatile or Essential Oils. The importance and excellence of the work are so well known that no comments in regard to details seem to be necessary, except the statement that the present volume appears to maintain the high standard of the preceding ones. H. L. w. 7. The Absorption Spectra of Solutions ; by Harry C. Jones and W. W.Srrone. Pp. 159 with 98 plates. Publication No. 130, Carnegie Institution of Washington, 1910.—This investi- gation is a continuation of the work of Jones and Uhler which was begun in 1905 and subsequently greatly extended by Jones and Anderson. (See Carnegie Publications Nos. 60 and 110; also this Journal, vol. xxviii, page 78.) The first chapter of this monograph is devoted to a brief but excellent classification of the various types of spectra, to some of the theories of spectra, and to the methods best suited for study- ing specific spectroscopic problems. Chapter II deals with the apparatus used in the present investigation. The remaining ten chapters give in detail the variations in the absorption spectra of solutions of certain salts of potassium, cobalt, nickel, copper, chromium, erbium, praseodymium, neodymium, and uranium which are caused by chemical agents and by changes of temper- ature. The quantity of work done is very great, since about 3000 solutions were investigated. The quality of the work is of the first class, as can be readily seen both from the text and from the 98 plates. The results obtained are too numerous to be recorded in a brief review. Suffice it to say that well-defined “ solvent- bands” have been discovered for water, the alcohols, acetone and glycerol, and that all of the results seem to confirm the hypoth- esis of solvation as developed and emphasized by Jones. Ei SsUr II. Grorocy anp Mrneraroey. 1. Indiana. Department of Geology and Natural Resources. Thirty-fourth Annual Report. W. 8. Brarcuney, State Geol- ogist. 1909. Pp. 392, with numerous maps, photographs, tables, ete. Indianapolis, 1910.—The soil survey of Indiana, begun in 1907, was continued during 1909 by A. E. Taylor and C. W. Shannon, and the statement of the important results of their work fills the larger part of the present volume. During the three seasons in which the work has been carried on, thirty-three counties in the southern part of the state have had their soils classified, mapped, and treated in detail. It is proposed to go on until the entire State has been similarly treated. The Survey 334 - Scientific Intelligence. during 1909 also gathered detailed data in regard to undeveloped water-power sites of the larger streams ; this work has been under the charge of W. M. Tucker. The volume contains, further, a report in regard to the natural gas, by B. A. Kinney, and another by the State Inspector of Mines, J. Epperson. 2. West Virginia Geological Survey. Volume Five, Forestry and Wood Industries ; by A. B. Brooxs. I. C. Wuire, State Geologist. Pp. xiv, 481, with an envelope map of West Virginia and numerous illustrations. Morgantown, 1910.—This fifth vol- ume of the West Virginia Geological Survey, carried on under the guidance of Professor I. C. White, is devoted to a discussion of forestry by A. B. Brooks. This is a subject of vital interest at the present time, and one which fortunately is attracting the attention of those whose political position is such as to enable them to improve the existing situation. In the early days, the State was covered by an almost unbroken forest of more than fifteen and a-half million acres in extent. Now the virgin forest has been reduced to one and a-half million acres. It is estimated that if the large number of saw mills engaged in the State were to continue in active operation as at present, they would cut all the timber in a little over sixteen years. This outlook is sufti- ciently serious ; but it is aggravated by the very large waste that comes in in various ways, particularly through loss by fire. The volume by Mr. Brooks discusses the subject thoroughly from vari- ous standpoints, showing the great importance of the forests in their indirect relations, as well as a source of usable timber. An interesting chapter is devoted to the agents destructive to trees, including a great variety of destructive fungi and insects. A summary is given of the forestry work now being carried for- ward in some twenty States; an exhaustive list of the native trees is also added. The volume is made attractive by a large number of excellent reproductions of photographs. The Survey has also recently issued a map of oil and gas fields, and the structural contour of Wood, Ritchie and Pleasants Counties. 3. Geological Survey of Tennessee; Grorcn H. Asutey, State Geologist.—The State Geological Survey of Tennessee was established in 1909. The special purpose, scope and methods of the work planned are detailed by the State Geologist in a pam- phlet, 1—-A, of 33 pages, extracted from Bulletin No. 1, on Geo- logical work in Tennessee. Other pamphlets, recently issued, are 2—A, giving an outline introduction to the Mineral Resources of Tennessee, and 3—A, on drainage problems ; these are taken from Bulletins No. 2 and No. 3, respectively. 4. Atlas Phographique des Formes du Relief Terrestre.—At the ninth Geographical Congress, held in Geneva in 1908, it was voted to adopt the plan proposed by MM. J. Brunhes and E. Chaix for the preparation of a collection of views, showing the different forms of terrestrial relief. A commission of eleven was appointed to report a complete plan for this enterprise at the Geology and Mineralogy. 335 next Congress and to take steps to inaugurate the work. A preliminary circular has been issued, giving the plan as thus far developed. The photographs are to be published in quarto form, as separate plates, with one to two views on each ; a leaf of the same form will accompany each plate, giving a brief expla- nation, a chart showing the point of view, etc. It is estimated that some 500 to 600 plates will be needed to print all the types of relief as classified according to the provisional scheme adopted. If, in addition to these, regional series are arranged for, the number may be increased two or threetimes. The price proposed is one frane for a single plate or half that sum when a series of 100 are ‘subscribed for. Of the various series planned the Committee wishes to go forward first with those dealing with the forms determined by tectonic conditions (faults, folds, etc.) and those connected with glacial action. Photographs relating to these two subjects are solicited and at the same time subscriptions are asked for parts of the work. Communications may be sent to one of the Executive Committee: Prof. J. Brunhes, Fribourg, Switzerland ; Prof. K. Chaix, 25 Avenue du Mail, Geneva ; Prof. Em. de Martonne, 248 Boulevard Raspail, Paris. 5. The Illinois Oil Fields in 1910 —R. 8. Buatcuitey, of the Illinois Geological Survey, has issued a circular giving an account of the production of oil in the state during 1910. The estimated amount produced for the year is 35,000,000 barrels, while that of 1909 was not quite 31,000,000 barrels. Extensive explorations by drilling have been conducted at numerous points. 6. Physical Notes on Meteor Crater, Arizona.—The interest- ing but as yet unsolved problems connected with Meteor Crater in Arizona (see vol. xxx, 427) have been discussed in a paper by W. F. Maciz in the Proceedings of the American Philosophical Society (vol. xlix, pp. 41-48). After describing the locality and the general character of the specimens found, he goes on to give the results obtained from some magnetic experiments with a cylinder of the Canyon Diablo iron. Briefly stated, these showed it to have a magnetic permeability about one-half that of a cylin- der of Norway iron. Experiments were also made with the shale balls: one of these, nine inches in diameter and entirely oxidized, was examined as it lay in the pulverized sandstone of the outer rim of the crater. It showed strong local poles over the surface, with, in general, south polarity on top and north at bottom. Another piece of a shale ball, which at the crater showed distinct polarity somewhat irregularly distributed, after arriving at Princeton in a box with other specimens, had lost its polarity and behaved like soft iron. Much of the shale is so feebly magnetic as hardly to affect a needle even close to it. The observations of Baker in 1891 showed no evidence of a local magnetic field within the crater, while the whole variation in the dip of a vertical needle was such as might. have been due to errors of observation or perhaps to local conditions, such as the presence of iron pipes in the drill holes. Whether the size of the 336 Scientific Intelligence. meteor, to which it is assumed that the crater is due, was 750 ft., or 250 ft., in diameter, it should, at a reasonable depth, seriously affect the magnetic field, and a difference of dip between extreme stations of 30’ should exist even for the smaller sized sphere. “That no such difference was found argues that the meteor was broken and scattered by the impact, or more probably, as Mr. Barringer strongly argues in his latest paper, was a cluster or swarm of small masses of iron, mostly of the shale ball variety. The possible intrinsic magnetism of these masses, coupled with the possibility that they have gradually oxidized in the depths of the crater, would account for the absence of any observed mag- netic field.” e The symmetry exhibited by the crater, both in the tilting of the strata of the rim and the distribution of ejected material, is such as can plausibly be referred to the impact of a body rather sharply inclined to the vertical. Some experiments were tried with a lead ball shot from a high-power rifle at an angle of 30° from the vertical into a level floor of smooth, densely packed silica. The distribution of ejected matter on either side of the surface was distinctly like that observed in the crater. An approximate estimate is made as follows of the energy with which the meteor may have struck the earth. The work done in the excavation of the crater involves the probable ejection of some 330 million short tons. To lift the mass up and clear of the hole would have required about 1610” foot-tons; while if allowance is made also for the work of tilting back the strata and lifting the unbroken rock masses around the rim, some 20x 10" foot-tons in all may have been required. The larger part of the energy, however, was spent in breaking up the rock and reduc- ing the sandstone grains to the condition of finely pulverized silica. It is estimated that over 500 million tons of rock were broken up, and one-fifth of this converted into the pulverized form. The work done against friction resulted in heat, but in general the temperature did not rise to a point sufficient to melt the quartz, as only a small amount of melted material is found. Taking 2500° C. as the outside limit of temperature, and suppos- ing all the silica heated to that extent, the heat developed would be equivalent to 9°25 x 10“ foot-tons. If, however, a general temperature of 625° is assumed, that is, below the point at which the Widmanstitten figures would disappear, the heat developed would be equivalent to 2°3 X 10° foot-tons. In addition to the work on the silica, a layer of hard limestone, 300 feet thick, was also broken up, and much of it pulverized. It is estimated finally that the total amount of energy expended may easily have been 60 X 10” foot-tons. The velocity of the meteor may be put somewhere between the outside limits of 3 and 48 miles per second. If 18 to 20 miles per second be assumed as the most probable velocity, it is found that a mass of 400,000 tons would have had the amount of energy estimated as necessary for the work done in connection with the formation of the crater. Geology and Mineralogy. 337 7. Minéralogie de la France et de ses Colonies. Description Physique et Chimique des Minéraux; Etude des Oonditions Géologiques de leurs Gisements; by A. Lacrorx. Vol. IV. Part I, pp. iii, 1-360; Part II, pp. 361-924. Paris, 1910 (Ch. Béranger).—The author is to be congratulated as well as the mineralogical public, in that the great work which he undertook some seventeen years since has now been brought to a successful conclusion. The four volumes which have been published give an admirable account of the minerals of france and the French Colonies. The work of the author is always thorough and orig- inal, and the description of species contains much that is import- ant and new. Volume four, which has recently been issued in two parts, is devoted for the most part to the sulphates, phos- phates, and related compounds. There is also a Supplement of more than two hundred pages, made necessary by the discoveries of recent years. The work is well printed and fully illustrated. 8. Practical Mineralogy Simplified for Mining Students, Miners and Prospectors ; by Jessk Perry Rowe. First Hdition. Pp. 162. New York, 1911 (John Wiley & Sons).—This little book kas been prepared with the object of putting before practi- cal miners a simple and elementary description of the commonly occurring minerals and ores, with a means for their investigation. The species are grouped according to the prominent metals present, and two large tables give a summary of the characters for non-metallic and metallic minerals respectively. 9. Caleites of New York ; by Herserr P. Wurtiock, Pp. 190, 27 plates. New York State Museum, John M. Clarke. Director. Memoir 13. Albany, 1910.—The species calcite is an almost inexhaustible subject from the standpoint of the crystal- lographer, and many important monographs have been published describing the forms from different localities. The author has now taken up the calcites of New York, which have been obtained from a wide range of different. localities, among which that of Rossie has been famous for nearly 100 years. The thoroughness of this investigation will be appreciated from the fact that some twenty-five plates are needed to show the different types of crys- tals with their wide range of forms ; among these a considerable number of new ones are noted. 10. Les Minéraux des Pegmatites des Environs d’ Antsirabé a Madagascar ; par li. Duparc. Mem. Soc. Phys. et d’Hist. Nat. de Genéve, xxxvi, fasc. 3, pp. 283-410, 1910.—In this work the author gives an account of his investigations of the pegmatite dikes of a part of Madagascar, which in recent years have fur- nished the markets of the world with beautiful gem-stones and fine crystals for mineral collections. In the laboratory study of the material collected the author was aided by his assistants, MM. Wunder and Sabot. A short sketch is given of the geology of the island ; there follows a general description of the occur- rence of the dikes, accompanied by a detailed account of the separate localities visited, where work of exploitation is going 338 Seientifie Intelligence. on; the memoir closes with a chemical and physical study of the minerals collected. Some attention is also devoted to the occurrence and petrography of a number of basaltic cones and extrusions of lava encountered in his journey. The pegmatites occur as a series of seams, dikes, stringers, and lenses, penetrating granites, gneisses, and schists, and their general character appears to be quite similar to what has been observed elsewhere in such cases. Here also they often attain gigantic dimensions in the size of the individual crystals. The surface of the country, the hill slopes and ridges, as well as the valleys, is nearly everywhere covered by a thick deposit of laterite, resulting from decay of the rocks in place. The pegmatites are likewise changed on the surface, but their presence beneath is revealed by the large unaltered masses of quartz, which form lines of bowlders in the soil. Some of the gem material is found in the soil, but the best occurs in pockets, etc., in the yet unaltered dikes. The chief minerals accessory to the pegmatitic quartz, feldspar and mica which furnish the gems are tourmaline, beryl, and garnet with some spodumene. The tour- malines are mostly black, but splendid rubellites and green, yellow, and brown varieties occur, and, as in other localities, the same crystal often shows zones and bands of several colors. The beryls are sometimes of enormous size and furnish aquamarines as gems; sometimes they are of a sky-blue color; there occurs also more rarely a beryl of a peach-blossom color, with a different crystalline form, which makes fine gems (see this Journal, xxxl, 81). The garnet is a yellow variety of spessartite and has yielded some small gems. The spodumene is sometimes lilac (kunzite), sometimes yellowish green, or white. A large number of chemical analyses and determinations of the optical and phys- ical properties of these minerals have been made and the results are given. The whole is a valuable and extensive contribution to our knowledge of the mineralogy of. Madagascar. Ts avi tba 11. Production of Phosphate Rock in Florida during 1910.— E..H. Settarps, State Geologist, states in a preliminary circular that the production of phosphate rock in Florida during 1910 exceeded that of any preceding year, the output having for the first time exceeded 2,000,000 tons. The total production for 1909 was 1,862,151 tons, while for 1910 the production was approxi- mately 2,029,797 tons. This amount includes a production of 392,088 tons of hard phosphate rock and 1,637,709 of pebble rock ; the increase noted, therefore, is in the latter variety, the former having diminished from a total of 527,582 tons in 1909. 12. North Carolina Geological and Economie Survey ; JosepH Hype Pratt, State Geologist.—The following publica- tions have been recently issued : Economic Paper No.19. Forest Fires in North Carolina during 1909; by J. 8S. Hotmus, Forester. Pp. 52, with 9 plates. No. 20. Wood-Using Industries of North Carolina ; by RocER E. Srmmmons, under the direction of J. 8. Hoxumzs and H.S. _ SackEeTT. Pp. 74, with 6 plates. Miscellaneous Intelligence. 339 13. Wote on the parietal crest of Centrosaurus apertus and a proposed new name for Stereocephalus tutus; by LawrEeNcE M. Lamer. The Ottawa Naturalist, vol. xxiv, pp. 149-151 with plate III, Dec, 1910.—Centrosaurus is one of the Ceratopsia from the Judith river (Belly river) formation of Alberta and was first described by Lambe in 1902 as Monoclonius dawsoni and in 1904 made the type of a new genus and species. The remains con- sisted of a parietal frill and a supposed nasal horn core, but now the latter is discovered to be in reality a portion of the crest itself, being a curious projecting process of the parietal bar which forms the rear margin of the fontanelle on the right-hand side, extending obliquely forward and slightly upward over the fon- tanelle itself though not in contact with any of its forward border. Lambe supposes the entire structure to have been covered with a common integument. It seems from the drawing, however, to be a process similar to the curious hook-like projections from the rear of the frill and is probably an instance of the development of spinescence accompanying racial old age. So far as one may judge from such fragmentary remains, Centrosaurus does not seem to the reviewer to be antecedent to either of the Laramie ceratop-. sian genera, Triceratops or Torosaurus, but to represent the ter- minal member of a side branch of the Monoclonius-Triceratops phylum, occupying a place among the Ceratopsia similar to that of Stegosaurus among the armored dinosaurs. The proposal of the new name Euoplocephalus to replace Stereo- cephalus (preoccupied), which was also described by Lambe, is but natural, although, in view of the necessary revision of all of these genera in the Stegosauria monograph now under preparation by the reviewer, it may perhaps only add to the burden of an already great synonymy. R, S. L. III. Miscerzranxous Screntiric INTELLIGENCE. 1. Lhe Carnegie Foundation for the Advancement of Teach- ing. Fifth Annual Report of the President, H. 8. Prircnerr, and of the Treasurer, R. A. Franks. Pp. vi, 113. New York City, October, 1910.—The Carnegie Foundation completed its fifth year on September 30th, 1910. At that time, the original gift of $10,000,000 had been increased by something more than $1,100,000, from the accumulated surplus. This increase obviously adds much to the extent of the work which can be accomplished ; the fact, however, that of the total income for the last year, $543,880, all but $5,700 was expended, seems to indicate that an increase of capital from this source is hardly to be looked for in the future. The list of accepted institutions is now seventy-one, having been increased the past year by the addition of the Uni- versity of California, indiana University, Purdue University, and Wesleyan University. Retiring allowances were given to sixty- four teachers, forty-six of whom were in accepted institutions and eighteen in institutions not on the accepted list. Twenty- 340 Scientific Intelligence. three pensioners died during the year. It is interesting to note that two institutions not included in the Carnegie list have under- taken to provide a pension fund for themselves. The Second Part of the Report is devoted to an interesting discussion of the relations of colleges to the secondary schools. The difficulty of bringing the high school and the college into close relations is one that has been felt for many years, and the importance of it is especially recognized at the present time. Dr. Pritchett takes up the various points involved, and indicates what he would regard as a promising method of removing the difticulty. The closing g pages of the Report are devoted to “brief obituary notices of the teachers who have died during the past year. ' 2. A Teat-Book of General Bacteriology ; by. Epwin O. Jorpan, Ph.D. Second revised edition, pp. 594, illustrated. Philadelphia and London, 1910 (W. B. Saunders Company).— While the book is adapted primarily to the needs of students of medicine, it is general in its scope, as indicated by the following subjects treated: Methods of studying bacteria, their structure and development ; effects of physical and chemical agents, effects produced by bacterial growth, the relations of bacteria to dis- eases of animals and plants, bacteria in milk, air, soil, water, etc., their importance in the arts and industries, and other topics. A good bibliography is given. The subject matter is presented clearly, the illustrations are good, and the book should be of much value to students of gen- eral bacteriology. L. F. R. 3. Catalogue of the Lepidoptera Phalene in the British Museum, by Sir Grorce F..Hampson. Volume X. The Noc- tuidee. Pp. xix, 829, 214 figures.—The tenth volume of the British Museum Catalogue of Moths, earlier parts of which have been repeatedly noticed in this Journal, has recently appeared. It is devoted to the Noctuid subfamily Erastriane. This is represented here by 1222 species belonging to 136 genera. This sub-family is largely confined to the more arid districts of tropical and warm temperate regions, and has few species in the colder zones. The plates belonging to this volume are promised at an early date. Mécanique Sociale ; par Sprrvu C. Harer. Pp. 256. Paris (Gauthier— Villars); Bucarest (Ch. Gobl), 1910. OBITUARY. Dr. Henry Pickrer1nc Bownrrcu, Professor of Physiology in Harvard University, died on March 13 at the age of seventy-one years. Professor, J. H. van’r Horr, the distinguished Professor of Physical Chemistry at the University of Berlin, died on Febru- ary 1 in his fifth- pinth year. Professor Jutius Wituretm Brin, the German chemist, died at Heidelberg on February 5 at the age of sixty years, hes aera +: - f. _— ie VOL) XXX. MAY, 1911. ae Established by BENJAMIN SILLIMAN in 1818. THE AMERICAN JOURNAL OF SCIENCE. Epitrorn: EDWARD S. DANA. ASSOCIATE EDITORS Proressors GEORGE L. GOODALE, JOHN TROWBRIDGE, W. G. FARLOW anp WM. M. DAVIS, OF CAMBRIDGE, Proressors ADDISON E. VERRILL, HORACE L. WELLS, L. V. PIRSSON anp H. E. GREGORY, or New Haven, Proressor HENRY S. WILLIAMS, or Iruaca, Proressorn JOSEPH S. AMES, or Baurimore, Mr. J. S. DILLER, or Wasuineron. FOURTH SERIES VOL. XXXI-[WHOLE NUMBER, CLXXX1] No. 185—MAY, 1911. WITH PLATES I-IItl. NEW HAVEN, CONNECTICUT. PIO ak THE TUTTLE, MOREHOUSE & TAYLOR CO., PRINTERS, 123 TEMPLE STREET. Published monthly, Six dollars per year, in advance. $6.40 to countries in the Postal Union ; $6.25 to Canada, Remittances should be made either by money orders, _ registered letters, or bank checks (preferably on New York banks). Important Announcement. NATRAMBLYGONITE. Having secured a small lot of this new mineral as described in this Jour- nal, No, 181, January 1911, by W. T. Schaller, I am now in a position to furnish desirable specimens of same, at very reasonable prices. RUBY CORUNDUM. It has been some years since the Buck Creek deposit has produced material of specimen quality; having been fortunate in securing a lot of these double terminated crystals, the result of considerable work which did not pay the miners for their trouble, the mines were again closed. This is an opportunity for collectors to secure very desirable crystals, single or in sets, showing the variety of occurrence of form and colors; prices very reasonable. — HIDDENITE. We have secured the balance of the former lot of Hiddenite crystals which sold so rapidly last month ; we are now in a better position to furnish sets showing the different effect of etchings, form and color. These specimens have the deep emerald green color so desired by all collectors. When made in sets mounted on sheet wax, they present a beautiful contrast for show and study and should have a place in every collector’s cabinet. RUTILE. From another source was received a lot of trilling in this most unique mineral, all of good crystal quality, a number of them quite transparent and possessing a very brilliant luster ; prices very low. PRECIOUS AND SEMI-PRECIOUS STONES. Having secured from a receiver’s sale a very large stock of cut stones all colors, size and qualities, I take this opportunity to offer same at 75 per cent of their real value, so as to dispose of them rapidly. This is the chance of a lifetime for anyone interested in precious and semi-precious stones, carvings, mosaics, cameos, etc. Ask for an assortment, which I will gladly send on approval prepaid, for your selection. You can return at my expense anything you do not desire. REMARKABLE COLLECTION. I have just received for sale a remarkable collection which was in the possession of a well known mineralogist, who’ was noted for the intense interest he had in fine, choice specimens ; this collection represents years of diligent and presevering collecting, the specimens representing nearly all he old localities and most of the recent ones. Some of the specimens are the finest found and this is an opportuuity which new collectors should take to get possession of some of the choice things of past localities. © Tam now ready to furnish a complete list of this collection to anyone upon request and will advise early correspondence on same. A. H. PETEREIT, 81—83 Fulton Street, New York City. Phone Beekman 1856. THE AMERICAN JOURNAL OF SCIENCE [FOURTH SERIES. |] tos Arr. XXX.—The Melting Points of Minerals in the Light of Lecent Investigations on the Gas Ty hermometer ; by Arraur L, Day and Rozsrrt B. Sosmay. Wuew the work of the Geophysical Laboratory was begun, no temperatures above 1150° ©. had ever been accurately ~ measured with the nitrogen thermometer. It had been the custom of the Reichsanstalt to interpret the readings of a thermoelement above 1150° by first calculating the curve of temperature and thermal electromotive force of the element for the region between 400° and 1150°, and then extrapolating this curve. The fixed temperatures commonly printed on high- temperature measuring devices, like the Siemens and Halske direct-reading galvanometers, are based on this extrapolation. The absolute accuracy of the Reichsanstalt scale (400° to 1150°) was estimated to be about 3° at 1150°, and the extra- polation above 1150° with the thermoelement is certainly in error by more than ten times this amount at the melting point of platinum. It was therefore deemed necessary to undertake a new investigation of the high temperature region with the gas ther mometer, and im particular to extend its range for a considerable interval above 1150°, in order that a sound basis might be established for the mineral work of this laboratory. Accordingly, such an investigation was begun in 1904, and the final results were published within the past year.* It proved possible not only to attain higher accuracy in the region between 400° and 1150°, but to extend the fundamental meas- urements to 1550° with an accuracy estimated at 2° at the latter temperature. In the meantime, the temperature measurements made in this laboratory with thermoelements had been interpreted in * Preliminary publication : Day and Clement, this Journal, xxvi, 405-463, 1908. Final publications: Day and Sosman, this Journal, xxix, 93-161, 1910; R. B. Sosman, ibid., xxx, 1-15, 1910. Am. Jour. Sc1.—FourtsH Series, VoL. XX XI, No. 185, —May, 1911. 24 342. Day and Sosman—Melting Points of Minerals. the old way, by extrapolating the curve of temperature and thermal electromotive force. These results now require to be corrected by the amount of the difference between the old tem- perature scale and the new, which makes it necessary to recal- culate the existing temperature data of this laboratory in terms of the new séale. This paper will present a summary of the values resulting from this recalculation. 1. Melting and Inversion Points. The “ melting point” of a pure substance may be defined as the temperature at which the crystalline and liquid substance * can remain side by side in equilibrium ; an “inversion point,” as a temperature at which two different seni forms of the substance can remain side by side am egudlibriwm. Only the addition or withdrawal of a quantity of heat will cause the disappearance of one of the two forms in contact. Both melt- ing and inversion are therefore characterized by two concur- rent phenomena, the appearance or disappearance of a particular crystal structure and the appearance or asap pearance of a quantity of heat. 2. Melting Intervals. The above definition applies to pure compounds. If the material is a mixture or a solid solution, it will have, not a melting point, but a melting eterval, with (theoretically) definite temperature limits. The so-called melting interval of slow melting compounds is of entirely different character. Tammann,* and more recently Dittler,+ have criticized certain of the melting-point determ1- nations made in this laboratory, more particularly those of anorthite and diopside, as being too high because of inade- quate furnace control and super rheating.. This is plainly due to oversight in the consideration (1) of the phenomenon of melting in silicates, and (2) of the records of our measurements. ieee substance always undercools before crystallizing, the tem- perature of crystallization will vary with the rate of cooling, 1. e., will be a random temperature depending upon the condi- tions of experiment and not a physical constant characteristic of the substance. If the same substance mélts so slowly that it readily becomes superheated while melting, the temperature of complete liquefaction will vary with the rate of heating and is also a random temperature dependent solely upon the condi- tions of experiment. Pure quartz, albite and orthoclase are such substances, and no doubt others will be found. The temperature-time curve is not a competent method with which to determine the temperature of change of state of such sub- stances. * Zts. physikal. Chem., Ixviii, 257-269, 1909. + Zts. anorg Chem., lxix, 273-304, 1911. Day and Sosman—Melting Points of Minerals. 348 Tf, on the other hand, the substance yields the same constant melting temperature with widely different rates of heating, the melting temperature is a physical constant which is char- acteristic of the substance and does not depend upon the con- ditions of experiment. Diopside and anorthite are such substances, and the published melting points of pure anorthite and diopside to which criticism has been directed, were obtained in this way. (See references 1, 4, 32, 44, 46.) If it be contended (as has recently been done by Dittler) that the thermoelement does not give the true temperature of the mineral at any time during the heating, or that the energy change in the system is not “contemporaneous with the dis- appearance of crystal structure, experimental proof is readily obtainable with the “quenching” furnace, which fixes the matter beyond all doubt. Suppose the melting point of diopside, for example, to have been observed, in the manner described abov e,at 1391°. Place a portion (2 orams) of this erystalline charge in a quenching furnace, heat it slowly to 1388° and hold the temperature con- stant at that point for an hour, or until there can be no further question that the thermoelement and all parts of the charge have the same temperature ; then remove the bottom of the furnace and drop the crucible with the diopside suddenly into a basin of mercury, which has the effect of cooling it almost instantly without giving the slightest opportunity for any fur- ther change within the substance. If the crystalline structure of the diopside, upon examination, is found to be unchanged, 1388° is below the melting temperature. Place the same mate- rial in the furnace once more, heat to 1396°, hold this temper- ature constant for some time and drop the charge quickly into mereury as before. If the diopside now appears as a clear glass, the change of state and all its attendant phenomena must have occurred between 1388° and 1396°. By way of offering a visible record of this particular case, this experiment has “been performed in this laboratory with both diopside and anorthite (chemically pure), and photographs of the substances as they appeared after removal from the furnace are reproduced in figs. 1 to 4. These photographs with their accompanying data give absolute proof that the melting point of the diopside was between 1388° and 1396°, and the melting point of the anorthite between 1547° and 1561° (new nitrogen-thermometer scale). The so-called melting interval of slow-melting compounds (quartz, albite, orthoclase) should be carefully distinguished from the we melting interval” of mixtures (lime-soda feldspars, impure natural minerals). The former melt and become amor- phous very slowly, but their melting interval is an interval of time, whereas the melting interval of mixtures is an inter val of B44 Day and Sosman-—-Melting Points of Minerals. Fie. 1. Fie. 2. eliiel, By, Fic. 1. Chemically pure diopside. Element R, 14100 microyults = 1388". Time, 1 hour. Crossed nicols. 200 diameters. Crystalline. The interfer- ence colors show in bands around the edges of the crystal fragments. Fic. 2. Chemically pure diopside. Klement R, 14200 microvolts = 1396°. Time, 14 hour. Photo with crossed nicols. Magnified 52 diameters, Glass. Fig. 3. Chemically pure anorthite. Element R, 15950 microvolts = 1547°. Time, % hour. Crossed nicols. 200 diameters. Orystalline. The twin- ning shows distinctly. Fic. 4. Chemically pure anorthite. Hlement R, 16100 microvolts = 1561°. Time, 44 hour. Crossed nicols. 100 diameters. Glass. temperature. The pure soda feldspar (albite), if held at a tem- perature slightly above its melting point for a sufficient length of time, will melt completely ; the mixture, on the other hand, if held at a temperature within its melting interval, will melt in part, come to equilibrium, and remain indefinitely part erys- talline and part liquid. Day and Sosman—WMelting Points of Minerals, 345 The solidifying points and intervals of silicates, unlike those of most metals, are not as accurately determinable as the me/t- ing points and intervals, because of the tendency to under- cooling. ‘The extent of the undercooling varies widely, and is affected by numerous incidental conditions. It may amount to only one or two degrees under one set of conditions; and for the same substance may become so great under other con- ditions that the material is obtained at ordinary temperatures in the form of glass, which will remain amorphous indefinitely. 3. Principal Temperature Determinations at the Geophysical Laboratory, 1905-1910. In Table I the principal fixed temperatures that have been determined during the past five years in this labora- tory are summarized, and restated in terms of the revised temperature scale. Zhese are not new determinations. The original thermoelement or optical data have simply been inter- preted in terms of the Day and Sosman scale of 1910. All of the early thermoelectric measurements of this labora- tory were interpreted by calculating a curve of the form: e=a+bt+ct* from the melting points of zine, silver, and copper, and using this curve for extrapolation, as has been the general custom since the publication of the work of Holborn and Day in 1899. Occasionally a function of the form t=a + be + ce’ has also been employed. The curve ¢=/(¢) gives temperature values that are about 16° low at 1500°, and the curve ¢=f(e) gives values that are about 33° low at 1500°. At 1100° the correction is only 1°5°, so that temperatures up to 1100° remain practically unchanged. The data of Shepherd and Rankin on Binary Systems of Alumina with Silica, Lime, and Magnesia, have already been revised and published inthe German translation of that article.* The revision of the other data is here published for the first time. The melting points which form the basis of reproduction of the present scale are as follows : + EGAN CREM tse Te Pe atin eee ect Oo Mea Naunt Ue ARGO ANNO MNOMWY ines Ben cE a eee es ee OOM Silivermeee cece phi Sekirei Mae OGIO: () GON ee ee eG oS es. 8 NO GOLA Coppel ie Aa ran aaah en NOSDG JDO DSCC] ieee ia eo) a eee == agile ieallaoituim = 25 Seen os re ae 1549° PAM ORtMIGC: eee en sees Ue ae OS 2 Vavt imu eae en eA ole We * Zts. anorg. Chem., Ixviii, 370-420, 1910. + Day and Sosman, loc. cit., p. 161. 346 Day and Sosman—Melting Points of Minerals. TaBLE I,—Principal Temperature Determinations at the Geophysical Laboratory, 1905-1910. Substance Formula Transformation pera- | Observer | Reference ture a i a to @, and 65 Wright & |39, 64 Quantz SiO. reverse 575" | Larsen (1909-10) Aluminum ’ , saan : lnraes Shepherd |41, 65 Tatar onte) nes panos 1816 |e Rankin |(1909-10) : | Allen a-Magnesium | : “ ey STE evi ‘MgSiO; 1554 |Wright & 18 £1906) Clement & ==» |Allen & |. ue oe | 1557 White 32 (1909) a-Caleium | | metasilicate : sf Allen & |. (Pseudo- CaSiOs | 1540 | White G96) wollastonite) a-Calecium ; eg! >, |Day & 8, 11), 19 orthosilicate |C225!0« | 2130 Shepherd (1906-7) Eutectic between) | -calci : metasilicate CASO» 77 4** 146 |“ c and cristoba- SiO, 28 % lite Eutectic between a-caleium : metasilicate [CasiOs, 66 % do 1440 Ge Ob and a-calcium | CaxSiO. 34 % orthosilicate | Eutectic between) a-calcium 'CasSiO, 93 4 Viet. a és orthosilicate CaO 74% | 2065 and lime | Tricaleium | ‘Dissociation into Shepherd 41, 65 aluminate B20. A105 CaO and liquid) 1937 |& Rankin (1909-10) 5:3 compound | | | v lime and 5CaO.3A1,05 \Melting 1382 ve a8 alumina | = | Rees \Ca0.Al20s Ie ey eee Q a 3:5 compound | \Dissociation into of lime and [|3Ca0.5A1.0O; Al,O; and 1700 ih i alumina liquid | Yel Eutectic Bree oahu 4 Melting 1378 * | ‘6 F a of f Eutectic revauriee oe ss 1378 pe bc . \CaO. Al.05 1G) % ‘ ‘ Hates | 8Ca0.5A1,0,27¢ | | Selby & | : Magnesium-cal- | cium metasili- |MgSiO;.CaSiOs « 1391x/Pay © 46 (1910) cate, (Diopside) Syeesees | 66 | Allen & . Bh Ob 1391 | Wite i (1909) ** Percentages by weight. Day and Sosman—Melting Points of Minerals. 347 TABLE I.—Continued. a= —— “= = = = == SSS ror — | —— Substance Formula Transformation pera- | Observer | Reference ture Eutectic between) | 2 solid solu- 95.54 56) {62% diops. tions of diop- Ye 80 | 384M egSiOs : ,|Allen & |. side and Motel 49 diops. Melting 1385 White Be (1909) a-MgSiO, in rae. | 962M gsiO; each other Eutectic between) . : diopside and ne 60 & 66 1357 a 66 a-CaSiOs Bomtan 1 | Calcium alumi- | j | ae CaAl.Si2O; or Day & \1, 4 licate. | aaa Pra “ Me ? (saeethitey * | CaSiOs. AlsSiOs 1552 alien —_—(1905) i os “ 1550* 2oy & 46 (1910) ; Albite, ‘Day & 1,4 Bytownite meee “ 1516+ a then (1905) mae yities Albite; fs ie ‘ Labradorite Nnorthites : 1477+ : e Andesine- Albite; p ‘ Labradorite Anorthite; : 1430+ G ae Albites eeu ; ee Anorthite, hate 1375¢, Calcium Dissociation pres- carbonate CaCOs ssocratin Pre | 898 Johnston |56 (1910) , o . Day & 1,4 is NE as meting 741 Vatien —_‘((1905) SOHEREE NaCl ss | 800 White 44 (1909) Sodium x ; : sulphate Na»SO. : 884 : § In the first column of Table I is the name of the compound or mineral; in the second, its molecular formula; in the third, the transformation or reaction which takes place at the tem- perature given in the fourth column. The last two columns contain the names of the observers, and the references. The numbers are those in the published list of papers of this labo- ratory, and the references are given at the end of this article. A number of melting and transition temperatures were only approximately determined, on account of the sluggishness or slow rate of the change. These are summarized in Table II. * Direct comparison with the nitrogen thermometer. } Being isomorphous mixtures of two compounds, these substances have no melting point, but a melting imterval. The temperature limits of this interval are probably narrow, and its existence is entirely covered up by the slowness with which these silicates reach equilibrium. See full discussion in paper 1 (pp. 60-69), or 4 (pp. 40-50). Day and Sosman—WMelting Points of Minerals. TaBiE Il.—Approwimate Temperature Determinations. Substance Formula Transformation Te Observer | Reference Cristobalite : : About Day & 8, 11, 19 (from quartz) Sia Melting 1600 Shepherd (1906-7) Day & 8, 11, 19 Quartz SiO Inversion to Above Shepherd |(1906-7) 3 eristobalite 800 Wright & |39, 64 Larsen (1909-10) Eutectic* : between cristo-|SiO. 80% Sree) A rite |Shepherd 41, 65 balite and AlSiOs 20% el Foe 1600 |& Rankin (1909-10) sillimanite ; Hateoliy belmee AI Os 84 ere About |Shepherd 41, 65 CTT She | hd ORES LOR PON atl gaan tae 1810 | & Rankin (1909-10) sillimanite | Inversion a to | Magnesium | F B (clino- About Allen & ja5 metasilicate | MgSiOs enstatite), and| 1375 [White >» (1909) reverse Inversion to le B-Calcium a-CaSiOs B 6, 82 metasilicate, CaSiO, (pseudo- 1190 * 4906-9) (Wollastonite) wollastonite), f and reverse Day & 8 11. 19 Caleium : Inversion a to Sa alVeiayaya. orthosilicate Ca2Si0. 6 and reverse te Shepherd |(1906-7) Calcium ; Inversion ? to ye ‘ ; orthosilicate \CasSi0. y and reverse Blio , ns Nes | Formation of 2 Ser Ene '3C'a0.Si0; phases by | yeuyt (een ten ein cigtay | dissociation ete Eutectic of 41, 65 spinel and peer Ee: poe Melting oe ee (1909-10) periclase peso ge | : : Below Day & 1,4 aoe Ree 5 1200t Allen —(1905) igoclase- 1te3 ¢ 66 | “c Andesine Anorthite, : EAD) Microcline | KAISis0¢ “ erhin te) oy dager c | : Change to 1300 and |Shepherd 41, 65 ee noe sillimanite above** & Rankin (1909-10) Cyanite or : 1300 and Z rj disthene ALSIOs i | above** a'-Magnesium Chane em : Allen, metusilicate, MgSiOs to BMocio, (L800%* |Wright & |13' (1906) (Enstatite Gi | to P-MgSiO; | Clement a (clino-enstatite) 6'-Magnesium e mains esi aH eee amphibole$ y'-Magnesium | orthorhombic /M&SiO. 4 eee amphibole§ grape | * The eutectic compositions are given in percentages by weight of the two compounds named. + See note (+) to Table I. Day and Sosman—NMelting Points of Minerals. 349 {The purest natural albite showed signs of melting when heated a few minutes at 1200°, and again when heated for four hours at 1100°, Small amounts of certain impurities might, however, lower the melting point con- siderably, while others would raise it. No determinations of melting point on chemically pure albite or orthoclase have been made, The best statement that can be made, therefore, is to say that the melting point is probably below 1200°. The facts concerning microline are similar. ** These temperatures differ from the others in the tables, in not being fixed physical points. The change is from an unstable into a stable form and is not reversible. The figures merely represent the temperatures at which the change is rapid enough to become observable within a reasonable length of time. § Both of these have been called kupfierite by different authors. References. 1. The Isomorphism and Thermal Properties of the Feldspars. I. Ther- mal Study ; by Arthur L. Day and E. T. Allen. II. Optical Study ; by J P. Iddings, with an introduction by George F. Becker. Publication No. 31, Carnegie Institution of Washington (1905). 4. Der Isomorphismus und die thermischen Higenschaften der Feldspate. Arthur L. Day and E, T. Allen. Zts. physikal Chem., liv, 1-54, 1905. 6. On Wollastonite and Pseudo-Wollastonite, Polymorphic Forms of Cal- cium Metasilicate. E, T. Allen and W. P. White, with Optical Study by Fred. Eugene Wright. This Journal (4), xxi, 89-108, 1906. te Quartz Glass. Arthur L. Day and E. S. Shepherd. Science, NSIS; Vol. xxiii, No. 591, pp. 670-672, 1906. 8. The Lime-Silica Series of Minerals. Arthur L. Day and EK. 8. Shep- herd. Jour, Am. Chem. Soe., xxviii, 1089-1114, 1906. 11. The Lime-Silica Series of Minerals. Arthur L. Day and EH. S. Shep- herd, with Optical Study by Fred. Eugene Wright. This Journal (4), xxii, 265-302, 1906. 13. Minerals of the Composition MgSiO;; a Case of Tetramorphism. H. T, Allen, Fred. Hugene Wright, and J. K. Clement. This Journal (4), xxii, 385-488, 1906. 19. Die Kalkkieselreihe der Minerale. Arthur L. Day, E. T. Allen, E. S. Shepherd, W. P. White and Fred. Eugene Wright. Tschermak’s Min. Petr. Mitt., xxvi, 169-232, 1907. 32. Diopside anc its Relations to Calcium and Magnesium Metasilicates. EK. T. Allen and W. P. White; with Optical Study by Fred. Hugene Wright and Esper S$. Larsen. This Journal (4), xxvii, 1-47, 1909. 39. Quartz as a Geologic Thermometer. Fred. Eugene Wright and Esper S. Larsen. This Journal (4), xxvii, 421-447, 1909. 41. The Binary Systems of Alumina with Silica, Lime, and Magnesia. H. S. Shepherd and G. A. Rankin; with Optical Study by Fred. Eugene Wright. This Journal (4), xxviii, 298-333, 1909. 44. Melting Point Determination. Walter P. White. This Journal (4), xxviii, 403-473, 1909. Melting Point Methods at High Temperatures. Walter P. White. Ibid. (4), xxviii, 474-489, 1909. 46. The Nitrogen Thermometer from Zinc to Palladium. Arthur L. Day and Robert B. Sosman; with an Investigation of the Metals by Eugene T. Allen. This Journal (4), xxix, 93-161, 1910. 54. The Platinum-Rhodium Thermoelement from 0° to 1755°. Robert B. Sosman. This Journal (4), xxx, 1-15, 1910. 06. The Thermal Dissociation of Calcium Carbonate. John Johnston. Jour. Amer. Chem. Soc., xxxii, 938-946, 1910. 61. Preliminary Report on the Ternary System CaO- Al.0;-SiO.; a Study of the Constitution of Portland Cement Clinker. E. S. Shepherd and G. A. Rankin; with Optical Study by Fred. Eugene Wright. J. Ind. and Eng. Chem., iii, 211-227, 1911. : 64. Quarz als geologisches Thermometer. F. E. Wright and E. S, Lar- sen. Zts. anorg. Chem., lxviii, 388-369, 1910. 65. Die binairen Systeme von Tonerde mit Kieselsiure, Kalk, und Mag- nesia. H. S. Shepherd and G. A. Rankin. Nebst optischen Untersuchun- gen von fF. E. Wright. Zts. anorg. Chem., Ixviii, 370-420, 1910. (laa yee aca Tha) Nee bae, Alacra evetyay iene ee aes A900 Ys RO er ae Ve 1 Lo 350. J. Roberts—Cerium by Potassium Permanganate. Arr. XXXI—On the Separation of Cerium by Potassium Permanganate ; by Evwin J. Roserrs. [Contributions from the Kent Chemical Laboratory of Yale Univ.—cexviii. ] Onx of the most useful methods for the separation of cerium from the rare earth metals which accompany it depends upon its oxidation by potassium permanganate according to the equation : 3Ce,0,+2KMnO,+H,0 = 6CeO,+2KO0I+2Mn0,. This action takes place in a cerous nitrate solution provided some neutralizing agent is present to take up the nitric aeid set free. Stolba, in "1878, * used zine oxide for this purpose, and other investigatorst have since used the method. Dross- bach,t Bohm,§ Muthmann and Weiss,| Meyer‘ and others have used alkali hydroxides or carbonates in place of zine oxide. More recently, Meyer and Schweitzer** have studied this process, using sodium carbonate. They give the equation : 30e(NO,), -+ K MnO, +4Na,CO, +8H,0 = 3Ce(OH), + Mn(OH), +8NaNO, + KNO, +4€0,, and use two different solutions for the precipitation. One, for the complete removal of the cerium, contains one molecule of permanganate to nearly five of sodium carbonate; the other, for the purification of the cerium, contains one molecule of permanganate to less than three of sodium carbonate. It is obvious from the above equation that in the first case large amounts of other earths will be precipitated with the cerium by the excess of alkali, while in the second case the nitric acid will be in excess at the end of the process, leaving consider- able amounts of cerium in solution. The object of the present work was to study the process carefully and to produce, if possible, a sharper separation of the cerium. When a little potassium permanganate solution is added to a hot neutral solution containing cerous nitrate, the red color of the permanganate is instantly ‘pleached, and a brown precip- itate appears. If the addition of the permanganate 1s con- tinued, the color is bleached more and more slowly, the liquid becomes distinctly acid to litmus, and finally the red color becomes permanent. If a little alkali is now added, the color is again bleached, and if the liquid is kept neutral, or very nearly so, the red color will not be permanent until all the cerium is precipitated. The action seems to be as follows: the cerous nitrate is oxidized by the permanganate, probably according to the equation: * Jahreshber., 1878, 1059. + Muthmann and Roélig, Ber. chem. Ges., xxxi, 1718; James, Jour. Am. Chem. Soc., xxx, 982. t Ber. chem. Ges., xxix, 2452. § Zeit. angew. Chem., 1903, 1129. || Ann. Chem., cccxxxi, 9. “| Zeitschr. anorg. Chem., xxxvii, 378. ** Zeitschr, anorg. Chem., liv, 104-120. E. J. Roberts— Cerium by Potassium Permanganate. 351 3Ce(NO,),+ KMnO,+3H,0 = 2Ce(NO,),+Ce(OH),+KNO,+H,MnO,. (1) The ceric nitrate formed then hydrolyzes: Ce(NO,),+4H,O =~ Ce(OH,) + 4HNO,. (2) If no alkali is added, the accumulation of nitric acid in the liquid checks the process by reversing the action in equation (2), and also in equation (1), for when a little of the mixed precipitate of hydrated ceric and manganese dioxides was treated, after washing, with dilute nitric acid and warmed, a strone permanganate color was produced. The addition of alkali at this point of course simply removes the free acid and allows the action to become complete. The fact that the oxidation and precipitation take place even in the presence of appreciable amounts of free nitric acid makes it possible to secure a better separation than would otherwise be the case, for if the liquid is kept perfectly neutral the small amounts of trivalent earths which are locally thrown down by the alkaline precipitating solution will not be re-dissolved, while if the liquid is distinctly acid during the whole process they will tend to re-dissolve on stirring. These considerations give the theo- retical conditions for the process. As to the choice of a neutralizing agent, it is of course desirable to use no compound of a metal which is precipitated by oxalic acid or which is not easily separated from the rare earths, as, for example, zinc oxide or calcium carbonate. Magnesium oxide was tried, but does not give as good results as the alkali carbonate, and is not so easily controlled. Sodium carbonate seems to be the simplest and most convenient neutralizing agent for the purpose. The solution used for the precipitation is made by dissolving 15805 grams (one mol.) of potassium permanganate and 424-00 grams (four mols.) of dry sodium carbonate in any con- venient amount of water. The solution should be kept in glass-stoppered bottles and should not be allowed to come in contact with organic matter. Solutions of pure permanganate amd of pure sodium carbonate, of any convenient strength, are also needed. The process is carried out as follows: The rare earth solution, which should not contain other salts than nitrates, is heated to boiling in a large porcelain dish, and, if not already neutral, is neutralized with the sodium carbonate solution. The solution of permanganate is added in small quantities, until the red color just begins to be permanent, and the mixed solution of permanganate and alkali is then added very slowly, with constant stirring, the liquid being kept nearly at the boiling point during the whole process. A faint color of permanganate is maintained all the time in the liquid, a little of the pure permanganate solution being added if at any time the color is entirely bleached. This is important, as the constant acidity of the liquid is thereby insured. When the 352 LE. J. Roberts—Cerium by Potassium Permanganate. cerium is nearly all precipitated, the color is bleached more slowly after each addition of the precipitant, and the efferves- cence is less noticeable. The acidity of the liquid should now be tested from time to time, which may be done with litmus paper if only a slight excess of permanganate is present. Small amounts of the mixed solution or of sodium carbonate are added, until the liquid is nearly neutral to litmus, and still is faintly colored with permanganate. The whole is heated and stirred for about ten minutes, and filtered hot. The precipi- tate is washed with boiling water till the washings give no precipitate with ammonia. If the liquid at the end of the precipitation is faintly acid, the filtrate usually contains a trace of cerium giving a faint yellow color with ammonia and hydrogen peroxide, while from’ the precipitate a preparation of cerium chloride may readily be obtained which shows no absorption bands in a thickness of 15°" of very concentrated solution. The presence of a little cerium in the filtrate, where the earths in the latter are to be subjected to fractional erys- tallization, is usually not objectionable. To insure the complete removal of the cerium, the liquid at the end of the precipitation must be made absolutely neutral, when a filtrate is obtained which gives no test for cerium with hydrogen peroxide. Before testing, the excess of perman- ganate is removed from the filtrate by boiling with a few drops of alcohol. After re-filtering, the other earths are precipitated by oxalic acid. The precipitate of cerium and manganese hydroxides is dissolved in strong hydrochloric acid, the solution diluted and precipitated with oxalic acid and the cerium oxalate re-converted to nitrate. On repeating the pro- cess with this material, a pure cerium preparation is obtained, which gives an oxide of the characteristic pale yellow color. The process was thoroughly tested on several different kinds of material. A mixture containing as much as thirty-five to forty per cent of praseodymium and neodymium yielded pure cerium in two operations. Attempts were also made to extract the first cerium and manganese precipitate with dilute nitric acid, in order to remove the last traces of praseodymium and neodymium without recourse to a second permanganate treatment. Tho precipitate obtained in the usual manner was_ thoroughly washed on the filter, then transferred to a large dish and heated nearly to boiling with a large volume of water. Nitric acid was then added carefully until the liquid began to show the permanganate color. The whole was heated for about fifteen minutes, and the precipitate filtered off and washed. This cerium preparation, when freed from manganese, gave a pale reddish oxide, showing that the neodymium and praseodymium are not entirely removed from the first precipitate by extrac- tion with dilute nitric acid. Handlirsch—New Paleozoic Insects. 3538 Arr. XXXI1.—WNew Paleozoic Insects from the Vicinity of Mazon Creech, Illinois; by Anron Hanpuirscu, Imperial Natural History Museum, Vienna. [Continued from p. 326.] Order PROTOBLATTOIDEA Handlirsch. Family EOBLATTIDA Handlirsch. ANEGERTUS, new genus. Anegertus cubitalis, new species. Fig, 32. An insect of about 60", the anterior wings of which measure 45™™ and are nearly elliptical with a somewhat broadened Fie, 32. eat A mils bo we | ny V2 ey Dene Co Se ON Sl a muvee ~ _— Fic. 32. Anegertus cubitalis (negative). x2. 354 Handlirsch— New Paleozoic Insects from the apical half. Costa marginal and equally curved; subcosta long, nearly straight and reaching almost to the tip. The costal area comparatively broad, with numerous oblique, ramous veins, bridged over by cross veins. Radius straight, parallel to and very near the subcosta, its sector rising very near the base, diverging but little from the radius and sending off only 2 short branches near the end. The media is likewise somewhat reduced and cleft in the apical half into 2 branches. The cubitus on the contrary shows an extraordinary develop- ment occupying nearly half the wing and is cleft into 2 main branches at about a quarter of the wing length. ‘The anterior of these branches splits by repeated forking into 6-7 twigs; the posterior sends forth partly forward, partly backward, a number of forked twigs. The anal area, which occupies more than a third of the hind margin, is limited by a strongly curved vein and filled by about 6 veins which are sometimes forked and run in a regular vault toward the posterior margin. Cross veins numerous. Prothorax comparatively large and disciform; the head broad and free, with strongly vaulted lateral eyes. One of the legs, probably of the second pair, is preserved; it is compara- tively long and robust. If I am not mistaken, the posterior legs are directed backward in the middle line of the fossil, close together, a situation presupposing large and approxi- mated coxe. I have no doubt about the close relationship of this fossil to Hoblatta from Commentry, which has likewise a reduced sector and media and an expanded ecubitus. Holotype in Peabody Museum, Yale University, Cat. No. 43. Family ASYNCRITIDA, new family. Comparatively clumsy with a short, stout body, covered and exceeded in length by the singularly specialized wings, which are strongly dilated in their apical half. The small disk-like pronotum is pear-shaped. Meso- and metanotum fused to a solid complex, each of them broader than long. Coxz very large, femora and tibize of the posterior legs very short. The segments of the abdomen much broader than long. Anterior wings with a broadly rounded apical margin, a marginal costa and a long subcosta, fusing near the tip of the wing with the costa. Sector rising at a quarter of the wing length, diverg- ing widely from the radius and giving off but few branches, directed forward, parallel to the radius. Media free, forked only in the apical half. Cubitus free, with broken branches forming a polygonal network. The smaller interstices with straight cross veins, the broader ones, like the costal area, filled by a polygonal network. Anal area reaching nearly halt the length of the wing, containing a few veins regularly curved Vicinity of Mazon Creek, Illinois. 355 backward toward the apical margin. The posterior wings seem to have been quite similar to the anterior pair with the exception of the fan-like anal area, which attains to about half their length. ASYNCRITUS, new genus. Asyncritus reticulatus, new species. Fig. 33. Length of the thorax and abdomen 28™™. Length of the wings 28". Pronotum but a little longer than broad. Pos- terior femora only as long as the pronotum. Anterior wings Fig. 33. Fie. 33. Asyneritus reticulatus. x 2°2. about two and a half times longer than broad, their greatest breadth falling in the third quarter. Sector running obliquely to the middle of the apical margin, sending forth but 2 branches directed forward, their interstices being filled by an irregular network. Costal area with 2 rows of cells. Radio- medial space with 2 rows of large polygonal cells. Media with a simple fork. Oubitus split into about 8 branches, all angularly broken, some of them falling in the anal furrow. By the aid of the cross veins the whole cubital space takes on the aspect of a network. One sees only 4 anal veins, but 2 others may have existed. 356 Handlirsch—New Paleozoic Insects from the By this singular specialization of the venation Asyneritus differs from all other Protoblattoidea. It is so characteristic that the establishment of a new family is justified. Holotype in Peabody Museum, Yale University, Cat. No. 44. Family EPIDEIGMATID A, new family. This family is based on a slender Protoblattoid with a nearly semi-elliptical disciform pronotum, long and somewhat waved Fig. 34. er, Fic. 34. Epideigma elegans (positive and negative combined). x3. anterior wings, the subcosta not uniting with the costa but with the radius as in some of the Protorthoptera. The anal area strongly shortened and limited by the characteristic Vicinity of Mazon Creek, Lllinois. 357 strongly curved furrow. Costa marginal. Sector rising in the middle’ of the wing, giving off but a few branches, directed backward toward the apical margin. The media free, sending some branches obliquely backward, the last of which reaches the posterior margin: COubitus of a moderate development, splitting into numerous irregular branches which form, by the aid of cross veins, an irregular network and often take the character of intercalary veins. EpipEIeMA, new genus. Epideigma elegans, new species. Fig. 34. Prothorax somewhat longer than broad, of a nearly semi- elliptical shape. Head comparatively small, not quite con- cealed under the pronotum and bearing large, strongly vaulted, lateral eyes. Antennee long and slender. Anterior wings 28"™™ long with a somewhat waved costal margin and with a broad costal area, filled by oblique branches of the subcosta and of the radius. Radius reaching nearly to the tip. Sec tor rising at about one-third of the length, diverging com- paratively far from the radius and giving rise to only 2 branches. Media. independent of the radius and cubitus, slightly waved and traversing the middle of the wing. Of its 3 branches, which extend obliquely backward, the first is forked and has, like the second, a few short apical veinlets with the aspect of intercalary veins. The cubitus is cleft near the base inte 2 main branches, both splitting into numerous twigs, running obliquely backward and taking part in the ‘formation of a network. All main interstices are filled by a more or less regular polygonal network. The very short anal area hardly reaches one-fifth of the length of the wing and is limited by a sharply curved furrow. Holotype in Peabody Museum, Yale University, Cat. No. 45. Family CHELIPHLEBID At Handlirsch. By the discovery of a new species this family loses in some degree its provisional character and can with more certainty be added to the Protoblattoidea. It seems to be most nearly related to the Euceenide, having like those broad and short front wings with a broad costal area, the veins of which are not arranged in regular comb-like manner, but are irregular and ramified. The subcosta unites with the costa; the radius remains simple and its sector sends forth but few branches, directed obliquely backward. The media free and splitting into only few branches. The cubitus more richly ramified. Pronotum disciform and comparatively small, equal in length and breadth. Am. Jour. Scl.—FourtH SEries, Vou. XXXI, No. 185.—May, 1911. 25 3} or (os) Handlirsch—New Paleozoic Insects from the Cheliphlebia mazona, new species. Fig. 35. The only parts present are the nearly circular prothorax of about 8™" in length and a 34™" long portion of an anterior wing, which probably attained a length of 40™". Costa strongly Fie. 35. Fic. 35. . Cheliphlebia mazona. x1°8. curved ; subcosta nearly straight, with irregularly waved rami- fied branches that attain about three-quarters the length of the wing. Radius very near the subcosta and not reaching the tip of the wing, simple and uniting with the costa at a short dis- tance below the subcosta. Sector rising at one-third of the length, giving off only 2 oblique branches. Media cleft before the middle into an anterior simple and a posterior forked branch. Cubitus and anal area not preserved. Holotype in Peabody Museum, Yale University, Cat. No. 46. Family EHUCAXNID 4 Handlirsch. The finding of new specimens allows a more exact descrip- tion of this family to be given. What I had previously held Vicinity of Mazon Creek, Illinois. 359 for the whole radial system is only the radius (s. str.) without the sector, giving off some branches forward. What I thought to be the media evidently is the long radial sector, rising near the base and sending forth some long branches, directed obliquely backward. A part of the veins designated by me as belonging to the cubitus must therefore be attributed to the media, which is not abundantly forked. The cubitus, too, does not show a conspicuously rich embranchment. In the posterior wing the sector seems to be much more expanded and in consequence of this the media and cubitus are strongly reduced ; the anal area is large and fan-like. Coxe of the hind legs very large. Kucanus Scudder. Euceenus ovalis Scudder. Figs. 36-39. The Yale collection contains a male and two females of this species which may be easily distinguished by the presence of Fig. 36. Fie. 37. ~ Fie. 36. Hucenus ovalis.? x8. Fre. 37. Euceenus ovalis.2 x3. 360 Handlirsch—New Paleozoic Insects from the Fie. 38. Fic. 38. Euceenus ovalis. 6 x3 Fie. 39. ————— oot he R Fic. 39. Huccenus ovalis (reconstruction), x3. — Vicinity of Mazon Creek, Lllinois. 361 an ovipositor. The cerciare short. The ninth segment of the male is of the same size as the preceding. All the segments are laterally produced into blunt lobes. Pronotum compara- tively smaller in the male than in the female. Coxee of the posterior legs large, the femora robust and short. Ovipositor extending but little beyond the end of the abdomen. Anterior wing of the ¢ 238™™ long, of the 9 24-25™™. Besides these three individuals, the collection contains another female which does not have the wings lying over the abdomen as is usually the case, but they are spread laterally, therefore permitting of a more exact determination of the veins. Unfortunately in this individual only the basal half of the wings is preserved. This specimen is somewhat smaller than usual, having a wing length of scarcely 22™™. As far as I can see, the cubitus is cleft in one of the wings into but 4 Fie. 40. Fie. 41. Fie. 40. Eucenus minor (lof wing). @ 2-5, Fie. 41. Hucenus minor (negative of right wing). @ x2" Fie. 42. Fic. 42. Euccenus mazonus. x2°5. 362 Handlirsch—New Paleozoic Insects from the branches and it seems possible that this individual may belong to a species distinct from ovalis Scudder or at least to a variety, for which I propose the name /. minor m, (figs. 40, 41). Plesiotypes in Peabody Museum, Yale University. Cat. Nos. 47-50. Eucenus mazonus Melander. Fig. 42. Much smaller than ovalis Scudder with a narrower pronotum and a very stout body. Length of thorax and abdomen 15™", equal to the anterior wing. Costal area of the anterior wing comparatively narrow. Radial sector of the posterior wing with 4 large branches, occupying nearly the whole anterior half of the wing, pushing to one side the media and cubitus, which are reduced to 3 parallel veins. The large anal area fan-like, with numerous veins. Pronotum one and two-thirds times longer than broad. Anterior leg delicate, the tibia as long as the pronotum. Probably a male. Plesiotype in Peabody Museum, Yale University, Cat. No.'51. Eucenus pusillus, new species. Fig. 43. Resembling /: mazonus, but of a smaller size. The abdomen more slender, the costal area of the anterior wing broader. Fie. 43. Fic. 48. Hucenus pusillus. x4. Radius cleft at the end, the sector with (?) 5 branches. Length of the pronotum nearly 5™™, of the wing 11™. Holotype in Peabody Museum, Yale University, Cat. No. 52. EHucenus rotundatus Handlirsch. Fig. 44. The experience gained by the study of the new material has led to another determination of the veins of /. rotundatus Vicinity of Mazon Creek, Lllinois. 363 Handlirsch, as is shown in fig. 44. What I previously had called media evidently is the sector and the complex of veins formerly regarded as cubitus contains also the media. Holotype in the U. 8. National Museum. Fic. 44. Fie. 44. Eucenus rotundatus (new reconstruction). Family ANTHRACOTHREMMIDA) Handlirsch. Three of the fossils in the Yale collection seem to be of a near relationship with Anthracothremma Seudder. A com- mon characteristic of these forms is a greatly reduced and shortened anal area, a shortened subcosta and the tendency of all veins to stretch in the direction of the longitudinal axis of the wings. In all these species the media is of a comparatively small extent. PERICALYPHE, new genus. Pericalyphe longa, new species. Fig. 45. A slender insect with anterior wings 50™™ long which are at least four times longer than broad. The subcosta is strongly reduced and unites with the radius. Costal area narrow, traversed by short, stiff veinlets. The simple radius not quite attaining the apex; its sector rising beyond the first third of the length and sending off 9 simple and very regular parallel branches, directed toward the apical border. The quite inde- pendent mediai vein remains undivided. The cubitus, running in a large curve toward the end of the hind margin, sends off 6 long branches, some of which are ramified. All of these take the direction of the sector branches and of the media. Only one or two short branches near the base run obliquely to the posterior margin. The anal area, limited by a curved furrow, reaches but one-fifth of the wing length. Very indis- tinct cross veins probably fill all the interstices. In the posterior wing the subcosta is produced almost to the tip but likewise unites with the radius. The disproportionately small pronotum forms a semicircular disk of scarcely 5™™ in length. Meso- and metathorax are of a 364 Handlirsch—New Paleozoic Insects from the size proportional to that of the whole insect. The posterior leg is long and robust, its femur 15™" long and not thickened toward the base. This fossil seems to have its place very near 'S o.98 Se oOo oO ae ere M. Talbot—Podokesaurus holyokensis. 471 Description. In the description which follows these features are to be noted: Light construction—hollow bones; slender, straight femur ; ; position of fourth trochanter ; position of fibula, lying close to the tibia; great length of tibia and metatarsals ; ; small humerus ; narrow shaft of ischium; great length of pubis; length of vertebrae. Foreimb.—The humerus (fig. 1, H.)* is a very delicately shaped bone, £2"™ long (not quite half the length of the femur), slender and well rounded, with high radial crest. The proxi- mal end shows a slight constr iction above the crest and below the bone tapers gradually to a diameter of 3™™ at the distal end. Measured through the crest itis 8™". There is a trace of the impression of the radius or ulna just beyond the distal end. In the scapular region is a broad, flat bone, lying vertically in the rock, but twisted on its long axis at a right angle, mid- way of its "length (fig. 1, S.). This bone has a length, as shown, of 20™™ ‘and a width of at least 3™, Lying near its pr oximal end is another flat bone, 8™™ long by 5™ wide. These may be three separate bones, more or less firmly united in the living animal. Further development is necessary, how- ever, to bring out their outlines and their relation to each other. Hindlimb.—The femur (fig. 1, R. F. and L. F.) is slender and nearly straight with thin walls. The bone is expanded on the back’ side at the distal end. The length is 86™™ and the diameter, just distalwards from the fourth trochanter, is 6°5'"™. The fourth trochanter (fig. 1, T.) is 18™" long and about 2™™ high and is situated just beyond the middle of the shaft, toward the distal end. Only the proximal end of the right tibia (fig. 1, R. T.) is exposed and there the bone is well rounded. This, however, may be only a small part of the proximal end, as the bone is embedded in the rock. The left tibia (fig. ik L. T.) is split lengthwise, part of the bone lying in each half of the bowlder. It is an almost straight, narrow shaft with the surface lying uppermost bent slightly at the proximal end, due, probably, to the expansion of the bone. In the position in which it lies, the bone is of nearly the same diameter throughout, about 7™™. Its length is 104", Lying close against ‘the tibia and of almost equal length is the extremely thin fibula (hoses: Unie): There is a small, convex bone, 4"™ by 6™™, lying where the *The first four text-figures have been drawn by Miss Clara Gould Mark of Mount Holyoke College. For the privilege of using the restorations of Professor Marsh I am indebted to the editor of this Journal. The photo- graphs were taken by Mr. Asa 8. Kinney of Mount, Holyoke College. 472 M. Talbot—Podokesaurus holyokensis. right tibia and metatarsals should meet, that may be the astrag- alus (fig. 1, A.). It shows no sign of an ascending process. Two of the right metatarsals (fig. 1, Rt. Met.) lie in position. Their diameter is 2°5"" and 3°5™™, ‘respectively, along the shaft, but at the distal end there is an expansion to a diameter of 5™". In position, at the distal end of the left tibia there is one metatarsal (fig. 1, L. Met.) slightly curved and 65™" long. Alongside are traces of a second. “One of the digits (fig. 1, D.), whose divisions are indistinctly shown, lies between two ribs not far from the right metatarsals. It is 20" long and 1:5™™ in diameter and is terminated by an ungual 7™™ long. Cross sections of two bones near this one look like unguals. < Pelvis.—The pelvic bones are partly covered by the right femur and their outline is not distinct. What is probably the Fic. 2 Fic. 2. Pelvic bones. x1. F, Femur. Jl. lium. Js. Ischium. P. Pubis. pubs (fig. 1 and fig. 2, P.) is a remarkably long, thin bone, oo long, expanded at the distal end. The bone scems to be in position and makes an angle of about 40° with the line of the vertebral column. Its length is comparable with that of a new, undescribed form from northern Wirttemberg. The ischium (BI | fig. 2. fie cL vamd) sie 32, ‘Is.) is well rounded, anteriorly, aud “has a shaft 4™ wide of which a length of 30™™ is exposed.* The distal end is embedded in the rock. * The illustrations seem to indicate the presence of two bones but this may be due to the presence of a ridge on the ischium, as is seen in Comp- sognathus. M. Talbot—Podokesaurus holyokensis. 473 The contact with the ilium and the acetabular edge is obscurely visible. There is a bone running posteriorly from the head of the right femur which may be the posterior process of the ilium (fig. 1 and tig. 2, I1.), the anterior process being covered by the femur. This posterior extension can be traced indistinetly for 27™™ and either points upward or continues the line of the posterior part of the vertebral column. Vertebrw.—Of the vertebre there are visible seventeen pre- sacral (fig. 1, V.) and thirteen caudal (fig. 3, V.), all very hight Fie. 3. Fie. 3. Caudal vertebra and skull bones (?), x14. A, B. C. Bones of the skull (?). D, LE, F, G, V. Caudal vertebre. and hollow, and some, at least, slightly concave at each end. The presacral vertebree are slender, the measurements for the sixth presacral being 4™ through the middle of the centrum, 6™™ at the ends and 15™" from.end to end. A strong, grace- fully curved neural spine (fig. 1, N.S.) rises from the vertebrze in the dorsal region, about 10™™ high and 12™™ long at the base. The first two or three presacrals are a little larger than the others and those at the anterior end of the column, much stronger. One of the latter measures 10"™ at the end and the diameter through the middle of the centrum is only a little less. They are not so long, however, measuring only from 12"™ to 13™". One of these vertebre is shown in cross section at a thin edge of the rock and has a transverse diameter through 474 M. Tulbot—Podokesaurus holyokensis. the cavity of the centrum of 4" while the height of the cav- ity is 55mm (tig. 4). The caudal vertebre are only a little expanded at the ends and are very slender throughout their length. A typical one is 17"™ long witha diameter of 4™™ at the ends. The neural spines, if such they are (fig. 3, D.-G.), are of different shapes. These caudal vertebra are so nearly of a size, one with another, that there is no apparent tapering of the series and it is not clear which is the proximal end of the tail nor is it possible, as yet, to estimate its length. v2bs.—Quite a number of ribs (fig. 1, R.) are visible, all very slender and hollow, but the proximal end is not exposed save in one instance where the bone is so broken that the outline is not distinct. Near this proximal end, however, the bone is somewhat concave and expanded as if it might be bifurcate (fig. Hic. 2 Cross 7 RR’). he largest rib umcoveredais 925" section of cervical 1 Gomal orde tom ance Jopeiaeall eve ong and 2™" wide toward the proximal end, vertebra showing r F 3 hollow construc- while in no place does the thickness seem to ee xe, i be more than 1™™. The most anterior of the trum. NV Neural cervical vertebrae preserved have long ribs. canal. These are on another piece of the bowlder and are not figured. Just anterior to the distal end of the pubis there is a small cluster of gently curved abdominal ribs (fig. 1, A. R.), exceed- ingly slender and dove-tailed as’ they lie in the rock, the position due, probably, to slipping. This mass of interlacing ribs covers a space of 40" by 23". There are at least eleven of these ribs on each side of the median line, the largest of which is 18" long and so small in diameter that the bone looks like a mere thread in the rock. Slender as are these ribs, they seem to be hollow. Sternal element (?).—In the center of this mass of abdominal ribs is a small body that responds to dilute hydrochlorie acid, as do’ all of the bones. The part exposed measures 5™™ by 3™™. This may be one of the sternal elements displaced. Gastrolith.—Lying 10™™ away, and still among these ribs, is a small piece of quartz, a flat, well-rounded pebble, 1™" thick and 10°" long. The width exposed is 4™™, but more of the pebble is embedded in the rock. There are no other pieces of quartz larger than a grain of sand visible in the bowlder, and considering this fact, considering, also, its smooth, polished surface* and its position, the writer concludes that this must * Dinosaurian Gastroliths, G. R. Wieland, Science N. S., xxiii, pp. 819- 821. M. Talbot—Podokesaurus holyokensis. 475 be a gastrolith. This would seemingly be the first record of gastroliths found with carnivorous dinosaurian remains. Comparison with other forms. Herbivorous dinosaurs.—Compared with Vanosaurus agi- lis Marsh, the oldest known predentate dinosaur, the following points may be noted :* Points of similarity—The femur is shorter than the tibia, the ribs are very delicate, and the poste- rior extension of the ilium seems to have the same position. Points of dissimilarity—tIn this form the femur is more slen- der and is nearly straight, while in Vanosaurus it is distinctly eurved. In this form, too, the fourth trochanter is much nearer the distal end of the femur, the metatarsals are more slender and probably longer, the humerus is relatively much smaller, the shaft of the ischium is narrower, the pubis has a long ante- rior extension, and there is no postpubis. In Laosaurus consors Marsh,t from the Jurassic, the femur is much curved and is more nearly equal in length to the tibia, while the fourth trochanter is far up toward the proximal end ; the fibula is more bent, curving away from the tibia; the metatarsals are very much shorter; the prepubis is short and pointed ; and a postpubis is present. Laosauwrus has a short humerus as this form has but the shape is not the same. Hypsilophodon foxii Huxley (fig. 5), of the Wealden, Eng- land. Here we notice that the main points of difference are in the length and position of the pubis, and the presence of a postpubis in Hypsilophodon. In Hypsilophodon the femur is curved, but the fourth trochanter is situated more nearly as this one is, toward the distal end; the humerus is stouter and larger in every way with the radial crest much farther from the proximal end; the metatarsals are much shorter; the ribs are very much stronger; the neural spines are of an entirely dif- ferent shape. Carnivorous dinosaurs.—Anchisaurus colurus Marsh,t a very slender, long-limbed carnivore from the same Connecti- eut valley Triassic, with which this form seems to compare quite closely in general outline, shows the following differences. There is a decided difference in the relative lengths of the femur and tibia, the femur of the large form being much longer than the tibia. Marsh points out that the long femur is found in the larger animals of both carnivorous and herbivo- rous types, while the smaller birdlike forms of both types have the tibia longer than the femur.§ This small form certainly * Neubeschreibung des Originals von Nanosaurus agilis Marsh, F. von Huene and R. 8. Lull, fig. 1, p. 185. + Dinosaurs of North America, O. C. Marsh, pl. Lv. t Ibid., pl. tv. §Ibid., p. 201. 476 M. Talbot— Podokesaurus holyokensis. Fic. 5. Restoration of Hypsilophodon foxii Huxley, after Marsh. x1/8 natural size. Fic. 5. M. Talbot—Podokesaurus holyokensis. AV helps to confirm his observation. In Anchisaurus, the femur is much bent andthe fourth trochanter is high up toward the proximal end ; and all the limb bones are much stronger. The pubis of Anchisaurus is shorter, and runs at almost a right angle to the vertebral column; nor is it expanded at its distal end, as it is in this form. On the other hand, there is a general similarity in the size of the corresponding vertebra, though those of the caudal region of Anchisaurus are not quite so long and slender. Compsognathus longipes Wagner (fig. 6), from the Jurassic of Bavaria, corresponds quite closely with this form in length and general proportions of the limbs, in the shape of some of the neural spines, especially those just anterior to the scapu- lar region, and in the general shape of the ischium. The pubis of Compsoynathus is much shorter, proportionally, however, makes a larger angle with the vertebral column, and is much more expanded at its distal end. The shaft of the ischium is not so slender nor so uniform in width. Classification. No attempt at a definite placing of this fossil will be made in this paper, but certain conclusions reached by the foregoing study will be stated. The fossil is interpreted as that of a ~ carnivorous dinosaur because of the length, shape, and position of the pubis and the absence of a postpubis. Since no jaw has been found and there is no proof of the presence or absence of an ascending process from the astragalus, the determination of its position among the dinosaurs depends on the character of the pubis and the presence or absence of a postpubis. The position and length of the pubis are more nearly like those of the carnivores than those of the herbivores, and there seems to be no postpubis. The ischium is shaped somewhat like the bone in Scelidosaurus harrisont Owen, that von Huene describes as the ischium, with a query, however, stating that it may be the pubis, but not accepting that interpretation because he can find no obturator foramen.* The short, slender humerus and the long, straight hind- limb bones, together with the weil-developed fourth trochanter, are indicative of bipedalism. The length of the tibia, much greater than that of the femur, the extreme length of the meta- tarsals, over half that of the tibia, and the very light construc- tion of the skeleton, indicate rapid locomotion, which Lull uses as an indication of adaptation to climatic conditions, argu- ing that this animal must have been able to travel fast and * Ueber die Dincsaurier der Aussereuropzischen Trias, F, von Huene, p. 57. 478 M. Talbot—Podokesaurus holyokensis. Fic. 6. Fic. 6. Restoration of Compsognathus longipes Wagner, after Marsh. x 1/4 natural size. ; M. Talbot—Podokesaurus holyokensis. 479 far for water in a semi-arid climate. In support of this inter- pretation, also, may be added Barrell’s suggestion that the rock in which the fossil is embedded indicates deposition on a flood- plain in asemi-arid region. It is significant that many of these Connecticut valley sandstones that have dinosaur footprints have also raindrop impressions. Through the courtesy of Professor Schuchert the specimen has been sent to the Peabody Museum of Yale University for development, where further study will be given it by Pro- fessor Lull, to whom my thanks are due for many suggestions in regard to the interpretation of this fossil. Mount Holyoke College, April, 1911. 480 Branner—Minerals Associated with Diamonds and Art. XXXIX.—The Minerals Associated with Diamonds and Carbonados in the State of Bahia, Brazil; by J. C. BRANNER. SEVERAL persons have already studied and written upon the minerals associated with diamonds in Brazil.* Most of the papers, however, have been based upon the concentrates found in washing for diamonds in the Diamantina district in the state of Minas Geraes. Comparatively little has been said of the general geology of the Brazilian diamond region, and that little has not always helped to clear up the questions that arise in connection with the origin of the diamonds. Indeed the geology of the diamond district of Minas Geraes is far from being simple, and no reconnaissance has yet been made that shows the general structure and relations of the rocks and minerals. The diamond district of Bahia is north and a little east of Diamantina and about 650 kilometers distant on an air line. The paper published by Damour in 1853, two by Gorceix published in 1884, and one by Hussak published in 1899, deal with the minerals found with the diamonds in Bahia, but none of these papers contains anything on the stratigraphy of that district. Indeed, up to the year 1905 practically nothing was known in regard to the general geology of the Bahia diamond region. In that year Professor Derby made a trip into the region for the state of Bahia, and prepared a short report to the government on its geology. That paper was published in Portuguese at Bahia, and, on account of its importance, it was translated into English and republished in Hconomic Geology (vol. I, pp. 1384-142, Dec., 1905), and again in the Smithson- ian Report for 1906 (pp. 215 to 221). The paper by Derby represents our knowledge of the geology and mineralogy of the diamond regions of Bahia up to 1907, when I visited and traveled through that region for some months. General geology of the diamond regions of Bahia.—I have already published brief sketches of the geology of the Cha- pada Diamantina of Bahia, + and only enough will be repeated in this place to give an idea of its broad features. * Boué, Bovet, Brongniart, Damour, Derby, Dufrenoy, Eschwege, Gorceix, Hussak, Moissan, and others. +The economic geology of the diamond-bearing highlands of the interior of the state of Bahia, Brazil. Engineering and Mining Journal, Ixxxvii, 981, 1031. May, 1909. Outline of the geology of the black diamond region of Bahia, Brazil. Australasian Assoc. for Ady. Sci. 1908, 324-328. Brisbane, 1909. Carbonados in the State of Bahia, Brazil. 481 The table following shows the order of the rocks, their approximate thickness, and ages, in so far as the ages are now known. In the adjoining state of Sergipe, Cretaceous fossils have been found in abundance, and some have been found in the vicinity of the city of Bahia, but below the Cre- taceous nothing has thus far been discovered that makes it possible to determine the ages of the beds with certainty.* The divisions mentioned in this paper are based solely upon stratigraphie and lithologic characters and physical breaks, and must therefore be regarded as tentative until something more trustworthy can be found. It should be added, however, that these divisions when put to the test in the field appear to be legitimate, and, for the most part, widespread. TABLE OF GEOLOGIC FORMATIONS IN THE INTERIOR OF BABTA. Names Thickness in meters Ages Ala g G as senlesmame x iepe gee tee SUE eS ie Ee Tertiary DELalperSerlese se eee Ss) eR is) Secs Cretaceous Salitverlimestomesss9- 22522 223505... 25 ace: Jura ? Estanciasred: beds 42-5222. 3.24 3 (0) Senter Mer seer cate Trias ? Lavras series (diamond bearing)700-_--.--------- Carboniferous Cambao quartzites __..----..-100-(possibly part of the Lavras). Caboclo shales (Paraguasst of Werby. Wess 2 Gaye Si DO Oe seep es Devonian ? JAcuIpS AMS a See: 100 (possibly part of the Caboclo). Tombador sandstones ---- ---- A Oe ae etre os Silurian ? Minas (or Jacobina) series ...1000......---.-.-. Cambrian ? Crystalline complex ..-. ------ a EMA Pre-Cambrian in part ? The Cambao quartzites and the Jacuipe flints appear to be local deposits that may belong to the larger divisions either above or below them. Attention is directed to the fact that the diamonds and car- bonados are found in the Lavras series, and to the further fact that the rocks of that series are everywhere of sedimentary origin, strongly false-bedded, and for the most part gently folded. The demarcation between the Lavras series and the rocks above and below is clear and well defined, with the pos- sible exception of piaces where the Cambao quartzites occur. Eruptives have been reported to me as occurring in the rocks of the Lavras series only in the region south of the Serra do Assurua about Fundao, and Guariba. These rocks I did not see personally, but they are reported by H. E. Williams, assist- ant of the Servicgo Geologico do Brasil, who, as my assistant, examined portions of the diamond regions. Mr. Williams says *JTt is reported that Carboniferous fossils were found in the interior of Pernambuco in 1910 by Dr. José Bach, but I have not been able thus far to get direct confirmation of the discovery. 482 Branner—Minerals Associated with Diamonds and these eruptives are dikes of diabase and that they break through the Lavras quartzites. It should be noted in this connection that the principal dia- mond deposits extend over a wide area in the interior of Bahia, and that nowhere have eruptive rocks been found cutting the diamond-bearing sediments except at the places reported by Mr. Williams. In the erystaliine complex, however, that underlies all of the sedimentary beds of the interior of Bahia, there are many kinds of eruptive rocks. So far as observed, the eruptives in the crystalline series are all old; at least they have not been found to pass up into the sedimentary beds in which the-dia- monds are found, with only the exception just mentioned. Of the sedimentary rocks the Minas (or Jacobina) series, as shown in the Serra de Jacobina, has its beds generally stand- ing at a very high angle. Whether the position of these beds is due to folding or to faulting is not yet entirely clear, though the weight of evidence seems to be in favor of the theory of faulting. ‘4 The Lavras quartzites.—Of the overlying Paleozoic beds the diamond-bearing Lavras series was most studied. The beds are usually more or less folded. The folds vary greatly: in some places they are so gentle as to be almost imperceptible, while in others they are highly distorted. In the mountains west of Gruna the beds are much contorted, broken, and faulted. Over most of the area, however, ane folds are not closely pressed, the dips varying from ten to forty degrees. This structure in connection with the sharp separation of the Lavras series from the underlying and overlying beds, makes the working out of the distribution of the diamond-bearing beds comparatively easy. The rocks of the Lavras series are nearly all pinkish quartz- itie sandstones and conglomerates. The pink color appears to be characteristic of the series everywhere except in places where weathering has permitted the leaching out of the color- ing matter. Analysis of the Lavras Quartzite Collected by J. C. Branner near Andarahy, State of Bahia. L. R. Lenox, analyst. sien (si\O))) 38 oo as ee pee SrA. SLOAN, Oxides of iron and alumina (Fe, 0, and Al 0 se = Aloe Lime (CaO) . Be ce in Do pe ee yee OG Magnesia (MgO) .- at A rages Smee ea Si! Sy SN trace 99°92% * This Journal, xxx, 390-391, Dec. 1910. Carbonados in the State of Bahia, Brazil. 483 The composition of the quartzite as shown by the analysis is about what might have been expected. The pink color is evidently due to the small percentage of iron present in a high state of oxidation. It has already been said that the beds are almost everywhere strongly false-bedded. It is a striking fact that the false dips are not variable in direction as we are accustomed to see them in such rocks, but they are remarkably constant. In the north- ern and eastern parts of the diamond district these false dips are almost invariably toward the north. So nearly universal is this rule that I venture to say that, in the region north and east of Morro do Chapeo, fully ninety-nine per cent of the false dips is toward the north. About Lengoes and south of there south dips are much more common. Occurrence of the diamonds.—Most of the diamond wash- ings in Bahia are in stream beds, either actual or abandoned. In other words, the diamonds are found chiefly in alluvial deposits. The position of these deposits, however, shows clearly that the diamonds come directly from the Lavras series. In a great many places the diamonds and carbonados are found in alluvial deposits along streams that flow over the Lavras beds only. Such is the case at and south of Morro do Chapeo and at Campinas. In some of the most productive areas the alluvial deposits have been long exhausted, and the miners now obtain the stones directly from the disintegr ating Lavras quartz- ites themselves. Between Lengoes and Andarahy the quartz- ites have been found so productive that the miners have removed the disintegrated rock down to where it is so hard that it cannot be removed with the hoes, mattocks, and almo- cafres used to scrape it away. At Ventura the diamonds and carbonados are found in grav- els that rest upon the Caboclo shales, but the valley is very narrow and the Lavras quartzites cap the hills to the north and west, while the stream itself rises in and flows for many kilo- meters over the Lavras beds, so that there seems to be no rea- sonable doubt that they have the same origin at Ventura as they do at Campinas, a few kilometers to the north, where they are taken directly from the disintegrated quartzites. Personally I have never seen a diamond or a carbonado in the original quartzite, but it seems to be a matter of common information that they have been so found. Professor Derby tells me that he himself has seen one such specimen. The evi- dence, therefore, all points to the Lavras sedimentary series as the immediate source of the diamonds and. carbonados in the state of Bahia. The minerals associated with the diamonds and carbonados. —The question naturally arises whether the diamond originated 484. Branner—Minerals Associated with Diamonds and in the Lavras quartzites, or whether they originated elsewhere and were washed out of their original matrix and deposited along with the other sediments in the Lavras beds. Perhaps this question might be readily answered under ordinary cireum- stances, but in this case it is not so simple. It was hoped that a study of the minerals associated with the diamonds might settle the question definitely; and it was in this hope that the present examination was undertaken. In order to study the mineralogical association of diamonds and carbonados in the diamond region of Bahia, I have obtained thirty-five samples of the concentrates from different parts of the Bahia fields, and have had the minerals in them separated and identified. Some of the samples were collected by myself, but most of them were obtained for me through the kindness of my Bra- zilian friends, Dr. Alencar Lima of Bahia, Col. Dias Coelho of Morro do Chapeo, and of Mr. Arthur R. Turney of Lencoes. The samples consist of the fine heavy materials jeft in the bot- tom of the wooden bateas at the final “clean-up” in washing for diamonds at various places in the state of Bahia. The different minerals were separated by the use of Thoulet heavy solutions, and the percentages were determined by weighing before and after the separation. The weights are not alto- gether precise owing to slight losses, one of which came from the solubility of the gold in the solutions used. Notes on the mimerals.—The table shows that the minerals are unevenly distributed, some of them being very abundant in one locality and lacking at another. As was to have been expected, quartz is by far the most abundant at all localities. The quartz grains nearly all have secondary enlargement ; that is, the cementing material of the diamond-bearing quartzites is partly quartz in optical continu- ity with the original sand grains. * Samples 1 to 5, inclusive, are from Mosquitos near Lengoes, and were given me by Mr. Arthur R. Turney of Lengoes, Estado da Bahia. +Samples 6 to 14, inclusive, were obtained for me by Col. Dias Coelho of Morro do Chapeo. The precise localities are not stated, but it is under- stood that they are from washings about Morro do Chapeo, Campinas, and Ventura. +15, from Col. Dias Coelho, is marked Tres Moitas, Jacobina Nova, which is on the west side of the Salitre Valley about 130 kilometers west of Villa Nova. $16, from Col. Dias Coelho, is marked Rio Cambio, Jacobina Nova. The locality is within the Salitre drainage west of the Serra de Angico and 160 kilometers west of Villa Nova. 17, from Col. Dias Coelho, has no locality label. “| Nos. 18 to 35, inclusive, were sent me by Dr. Alencar Lima of Bahia. The exact localities are given in the table. The Chique-Chique mentioned is near Mucugé and should not be confused with a city of the same name on the Rio Sao Francisco. GOS. a Neg en een eee SOquIOg SOp oSe10g Gy CUT Gi | eee oe |e gee aca enbiyp-enbiypO ‘svriesueiey Fe EG Gled | Pin eaieect [see hte rae oe enbiyO-enbriyp ‘etieqis ¢e GIE| G& 8-09 | ~~ |" ~~enbryp-enbryp ‘eurp ep preg mog gg 8 8 ' 1 ' ' 1 ' Lo SH SHHOSOHIOID ra lea ee | oi te | IE, eS Betets | UIE | Quake fe Pesan enbryO-enbiyp “vog esnop Tg eee il kin, | OO ee le eee: ORO Sa Ica eOs0cr| ela nily ot oe enbiqo-enbryp ‘opreursey 0g of ety sfi = 2 ESTP eral cl i ene | pcre [Pel | scale al eee leecher |u| hh a ee enbryQ-enbiyp ‘preg mog gz “> | @9F | ~~-~ |-enbryp-enbiypD ‘nssensereg ‘etteqig 9¢ ri a) ' ‘Ost ou 1SRR ' ' R 1 ta 2 > i) ' ‘ 1 t eee | eal abo x Be cr eee ut ev eCea te OC 0Ge nn S| ge eee = SOPURS SOP ZMT 1G ey 0-8 || =| © | 89 22) etre |---| 2-08) 8-71] soe | =-7- |---- enbryo-enbrqo ‘sunzeyy onsjop 9¢ ST Sc Pe Nie ie br | eg ta i Uta | Vera Re ia col fie ha 9.8 x |9.T i sie ast 6-6 a2 Po Ce ee enbiyp-enbiyp ‘oxleg op ovdep cS 2-0 ee) aa | G0 \eor | =_ | e0r| = | See | =F |----enbiyg-onbiqg ‘woeung op yeurp 7¢ ce 8-01 =-|--|--|=-|--|--|--|--|--|--]----]----| x | ---|----]----]---- SBOE ial peso aller 9-18} -¢9 | ~~~ |“ onbryp-enbryp ‘oxteg op preg miog gg --|--|--|--|--|--]--|--|--]--|----]----|---- |----|----]----]---- |----|---+ SESS SSC Rea Borie = 00) Cal seit eae enbiyp-enbiy9 ‘vog esnoy gz --|--|--|--|--|--]--|--|--|--| ----]----]---- |----]----]----]---- |----]---- PSS eanionh 5. a5 FOG lest | ae Se ae enbiyp-enbiy,) “eog esnog TZ oa) a eH | ia SN ri) eal ec Fert | irae CHO Me ee ae ee ee OT “COMI So: ice ke enbigo-enbiyp ‘eog vsnoQ 0g --|--|--|--|--|--|--|--|--|--]----|----]---- |----|----]----]---- $295) SSSCa BS roa elena. | oo {OOM tsae: cblcka sae enbiyp-enbryy op eaeidnrg 6] --|--|--|--|--|--]--|--|--|--]----|----|---- |----]----]----]----.|----|---- coec |josos x Sane -00L | “77> |7enbryp-enbiyoD ‘eog’esnop odumiey TL ==|==|--|=-\--|-- x |==|--|--] == aol ee HST ERTS = Seis iia Se KS ES == GG | ame eam |e elie | ee ae ee ae oodeqy op ox1oyy Jo AqrutorA pT | T g ---- 8.96 pa RG fo ogc iat ry en ae CUS CAON BULQOOBL “eIlezesuy epuezeg ‘ouquiey ony 91S Fo | ese quay | ogee IJe= asses se. BAON vUIqooEr ‘sezIoyy Seay, Cyt |) PoRs e(R.6) |] Seno. pses=esccssaceecacs oadey) op ono FT Poel eee |) Se Onecare ee Soo elec [et = re eee eae OO CeOG | einen al eej ee (=| | & SSG g len 4) oil > eek: eases == 90- | ~~" | ¥0-L | F0- G-L6 | 6-16 ; : B55 jlomacyooce Meg [ar olose=\i sess -00L -0g¢ Se: ee vinjueA pue seurduep i seal ameepa waa ee Wee 120 0 lee ee CO are eee Were = sla nO seon | Or snce ‘oodeyo op oatoyy Jo Aqruror A meiosis (2 |cs oe (08k 571g 16840. |= 61-0 l0¥-0 |-=-= | ¢2-0 | ---- | 08-96 | 29-c17 a EE Relea | > |. -= 08:0 |SS-0 |" | | «6-0 6:86 | GSE-7&6 os Scot is a8 en ane SP-0/40-0 |_| |_| 8 | |e H9 | GL-Gs J ene) ft | |= 2-1 - | x 1-==- o.ggle se | es hae een clei eee mele ice,| |) aoe fore me cee |) =-|-- Teed) 73 Glee mae) | | meee (1-7) (mae |G (pees | ete CCC al teen len QC an lata | ‘ --|--]-==-]----| @.ggl----|---- Per a ROPE aed Cel eemelee se eae aos lence crc gee « qeou ‘soznbsoyy J <> &= _ || Ste || Bat SI Fa ese See Uo = 6% DM eee ee SO ee ee SOM cnn ee eSen aco SN TS SE RS Bi) ORE RS SS OR Pee URES LOBES ee | HN oO tH 10 * ‘hg |, a | sured “ALITVOOT “a TANVS “ /91310 \\ "oqrueary, | mcegayejeian core) ‘oyrypoorg OJIZVUOT ‘oq TUOUNT] “oq TToIneyg ‘op QRUIO Ten. Joule “ou TQOUSRTL m00 ‘opETpoyeyzoO oull oy ‘OUI[BULINO J, ‘punog sjowamypy fo abnjuaowg—Uzvag ‘VIHVG JO DLVLG AHL NI SOGVNO@AVO GNV SGNONVI(] HLIM GALVIOOSSY SIVUGNI, 40 a1avy, .—FourtH Series, Vou. XX XI, No. 186.—June, 1911. 33 Am. Jour. Scr 486 Branner—Minerals Associated with Diamonds and The corundum repor ted is of the ruby and sapphire varie- ties, and is of a pale gray color. The magnetite is sometimes found partly altered to hematite and the large amount of hematite and martite found with the diamonds is to be referred to this source. Metallic copper was found in several of the samples, but as the pieces are very small, and invariably curled up like filings, it is supposed that they came from the pieces of sheet copper often nailed in the bottom of the bateas to mend holes, and that these small fragments have been scratched off by the gravels while the washing was going on. Dr. Aw Rog ers gives me the following note upori the Savas and diamonds found in these concentrates : “ The phosphate ‘favas’ are rounded grains with faint greasy luster and pitted surfaces. The colors are quite variable, faint yellow, bluish gray, and dark brown. Under the microscope fragments are almost opaque, with a faint action on polarized light and have an aggregate structure. The index of refraction is about 1°62 and the hardness about 5. In composition the ‘favas’ are hydrous phosphates of aluminum. They probably vary in composition as they do in specific gravity, some being a little greater than 3°3, and some a little less. “ Diamonds were found in two samples,—a single smoky cleavage in one sand, and thirty-five crystals and cleavages in another sample. These crystals are colorless, yellowish, green- ish, and pale wine-colored. The crystals are the common iso- metric forms, usually octahedral in habit, the faces being the octahedron, the dodecahedron, and the trisoctahedron, Some crystals are fairly sharp octahedrons with plane faces and grooved edges, but most of them have curved faces in oscilla- tory combination with each other. Several spinel twins of tri- angular shape were noticed. ‘The crystals vary in size from 1 to 2"™ in diameter. One carbonado was found in this sample.” Four microscopic slides have been made from the diamond- bearing quartzites collected south of Lengoes for the purpose of ascertaining whether any of the minerals that are so charac- teristic of the metamorphic rocks oceur in them in such manner as to show that they originated in the quartzites themselves. Unfortunately the only minerals thus far found in the micro- scopic slides are quartz, chalcedony, and tourmaline: quartz forms the mass of the rock, the cementing material is partly quartz but largely chalcedony, while the tourmaline appears to have been carried in solution into small irregular cavities and there crystallized out. The dark greenish bands made by the tourmaline are quite visible to the naked eye in hand speci- mens. ‘This particular microscopic examination does not, there- fore, throw any positive light on the subject. But in view of the ‘scarcity of the diagnostic minerals, no other result was to have been expected from the examination of so few slides. Carbonados in the State of Bahia, Brazil. 487 One other mineral which is not mentioned in the table is found in the diamond-bearing beds in Bahia. This mineral is a jet black, opaque, hydrocarbon having a lustrous conchoi- dal fracture, a hardness of 2°2, specific gravity of 1°51, and is related to asphaltum, but is not certainly identifiable with any known form. It was thought at first that this mineral might be especially interesting on account of its association with the diamonds and carbonados, but it occurs in large lumps and seems to have no apparent relation to them. Comparison of the minerals of Bahia and Minas.—In order to compare the minerals found in different districts I have brought together here the Bahia lists published by Damour, Goreceix, and Hussak,* and I have made up from various sources a list of all minerals reported from the diamond washings of Minas Geraes. In comparing the minerals from the Minas diamond wash ings with those at Bahia, it should be kept in mind that in Minas the diamords are found for the most part in river beds, either actual or abandoned, and that the minerals associated with them are the result of a long and high degree of concen- tration. In one instance to my own knowledge the Minas dia- monds were obtained from a metamorphosed rock that was broken up by weathering. In the Bahia region this long and complete concentration has occurred in some instances, but in others there has been practically no natural concentration since the minerals were set free from the quartzite matrix. Conclusion regarding the relations of the Bahia and Minas districts.—Taking the Bahia list of minerals as a whole and comparing it with the Minas list, we must conclude that the only difference between them that seems worthy of note is the presence of the carbonado in Bahia and its absence from Minas. The other differences, such as the finding of ilmenite, monazite, topaz, and spinel in the Bahia deposits, and the occurrence of tantalite, euclase, fibrolite, titanite, chrysoberyl, chromite, and klaprothine in the Minas gravels I regard as purely accidental, and lable to disappear with further search for those minerals in the regions from which they have not yet been reported. The tale and graphite in the Minas deposits may be regarded as purely local, as is that of asphaltum in the Bahia deposits. The diagnostic minerals—A study of this table at once shows that most of the minerals mentioned have no diagnostic value. Quartz, for example, may occur in rocks of any age, and in sedimentary, eruptive, deepseated, or metamorphic rocks, in pegmatites, or in ordinary veins. It cannot therefore be re- * Hugen Hussak.—Ein Beitrag zur Kenntnis der sogenannten ‘‘Favas” der brasilianischen Diamantsande. Tschermak’s Min. u. Petr. Mitteilungen, Xviii, 334-359, Vienna, 1899. 488 Branner—Minerals Associated with Diamonds and Minerals associated with Diamonds and Carbonados in the States of Bahia and Minas Geraes, Brazil. Bahia, by Bahia, Minas Geraes, various by by various authors, Branner. authors. heeBery) te hea Oa HEE ae (0) De Brookite = 2s see Ol Due eee Oi FRET eae (0) 8. Carbonado _-.--- On TER ed See Of ste Ree (0) 4. Cassiterite ...-- fe) 5. Chalcedony ---- ia de UE ee. 6. Chioritesseoeeee 0) 7. (Chromites sos see Pe Se A, Me ane my ce By oe Ce) 8. Chrysobery] 2u-'50) Si Sed 3 ai eee ) 9. Cinnebar.__..-- (0) - 10. Columbite ....~- (0) 11. Corundum -__.-.- OF sng sala hee ee ee ea Co) 12. “Cy anNte eae ee Oat Sek he Say OF on, pee eee ) 13. Diamond _____-_- Oi sete gare pee oe Ope ee ae (6) 14. Diaspore 2... ) 5s Huclase se] soe (6 ele So ees Rh A ae SE ot fo) 16. “Favas” ee PIO TERE eee ek Se ORR: Wale Co) 7 Ep rolite 2 ets oO. See aie ee BAM oe bse (6) 1esiGarnet@e ests On : Ce Ole: BEG EUS a are TO. Goldeee es piss Ont, © BO ee ee ae ro) 90. Hematite____--_- Spee s ee cee ho (Ovant te AL) Renee te) Ole imenitew sea OE ee ae oh at fe) po. Klaprothiney<.2.' \4). ii pede: teise win eee C) 23. Limonite ._...-- COTES tet Ao ee ©. gtr eee fo) 24, Magnetite ---- -- Onn cI ae eee Oy Aeon ete ) 95. Manrtite ses. Ot Reet eee eee Oe eee eee fo) OG. Monazite s4-6,625), HOu jae ere Co) OH, INUOVINC ae sees oS es eee On Nato ees 98° ‘Octahedrite. 55 00 22 ase Oi ih gee eet ae) 29: Perovekite.--c8. 1 eee ee eee (0) SOME Vier =.= eee OL ee ge pare ee OF ae ce nee ) Sls QI aeons soe ON e as Sere see O) ae Uae ee (0) 89. Rates See eS POT aR are oa ee eee One ne EO) 33.) Spinel see CO Gort os ) 34. Staurolite ___--- ORL. 2 ee ee (0) 35) Mantalites eee O' UGE AR SES De eee ee ee (0) 365 Hhitanitese ee ae 0} Se eee OGhS ANE to) 37. Lopame eee re: Saud Wicesk sO ieee: Soir bile: co) 384 Tourmahines 425 «Ou ce SO Wie ee Pe) 39. Xenotime __-.--_- OF Gok se 5 Sa ed ee (0) AM), PAKEGOM 222 54 5-55 Opis coe eee OMS oo ae nO Minerals found in the rocks, but not generally in the concentrates. AT, -Asphaltnm eee aa | S22. peace 0 AD: (Calcitesens eae MEA oe TG) AS) Heldspaneeaeee (0) a ce ae eee Meascrcs a: O) 44, Graphite: 22.15.01 fap Wifie eee = SoA ere cea ) A5. Kao line eee OM Lee oct a fo) AG Mica) eee (oy ethos ee Ooo sect (0) Mig lesions sos O. sosscooascae O} og ans hs ee O 48, Pyrolusite 22-22) 9) epee eee oO 49, "Tale. 322 {oo ee oe eee 0) Carbonados in the State of Bahia, Brazil. 489 garded as helpful in determining the Brazilian association of diamonds. x Omitting from the list all minerals of common or wide occur- rence, I have made up the following table of those which I regard as diagnostic in the state of Bahia, and the columns indi- cate their genetic rock associations : Table of Diagnostic Minerals associated with Diamonds and Carbondos in Bahia, Deep- Erup- seated Metamor- Pegma- tives granites phies tites Beye. tee ake 8 = ii x x IByRoOi® 25252552252 x x He ee Calssiterite: s2s. 20225” Ss Se re x Columbite..__....-.-- a re te x Corundum -_-_-_--_- ae ings Ks Ke os Cyanitersno sm ses a = x ae Diaspoves 2 22 s22- 2 - fh x ah WIOMEVATIO= 2. os oes aia Ke x od Octahedrite 25 22522 ie we x E Shiaimurolitk <2 2 55 ee ae - x? x ais Mambahiteers Ge owas ah oe ae x iganniterte ke 2 oi eo. a! x xX ep Nenotime:s sss 52 uae = Lae x we x ANE COME Mees ee ee a x x ae This table brings out the fact that the genetic relations of these minerals are plainly with deep-seated granites, metamor- phics, and pegmatites rather than with eruptives. In view of the widespread occurrence of diamonds in the Lavras quartzites in Bahia, and in view of the absence of eruptives from the vicinity of the diamond washings, one of the following conclusions seems to be warranted : First, either the Bahia diamonds or carbonados originated in the old granites and metamorphies that underlie the Paleozoic beds in Bahia and have passed down mechanically to their present position, or Second, they are one of the results of metamorphism of the quartzites themselves and have originated where we now find them. : Possible origin of the diagnostic minerals.—If the diamonds and their associated diagnostic minerals were derived from older metamorphic or eruptive rocks, where did they come from? Kast, west, and south of the diamond regions of Bahia are extensive areas of the crystalline rocks:mentioned at the beginning of this article, consisting of old, probably Archean, granites, gneisses, schists, and eruptives of many kinds, and similar to those upon which subsequent deposits have been laid down. If, therefore, such a source is assumed for the diamonds 490 Branner—Diamonds and Carbonados, Brazil. and their associated minerals, there is no great geographic dithculty in accounting for them. But if they originated i in the ancient rocks and passed down with their associates through ages to the Lavras series, the hardness of the minerals must have been a factor of eonsider- able importance in their history. The associated diagnostic minerals, including the diamonds, are arranged with reference to hardness § in the following tables: : Table showing the Hardness of the Minerals associated with Diamonds in Bahia. Minerals ‘ Hardness 1: Diamonds carbonados).: 44-225 acess: eae ele 2° \Corm dina se ee Ae eee ae re 9 34 ‘Beryl spinel)... 2) ied ee ene eee 8 46 ZATCOMe 2. ee eee ee pee ee 75 5. Cyanite, staurolite, cassiterite, diaspore be omen 7 65 Mamntaliters. << <2) ea} eee re eee cu 6°5 lee Broolstexscolumibitee sae a eee 6 8. Monazite, octahedrite, titanite, “favas”_. ..-- 5'5 Di, DRETIOTUENGHS Sel CANE MAE Nic isl eis ia ly att 5 The hardness of the diamond (and carbonado) is a strong point in favor of the theory of the possibility of its having been passed down through several geologic ages, but the chances of the survival of a mineral having a hardness below 6: seems rather small though not at all impossible. The value of this evidence, therefore, seems to be doubtful. Conclusion.—It is fully realized that the results of this study are mostly negative. There is no evidence, however, that the Brazilian diamonds are of eruptive origin. And they certainly were not brought into their present position in the Lavras quartzites by eruptive action of any kind. I am strongly inclined to believe that the diamonds of Bahia have their origin in the quartzites where they have been occa- sionally found in place, and that they are associated with min- erals characteristic of metamorphic rocks for the reason that those minerals also originated in the quartzites under the same conditions as the diamonds themselves.* *The occurrence of diamonds in place in South Africa and in Arkansas, and associated in both instances with peridotites, appears to have thrown the burden of proof upon any geologist who ventured to suggest any other than an igneous origin for diamonds in other parts of the world. But even this stronghold of the believers in the igneous origin of diamonds is seriously shaken by two important facts:— First, that diamonds are found in only a small percentage of the South African eruptive pipes. Second, that diamonds have been found in the garnets of the eclogites brought up into the diamond-bearing pipes by the eruptives. So that after all, a metamorphic rather than an eruptive origin of even the South African diamonds seems to be a possibility, if not a probability. [A. L. du Toit.—The diamond-bearing blue-ground and allied rocks of South Africa, Trans. Edin. Geol. Soc., 1x, 361-362, Edinburgh. | Berry—Engelhardtia from the American Eocene. 491 Art. XL.—An Engelhardtia from the American Kocene ; by Epwarp W. Bzrry. Tue walnut family (Juglandacez), which in the popular mind is fully rounded out by the enumeration of the walnut, butternut, hickory and pignut, consists of six or seven genera and about forty species scattered throughout the warmer parts of the north temperate zone and penetrating some distance south of the equator along the Andes in South America and in the East Indies. The Juglandacez are of considerable interest for a variety of reasons, chief among which, aside from their great economic importance, are their line of ancestors reaching back to the mid-Oretaceous, and because of the much discussed question as to whether their morphological characters shall be interpreted as primitive or as mere simplifications of a more highly organized stock. All of the genera have not adopted the same methods of dissemination and certain tropical and sub-tropical genera have kept the seed part of their fruits comparatively small and light, thus enabling them to produce large nnmbers of seeds with the same expenditure of energy required for a single wal- nut. Furthermore, instead of depending upon chance for the distribution of their latent progeny, the bracts which are nor- mally present throughout the family have developed enor- mously and serve as wings. ‘This is especially true of the genus Engelhardtia, a recent addition to which is the occasion for the present brief note. The genus Engelhardtia was described by Leschen in 1825 and contains about ten species of the southeastern Asiatic region. These range from the northwestern Himalayas through farther India and Burma to Java and the Philippines. The pistillate flowers are small and are grouped in paniculate spikes. They develop into small drupe-like fruits, each of which is connate at the base to a large expanded tri- alate inyoluere. A single little known species rarely represented in even the larger herbaria occurs in Central America and is the type and only species of the genus Oreomunnea of Oersted. This is much more restricted in its range than are its kin beyond the Pacific. Oreomunnea is very close to Engelhardtia, and for the purposes of the paleobotanist the two may be considered as identical since they represent the but slightly modified de- scendants of a common ancestry which was of cosmopolitan distribution during the early Tertiary. The present isolation of Oreomunnea furnishes a striking illustration of the enor- mous changes which have taken place i in the flora of the world 492 Berry—Engelhardtia from the American Eocene. in the relatively short time, geologically speaking, which has elapsed since the close of the Cretaceous. The principle has frequently been enunciated that when closely related forms are found in the existing flora of the world, restricted in range and isolated from their nearest relatives, or when the existing genera are monotypic, it is quite safe to predict an interesting “and extended geological history. Engelhardtia proves to be another illustration of this princi- ple, for its peculiar three-winged fruits have been known in the fossil state for almost a century. They were long unrecog- nized, however, and the earlier students who described them compared them with the somewhat similar winged fruits of the genus Carpinus (Betulaceze). With the botanical exploration of distant lands in the early part of the 19th century, specimens of Engelhardtia began to be represented in the larger European herbaria, and Baron Ettingshausen,* that most sagacious of paleobotanists, as long ago as 1851 pointed out that certain supposed species of Carpinus were really fruits of Engelhard- tia. He returned to the subject in 1858+ without, however, actually changing the names of any of the supposed species of Carpinus nor does he seem to have been aware of the existence of a living species of Engelhardtia in Central America. Since Ettingshausen’s announcement a dozen or more fossil species have been described. The oldest known occurs in the upper Eocene or lower Oligocene (Ligurien) of France and the species become increasingly abundant throughout southern Europe especially toward the close of the Oligocene and the dawn of the Miocene, Saporta stating that the “slabs from the leaf-beds at Armissan in southeastern France are thickly strewn with their peculiar fruits. Fossil forms continue in Europe throughout the Miocene and Pliocene and specimens of late Miocene or early Pliocene age are recorded from Spain, France, Italy, Croatia and Hungary. The accompanying sketch map of the world (fig. 1) shows the existing distribution of Engelhardtia and Oreomunnea some- what generalized and exaggerated in order to be seen on so small a scale map. The Tertiary occurrences of Engelhardtia are indicated by stars, a single star covering all of the records in a single area, as for example, southeastern France, from which one Liourien, three Tongrien, five Aquitanien and one Pontien occurrences have been recorded. No fossil American species have been previously known with any degree of cer- tainty. Lesquereuxt in 1883 recorded Engelhardtia oxyptera * Ettingshausen, Die Tert. Fl. von Wien, Abhandl. k.k. geol. Reichsan- stalt, Wien, xi (8), p. 2, 1851. + Ettingshausen, Beiter. z. Kennt. foss. Fl. von Sotzka, Sitzungsb. k. Akad. Wiss. Wien, xxviii, 1858, p. 12, pl. iv, fig. 4; pl. v, figs. 1-3. ie ee Cret. and Tert. FL, U.S. Geol. Surv, Terr., vol, viii, p. 192, Berry—Engelhardtia from the American Eocene. 493 Saporta* from the Miocene of Florissant, Colorado. I have been unable to get track of this specimen and as no specimens of Engelhardtia have been detected in Professor Cockerell’s recent and extensive collections from this locality and as Carpinus is not at all uncommon, it seems probable that Les- quereux’s determination was based upon material of the lat- ter genus, particularly as the Florissant yen beds are consider- ably younger than the type locality for Hngelhardtia oxyptera, Fie. 1. 0 reomunnea wth i Fic. 1. Sketch map showing the existing geographical distribution of Engelhardtia (vertical lining) and fossil occurrences (stars). which was Armissan, France. In any event the Florissant material differs markedly from the present species from Missis- sippi, having a different venation and being only half the size of the latter. With these introductory remarks we may proceed to the de- scription of an unmistakable species of Engelhardtia recently recognized from the Eocene deposits of northern Mississippi. The single specimen of the fruit shown in the accompanying text figure was collected in 1889 by W. J. McGee while engaged in his studies of the so-called Lafayette formation and has lain unstudied in the collections of the U. 8S. National Museum since that date. *Saporta, Etudes sur la Végét. du sud-est de la France & l’époque Ter- tiaire, xi, p. 344, pl. xii, fig. 2, 1866. 494. Berry—Engelhardtia from the American Hocene. The writer has collected leaves of Engelhardtia in these deposits, but since their exact relationship with the fruit is unsettled they will not be considered at the present time. The species, which is new, may be called Engelhardtia (Oreomunnea) mississippiensis sp. nov. Description.—Involuere large, trilobate, somewhat reflexed. Ale widely spreading, the angle between the median and Fia. 2. n wry ' ' ' ' ’ ' ' ‘ o im - Ss Bh 5 } SSS Ss, ~> p 4 8 SSE Yai oy =a or oa LS. fee eA a LEONA MOTT OY aril, 4, ey Aes. \ a UX ay aos « NY Ss) SS | YA Y Mee Rann A z os Fic. 2. Engelhardtia (Oreomunnea) mississippiensis, sp. nov. from the Eocene of Mississippi (nat. size). lateral wings being 70° to 80°. Sinuses correspondingly open, rather straight-sided, rounded at the angle, which is 1:5" from the extreme base of the specimen. The median wing is the longest of the three and is equilateral, spatulate or oblanceo- late in outline, expanding gradually distad from a basal width of 8™™ to a width of 18™", where the distal portion is broken off, 5°" above the base. Since this apical part is missing the total length is estimated at 6°5°", which is a minimum rather than amaximum estimate. Lateral wings slightly inequilateral, the outer part of the lamina being a trifle wider than the inner. Apex rounded. Length 5°". Greatest width, which is above the middle, 11™™. Least width proximad, 7™™. Primaries three in number, one median primary being present in each wing. The primaries are relatively very stout and continue with but slight attenuation to the tips of the wings. No sub- ordinate primaries or discordantly directed secondaries are present as in some of the European Tertiary species. Berry—LEngelhardtia from the American Eocene. 495 Secondaries numerous, thin, more or less parallel, about twelve to fifteen pairs to each wing, alternate. The secondaries branch from the midvein at a wide angle which becomes progressively less distad, where they are placed at more frequent intervals and are more regularly curved, camptodrome throughout. Ter- tiaries extremely fine, forming small arches just inside the margin and more or less rectangular meshes within the spaces bounded by the secondaries. Margins strictly entire throughout. The essential portion of the fruit is poorly preserved and partially broken away, as is usually the case in the fossil species of this genus. It appears to have been of considerable consistency, and the whole fruit having fallen face downward the reflexed wings raised the peduncular portion, which either rotted away before fossilization or, what is more probable, was broken off when the specimen was collected. ‘Among previously described Tertiary forms the present species is most similar to Engelhardtia Brongniurti Saporta, a species recorded from Spain, France, Italy, Germany, and Austria-Hungary and supposed to range from the Oligocene to the Pliocene. The American species is somewhat larger than the usual size of Hngelhardiia Brongniarti, although Unger has figured forms of the latter which do not differ much in size from Sotzka in Styria. The wings are more spreading and the outlines are much more elegant. In the European form the wings are rounded apically as in the American species but they are approximately the same width throughout and do not taper downward as they do in Engelhardtia mississippi- ensis. The secondaries, instead of being regular and camptdo- drome as in the latter, are less numerous and more irregular in position, several in each wing ascending from the base for considerable distances approximately parallel with the midvein. Among the existing species with which it has been compared Engelhardtia Mmississippiensis is very similar to most of the described oriental forms, perhaps resembling Hngelhardtia spicata Blume more closely than the others. The latter ranges from the northwestern Himalayan region through Burma to Java and other East Indian islands. Comparative material of Oreomunnea is very scarce. A single fruit in the National Herbarium is closer to the fossil than are any of the Asiatic species, but in the absence of more material the limits of vari- ation in Oreomunnea are unknown. In a general way Engelhardtia fruits are not unlike those of Carpinus, as has already been mentioned. There seems to be little occasion for confusion, however, even in poorly pre- served fossil material. The fruit proper is decidedly dif- ferent, although this is seldom well enough preserved in 496 Berry—Engelhardtia from the American Eocene. fossils to be decisive. The involuere is also markedly differ- ent in the two genera. Carpinus involucres are usually smaller with the median wing much wider and longer than the lateral wings and with somewhat different venation. The margins are also toothed while in Engelhardtia they are always entire. Ihave examined fruits of all of the existing species of Carpinus and experience no difficulty in readily dis- tinguishing them from those of Engelhardtia, the American species of the former being especially different in appearance from those of Engelhardtia. I have seen involucres of the old world Carpinus betulus from trees cultivated in this country in which the wings had entire or nearly entire margins, but the aspect of the specimens as a whole, because of their different proportions and venation, was markedly unlike Engelhardtia, and if they had been found as fossils no com- petent paleobotanist would have been at a loss regarding their botanical attinity for a single instant. Engelhardtia mississippiensis was collected from a locality about one mile southeast of Early Grove in northeastern Marshall County, Mississippi, a few miles from the Tennessee border. The age of the beds is Wilcox Eocene, as indicated by the large flora associated with the present species. Johns Hopkins University, Baltimore. Gooch and Kuzirian— Use of Sodium Paratungstate. 497 Arr. XLIL—The Use of Sodiwm Paratungstate in the Deter- mination of Carbon Dioxide in Carbonates and Nitrogen Pentowide in Nitrates by Loss on Ignition ; by F. A. Goocr and S. B. Kuzrrran. [Contributions from the Kent Chemical Laboratory of Yale Univ.—cexx. | From certain carbonates, like those of magnesium, zine and cadmium, carbon dioxide may be expelled by simple ignition at a moderate temperature, leaving an oxide in definite and weighable condition. In the case of calcium carbonate, this process of decomposition is completed only at the high heat of the blast lamp, and the reaction, being easily reversible in the atmosphere containing carbon dioxide evolved in the igni- tion or produced in the source of heat, may leave the oxide not quite pure. Strontium carbonate and barium carbonate are not entirely broken up by simple ignition under the conditions ordinarily available in analysis; nor are the alkali carbonates. For the decomposition of refractory carbonates it is customary to make use of a suitable flux which, by combining with the oxide, will aid in the expulsion of the carbon dioxide. Anhy- drous borax,* silicon dioxide,t potassium dichromate,t and recently, sodium metaphosphate§ have been thus used in the analysis of carbonates, and they are applicable similarly to the determination of nitrogen pentoxide in nitrates which leave definite oxides on ignition. Such fluxes, moreover, serve the very essential end of conserving the residual oxides in definite and stable form for weighing under the ordinary atmospheric condition of the balance room. Of those mentioned the first two, borax and silicon dioxide, require prolonged ignition to bring them to constant weight before making use of them to react with the carbonate or nitrate ; and generally they yield in the fusion process a pasty magma, so that prolonged heat at a high temperature is necessary to the complete expulsion of the gaseous product. Sodium metaphosphate, though more fluid in fusion, also demands prolonged care in the preparation. Potassium dichromate is too easily decomposed with loss of oxygen to be employed in exact processes demanding long con- tinued fusion or heating to temperatures much above its fusing point. These are points which have been sufficiently emphasized in the work to which reference has been made. * Fresenius : Zeitschr. anal. Chem., i, 181. + Rose: Ann. Phys., exvi, 635. Fresenius: Zeitschr. anal. Chem., i, 184. Richards and Archibald, Proc. Am. Acad., xxxviii, 443. t¢ Rose: Ann. Phys., exvi, 131. Fresenius : Zeitschr. anal. Chem., i, 188. § Lutz and Tschischikof: Chem. Zentralblatt, 1905, i, 564. Bdttger: Zeitschr. anal Chem., xlix, 487. 498 Gooch and Kuzirian— Use of Sodium Paratungstate. In sodium paratungstate of composition corresponding approximately to the formule 5Na,0.12WO,, or Na,,W,,O,, we have material very easily prepared, stable in fusion, and well suited for use as a flux in the rapid determination of the loss of carbonates and nitrates on ignition. For the following experiments the sodium paratungstate was prepared by dehy- drating and fusing over the blast-lamp a known weight of normal sodium tungstate, Na,WO,.2H,O, adding an equal weight of tungsten trioxide, WO, (previously ignited with care to remove all ammonia and to insure complete oxidation), and heating to clear fusion. The cooled mass, which is very easily pulverized, was ground in a mortar and bottled. From this material kept in a desiccator over sulphuric acid (though not more than ordinarily hygroscopic), portions were weighed for the analytical determinations. Approximately half the weight of the paratungstate is tungsten trioxide (molecular weight 232) and this should be capable of expelling carbon dioxide (molecular weight 44) to the amount of one-fifth its own weight, and of nitrogen pentoxide (molecular weight 108-02) to an amount one-half its own weight. The weights of paratunestate used were, therefore, always in excess of ten times the weight of carbon dioxide and four times the weight of nitrogen pentoxide to be expelled. In the following deter- minations it was the practice to weigh a platinum crucible, introduce the dried carbonate and weigh again, add a suita- ble amount of the prepared sodiuin paratungstate, stir carefully with a platinum wire with care to avoid mechanical loss, and weigh again. The crucible was heated over a bunsen burner, first at very low heat and then to fusion of the mixture for five minutes, cooled in a desiccator over sulphuric acid, weighed, and re-ignited to test the constancy of weight. ‘The constant weight was usually got in the first ignition. In Table I are given the results of the estimation of carbon dioxide in calcite. TasreE I. Analysis of Calcium Carbonate ( Calcite). CaCO; NasoWi201. Loss on Theory taken taken ignition for COz Error grim. erm. grm. erm. grm. 0°5000 2°5 0°2195 0°2198 —0:0003 , 075000 2°5 0°2206 0°2198 —0°0008 0°5000 2°5 0°2200 0°2198 —0°0002 05000 2°5 0°2203 0°2198 —0°0005 0°5000 275 0°2200 0°2198 —0°0002 0°5000 2°5 0°2204 0°2198 — 00006 0°5000 2°5 0°2190 0°2198 —0'0002 0°5000 2°5 0°2200 0°2198 —0'0002 Gooch and Kuzirian— Use of Sodium Paratungstate. 499 The strontium carbonate used in the experiments given below was prepared with great care. Strontium chloride, C. P., was partially precipitated by strong hydrochloric acid, washed with hydrochloric acid, dissolved in water and added in dilute solution drop by drop from a stoppered funnel to a saturated solution of ammonium carbonate heated to the point of decomposition. The precipitated carbonate was carefully washed, dried, washed again with ammonium carbonate and then with boiling water, and dried below red heat. In Table II are the results obtained with this preparation of strontium carbonate. Tasxe II. Analysis of Specially Precipitated Strontium Carbonate. NayoW12041 SrCO3 taken Loss on Theory for taken (approx. ) ignition CO, Error grm., grm. grm. grm. grm. 0°5000 2°5 01488 0'1490 —0°0002 0°5000 2°5 0°1494 0°1490 + 0:0004 0°5000 2°5 0°1494 0°1490 + 0°0004 0°5000 2°5 0°1490 0°1490 00000 0°5000 2°5 0°1496 0°1490 +0:0006 0°5000 2°5 0°1486 0°1490 —0:0004 With barium carbonate prepared from the chloride by par- tial precipitation with strong hydrochloric acid, solution of the precipitated chloride in water, gradual addition of the solution with constant stirring to a hot solution of ammonium carbonate, careful washing, and drying, the following results were obtained : Taste III. Analysis of Specially Precipitated Barium Carbonates. NayoW12041 BaCOs taken Loss on Theory for taken (approx. ) ignition CO, Error erm, grm. grm. grm. grm. 0°5000 2°5 01120 O°1115 + 0°0005 0°5000 2°5 0°1125 071115 + 0°0010 0°5000 2°5 0°1109 O'1ll15 — 00004 0°5000 2'5 O°1113 O'1l1l15 —0°0002 075000 2°5 0°1123 0-1115 — 00008 The similar use of sodium paratungstate for the determina- tion of the nitrogen pentoxide in nitrates is shown in the following table : 500 Gooch and Kuzirian— Use of Sodium Paratungstate. Tasie IV. Analysis of C. P. Nitrates of Commerce, after Drying. Nitrate NajoWi2041 Loss on Theory for taken taken ignition N20; Error grm. grm. grm, grm. ent, KNO, 0°5000 1.5 0°2668 0°2670 — 00002 075000 15 0°2678 0°2670 + 0°0008 0°5000 1:5 0'2674 0°2670 +0°0004 0°5000 15 0°2672 0°2670 +0°0002 0'5000 15 0°2675 0°2670 +0°0005 Sr(NO,), 0:5000 2 0°2544 0:2543 +0:0001 0°5000 3 0°2546 0°2543 +0°0003 Ba(NO,), 0°5000 3 0°2073 0°2067 + 0:0006 0°5000 3 0°2076 0°2067 +0°'0009 From the results of the experiments described above it is obvious that sodium paratungstate, easily prepared and stable, makes an excellent flux for use in the rapid determination of carbon dioxide and of nitrogen pentoxide by loss on ignition. Johnston and Adams—Melting Points of Metals. 501 Arr. XLI1.—The Influence of Pressure on the Melting Points of Certain Metals; by Joan Jonnston and L. H. ADAMS. Tue authors have been engaged in developing methods and apparatus by means of which it will be possible to investigate the effect of high temperatures and pressures on certain systems and reactions, and especially those in which water plays an impor- tant part. The work has progressed until now we are able to introduce into the bomb current leads and thermoelement wires in such a manner that the wires are all thoroughly insu- lated electrically, and the joint remains absolutely pressure tight. Thus, it is possible to heat a substance to somewhat over 400°,* under pressures up to 2000 atmospheres ;+ and to measure both temperature and pressure with precision. More- over, the whole system, by reason of the special methods of construction adopted, is absolutely free from pressure leaks, even when the bomb is repeatedly closed and opened, discon- nected from, and reconnected with, the remainder of the high pressure system. or instance, on one occasion heating was continued for 30 hours continuously at a pressure of 1800 atmospheres, without sensible loss of pressure in the whole interval. Before proceeding further in the main purpose of the work, it seemed desirable to make a rigorous test of the precision of temperature measurements under these conditions. For this purpose, determinations of metal melting points are suitable by reason of their definiteness. Accordingly it was resolved to investigate the effect of pressure upon the melting point of tin, bismuth, lead and cadmium, the four metals which lie within our present range of temperature. The results are eminently satisfactory, for, as we shall see, a precision in the temperature measurements of about 0:02° was reached, using a thermocouple of copper-constantan, for which an electro- motive force of one microvolt corresponds to about 002°. It may here be said that the results of this investigation are in entire accord with what was to be expected from the magnitude and direction of the heat change and volume change which accompany the process of melting; and that the small discrep- ancies between the observed and the calculated values are to be ascribed mainly or entirely to uncertainty in the experimental determinations of the latent heat of fusion and of the change of density on melting. * The highest temperature the oil will stand without serious charring. + The limit of the present compression pump. Am. JouR. Se —Fourti Srerizs. Vou. XXXI, No. 186.—Junz, 1911. 502 Johnston and Adams —Influence of Pressure on the Briefly, the method pursued in this investigation is as fol- lows: A charge of the metal to be investigated was placed in a suitable apparatus for heating it under pressure ; its freezing point or melting point was then determined, at constant pressure, using the Frankenheim method. A small electric resistance furnace supplied the heat; a thermocouple was used to measure the temperature of the metal; and a high boiling paraflin oil served to transmit the pressure. Description of the Apparatus. The essential parts of the apparatus are: the bomb, in which the substance to be heated under pressure is placed; the ptunp, with which to supply the pressure; and a gage for measuring it. In fig. 1 is shown the arrangement of these parts, together with the accessory apparatus consisting of valves, pressure- line connections, oil-reservoir, thermocouple, heating current wires, ete. In this diagram the bomb is drawn to seale but the remainder of the apparatus only approximately so. The bomb itself (the interior of which is shown in greater detail in fig. 3 was built up of a cylinder of machine steel AX, on to which were shrunk a ring of nickel steel 47 and a number of thin- ner rings of boiler-plate separated from one another by a space of about4™™. Water enters at W,, circulates between the rings and around the central cylinder AX, and leaves at W,.* The bomb is closed by means of two hardened steel plugs, G and H, held in place by a 500 ton hydraulic press, the platens of which are shown at PP. On the shoulder (LZ, fig. 3), of either plug is turned a V-shaped ridge, and on the adjacent shoulder of the bomb a groove to correspond. Between the groove and ridge lies a ring or “ washer” of thick sheet cop- per; by maintaining with the press a force of about 20 tons in excess of that exerted on the inside surface of the plug by the internal pressure, the bomb is effectively and easily closed. Opening the bomb again—after it has been removed from the press—is effected with the aid of a large nut which fits the heavy screw thread cut on G. Experience has shown that without some device for center- ing the bomb and plugs with respect to the press platens, good and certain closure of the bomb is not possible, time atter time. The arrangement used for this purpose is worthy of mention; it is very simple and consists merely of the curved surfaces of A and G on the one end, and of A and #& on the other. By this arrangement practically perfect alignment is *This water-cooling device leaves nothing to be desired as regards effi- ciency, but is somewhat difficult to construct and to make water-tight. It is not of our own design. the bomb having been reconstructed from one which had been designed and built by Dr. A. Ludwig, Melting Points of Certain Metals. 503 secured automatically, and a pressure-tight closure of the bomb thereby insured. Fie, 1. YZZZZZZLILE ZZZAUN 4 N PZZZZZZ2Z11k b WZZZZZZZZZ2 Qay WY Yi Fie. 1. General arrangement of apparatus. The bomb is drawn to scale ; the rest of the apparatus only approximately so. Pressure is supplied by the pump D, and transmitted through the system by means of a paraffin oil which has at the same time a high boiling point and low viscosity. Fisa steel bottle of about one liter capacity; it serves to increase the volume of the system. The valve V (shown in detail in fig. 2), is used to lower the pressure in the system. MM is a Bourdon gage graduated to 3000 atm. in divisions of 50. The connection block CO affords a 4-way connection between pump, valve, gage, and bomb. The latter consists essentially of a ring of nickel-steel RR, and a number of rings of boiler plate shrunk on to a cylinder of steel KK. On the outside of the rings is shrunk a thin cylin- der. The cooling water enters at Wi and leaves at W2. The bomb is closed by means of the steel plugs G and H, which are held in place by a hydraulic press the platens of which are shown at PP. The curved surfaces of A and B on one end and of A and G on the other constitute the essential parts of the device used for securing alignment of bomb and plugs. JJ is the electric furnace (for details see fig. 3), and SS soapstone blocks for thermal insulation at the ends. Thermocouple wires are shown at TT, and heating current ‘wires at Z. 504 Johnston and Adams—Influence of Pressure on the The electric furnace, depicted at //, and the arrangement of the interior of the bomb, will be described later on. The pressure in the system may be raised by means of the pump JP (fig. 1). It has a plunger 8™™ in diameter, and is of the type or rdinarily used with hydraulic presses. With this pump we have obtained without difficulty pressures up to 2000 atmospheres. It communicates through the pipe O with the oil-supply tank (not shown in the diagram) and forces oil out through the steel bottle #’ and the connection block C into the bomb. The bottle F’, which has a capacity of about one liter, serves to increase the volume of the system, which otherwise would be less than 100°, and thus makes for greater constancy of pressure during heating and cooling and during changes such as occur when the material under - investigation melts or solidifies. The connection block C affords a 4-way connection between the pump, the bomb, the gage, and the release valve V. The gage M, of the Bourdon type, is a new one manufactured by Scheeffer and Budenberg, and was calibrated by them. It is 8 in. in diameter and is graduated up to 8000 atmospheres in divisions of 50. The indications of Bourdon gages, as is well known, are subject to a hysteresis effect; that is, the reading corresponding to a certain pressure may vary slightly, depend- ing on whether the pressure is rising or fallmg. From various considerations, however, the writers are led to believe that the discrepancy in the present instance is small and that the readings of the gage are subject to a probable error no greater than + 5 atmospheres,—an accuracy quite sufficient for the present purpose.* The valve V serves to lower the pressure in the system or to release it entirely; it is shown in greater detail in fig. 2, which is drawn exactly to scale. The valve-stem, which ends in a cone fitting into a conical depression in the valve block, is kept in place by the nut m. The packing consists of three layers of leather surmounted by a steel disk. The pitch of the screw between stem and nut is identical with that between nut and block; by this means the valve is not jammed down into its seat, when the nut m is tightened. Fig. 2 serves also to illustrate the type of high pressure connection which has proved uniformly satisfactory for mak- ing all the connections necessary in work, such as the present, where high pressures are employed. On account of its sim- * We have under construction an absolute gage, of the piston type described by P. W. Bridgman (Proc. Am. Acad., xliv, 119-251, 1909), by means of which we shall be able to measure pressures with an accuracy of 0°1 per cent. In the present work its use was unnecessary by reason of the comparatively small influence of pressure on the melting points which we have investigated. Melting Points of Certain Metals. 505 plicity of construction and certainty of operation, it seems desirable to describe it more fully. The end of the pipe p is turned off at an angle of 60°; a small round nut is put on, using a left-handed thread, and the whole is held in place by means of the larger nut », which is undercut to receive the smaller nut. The end of the pipe fits into a conical hole also turned accurately at an angle of 60°. An essential part of the procedure is that before the joint is Rie. 2. Fic. 2. Details of high pressure valve and connections, drawn to scale. The main body of the valve is of best quality nickel steel. The conical lower end of the valve spindle s fits into a conical depression. The leather packing around the spindle is held down by the nut m. On either side of valve is shown the type of high pressure connection which has been used for joining various parts of pressure apparatus. The end of the pressure pipe p is turned off at 60° and fits pressure-tight into a conical depression. A small round nut is screwed onto the end of the pipe and the whole is held down tight by means of the larger nut n. assembled for the first time, the pointed end of the pipe should be given a blow with a hammer squarely on its end. This treatment enlarges the tip slightly and has the effect of making certain that when the end of the pipe is forced into its seat by screwing up the nut it is the extreme tip that binds on the surrounding metal. This type of joint may occasionally leak when pressure is applied for the first time; increase of pres- sure, however, makes the joimt tight by expanding the tip of 506 Johnston and Adams—Influence of Pressure on the the tube against its seat, and when this has once happened, it remains tight indefinitely. Joints such as this have been sub- jected to pressure up to 8,000 atmospheres, with both liquids and gases; when properly made, they show not the slightest leak whatever. This type of connection possesses the follow- ing distinct advantages over other types which have been used for high pressure work; (1) it may be taken apart and put together again repeatedly without impairing its efficiency ; (2) the joint is tight every time and thus there are avoided the troublesome delays so common in high pressure work ;* (3) the danger of the whole joint blowing out at high pres- sure—which may easily happen with ordinary butt joints—is practically obviated, because of the smallness of the bearing surface under all conditions. The tubing which has been used for all connections is of mild steel, cold-drawn, 12™™ in outside diameter and 2™™ inside diameter. It stands without rupture pressures of at least 8,000 atmospheres and can easily be bent or twisted if heated to redness. We shall now proceed to describe the disposition of the interior of the bomb, comprising the arrangements for heating, and for measuring the temperature of, the charge of metal ; the details are shown in fig. 3, which is drawn to scale. The electric furnace JJ (fig. 3) was made by wrapping a thin sheet of asbestos paper around a copper tube 25™™ in diam- eter, and winding on this “nichrome” wire (B and § No. 16; 1:°3"" in diameter), so that a coil 10° long with 5 turns per em. was formed. The whole was inserted in a cylinder of soapstone and the intervening space filled with “magnesite” mixture. Soapstone cylinders, SS, afford heat insulation at the ends of the furnace and also serve to fill up the space. One terminal of the heating coil is “grounded” onto the bomb through the brass ring 7; the other end is led through the lower soapstone plug 8, the steel plug A, and out along a slot cut in the top of the base plate 6. The thermocouple wires TT pass out through the top plug G. There are several ways in which good electrical connection may be made between the inside and the outside of the bomb. The method adopted in the present instance is as follows: a hole about 12™™ in diameter is bored in the steel part through which the wire is to pass and a cylinder of soapstone is turned to fit this hole. Through the soapstone is drilled a hole of the same diameter as the wire, which is then threaded through this hole. The soapstone cylinder is inserted in the steel, and *TIn this connection we desire to express our indebtedness to Mr. Geo, F. Nelson, to whom is principally due the success of this high-pressure joint, and without whose mechanical skill and ingenuity our work would not have been so far advanced. Melting Points of Certain Metals. 507 rammed firmly into position by applying a force of twenty or thirty tons; this force is communicated through a small steel eylinder of the same diameter, which is properly drilled or grooved so as not to injure or interfere with the wire or wires. In order to prevent the wires being actually squirted out through the holes in the stone, two procedures may be followed. Fic. 3. Electric furnace and interior of bomb, drawn to scale. JJ is the furnace, SS soapstone blocks for thermal insulation of the ends, and G and H, the plugs which close the bomb. Freedom from leaks is insured by the use of the rings or “‘ washers” of sheet copper shown at LL. Thermocouple wires TT are led into the bomb through the upper steel plug G. One termi- nal of the heating circuit is grounded onto the bomb at J; the other passes out insulated through H. The graphite crucible with thermocouple tube in place is shown suspended from the upper plug. The soapstone cylinder may be made in two or more parts and the holes for the wires staggered; or the internal part of the 508 Johnston and Adams—Influence of Pressure on the wire may be made somewhat larger than the outer part.* The latter method has been adopted i in the case of large wires, such as those of the heating circuit. The former method was used for the thermocouple and other small wires. The thermocouple wires are insulated in the horizontal portion of their path through the plug @ by sections of glass. tube, which are kept in place by short plugs of fibre. The heating current wire, after it has passed out through the soapstone, is ; likewise insulated from the steel plug H | by the insertion of a thin eylinder of fibre. The material to be investigated is contained in a graphite erucible of the form depicted in fig. 3. The erucible is held rigidly in position by small steel rods, which screw into the ends of larger steel rods, attached firmly to the plug G. The erucible lid, also of graphite, is held in place by means of a small screw-clip (not shown in the figure) fastened to one of the supporting wires. To the crucible lid is fastened by two steel pins the thermoelement jacket which is a porcelain tube, 7™" in diameter. The thermocouple wires are separated from one another, as usual, by means of a small porcelain tube slipped over one of them. By these methods of construction, motion of the thermo- element, or of its jacket, with respect to the charge, is abso- lutely prevented, so that we could always be sure that the junction was in the proper position with respect to the crucible, viz., located axially and about 6™" from the bottom of the charge. Temperature Measurement. To determine the temperatures, thermoelements of cop- per-constantan were employed. Three lengths were cut off a reel of No. 30 (B and 8) constantan wire (0°25"") and were joined separately to lengths of copper wire of the same diam- eter. Two of these elements were preserved as standards, and gave readings at all temperatures not more than 2 microvolts apart ; the third element was fixed in the plunger @ in the way already described. The differences between this element and the standards amounted to 20-30 microvolts, probably owing to strains set up in it while it was being made pressure tight “through G. “Before proceeding to the measurements, it was, however, nec- essary to calibrate the standard elements, since so far as we are aware, no satisfactory calibration of such elements over the temperature range 0° to 400° has yet been made. For this * An equivalent method which may be used with some wires is to tie a knot which shall be too large to pass through the small hole drilled in the soapstone block. Melting Points of Certain Metals. 509 purpose, we determined, on the standard element, the electro- motive force corresponding to the boiling points of water, naphthalene, benzophenone, and the melting point of. zinc ; while for the temperatures below 100° we compared it with a four-junction element belonging to Dr. W. P. White, the calibration of which is known with an accuracy of a few thousandths of a degree.* The values assigned to the standard temperatures are as follows :+ Naphthalene b.p.. ..-.------- 217°7°+0°057 (p—‘760) Benzophenonef (Merck) b.p..- 305°6° +0°063 (p—‘760) ZING MUNN eee Ss entre eRe ae ORD All temperatures given in the present paper are referred to the above fixed points; we also give the actual electromotive forces observed, so that the temperature may be easily cor- rected, if a change in the above reference points be rendered necessary by future work. In all the measurements, the cold junction was immersed in an ice-bath at 0°, and the electromotive force was measured by means of the usual potentiometric arrangement. By this means, an accuracy of 1 microvolt, corresponding to 0:02°, could be easily attained. The results thus obtained for the calibration of the standard element are: t e ue e (microvolts) ; (microvolts) 0 (0) 100°0 4927 24°92 975°7 ORV 10119 50°00 ZONI 305°6 15007 75°00 3096 418°2 21755 Attempts were made to obtain a single equation which should pass through all the points; but without success. Accordingly, two cubic equatious of the form e= A+ Bt+ Ct + Dt’ were used to calculate the electromotive force corre- sponding to any given temperature ; ¢ was computed for every 10° from 10° to 480°; the slight irregularities were evened out by adjustment of the successive differences. From the *See Phys. Rev., xxxi, 159. Its calibration curve is e = 155°311 ¢ + 0183 # — 0:00014 #. + Cf. Sosman, this Journal, xxx, 6, 1910; Day and Sosman, ibid, xxix, 98, 1910. It is to be noted that up to the present there have been only two gas thermometer determinations of the boiling points of naphthalene and of benzophenone—one by Crafts (Bull. soc. chim., xxxix, 277-89, 1883) and one by Jacquerod and Wassmer; the former does not claim an accuracy better than 1°, so that the only accurate data at present available are those of Jacquerod and Wassmer. {The b.p. of benzophenone (Merck), according to Waidner and Burgess (Bull. Bureau of Standards, vii, 6), is 0°2° higher than that of benzophenone (Kahlbaum); the b.p. of the latter is 305°4, according to Jacquerod and Wassmer. 510 Johnston and Adams—Influence of Pressure on the table thus obtained, giving e in terms of ¢, another and more convenient table was constructed, giving ¢ for each even 100 microvolts. This table is presented in abridged form—for every, 500 microvolts—in Table I, in which are also included the E.M.F.’s corresponding to the fixed reference points. Incidentally it may be noticed (1) that a quadratic equation will not express the relation between ¢ and e with sufficient accuracy; for such an equation when passed neuen the steam, benzophenone, and zine points misses the naphthalene point by 1°5°; (2) that the cubie equation of the above form fits the data with much greater accuracy than the inverse form of function, t=—Ae+ Be’ + Ce’. TABLE I, Abridged Standard Curve for Copper-constantan Elements. Microvolts t Diff. Microvolts t Diff, 0 0 11500 243°19 500 1298) 3525 PA@O00), | aeoawes a 1000 25°53 Ds - 500 261°30 aoe 500 37°77 ngs 13000 270:25 Aas 2000 49°72 nee 500 27914 aa 500 61°40 ue 14000 287°97 ere 3000 72°83 “a0 500 296°75 axa 500 84°05 ae 14.995 3054 4000 95:07 15000 305°48 aa 4227 10000 se 500 31415 eet 500 105°88 oa 16000 322-76 ee 5000 116°52 cee 500 331°31 aia 500 127-00 ORS 17000 339°80 Bid 6000 137°32 janie 500 348°24 ane 500 147°50 ane 18000 356°62 2a5 7000 157°55 age 500 364°95 aa 500 167°48 er ep ee 373°24 ape 8000 177°29 a0 500 381°49 oon 500 186-99 ase 20000 389°70 ai 9000 196°58 ANG 500 397°87 ate 500 206-07 alia: 21000 40600 aan 10000 215°47 500 41409 Bae 10119 217-7 ane 21755 418-2 500 224-79 924 22000 499°14 BD 11000 234:03 aie 500 430°15 non 500 243°19 23000 43812 This table may be used for any element with the aid of its deviation curve, which is obtained by plotting at two or more known temperatures the differences between the readings of the element in question and the standard curve. Its use thus saves much calculation and recalculation of thermoelement Melting Points of Certain Metals. 511 curves.* The uncertainty in the temperatures deduced in this way from the above table should, we believe, not exceed 0-1° on the temperature scale we have chosen ; and is much smaller than the present uncertainty in the absolute value of the temperature scale—an uncertainty which is about 1° at 450°. The readings may be immediately referred to any other scale of temperature by including in the deviation curve the differ- ences between the two scales. Temperature differences, how- ever, over a small range are probably accurate at least to 0:02°, regardless of slight discrepancies in the temperature scale, and of possible errors in the interpolation formula. For this reason the values of temperatures and differences in Table I are given to hundredths of degrees. Melting Point Determination. Briefly, the method pursued in the determination of the melting points of the various metals under pressure was as follows: A charge of the proper amount of metal having been placed in the graphite crucible, the whole was heated some- what above the melting point of the metal (a thin layer of oil on the surface of the metal prevented oxidation). The porcelain tube which serves to protect the thermocouple from the sur- rounding metal, attached to the crucible lid, was fitted into position. Crucible, steel plug (G, fig. 3), soapstone cylinder, etc., were then assembled and placed in the bomb, which was thereupon set in position on the lower platen of the press, and connected with the high pressure line (at Z) fig. 1). The upper press platen was forced down until the upper plug (@) slipped into place and centered itself; the electrical connections were made, the cooling watert turned on, and all was in readi- ness for the actual measurements. The pressure inszde the bomb was increased by means of the pump J), while simultaneously excess pressure in gradually increasing amount,{ was applied by means of the press, on the plugs G and H. When the desired pressure was attained, the freezing point of the metal was determined as usual by the Frankenheim method. No account was taken of the effect of pressure on the thermoelectric E.M.F. This influence is very slight in those * This matter is more fully treated in Sosman, loc. cit., q. v. +A supply of both hot and cold water was available; for most of our experiments hot water was used to ‘‘cool” the bomb. By the use of hot water in this connection a smaller heating current is required ; in addition, this plan makes for more uniform temperature distribution within the bomb. t¢ When the pressure was decreased again by releasing the valve V, a ? corresponding diminution in the external pressure exerted by the press was effected by suitable manipulation of the proper valves. 512 Johnston and Adams—Influence of Pressure on the cases which have been investigated: namely, for couples com- posed of platinum or other metals with mercury the difference of reading produced by a change of pressure of 1000 atmos- pheres is only about 0°2 microvolts.* It is therefore practi- cally certain that the error introduced by neglecting this variation does not exceed one or two microvolts, i. e., it is less, and probably much less, than 0:05°. It may be stated that we were able to get very sharp points ; it is obvious that the temperature did not stay long at the melting point, on account of the smallness of the charge and of the rapid rate of cooling, but the break in the curve was perfectly definite and reproducible. This is shown by the results, which have been brought together in chronological order in Table II, for the four metals investigated, namely, tin, lead, cadmium, and bismuth. In the second column is given the pressure in the system in atmospheres at the instant of freezing. The third column shows the observed reading of the thermoelectric E.M.F. at the freezing point; while in the last column these latter values are converted into degrees (uncorrected, that is, taken directly from Table I). Freezing points rather than melting points were taken since the former were with our apparatus much the sharper. Melting points were taken in a few cases, however, and found to agree well with the freezing point under the same pressure. A study of the results in Table II shows conclusively (1) that the melting point of a given metal under a given pressure is reproducible, since measurements made on different days, with different charges are in good agreement; and (2) that since the same results were obtained with increasing and decreasing pressure, the hysteresis effect of the gage was not large enough to interfere seriously with the accuracy of the measurements. Relation Between Melting Points and Pressures. It was expected that the change in melting point, Az, would be a linear function of the: pressure P. Accordingly, the straight line, ¢ = a@+bP, which best fits the data, was calcu- lated by the method of least squares, from the observations with each of the four metals used. The results of these caleu- lations are shown in Table III. In the first column of each table is given the pressure arranged in ascending order of magnitude. Columns 2 and 3 show respectively the observed change of melting point and that calculated from the least square curve. The differences between observed and *H. Wagner, Ann. d. Physik., xxvii, 955, 1908; Hérig, ibid., xxviii, 371, 1909. Melting Points of Certain Metals. 513 Tasue II. Direct Experimental Results. Freezing point (om =" Pressure in microvyolts in degrees* Date in atmospheres (uncorrected) TIN Ie te Diane ours ete 1000 10995 233°89 750 10952 233°09 2000 11174 237°18 1490 11081 235°47 1000 10994 233°87 300 10904 232°20 2000 11174 237°18 Fant le. fe 22 500 10907 232°26 1000 10999 233°96 1995 11170 23711 LEAD Yank lO ees = 2 500 16399 329°26 1000 16641 333°38 1490 16876 337°35 2000 17114 341°38 770 16536 331°59 505 16409 329-43 350 16332 328°12 250 16293 3827°46 150 16238 326°53 CaDMIUM Alavi Gull eae Be Ss 1500 16331 32811 1500 16333 328°14 2000 16524 331°39 2000 16526 331°42 Mane Gee ee 1515 16352 3828°47 1010 16162 325 24 512 15976 322°06 350 15912 320°97 BISMUTH Deamon eae ss 22 500 12860 267°71 1000 12746 265°67 998 12748 265°71 1550 12655 264:°04 2010 12551 262°18 1480 12662 264°17 * These temperatures were read directly from the standard curve, no account being taken of the differences in electromotive force of the element actually used and the standard element. Since, however, these differences were always small and correspond to less than 1°0°, and since the standard eurve changes in curvature very slowly, no error is introduced thereby into the change of melting point caused by pressure, the quantity in which we are chiefly interested. 514 Johnston and Adams—Influence of Pressure on the TasLE III. Change of Melting Points with Pressure. Change of melting point Pressure = a” = in atmospheres observed * calculated Difference IN P At=t’—230°61° dt=0:003275 P Ar—or 500 1°59 1°64 — 0:05 500 1°65 1°64 ae (Oil 750 2°48 2°46 + 02 1000 3°26 3°28 — 02 1000 3°28 3°28 00 1000 3°35 3°28 + 07 1490 4°86 4°88 OZ, 1995 6°50 6°54 — 04 2000 6°57 6°55 + 02 2000 6°57 6°55 + 02 BIsMUTH At=269:°37—t’ 6t=0'003548 P 500 1°66 ETT — 011 998 3°66 3°54 + 12 1000 3°70 3°55 + 15 1480 5°20 5°25 — 05 1550 5°33 5°50 eat la 2010 7°19 7°12 + 07 CADMIUM At=t' —318°81 6t=0-006288 P 350 2°16 2°20 — 0°04 512 3°25 3°22 + 03 1010 6°43 6°35 + 08 1500 9°30 9°43 — 183 1500 9°33 9°43 — 10 1515 9°66 9°53 + 138 2000 12°58 - 12°58 00 2000 12°61 12°58 + 03 LEAD At=t' —325°35 6t=0-'008026 P 150 1:18 1°20 — 0:02 250 2 2°01 + 10 350 DTT 2°81 — 04 500 3°91 4:01 — 10 505 4:08 4°05 + 03 770 6°24 6°18 + 06 1000 8°03 8°03 00 1490 12°00 11°96 + 04 2000 16°03 16°05 — 02 calculated values, as may be seen by referring to the fourth column, are on the whole quite small; indeed the deviation Melting Points of Certain Metals. 515 of the results from a straight line is no greater than the prob- able error of the observations. In other words, no indication of a tendency of the pressure-temperature curve to bend toward the pressure axis can be observed. This is confirmed by inspection of the graphs reproduced in fig. 4, which were obtained by plotting t or A¢ against P for each metal. TPREE ae Habe wld Oo 1000 2000 Fie. 4. Diagram showing the relation of the change of melting point AT to the pressure P (in atmospheres) for each of the four metals studied. The relations between A T and P, it will be noted, are represented with great exactness by straight lines. The melting point of bismuth decreases with pressure while that of the other metals increases. It was thought that it would be of interest in this connection to calculate from the Clausius-Clapeyron equation the change of melting point with pressure. We may write the equation Diy ws tae LOGE Vache phon Was nee =! BIRO y where dé is the change in melting point for the change in pressure dp, ZT’ is the absolute temperature of melting Vi 516 Johnston and Adams—Influence of Pressure on the and V, are the volumes of 1 gram of the metal at the melting point in the liquid and solid states respectively, and g is the latent heat of melting in calories per gram. A number of measurements not agreeing among themselves are recorded in the literature on the latent heat and volume change on melting of Cd, Pb, Bi, and Sn. We have made use of the data of Person* for the latent heat of melting, and of Vicentini and Omodeit for the change of volume at the melt- ing point, to calculate by means of the formula, the change of melting point per 1000 atmospheres for each of the four metals. These data were chosen for the sake of uniformity and because we believed them to be the best available. -The results of the calculation follow : I II Til IV Vv ott per 1000 Latent Vol. change dt per 1000 atm. cale. from Probable heat: cal. onmelting atm.cale, from obs. change of error§ of per gram. ccpergram. vol.changeand wm.p. with coefficient Metal qd Vi-—V; latent heat pressure in col. IV Sn 14:25 0:003894 + 3°34 + 3°28 + 0:02 Cd 13°7 0:00564 + 5°91 + 6°29 + 0°04 Pb 5°37 =: 0003076 + 8°32 + 8°03 + 0:08 Bi 12°6 0°00342 — — 8°56 — 3°55 + 0°08 The agreement between the values given in columns III and IV is close—in fact, closer than might be expected when we consider the uncertainty in the latent heat and volume change of the metals involved. The change of melting point with pressure of tin and bismuth has been measured by Tammann.| He found for dt/dp per 1000 kg / em’, 2°2 for tin and 3°9 for bismuth. His caleulated values (by the Clausius-Olapeyron equation and using, as it happens, the same data for latent heat and volume change as those we have employed) are 3°3 and 3°3 for tin and bismuth respectively.4] It is worth while noting that Tammann made his temperature measurements with a thermoelement of Pt— Pt Rhand a direct-reading galvanometer. This fact is sufficient to explain the discrepancy between his results for tin and bis- muth and ours for the same metals. * Ann. Physik, lxxv, 462; Ixxvi, 482, 596, 597. +Beibl. Ann. Physik, xii, 176. ¢ These coefficients, are, of course, the dt (x 1000) of the third column of Table III. § For calculation of probable error of the coefficient cf. Merriman, Method of Least Squares, 6th ed., Chapter on the Precision of Observations. | Zs. anorg. Chem., xl, 54, 1904. { 3°3° per 1000 kg/cm? is equivalent to 3°4° per 1000 atm, Apparently, Tammann’s calculated value for bismuth is in error by about 2 units in the second significant figure. Melting Points of Certain Metals. oy Melting Point of the Metals at Atmospheric Pressure. The melting points of the samples of metals used in this investigation were determined with both copper-constantan and Pt—Pt Rh elements which were calibrated, as stated above, in steam, at the boiling points of naphthalene and ben- zophenone, and at the melting point of zine.* Metal Source Melting point* Sn Baker’s 231°0 Bi BG 270°7 Cd Kahlbaum’s 320°4 Pb Baker’s 326°7 These melting points are in very close agreement with those published by the Bureau of Standards, when the difference in temperature scale is taken into account. Incidentally, the melting points of Kahlbaum’s tin and lead were, measured. The result for tin was 230°5° and for lead, 326°7°. Summary. 1. There has been designed and built an apparatus suitable for studying chemical and physical reactions at temperatures up to 400° and under pressures up to 2000 atmospheres. Both temperature and pressure in the reaction zone may be meas- ured with fair accuracy. 2. The change with pressure of the melting point of tin, bismuth, lead, and cadmium has been measured; it was found to be a linear function of the pressure within the limits of experimental error. 3. By substitution in the Clausius-Clapeyron equation of the data of Vicentini and Omodei on the volume change at the melting point, and of Person on the latent heat of fusion, dt /dp was calculated for each of the four metals. The caleu- lated values show satisfactory agreement with those observed. 4, Incidentally, the melting points of tin, bismuth, cadmium, and lead were determined, and a standard curve for the cali- bration of copper-constantan elements at temperature from 0° to 425° is given. Geophysical Laboratory, — Carnegie Institution of Washington, March 25, 1911. *The values assumed for napht. bp., benzo. bp., Zn, mp. being 217°7, 305°4 and 418°2 respectively. Cf. ante p. 509, foot-note. Am, Jour. Sci1.—FourtH Smries, Vou. XX XI, No. 186.—Junn, 1911, 30 518 Van Horn and Cook—New Occurrence of Pearceite. Art. XLITI.—A New Occurrence of Pearceite ; by Frank R. Van Horn and C. W. Cook. Introduction. In the summer of 1908, Mr. R. B. Cochran, formerly super- intendent of the Compania Metalurgica Mexicana at Sierra Mojada, Coahuila, Mexico, presented ‘the Department of Geol- ogy and Miner aloey at Case School of Applied Science with seve- ral specimens of silver, copper and lead minerals from the Veta Rica mine at the locality mentioned above. We wish to take this opportunity of thanking Mr. Cochran, both for speci- mens and much useful information concerning the district.* One of the minerals was recognized as polybasite, but blow- pipe tests showed that it contained chiefly arsenic with little if any antimony, and must, therefore, be pearceite. This in itself was interesting, since the mineral has been found previ- ously only at four other localities, namely: Schemnitz, Hun- gary in 1833,+ Arqueros, Chile, in 1879 »{ Aspen, Colorado, i in 1892,§ and Marysville, Montana, in 1896. | The pearceite also occurred in well-defined crystal aggregates, and likewise appeared to be twinned, so that the specimens were sent to the University of Michigan, where their erystallographic proper- ties were investigated by the junior author. Geography and Local Geology. Sierra Mojada is the name of a town as well as that of a range of mountains which perhaps more properly might be called hills. The region is situated in the extreme western part of the State of Coahuila, Mexico. It is reached from Escalon on the Mexican Central railroad, a distance of 494 miles south of El Paso, Texas. From Escalon, the Mexican Northern railroad runs 78 miles northeast, and terminates at Sierra Mojada, which lies in a valley about three miles wide. This valley is bounded on the south by the range of hills called Sierra Mojada, and on the north by other hills called Sierra Planchada. Ore was first discovered in 1878, and is found for a distance of about three miles along near the base of Cretaceous limestone cliffs * A complete description of this region and geological occurrence of the ores was given at the Pittsburg meeting of the Geological Society of Amer- ica, Dec. 29, 1910, in the paper entitled ‘‘ The Occurrence of Silver, Copper and Lead Ores at the Veta Rica Mine, Sierra Mojada, Coahuila, Mexico,” by Frank R. Van Horn. +H. Rose, Pogg. Ann., xxxiii, 158, 1833. + Domeyko, Min. 393, 1879 & Penfield-Pearce, this Journal, xliv, 17, 1892. Penfield, ibid., ii, 18, 1896. Van Horn and Cook—New Occurrence of Pearceite. 519 which constitute the Sierra. The ore occurs at or near the contact of the limestone with a rock which is locally called conglomerate, although others have named it a porphyritic breccia.* However, the rock seems to be either a decomposed rhyolite or a rhyolite tuff, since it consists chiefly of fine-grained quartz and orthoclase, much decomposed. Along the contact of the two rocks there are many indications of faulting, such as breccias, slickensides, and clay selvages. It seems not improbable that the valley was made by faulting, which has left the cliffs as a fault scarp. There are about 19 mines in the district, which, even in 1900, produced about 200,000 metric tons of ore. Up to 1893, it was a silver-lead camp in which the predominant ore was argentiferous cerussite with small amounts of galena. In 1893, however, in the western part of the region, large bodies of silver-copper ore were found in the San Jose mine in addition to the silver-lead ore bodies. Simi- lar silver-copper ores were afterward found on adjoining prop- erties, of which the Veta Rica mine is one. The latter has proven to be the richest if not quite the largest mine in the camp. Ores and Minerals Found at Veta Rica Mine. Although Sierra Mojada is still predominantly a silver-lead camp, nevertheless the chief output of the Veta Rica, which is one of the largest mines in the district, is known as a siliceous silver lime containing workable amounts of copper. The chief ore is either a red or dark gray magnesian limestone, impreg- nated with quartz, cerargyrite, native silver, and sometimes barite. In the specimens of this type which were subjected to investigation, no well-defined copper minerals were observed, although it is said to contain from 0-6 up to 2 per cent of copper. In another part of the mine, along a fissure in the limestone, copper minerals containing silver are found. A rather inter- esting fact is that on the same level along this fissure are two ore bodies of this type, but one consists of sulphides while the other is entirely oxidized. Specimens from the former presented by Mr. Cochran consist of massive chalcocite, chalcopyrite, and a little covellite. Also small amounts of galena and sphaler- ite are said to oceur at this point. About 120 meters west of this ore body, but at the same horizon and on the same fissure, is the oxidized body, which consists of native copper, cuprite, azurite, and malachite associated with gypsum. The water channel which caused the oxidation of this body has evidently been prevented in some manner from reaching the sulphides. *The Sierra Mojada, Coahuila, Mexico, and Its Ore Deposits, by James W. Malcolmson, Trans. Amer, Inst. of Min. Eng., xxxii, 105, 1902. 520 Van Horn and Cook—New Occurrence of Pearceite. Occurrence of the Pearceite. In 1906, while working along the northern part of the silver- copper siliceous lime ore body, a fault was encountered having a displacement of about 40 feet. Following along the fault plane, silver-copper ores of great richness were discovered, along with considerable barite as gangue mineral. Minerals observed from this point were native silver, argentite, proust- ite, pearceite, and erythrite. The latter occurrence is rather peculiar, singe it is the only oxidation product, if native silver is excepted. Furthermore, there have been no other cobalt minerals noted either from this mine or the district as a whole, although it would seem as if some cobalt arsenides should be present. There were said to be about 200 pounds of pearceite erystals found, but on account of the heavier government tax on high grade ores, practically the entire amount was ground up and distributed through poorer grades. However, eight specimens were presented by Mr. Cochran, while a ninth aggre- ate was very kindly loaned to us by Senor Felipe Borrego, formerly with the Veta Rica mine, but at present foreman of the Guadalupe mine at Cerro de San Pedro, San Luis Potosi, Mexico. Crystallography. The crystallography of pearceite was studied first, in 1896, by Penfield,* who worked on crystals from the Drumlummon mine, Lewis and Clark County, Marysville, Montana. At this time he proposed the name pearceite for polybasites in which arsenic was in excess of antimony. He states that it crystal- lizes in the monoclinic system and possesses a rhombohedral sym- metry, due, in all probability, to twinning similar to that of the micas. On account of this rhombohedral symmetry, which was likewise exhibited by all the material under investigation, as well as the similarity on angles for several forms, and the imperfection of the crystals, definite orientation was found to be impossible. Also, an attempt to obtain etching figures, using nitric acid as the solvent, met with failure. Therefore, although several undoubtedly new forms are present, it is thought best to limit the present crystallographic report to a discussion of a new twinning law. : The pearceite occurs in aggregates of twin crystals, more or less perfectly developed, which show striations and vicinal planes. The individual crystals are tabular in habit with the basal pinacoid as the predominating form; the pyramids and domes occurring as very narrow faces. Unfortunately most of the edges of the crystals were broken off, as the mineral is very brittle. The basal pinacoid is characterized by the presence of *S. L. Penfield, this Journal, ii, 19, 1896. Van Horn and Cook—New Occurrence of Peareeite. 521 triangular figures. According to Hintze* the sides of these, figures are parallel to the faces (111), (111), and (101). The new twinning plane, which is practically present on all of the specimens, is likewise parallel to one of the sides on the tri- angle. It must, therefore, lie in either the unit prism-basal pina- coid zone or the ortho-basal pinacoid zone and the twinning plane must be either a pyramid or an orthodome. The examination of the twinning planes of a large number of monoclinic minerals shows that twinning in the ortho-basal pinacoid zone is very common, whereas the pyramids function but rarely as twinning planes. The new twinning plane on pearceite has, therefore, been tentatively assumed to be par- allel to the orthodome (702) on the basis of the twinning angle, which is 34° 42’. The average of the measurements, made on a number of different crystals, gives for the angle between the basal pinacoid (100) and the orthodome (702) a value which agrees quite closely with the calculated value, as may be seen from the following : Observed Calculated (100) : (702) Se eS 72° 39’ HOO BBY Fig. 1 shows the general character of the crystal aggregates as well as the presence of reéntrant angles. The triangular striations on the basal pinacoid are seen on numbers 1, 2, and 3. Parts of the best crystal measured are shown under 8, 9, and 10, but unfortunately the original was broken while deter- mining the specific gravity. Nevertheless the reéntrant angles are still visible on all three fragments. Number 7 is an aggre- eae of small thin plates and closely resembles some specular ematites in micaceous appearance. Chemical Composition. The empirical formula which was proposed in 1896 by Pen- field,t and was accepted, is (Ag,Cu),(As, Sb) S, or 9(Ag,Cu),S. (As, Sb),S,. This is analogous to the formula for polyba- site, (Ag,Cu),(Sb, As) S,, which was proposed by Heinrich Rose in 1829.t ? DF DOW WOH CLOT OT CD 6 tO A ps es 9 Cd 0 BIO SS WH DWH CLOTW ps = . e SHAHIAW UW CUWMNDAWOPR OCH PW OWROWP ee ed we MEO HE OLE OO MOO ARO E Crystallization of Parisite. 541 o>e slight. For erystals 1™™ thick the dichroism and absorp- tion are but slightly greater. Upon alteration the crystals become filled with a dusty product, are less transparent and often exhibit a brownish or brownish red stain of varying intensity. The indices of refraction were determined by the immersion method, using a barium-mercuric-iodide solution. The deter- minations were made on a number of perfectly clear, small erystals chosen on account of the uniform development of their prism zones; also upon one larger crystal (1"™ in diam.) ter- minated by a perfect basal plane which made it possible to orientate the crystal and cut a section parallel to the prismatic - axis. An attempt was made to measure the indices directly upon this crystal by means of the Abbe refractometer, but without suecess, owing to the small size of the section and its low degree of transparency. The fine striations parallel to the edge between the base and the prism stand out very sharply under the microscope and make it possible to orientate the erystals with great accuracy on the microscope stage. The values obtained with sodium light are given below, also those heretofore given for parisite as determined by Senarmont and those for synchisite according to Flink. Parisite, Quincy Parisite, Muso Synchisite, Greenland Warren Senarmont Flink e= 1757 1°670 1:7701 w = 1°676(+0°002) 1569 16742 eo = 0'081 0-103 0:0959 The Montana parisite, analyzed and described by Penfield and Warren (loc. cit.), also crystals from Muso valley taken from the mineral collection of Harvard University, were tested by the immersion method and their indices were found to cor- respond to the values given for the Quincy mineral. The older values given for the Muso mineral appear to be quite wrong. The ordinary rays for parisite and synchisite are almost iden- tical. The extraordinary rays appear to differ by 0-0131. While the extraordinary ray for the Quincy mineral is prob- ably not as accurately determined as the value for the ordinary, the error can hardly be as great as 0-0131 and the difference between the two minerals for this constant may perhaps be a real one, Chemical composition of Puarisite—About a kilo of fine- grained material recovered from the fragile lining of the cen- tral pockets was carefully washed and fractioned by means of screens, an electro magnet, and heavy solutions until a fraction was obtained weighing about ten grams and consisting largely of parisite mixed with more or less aegirite, anatase, feldspar and quartz. From this about three grams of clear yellow or amber-colored crystals were separated by hand-picking under 542 a powerful lens. Palache and Warren—Chemical Composition and Aside from a slight stain in a few erystals, the only impurities visible under the microscope were minute adhering grains of anatase and aegirite, hardly amounting to more than a trace. A partial analysis was also made on a few carefully selected erystal fragments of parisite from Muso valley taken from the mineral collection of Harvard University, to serve as a check on the earlier analyses made by Warren on crystals of the Muso mineral from the Brush collection, in New Haven, Conn. | i] | | ea a eee ie a ies 3 | 4 0) 6 meen mete) Mon- | Muso | Muso} Quiney| Muso | Synchi-| Synchi- | Cordy- _ tana | 1899 | 1910 | 1910 | Damour, site site* lite 1899 | & Fink Mauze|lius Warren anal yst | Deville Green- land Spee.Gr. | 41128 4:302 : 4°320| 4:358]) 3:902} 3°900] 4°358 Oz | 22°93 | 24°22 | 24:34) 24:16 | 23°48 | 26:54 | 25:99 | 23:47 Fluor, 590 | . 6°82 6°56 ya) 5°82 5°04 4°87] Ce205 | 2614 | 30°67 | 29°92) 30:94 | 44°21 | 28:14 | 21:98 3°72 (LaDi)2O; 28:46 | 29°74 | 28°75] 27:31 | 18°00 | 22°88 | 28°67 | 25°67 Yt.Os tr. if tr. 1:23 1:18 tr. Fe.O3 80 "25 32 FeO rill 1°43 CaO 10°98 | 10°70 |11°50] 11:40 | 10:10 | 17:13 | 16°63 1:91 SrO tr. ¢ BaO | 17°30 Na2O 69 20 30 “30 “19 | K,0 ‘19 ‘10 "22 20 12 | H.0 26 ‘10 tr. ald “80 Gangue 6°13 |(diff.) 1:02 2°58 102°48 | 102°65 102°21 | 101°34 | 102°05 | 102:00 | 102°05 O = 2F. 2°48 2°87 2°76 2°34 2°45 2°12 2°05 Totals 100:00 | 99-78 99°35 | 99:00 | 99-60 | 99:88 | 100-00 * Material somewhat altered and impure. + Material scanty, impure, and perhaps slightly altered. H.O 2°10; 1°56 expelled at 100°. The molecular ratios derived from the above analyses are as follows: CO, 1. Montana, 0°550 2. Muso, ’99, *D00 3. Muso, 710, D8 4, Quincy, B49 5. Muso,(D. &D.)— +553 6. Synchisite, 0) Flink : 7. Synchisite, 390: Mauzelius : 8. Cordylite, 580 | int, R.03 0°310: 0°166: 809: +183; ‘178 : 340: +178: 292: +187: 306: 160: 260: ‘189: "256: ‘181: CaO or CO, 0°196 =8: 1°79: ‘191 =3: 1:96; 2038 =3: ‘205 =3: 1:88: 180 =3: 1:65 306 =38: 1°52 =4; 2-02: 298 =3: 1°34: 412 117fe) 148 =3; 1:44; ewes MPwmoaqacoe =e F, R203 CaO Crystallization of Parisite. 543 The analyses were made in duplicate and agreed closely. Regarding the method of analysis it may be stated that the earths were separated from lime by precipitation with ammonia after the complete removal of fluorine by evaporations and fuming with sulphuric acid in order to ensure that no fluorides were precipitated with the earth-hydroxides. The cerium was separated from the lanthanum and didymium by precipitation of the cerium with chlorine in a potassium-hydroxide solution, a departure from the method employed in the earlier analyses of the Muso and Montana minerals. Care was taken to examine all filtrates from oxalate precipitations of the earths to recover, if present, any unprecipitated earths, a necessary pre- caution, as has been pointed out by Hillebrand and others. The results of these analyses are given above and in parallel columns are given for comparison the older analyses of pari- site ; also the analyses of synchisite and cordylite. For synehisite the ratios derived from Flink’s analyses are sharply—CO,: F:R,O,: CaO=4:2:1:2. This leads to the formula given by Flink, (RF),Ca,(OO,),. The analysis of Mauze- lins (7), though perhaps on less satisfactory material, leads to the same formula. From this it appears that the formula of synchi- site, as determined by Flink, differs from that of parisite in con- taining exactly one more molecule of calcium carbonate. To make the chemical compositions of parisite and synchisite iden- tical calls for a change of 2°3 per cent in the CO, and of 6-0 per cent in the CaO. ‘These are very serious differences, for they involve unallowable errors in the determination of the two constituents whose determination can be made with great accuracy. Small differences in the proportions of the rare- earth oxides are to be expected in minerals from different localities, and, indeed, analytical errors in the determination of the earths, unless considerable, would not seriously affect the ratios on account of the large molecular weights of these oxides. The close agreement in the lime and carbon dioxide determinations of the Muso, Montana and Quincy parisites (in all ten determinations have been made) leads us to place great confidence in their correctness, as well as upon that of the formula derived for the mineral from these localities. In discussing the analyses by Tschernik on the Manchurian parisite we are at a disadvantage, as the discussion appears in full only in Russian. The ratios derived from these analyses, however, do not correspond, so far as we can see, with the ratios given above for either parisite or synchisite, being far more complex. The presence of considerable amounts of water in all three analyses, as well as the variable and zonal character of the crystals analyzed by Tschernik, suggests strongly that the 544 Palache and Warren—Chemical Composition and material had undergone considerable alteration with consequent changes in composition. Ltelation of Synchisite to Parisite.—The synchisite of Flink differs in most respects but little from parisite, and was indeed, but on insufficient material, earlier described as such by Nor- denskidld. The present writers, in view of the results of this study of parisite, believe that no valid distinction exists between these substances, and that synechisite must stand as a synonym for parisite. Synchisite is rhombohedral with forms which permitted of only approximate measurement, on the basis of which they were interpreted in terms of the parisite axes. Most of these rhombohedral forms have now been found on parisite, but the new position chosen for the latter requires a readjustment of the symbols as given by Flink. This is made in the following table, it being understood that the angle given for synchisite forms, in each case, is that of the nearest then known form of parisite to the measured angle, none of the latter being given : Synchisite (Flink) Parisite equivalent (Palache) e (0001) ce (0001) m (1010) m (1010) m (1120) @ (1120) 2 (1121) 81° 33’ Zoi s3)) aes lied t (2029) 40 48 d (0338) 39 59 r (2023) 68 53 n (0665) 69 34 v (3034) 71 04 xX (0554) KO a9 p (1011) 75 34 (0553)) 74 59 _ >} not found s. (4043) +79 05 (O73), = 279) 109 a (3032) 80 13 ma (0552) 1A BO é (0115 Sipepil BY (013) 36 42 u (0229) 40 48 G (5.0.5.13) 40 46 q (0112) 62 46 J (7078) 62 56 y (0334) 71 04 M = (4043) ees B (0332 80 13 P(E) e806 y (0381) 85 06 I \(SS0510) 80 53 It is evident that the form series of synchisite fits into that of parisite perfectly, and as both are rhombohedral no dis- tinction of form can be made. The new determination of the optical constants of parisite has shown that the two substances have sensibly the same values. Parisite has been shown to have the same cleavage property which Flink showed to be characteristic of synchisite; the basal cleavage of both is vis- ible only in altered specimens, the fresh minerals breaking with conchoidal fracture. Crystallization of Parisite. 545 There remain but two sensible differences between the two substances; synchisite has a slightly lower specific gravity ; and its chemical composition shows the presence of exactly one molecule of CaCO, more than is present in parisite. It is hardly possible that this large difference in composition can rest in analytical errors; on the other hand, it is quite impos- sible for the writers to believe that two substances with such a profound chemical difference as this could be so nearly identical in all their physical properties. The following con- sideration is offered as a possible interpretation of the matter. In describing the crystals of synchisite Flink states that many of them show an enlarged central portion with forms and luster differing from the smaller crystals and with slightly altered optical character. This is a character common to the larger crystals. Now there is no statement in the paper of the quality of the analyzed crystals. Are we not justified in believing that the analysis material contained enough of this altered substance which might well be CaCO, to produce the differences found? The alteration would tend to lower the specific gravity, so that this difference too would be accounted for. A new analysis of perfectly fresh synchisite can alone settle the question. Microcline.—Microcline in well-formed crystals of orthoclase habit makes up the greater part of the porous material near the great central pocket. The crystals range from a diameter of two and a half centimeters downwards to mere crystal specks ; they are, however, very constant in habit, presenting a remarkably cuboid form due to the dominant development of the base, clinopinacoid and orthodome; prism and unit pyramid, the only other forms found, being very subordinate in size. The faces are smooth and give fairly good reflections of the goniometer signal. The albite twinning, shown by microscope study to be universally present, is not apparent on the exterior of the crystals; its presence makes the crystals sensibly mono- clinic however, and the measurements obtaimed approximate to those of orthoclase. Well-formed Baveno twins are seen in a few specimens, but most of the crystals are in clusters without apparent definite relation of the constituent individuals. The color of the microcline is white to pale ivory-yellow. On faces of the prism there is often a secondary coating of colorless glassy feldspar in parallel position to the main crystal, which the microscope shows to be also microcline, although its appearance strongly suggested the growths of albite so common on orthoclase from numerous localities. Orientated sections, cut from the freely developed erystals of the pocket lining and from some of the larger erystals with- out, show that the microcline is twinned after the albite law 546 Palache and Warren—Chemical Composition and only and thus lacks the grating structure characteristic of microcline in general. In basal sections the twinning is seen to be very finely polysynthetic. The individual lamelle appear as short strips slightly elongated parallel to 010. Their bound- aries are as a rule not sharp. The two sets of lamelle extin- guished symmetrically on either side of the trace of the twinning plane at an angle of 16 degrees (average of 12 meas- Fie. 4. Fic. 4. Micro-photograph of a basal section of microcline cut from a freely developed crystal of the pocket lining, Fallon quarry pegmatite. Shows twinning after the albite law only. One set of twin lamelle is extin- guished at an angle of 16° with the trace of 010. Crossed nicols. Magnification about 300 diameters. urements). The clearer growths of later age are in parallel position to the older crystal and in them the twinning lamellee are often longer and more sharply defined. In the small microclines throughout the finer-grained portions of the pipe the twinning is usually more sharply defined. The extinction of 010 sections was found from the average of ten measure- ments to be 5°. Figure 4 is a micro-photograph, using polarized light, of a basal section of a small microcline erystal. Crystallization of Parisite. 547 Riebeckite.—The riebeckite occurs in the form of elongated prismatic individuals, possessing a black color and a lustrous cleavage, scattered through the coarser-grained main portion of the pegmatite pipes. The associated minerals are quartz, microcline-albite-microperthite and aegirite with accessory zir- Fig. 9. Fic. 5. Micro-photograph of a basal section of a small microcline crystal in the fine-grained portion of pegmatite in the Fallon quarry. Shows twinning after the albite law only. One set of lamellz are extinguished at an angle of 16° with the trace of 010. Crossed nicols. Magnification about 250 diameters. con, fluorite and ilmenite. It varies in size from quite small individuals up to crystals one or two centimeters thick and ten or twelve long ; individuals a centimeter thick and five or six long being quite common. The erystals show a tendency toward a crystal outline consisting of the unit prism occasionally trun- cated by the clinopinacoid. The immediate boundaries are, however, always more or less intergrown with the surrounding minerals. Terminal planes are not observed. Measurements of the prismatic cleavage made on two individuals yielded identical values of 55° 05’, an angle considerably larger than that of common hornblende, which is 55° 49’, The mineral is 548 Palache and Warren—Chemical Composition and always to a greater or less extent intergrown with aegirite. While the latter is frequently included in the body of the former, its most common position is about the outside, parti- cularly the ends of the crystal, and although the riebeckite con- tains unorientated grains of aegirite, the usual mode of inter- growth is with the prismatic axes of the two minerals parallel. As the sides and particularly the ends of the erystals are approached, narrow strips of aegirite are interlaminated with the riebeckite, the amount usually increasing until the latter is entirely replaced. The riebeckite usually contains consider- able amounts of black dust, often arranged in wavy lines, as well as larger grains of black oxides, mostly ilmenite. Occasional grains of feldspar, fluorite and zircon are also included, some- times singly and again forming patches of varying size. Optical.—The deep color and strong absorption of the min- eral makes the determination of its optical properties, with any precision, very difficult. This difficulty is increased by the fact that it has been found impossible after many trials to obtain altogether satisfactory sections of the mineral across the cleavage owing to its extreme brittleness. By the study of finely crushed material and thin sections the following char- acters have been made out :— Ray near c=a. For) Ray || tob=c. For| Rayneara=b. For sections 0:03™™ or under, 0°03™" thickness, very | thickness under 0:03™™ deep blue to bluish,|dark smoky green to| yellow. For 0:03 and smoky green. For over almost black. over brownish yellow 0:03™™" nearly or quite | with a greenish shade. black. | | Absorption a 6043 73°97? Aegirite, Norway, Brogger 1:0975 : 1 : 6009 73 09 Acmite, 2 S09 96s sles mo Olio ical el Table of angles of Aegirite, Quincy. po=d472 go= 5762 e='3015 w= 72°27’ Calculated Measured Limits No. of ee aces Quality i) p Q p a 100 90°00' 90°00' 89°57’ §=90°00' 89°37'-90°04' 5 poor 6010 0000 90 00 00 37 90 00 1 poor m110 4335 90 00 43 33 90 00 43 00-43 55 27 good f 310 7042 9000 |. 7026 90 00 69 56-70 56 5 poor willl 5550 47 06 55 54 647 08 55 17-56 16 . 46 54-47 31 11 good s 111 -2315 3320 -2340 383 24 1 fair w 331 4821 69 52 48 29 69 50 47 45-49 13 ; 69 36-70 00 4 poor 2 331 -38747 6627 -3746 66 42 1 poor 6 551 4631 7710 44 46 77 57 44 02-45 31 77 80-78 25 2 v. poor 7 112 -5 32 1653 -5 51 16 50 516-6 19 2 16 40-16 57 5 good A 311 -66 44 5650 -6640 56 45 66 37-66 44 2 v. poor d131 2609 63 39 26 40 63 48 26 19 -26 52 63 34-64 03 3 good The forms w (831), (551), (112), and d(131) are new to aegirite although all are known on augite. The habit of the Quincy aegirite crystals is shown by figures 6 and 7. Optical.—Small crystals or crystal fragments show under the microscope the following pleochrism : a = pale to deep green, sometimes with a slight bluish tone. The color naturally varies with the thickness, but also varies quite widely in the same crystal. In fact portions of a erystal may be a very pale green to almost colorless, and the other portions medium to dark green, without there being, however, any other optical variation so far as can be told. b = pale yellowish green to almost colorless. = pale yellow to yellowish green ; almost colorless. In many crystals the whole or a part may show a brownish- yellow or even reddish-yellow color. This is often most pro- nounced about black oxide (ilmenite) grains and is believed to be a pigment stain of ferruginous character. There appears, at least, to be no regularity in the distribution of the brownish or reddish colorations. The, extinction, a A c’, is 6 degrees. Other optical characters appear to be as usual for aegirite. Crystallization of Parisite. 553 Chemical composition.—The almost universal contamination of the aegirite with other minerals made the obtaining of suit- able material for chemical analysis very difficult. By means of magnetic and heavy solution separations combined with hand- picking under the microscope, about three grams of material were finally obtained which showed as impurities only a little ilmenite and traces of octahedrite and quartz. The analysis made in duplicate averaged as follows: Aegir- ite, Fallon Quarry, North Common Hill, Quincy, Mass., U.S. A. (analyst, Warren). Per cent. Molee. ratios. IS) Oe NE Bee ee ae St 51°73 0°862 0862 NTO), Belo cel, ie Pen os “64 008 0-008 INTRO) Mee Ba ay. RE 1°91 ‘018 9) Fe,0, RR RAO MB 31°86 “199 Hee 10) ee OMe eee 87 "012 | MOREE a) aN Es 60 “008 | CAO n= sin Gheie esau 87 “O15 | 29 Opel gee ob « 14 003 f oe INO) es os ata aes ol 11°43 "184 | EO ep ENS UD “40 “004 J NGL O ees Grice ee 2 Maree "20 eae eee Pee enone Mogae Shee Wee =100°65 Specwolaviat ton Cee eee bees =3°499 Although a portion of the TiO, was probably present as TiO, (octahedrite), most of it is combined with RO as ilmenite, and after deducting the TiO, and the proportionate amount of RO as ilmenite, the combined ratios are: SiO,=0°838, R,O,= 0-217, RO + R,O = 0-218; or very nearly SiO,: R,O,: RO+ R,O=4:1:1, leading to the formula (R’’,R) Fe,Si,O,,. So far as known to the writers, the ferrous iron is lower in this aegirite than in any other hitherto analyzed, and the compo- sition approaches very closely to the theoretical composition of the compound, Na,Fe,Si,O,,, which is SiO,, 52 per cent; Fe,O,. 34-6 per cent; Na,O, 13°4 per cent. Fic. 8. Fre, 9. Fies. 8, 9. Ilmenite. Iimenite. — I\menite occurs in moderate abundance in both the Ballou and Fallon pegmatites. It appears to have 554. Palache and Warren—-Chemical Composition and been of rather late formation and is particularly associated with aegirite; as embedded xenomorphie plates, and groups of tiny crystals implanted on crevices of fractured aegirite erystals (Ballou quarry); and as clusters of larger crystals upon the walls of cavities left by the destruction of such erystals b magmatic resorption (Fallon quarry). The crystals are anil not exceeding a diameter of 2", and are always very thin tab- ular in habit. A dull-black coating of manganese oxide com- monly gives them a lusterless appearance, but in two specimens brilliant crystals were obtained which, despite minute size, gave good measurements on the goniometer. Octahedrite is almost always sparingly present with ilmenite. fe The forms observed are as follows: ¢(0001), m(1010), @1120), (2130)* 6(1019), (1014), (2095), (1011), K(0.7.7.20), €(0112), A (0445)*, s(0221), »(0552)*, g (0.3.3.11)*, k(0.3.3.10)*, m (1123), (2248). The crystals from Ballou quarry showed the forms c¢, 7,a, 6, 7, and 7,, the prism zone being well developed and the base large and very brilliant. The prism is new to ilmenite. Crystals from Fallon quarry are dominantly rhombohedral, prism faces being reduced to mere lines. The crystals meas- ured showed the following combinations : Feeney eee Shes Cr i 5 CAT, is th, Ona 20. i, STS OT HR OO ND ES fo) . > 2 S Se eR As shown in the figures (figs, 8 and 9), flat positive rhombo- hedrons are largely developed on these crystals, recalling the description of one (the common) phase of ‘ Crichtonite” from Oisans by Des Cloizeaux.t That author considered the rhom- bohedrons which he measured (1015), and (1019), and (1.0.1.11) as negative, but left the determination of sign doubtful. The second of these forms is common to all the crystals from Fal- lon quarry and is certainly positive, so both the others should be likewise so considered. Several negative rhombohedrons new to ilmenite were observed and are based on the following data : Form Calculated Measured No. Limits of p. 9g p o p g (0.3.3.11); 30°00’ 23°34’ 30°00’ 23°461 3) 23°28 to 24. Ob! k (0.3.3.10) es 235) Ai ee Dy 85 A (0445 oe Dll -39 MG GL ye 2h) Vayil v8} uo) a2? (OKO) r (0552) pe easy ee w OG, ql 0 (2130) 10 53 90 00 10 30 90 00 4 10 36 to 11 34 * New forms. + Mineralogie, ii, 222, 1898. Crystallization of Parisite. 555 The presence on one crystal of the form 7, observed before only by Solly on a Binnenthal crystal, confirms this form. All the forms present gave angles agreeing very closely with the values calculated from the axial ratio of Koksharow as used by Dana. Chemical tests on the ilmenite from both quarries revealed strong qualitative reactions for manganese ; an analysis would be interesting, but it was not possible to separate enough of the fresh mineral for this purpose. Octahedrite.—Octahedrite is found chiefly in the large cen- tral pocket of the Fallon pegmatite, generally in close associ- Fic. 10. Fie. 11. ia Lu sae ation with aegirite and often formed posterior to the alteration of that mineral, since it is not infrequently seen on the walls of hollow casts of aegirite crystals associated with fluorite and ilmenite. Isolated crystals were also found implanted on feld- spar crystals. The erystals of octahedrite are small, of a deep black color, and of very brilliant luster. They show only the forms ¢(001), m(110), p(i11), #(112), and 2 (113), the two last the least common. These crystals are marked by two pecu- liarities ; they are in large part of prismatic habit with the first order prism dominant, a habit not before described for this mineral, and causing the crystals to be at first mistaken for 556 Palache and Warren—Chemical Composition and zircon ; and they occur in cruciform twin groups with the form (101) as twin plane. The twins are sometimes complete inter- penetrations of two equal crystals as shown in the figure ; some- times but one end of each is developed ; again a larger crystal has a much smaller one in twin relation to it. The groups are exquisitely sharp and leave no doubt as to the definiteness of the twinning since the two upper faces of the unit pyramid of each erystal : “and the two lower, parallel and opposite faces to these, reflect the signal simultaneously in pairs; thus the faces of (101) which are in zone with these unit pyramid faces must be parallel to the twin plane. This twin law has been observed but once before on this mineral, on erystals from the titaniferous calcite-quartz veins of Somerville, Mass.* There twins were extremely rare, while here they are sufticiently numerous to be considered as charac- teristic for the locality. Combinations of prism and unit pyra- mid are far the most common among these crystals. A few, however,’show the base as a tiny facet, and in a few the flatter pyr amids & or 2 replace the acute summit of the common form. Figs. 10 and 11 illustrate the habit of the octahedrite crystals. Fluorite.—Fluorite is distributed throughout all parts of the pegmatite masses. It is generally in small grains, but near the central pocket, especially in that part where crocidolite was abundant, the fluorite individuals were larger, one mass show- ing cleavage faces nine inches across having been found. Where wholly embedded in crocidolite the fluorite crystals are idiomorphic, octahedrons up to one inch in diameter thus occurring ; they are dull and somewhat rounded, the color a deep purple like all the fluorite of this locality, but occasion- ally there is a surface layer of bluish green color due to in- cluded fibers of the blue crocidolite. The hollow castes left by the solution of such crystals have already been described. In one or two cavities in the Fallon pegmatite there were seen tiny cubes of fluorite implanted on quartz, and in another such pocket, peculiar in containing also crystals of calcite, the cube was modified by two hexoctahedrons which appear to be new to fluorite. The measurements and derived symbols of these forms follow : No. of Calculated. Measured Limits read’gs ¢ p ? p d p { 3.10.16 16°42’ 33°07 15°45’ 33°20’ 4 3.16.10 10 388 58 26 10 36 58 33 10°30’— 10°42’ 58°30’ —58°36' 2 (10.16.38 32 00 80 58 32 04 81 05 31 52— 32 16 80 50 —81 20 4 ( 259 21 48 30 54 21 01 31 15 20 45— 21 18 3111 —3119 2 + 295 12 32 61 32 12 32 61 20 12 27— 12 37 61 23 —61 26 2 | 592 29 03 79 00 * On Octahedrite, Brookite, and Titanite from Somerville, Mass. C. Palache. Rosenbusch Festschrift, 1906, 511. Crystallization of Parisite. 557 Wulfenite.— Thin coatings of light yellow color as well as tiny erystals of wulfenite were found on smoky quartz in the crocidolite pocket. The crystals are in part model-perfect combinations of first-order pyramid with third-order prism n (111), and 7 (320), (see fig. 6, p. 990, Dana, System), in part cube-like combinations of a prism and the base. The amount of wulfenite is very small, and its presence is easily accounted for by the association, in the same region of the pegmatite, of molybdenite and galena. Boston and Cambridge, Mass., February, 1911. Art. XLVI.—Wotes on the Absence of a Soil Bed at the Base of the Pennsylvanian of Southern Ohio ;* by JESSE KK. .Hyne. Baitey Wixtis has recently urged, as one of the fundamental principles of paleogeography, the theorem that the failure of representation of many horizons and formations in an area where sedimentation has otherwise apparently been continuous, is to be explained by submarine erosion or failure of deposition due to current action, and that it is unnecessary to assume subaérial erosion in order to explain such a break in sequence.t E. O. Ulrich has since seriously questioned the validity of the grounds for Willis’ stand, and his discussion with Willis’ reply are briefly abstracted in the minutes of two recent sessions of the Geological Society of Washington.t Although these remarks are not a formal presentation of either side, it is felt that they open up to discussion one of the most important points in paleogeography and one about which it is apparent much remains to be learned. John M. Clarke’s eloquent appeal for more facts to support scientific theory, appearing in the same number of Science with these abstracts, is as timely to the present instance as to any of the unproved propositions. One of the principal arguments cited by Willis to support this theorem is the absence of anything resembling a soil bed or in any way suggesting weathering due to exposure to the atmosphere at the plane which marks such an omission from the column. It is maintained that if the bed whose absence it is desired to explain had been eroded while the whole stood * By permission of the State Geologist of Ohio. + Science, N. S., vol. xxxi, pp. 241-260, 1910. tScience, N. S., vol. xxxiii, pp. 312-316, 1911. 558 J. BE. Hyde— Notes on the Absence of a Soil Bed. above sea-level, some trace of the subaérial exposure would be visible at the contact. The following facts concerning one erosion surface are presented solely as evidence against Willis’ interpretation of the significance of the absence of soil beds or other evidence of exposure. The point which it is desired to make is that there may be no evidence of such a period of exposure at a contact which undoubtedly stands for several hundred feet of erosion, apparently under conditions highly favorable for soil formation. The explanation of the absence in this case cannot possibly be the one given by Willis, but no reason is, as yet, to be assigned. iat The base of the Pennsylvanian series, or Coal-measures, along the outcrop belt in central and southern Ohio usually rests directly on the Logan formations whose general age is early Mississippian. The latter were subjected to erosion which removed an unknown thickness and a relief of 200 to 800 feet was established before the Coal-measures were accumulated. At several points the Maxville limestone, late Mississippian in age, is found between the two, but it occurs only in patches. There was some erosion of the Waverly prior to the formation of this limestone, but it is very probable from the general field relations that by far the most of the erosion was in post-Maxville time. The irregularity in the distribution of this formation is to be explained, in large part at least, by the attitude of the post-Maxville erosion surface. This plane of disconformity extends entirely across Ohio. It has recently been followed in some detail from the center of the state to the Ohio river. In places, it cuts very abruptly into the fine-grained yellow sandstones of the Logan forma- tions, sometimes removing them entirely, and letting in coarse, massive Coal-measure sandstones, but usually the irregularities are gentle and observations of the contact may be necessary over a mile or more, in order to detect any considerable varia- tion. Differences in elevation of 100 feet are not uncommon, although most of them are less. Finally, it is the rule to find it standing uniformly high over considerable areas—several townships—and tending to be lower in adjacent townships. For example, in eastern Fairfield and southern Licking coun- ties it stands very high, about 300 feet of Logan being present at one point. To the northeastward and southwestward it sinks gradually until the Logan averages 60 or 80 feet and to the southwestward may be wanting entirely. At several points where the coarse Coal-measures sandstones are let into the Logan to a depth of 100 or even 200 feet, they take the form of a long, narrow mass which can be traced sometimes three or four miles, and clearly indicate the presence of a valley on the old Waverly surface. J. EL, Hyde— Notes on the Absence of a Soil Bed. 559 The Pennsylvanian rocks consist usually of coarse sandstone, frequently feldspathic and sometimes pebbly. One or two coal horizons are present near the base, but the coal is usually thin. Shale beds are not infrequent, but they are usually found higher up in the series, those beds near the contact being commonly sandstones, the so-called “Sharon conglom- erate.” Usually the poor outcrops and the short time at hand have prevented the determination of the effect of the Logan hills on the strata of the basal OCoal-measures, but at one point in Vinton county and again in Perry county it can be shown that the hills were gradually covered by the Coal-measures and that the first coal swamps were broken by “islands” of Logan, which were later wholly covered by the coarse accumu- lations. Apparently none or, at most, very little of the material was obtained by the erosion of the Logan; it is much coarser and has come from some other source. There is no evidence of marine influence in these basal beds. It was not until later, until the Logan eminences had apparently been entirely covered, that the first marine beds, of which there are several, were formed. With the exception of the occasional marine beds which are intercalated, the Pennsylvanian formations of the Appalachian provinee, including eastern Ohio, are now quite generally held to be the result of some phase of continental deposition. Fol- lowing the post-Maxville period of erosion, which developed the sub-Pennsylvanian topography, the simplest interpretation is to suppose slight subsidence sufficient to cause the cessation of erosion and the initiation of deposition, or a climatic change leading to the same result. The area which had been land in post-Maxville time would so continue, except that the topog- raphy would gradually be buried beneath the plain of accumu- eating material, over which coal swamps were spread, even while the tops of some of these hills were yet visible. In connection with the discussion mentioned earlier, it is to be noted that there are few erosion contacts where one would look for an old soil bed with better reason than at this one. ‘The old land surface, gradually buried under supposedly continental deposits in a climate of at least moderate humidity, it would seem, ought to carry some trace of surface weather- ing if not a well-developed soil bed. The actual contact is seldom seen because of poor outcrops. However, it has been seen at several points and there is no suggestion of pre- Pennsylvanian weathering in the rocks below the erosion plane. In three places the Coal-measures sandstones and conglomer- ates are fused on to the Logan. The writer has in his posses- sion one small hand specimen, one side of which is Logan and the other side the coarse Pennsylvanian sandstone with the 560 J.B. Hyde— Notes on the Absence of a Soil Bed. impression of a calamite stem. In this case, there was no ten- dency to break at the contact. At those localities where car- bonaceous shales or coals with their accompanying white clays rest on the Logan, the basal beds are clearly of the Coal- measures series and not a soil bed antedating the Pennsyl- vanian. It is not the purpose to here explain why there is no soil bed present. That seems to involve to a high degree the manner in which the basal Pennsylvanian beds were “formed, that is, whether they were truly continental or whether there was subsidence below a body of water and removal of the soil from the Logan hills by wave action. In support of the latter is the occasional occurrence in the basal beds of the Pennsylvanian sandstones of considerable numbers of fossiliferous limestone or chert pebbles which appear to have been derived from the breaking up of some part of the Maxville limestone, which is otherwise unrepresented in the immediate vicinity. However, the presence of the limestone pebbles and the absence of the soil beds are the only features which suggest that the basal Coal-measures are not truly continental or that the interval marked by the erosion plane has been one of greater com- plexity than simple subaérial erosion. Whatever may have been the intermediate stages, the occurrence is believed to show that a soil bed or evidence of weathering at an erosional contact is not essential to its inter- pretation as a plane of subaérial erosion, even when the next succeeding beds appear to be of continental origin, and that the absence of such a bed at such a contact may demand a much more complicated explanation than that of simple sub- marine or subaérial erosion. LE. A. Kraus—A New Jolly Balance. 561 Art. XLVII—A New Jolly Balance; by Epwarp H. Kraus. Tue first important modification of the spiral spring balance, devised by Jolly, was introduced by Linebarger* in 1900, whose improved balance has since been used rather extensively. The balance to be described has several new features, of which the recording of the elongations of the spiral spring and the reducing of the number of readings necessary to determine the specific gravity are the most important. The balance consists of the square upright tube T to which the fixed vernier M and the movable graduated scale X are attached, as shown in the accompanying figure. A second tube Z is movable within T by means of the milled-head A. The movable vernier N is attached to Z by the arm E. The screw B controls the rod R, movable within the tube Z. From R the spiral spring S and scale pans O and D are suspended, the thin wire rods W and V connecting the spring and scale pans with the pointer P, which swings freely in front of a small circular mirror. In using the balance for the determination of the specific gravity of a solid, for example a mineral, it is necessary that the graduated scale X, the verniers M and N, and the pointer P all be at zero, the lower scale pan being immersed in water. Since the vernier M is fixed, the zero positions of N and X will be opposite M. The pointer P is brought to zero by being made to coincide with the index on the small cireular mirror, the adjustment being accomplished by the screw B. A fragment is now placed on the upper pan © and the elonga- tion of the spring determined by again bringing the pointer P to the zero position. This is now done by turning the mill-head A, which moves the tube Z, the graduated scale X and the vernier N, upward. After the zero position of the pointer is obtained, the scale X is clamped by the screw Y. It is obvious that the reading at M will give directly the elongation of the spring due to the weight of the specimen in air. The fragment is now transferred to the pan D under water. The pointer P is again, for the third time, brought to zero; this time by moving the tube Z and the vernier N downward by means of A, the scale X remaining fixed, having been clamped by the serew Y. The reading at N gives at once the loss in the elongation due to the immersion of the fragment in water. Henee, if the weight in air be represented by W, the read- ing at M, and the loss of weight when immersed in water by * Physical Review, xi, 110-111, 1900. F. H. Kraus—A New Jolly Balance. 563 L, the reading at N, we have the simplified formula: Specific Bee) Aa gravity = The figure shows the balance* with the specimen immersed in water and indicates clearly that the chief advantages of this new form, over other balances in common use, are “to be found in the recording of the elongations so that they may be verified if necessary at the end of the operations, and further but two readings and a simple division are sufficient to deter- mine the specific gravity. I am under especial obligations to Mr. Ralph Miller, whose skill and experience as an instrument-maker were largely drawn upon in perfecting the balance. Mineralogical Laboratory, University of Michigan, February 15, 1911. * Made by Eberbach & Son Company, manufacturers and importers of scientific apparatus, Ann Arbor, Michigan. 564 C. Barus—Coronas of the Fog Layer. Arr. XLV II.—The Independence of the Coronas of the Thick- ness of the Fog Layer ; by C. Barus. 1. Lntroductory.—As an adequate theory of coronas is yet to be given, experiments with a definite bearing on the various features of the phenomenon are desirable. In my earlier work* I endeavored to elucidate the character of the interference phe- nomenon superimposed on the diffraction phenomenon, whereby the discs of coronas with white light eventually show a rhyth- mic succession of colors. A theory was suggested correspond- ing to that given by Verdet for the lamellar grating. Inferences so deduced were quantitatively in accord with facts. I showed that the reappearance of the same type of corona corresponds to a succession of diameters of fog particles in the order of the natural numbers n = 1, 2, 3, 4, etc. Finally, in a survey of coronas obtained with monochromatic light (mercury), it ap- peared that the disc and the first ring are alternately luminous in a way corresponding to the interference phenomenon in question. Finally, that in case of even the largest true coro- nas fog particles of an order of size greater than 10-** were in question, beyond which the corona degenerates into a mere fog. In another papert I touched upon the question of the inter- ference with each other of coronas due separately to two suc- cessive layers of fog particles normal to the line of sight, but the quantitative relations did not seem to be as promising as interferences inferred from the mere thickness of fog particles, already alluded to in comparison with the lamellar grating. 2. Hffect of Thickness: Apparatus.—From the point of view ot the elementary theory the effect of the thickness of the fog layer should be negligible; but it does not by any means follow that this is actually the case. In very many experiments with coronas, the thickness of the fog layer is not at the observer’s disposal; or cases of different thicknesses have to be compared. Hence the following experiment was devised, with the object of detinitely testing the question. In the figure, 7” is a cross section of the tog chamber, along rectangular trough of wood, cloth-lined and provided with two glass plates, g and g’, on the broad sides of the trough. The pool of water is seen at w. Two mirrors, J/ and JZ’, of plate glass (their normals at 2, 7’), horizontally hinged at / and — /’ and capable of being displaced parallel to their own plane by *See Carnegie Publications, No. 96, Part I, 1908; ibid., Part II, 1910. This Journal, xxv, p. 224, 1908 (axial colors) : ibid., xxiv, pp. 309-12, 1907 (cycles of coronas); ibid., xxvii, pp. 73-81, 1909 (mercury light). + Proc. Am. Philos. Soc., April, 1911. CO. Barus—Coronas of the Fog Layer. 565 virtue of the screw extension adjustment s and s’, are attached parallel to each other. The lower margin of JZ is somewhat above the upper margin of JZ’. Hence the observer on the left of the apparatus (in front) sees the direct rays AA’ from the source as well of the reflected rays 6 MW I'L’. By prop- erly adjusting the angle a, the small round distant source of white hght and its image in the mirrors (A’ and B’ respec- tively) may be made to coincide at the upper edge of AZ’. In such a case the corona due to the direct rays produced by a single thickness, d=15:5™ of fog layer, should exactly coincide with, i.e. be the complement of, the coronas due to the reflected rays and produced by a triple thickness d'=46'5™ of fog layer, if the variation of thickness in question is without effect. Other- wise the coronas should be dislocated at the margin of the mirrors. (7 3. Results and Summary.—These experiments were carried through in regular series, both for the dust nuclei of ordinary air as well as for an artificial nucleation due to phosphorus. The exhaustions were made systematically, every two minutes. The coronas due both to the A rays (direct) and the B rays (reflected) were read off as quickly as possible, after which filtered air was introduced to dispel the coronas by evaporating the fog particles. In this way about twenty coronas were suc- cessively compared, from the largest easily observable having an aperture of about 34°, to the small ones of vanishing size, the nucleation ranging from about 10° to zero (particles per Am. JOUR. Scil.—FourTH SERIES, VoL. XX XI, No. 186.—Junn, 1911. 38 566 C. Barus—Coronas of the Fog Layer. eubie centimeter), and the fog particles from a diameter of about 2x10-*™ to 107. In no case was there any dislocation of coronas, or of color, detected, though naturally the coronas in case of reflection from the mirrors were somewhat more yellowish in color (due to the reflecting surfaces) and less vivid (due to the reflections and greater thickness of fog layer); for it is hardly probable that the fog particles are quite of asize. The continuity of corre- sponding colored rings, however, was exact within the limits of observation. Hence thicknesses of 15° and over 45 produce identical coronas identical in aperture ; or the thickness of the cloud layer is without influence on the coronas. " Brown University, Providence, R. I. SCIENTIFIC INTELLIGENCE, * I. CueEemistry anp Puystcs. 1. The Radium Contents of some Uranium Minerals.—It has been shown by Boltwood and also by Strutt that in many miner- als a constant relation exists between the amounts of uranium and radium present, as would be expected from the theory of the formation of radium through the intermediate uranium-X and ionium. Recently Mlle. Gleditsch has found these relations to vary considerably (in the ratio 100 : 86: 68) in thorianite, Joachims- thal pitchblende, and autunite. Soddy and Pirret, while finding a constant ratio in thorianite and the pitchblende, have also found a deficiency of radium in autunite. MaArckwatp and RussELi have now confirmed the last mentioned result, and have made a further study of the autunite, which is a calcium-uranium phos- phate with the formula Ca(UO,),(PO,),.8H,O. This mineral, besides containing less than the expected amount of radium, is remarkable in containing no appreciable amount of lead as well as less helium than the required amount. From the amount of helium present, Soddy has calculated that the age of the mineral would be only about 30 years if none of this gas had been lost, whereas the radium corresponded to an age of many thousand years, so that he assumed that the radium in the mineral came from without and crystallized as an isomorphous replacement of calcium, instead of being a product of the uranium. This view of Soddy’s appeared improbable to Marckwald and Russell, so that they investigated the ionium contents of the mineral and found its amount to be only slightly less than normal in propor- tion to the uranium. Since ionium has an average life of not less than 30,000 years, it follows that the age of the ‘autunite must be Chemistry and Physics. 567 at least hundreds of thousands of years. The facts thus found lead to the view that the porous structure of the autunite, in comparison with the dense oxides, has allowed the escape of helium, and also the leaching out of the lead and a part of the radium. From a consideration of this and other minerals it appears that lead is more easily removed than is radium.—erichte, xliv, 771. H. L. W. 2. The Determination of Cane Sugar in the Presence of other Sugars.—Avo._F JoruEs has devised a new and very simple method for the determination of saccharose when other sugars are present. He finds that the other sugars, such as dextrose, levulose, maltose, mannose, galactose, arabinose, lactose, and rhamnose are readily decomposed by the action of dilute sodium hydroxide solution, while the saccharose may be left undecom- posed and may then be determined by the polariscope.