tee bem ae de heD tae Ghote be ec agin A chy la an ee hy Wath tr A nn nd tlt ihn AeA AD Sanat fhe Baste Ss oe ee Bo ircg tas Rn P= Ba Ma Mi Li nm As he Hh tn Sn oe Hh a am Fi nt Bh A OL me Nel eM Barn Lat Hoe Pace nae A He Ge bi Meta Tom Be Ree Meg DO he Daas the wiihn Mn hn Te ba Am Ban Pew Ng Bib «thar Ne Pm ae Hota Mt then Ne tane Rae tDia Hen Pao an Heit 4 the tn Ds (bok Se ee Rte te Wate? Phang Ran ste Byte Lo Male Tate Pate Aer) wR Ra a Re a Re Reset ne Bale Re bat Me Ain na Pinot tn ee tints ate a eek ee tan “ Nat Hin he Na te Ma Th Ne AAAI R AI PhS IN eo ee OCR Re dee ce ane > r See ee - Retr RRS, PON MAA nth = a aro — reer ae Perea eee te = - ares _* £ ra PP heat Mac McBeth Pr BD aR Be AAP ie Pees eet veri wo Pn PO Pine eA Beten IR Ne A a ~ ee ee > " nD RIANA Cig he te Ain ee oR Ale ten ee ata thm een a sp Ren hn Fentheneae M Ny Te Dot . 3 3 “ Bee Be Morty atngh. Aen Be ts e's tag etl en es eles Pair Mina Neg AO See Ma OG As Lhe TT Bas Nn By Re Me™ ane ot na Ne ee ee Meh Teknte afin kia Ne Nae TO ete SMe Tata, Gust Se a ea Re At tg en le Sg a A Rt CARA THE AMERICAN JOURNAL OF SCIENCE. Epirorn: EDWARD S. DANA. ASSOCIATE EDITORS Proressors GEORGE L. GOODALE, JOHN TROWBRIDGE, W. G. FARLOW anp WM. M. DAVIS, or Camsrince, Proressors ADDISON EE VERRILL, HORACE L. WELLS, LOUIS V. PIRSSON, HERBERT E. GREGORY AND HORACE 8S. UHLER, or New Haven, Proressor HENRY 8S. WILLIAMS, or Itnaca, Proressorn JOSEPH S. AMES, or Battimore, Mr. J. S. DILLER, or Wasuincton. FOURTH SERIES VOL. XXXVII-[WHOLE NUMBER, CLXXXVH}.. WITH NINE PLATES. NEW HAVEN, CONNECTICUT. i 1914. ‘THE TUTTLE, | MOREHOUSE & Ay LY L r NEW HAVEN CONTENTS TO VOLUME XXXVIL. Number 217. Page Art. I.—Determination of Mineral and Rock Densities at High Temperatures ; by A. L. Day, R. B. Sosman, and ea AE OSTE TDR Res ye eS es ye ee aa 2 ee ars 1 Il.—Notes on a new occurrence of Pisanite and Arsenopyrite, and some large Staurolite Crystals from the Ducktown District, Tennessee; by F. R. Van Horn --_---------- 40 IlI.—Diagrammatic Representation of Volcanic Phenomena; by tA. RErrer. «(With Plates to dV). 3 see 2 48 TV.—Nipa-palm in the North American Eocene; by E. W. PER Nee tae Se ee pec a ee ee as 57 V.—Modifieation of the Usual Method of Correcting Silica fon Included: Salts: by-S: B. Kuzirtam = 22222552252 - 61 VI.—Electrometry with the Displacement Interferometer ; YE Cov DAR Sitece 8 en ae nor Gh See eee OD VII.—Upper Devonian Deita of the Appalachian Geosyn- cline: byl Je -BARRELG a. 25 555 ie ee 87 SCIENTIFIC INTELLIGENCE. Chemistry and Physics—Behavior of the Radio-Elements in Precipitation Reactions, Fasans and Beer: New Source of Gallium, BAaRDEL and BovuLaNGER, 110.—Ammonium Peroxides, D’ANS and WepiIG: Poison of Toads, WIELAND and WEIL: Quantitative Analysis in Practice, 111.-- Constitution of Matter: Mechanics of Particles and Rigid Bodies, 112.— Elementary Treatise on Calculus, 113. Geology and Natural History—Conemaugh formation in Ohio, 114.—Cara- docian Cystidea from Girvan: A Contribution to the Paleontology of Trinidad, 115.—Formation of Coal Beds: Contributions to the Geology of the Nordingra Region, 116.—West Virginia Geological Survey, County Reports, 1913.—Iowa Geological Survey: Permo-Carboniferous Verte- brates from New Mexico, 117.—The Meaning of Evolution, 119.—First Principles of Evolution, 120.—Geology and Ore Deposits of the San Fran- cisco and Adjacent Districts, Utah : Nickel Industry, with Special Refer- ence to the Sudbury Region, Ontario, 121.—Uurious Lore of Precious Stones : Botanical Features of the Algerian Sahara: Wissenschaftliche Ergebnisse der Deutschen Zentral-Afrika-Expedition 1907-1908 : Labora- tory Manual and Notebook in Botany, 122.—Introduction to the Chemis- try of Plant Products: Modern Problems of Biology: A Laboratory Manual of Invertebrate Zodlogy: Heredity and Memory, 123.—Bibli- ography of the Tunicata, 1469-1910: British Parasitic Copepoda ; Cope- Bees Parasitic on Fishes : ’ Weathering of Aboriginal Stone Artifacts, Now, 124. Obituary—A. Fritscu: H. Poronrt, 124. iv CONTENTS. Number 218. Art. VIIL.—Geologic Reconnaissance of the Ayusbamba (Peru) Fossil Beds ; by H. E. Grecory IX.—Vertebrate Fossils from Ayusbamba, Peru; by G. F. Haron. | (With: Plates WV, Vil Vile s2 222: eee X.—The “Dam” at Cheshire, Connecticut ; by F. Warp_- 155 XI.—Progressive Development of Mechanics based upon a New Form of Fundamentai Principle of the Science ; by: HM: DamouRIAN =. 25 eee eee eT XII.—Temiskamite, a new nickel arsenide from Ontario ; by Pade WiALRER ee SSE Sot het See eee 170 XTIL.—Use of the Ammonium Salt of Nitrosophenylhydro- xylamine (‘“‘ Cupferron’’) in the Quantitative Separation of Titanium from Iron; by W. M. THornrTon, JR, --_-_- 173 XIV.—Contribution to the Optical Study of the Amphiboles ; by W. E. Forp SCIENTIFIC INTELLIGENCE. Chemistry and Physics—Action of Aluminium Carbide upon Solutions of Metallic Salts, Hmperr and Dirmar, 193.—Solubilities at the Critical Temperatures of Solvents, F. FRrEDRICHS: Organic Chemistry for Advanced Students, J. B. Conen: Industrial Chemistry for Engineering Students, H. K. Benson, 194.—Outlines of Theoretical Chemistry, F. H. Getman: Oxygen in the Sun, RunGE and PAascHEN, 195.—Fluorescence of the Vapors of Sulphur, Selenium and Tellurium, W. Steupine, 196.—The Series Lines of Neon, R. Rosst, 197.—Mass of Rapidly Moving Electrons, G. NEUMANN, 198. Geology and Mineralogy—Changes in Level in the Earth’s Crust, O. FISHER, 199.—Meteorites vs. The Earth, O. C. Farrineton, 200.—San Franciscan Voleanic Field. 202.—Devonian of Maryland : Stratigraphy and Paleontol- ogy of the Alexandrian Series in Illinois and Missouri, Part I, T. E. SavaGE: Important part played by calcareous Algz at certain geological horizons, E. J. Garwoop, 203.—Geological Survey of Oklahoma: Minéralogie de la France et de ses Colonies, A. LAcro1x : Handbuch der Mineralogie, C. Hinrzx, 204. Miscellaneous Scientific Intelligence—Semi-Centennial Anniversary of the National Academy of Sciences, 1863-1913, 204.—Annual Tables of Con- stants and Numerical Data, etc. : Fungi which cause Plant disease, F. L. STEVENS, 205.—Measures of Proper Motion Stars, 5S. W. BurnHam, 206.— Astronomy, H. Jacopy: The American Chemical Journal, 207. Obituary—S. W. Mircuect : B. O. Perce: S. C. CHanpLer: W. Upton: C. B. Ropinson: R. S. Batu: T. Lawrence: P. V. Bevan: W. P. BLoxam, 208. CONTENTS. Vv INiuiani er) 2309 Page Arr. XV.—Fossil Dolphin from California; by R. 8. Luu. CEATAVUG ON TN) NT 0 ee eS i a Pe A 209 XVI.—Note on the Occurrence of the Oriskany Formation on Parlin Stream, Maine; by L. V. Pirsson and C. SC EUW EME Re Tis air eR ene i Ue ae 221 XVII.—Upper Devonian Delta of the Appalachian Geosyn- Gllimes yi Vie EVN era nS ee aa es ea ee eee 225 XVilII.—Paleogeographical Affinities of the Alexandrian TTC Sia Da re lIN I oS Pay APSE oy eT i ces 254 XIX.—Use of Telluric Acid in the Determination of Bro- mine associated with Chlorine in Haloid Salts; by F. A. IGoOocH andskeel SAB nny © OTH) eee ae Se ee 957 XX.—Wilkeite, a New Mineral of the Apatite Group, and Okenite, its Alteration Product, from Southern Cali- fornia; by A. S. Eaxie and A. F. Rocrers...-.------- 262 XXI.—Rocks of the Cerro de Santa Ana on Paraguana, Wenezuelastbyeh As BENDRADT so. 2k Ne ke a is 268 SCIENTIFIC INTELLIGENCE. Chemistry and Physics—Rusting of Iron in Water, W. A. BRapBuRyY: Esti- mation of Periodates in the Presence of lodates and Iodides, MULLER and WEGELIN, 273.—Dictionary of Applied Chemistry, E. THorpr: Allen’s Commercial Organic Analysis, W. A. Davis and 8. 8. SADTLER: Chemis- try, Inorganic and Organic, C. L. Bhoxam, 274.—-Quantitative Chemical Analysis, CLowEs and CoLEeMAN: Industrial Organic Analysis, P.S. ARUP : Radioactive Origin of the Color of Zircons, R. J. Strutt, 275.—Suppres- sion of the Photoelectric Effect, K, FREDENHAGEN and H. Kistner, 276.— Effect of Temperature on X-ray Diffraction Patterns, DEBRoGLin, etc.: Measuring the Pressure Produced in the Detonation of High Explosives or by the Impact of Bullets, B. Hopkinson, 277. Geology and Natural History—Thirty-fourth Annual Report of the Director of the United States Geological Survey: Publications of the United States Geological Survey, 280.—Commission Internationale des Glaciers: Text- book of Paleontology, C. R. Easrman, 282.—Life of the Mollusca, B. B. Woopwarp: Victoria Memorial Museum, Bulletin No. 1, 1913, 283.—Atlas der Krystallformen, V.GoupscumipT: Paléontologie Végétale, F. PELOURDE: Fabre, Poet of Science, C V. Lecros, 284.—Eleventh Annual Report of the Bureau of Science, Philippine Islands: Diseases of Tropical Plants, M. T. Cook, 285. Miscellaneous Scientific Intelligencee—Report of the Secretary of the Smith- sonian Institution, 286.—Report of the Librarian of Congress, ete.: Cata- logue of the Lepidoptera Phalene in the British Museum, Volume XIII, 287.—Napier Tercentenary Celebration: International Electrical Congress, 288. Obituary—H. RosensBuscu: D. Gitu: T. TCHERNYCHEFF, 288. vl CONTENTS. Number 220. Art. XXII.—The Rodadero (Cuzco, Peru),—A Fault Plane of Unusual Aspect; by H. BE. Grucory 2-222 ae 289 X XIII.—Occurrence and Genesis of a Persistent Parting in a Coal Bed of the Lance Formation; by G. S. RoGrrs_ 299 XXITV.—About Climatical Variations; by H. Arcrowsni -- 305 XXV.—Late Paleozoic Glaciation in the Boston Basin, Massachusetts; by HY HL ann = 2) =e eee 316 XXVI.—Note on the Use of the Wilson Tilted Electroscope; by. Wo FARWELL s22i0o 5255s. al es eee ae 319 XXVII.— Mammut Americanum in Connecticut; by C. ScuucHert. With a note on the Farmington specimen by ReS: DuLts2) 2 Ss a2 a ee eg XOX VAT —Hydrolysis of Esters of Substituted Aliphatic Acids; by H.-W. Dean cio 20552 se eee ee XXIX.—Solid Solution in Minerals. V. The Isomorphism between Calcite and Dolomite; by H. W. Footer and WiMe Braprney 23522 ee 339 SCIENTIFIC INTELLIGENCE. Chemistry and Physics—Fused Magnesium Chloride as a Crystallizing Agent, K. A. Hormann and K. HéscuHeie, 345.—Carbon Sulphide-Telluride and Carbon Sulphide-Selenide, A. Stock, P. PRAETORIUS, and E. WILLFROTH: Double Bromides of Gold, A. Gursier and J. Huser, 346.—Gravimetric Determination of Selenium, J. Meyer: Active Nitrogen, E. TiepE and E. DomcKkeE: Rays of Positive Electricity and their Application to Chemical Analyses, J. J. THomson, 347.—Photo-Electricity, H. S. ALLEN, 348.— Introduction to the Mathematical Theory of Attraction, F. A. TARLETON: Chemistry of the Radio-elements, Part II, F. Soppy: Modern Seismology, 349.—Alternating Currents and Alternating Current Machinery: Mutual Repulsion of rigid Parallel Plates separated by a Film of Air, C. Barts, 350. Geology and Mineralogy—Third Annual Report of the Director of the Bureau of Mines, 351.—Canada Department of Mines, 852.—Report of Topo- graphic and Geologic Survey Commission of Pennsylvania, 1910-1912, 353.—Graphite Deposits of Pennsylvania, B. L. MILLER: Geological Sur- vey of New Jersey: Geologic Atlas of the United States, Niagara Folio, New York, E. M. KInpDLE and F. B. Taytor: Union of South Africa, Mines Department, Annual Reports for 1912, 354.—Inter-State Conference on Artesian Water: The Ocean: General Account of the Science of the Sea, J. Murray.—Descriptions of Land, R. W. Cautiery: Interpretation of Anomalies of Gravity, G. K. Gitpert: Mud Lumps at the Mouths of the Mississippi, 306.—La Face de la Terre, E. Suess: Water and Volcanic Activity: Zeitschrift fir Vulkanologie, 357.—Ignecus Rocks and their Origin, R. A. Dany, 358.—Bergalite, 8359.—Notices of recently described Minerals, 360. Miscellaneous Scientific Intelligence—Carnegie Institution of Washington, 361.—Annual Report of the Superintendent of the United States Coast and Geodetic Survey: Newcomb-Englemanns Populiire Astronomie, 363.— Milton’s Astronomy, The Astronomy of Paradise Lost, T. N. ORCHARD: Trigonometry, A. M. Kenyon and L. InGoup, 364.—Who’s Who in Science, International, 1914: Franklin Institute award of the Elliott Cres- son Medal: The Cambridge Manuals of Science and Literature: Tables and other Data for Engineers and Business Men, C. E. Frrris, 365. Obituary—J. Murray: E. S. Hotpen: W. W. Battery: G. MeRcALLI: A. GtntHeR: H. B. Woopwarp: A. R. CLARKE, 366. CONTENTS. vil Number 221. Page Arr. XXX.—Reef Formations of the Northeast Coast of Brazil; by G. A. Warine. (With Plate IX) .......- 367 XX XI.—Note on a New Meteoric Iron from Mount Edith, Ashburton District, West Australia; by W. M. Foorr. 391 XX XII.—Occurrence of Molybdenum in Rocks with special reference to those of Hawaii; by J. B. Ferauson.----- 399 XXXIIL—Fruits of a Date Palm in the Tertiary Deposits of Mastenmehexass by hi W BERRY 225 jo 2 02k es coe 403 XXXIV.—Separation of Titanium from Iron, Aluminum, and Phosphoric Acid with the Aid of the Ammonium Salt of Nitrosophenylhydroxylamine (‘‘Cupferron”’); by Wee ME HOR NTONG Rune re Sones coos ey ae 407 XXXV.—New Features in the Geology of Northwestern Spruzbergen by, O MorrepAnie 32028. 415 XXXVI.—Permian Geology of Northern Brazil; by M. A. sR UTS Oe ee rs ee Me ys CSN DS Ba 425 XXX VII.—Finger Lake Bed in Ashland and Wayne Coun- ties, Ohio, with Tilted Shore Lines; by G. D. Hupparp 444 XXXVUI.—Remnants of an Old Graded Upland on the Presidential Range of the White Mountains; by J. W. Cota D RHO AU eae Seon OR RRS Ue te en oe 451 XXXIX.—The Shawangunk Conglomerate and Associated — Beds near High Falls, Ulster County, New York; by dl Cre ISSO WEN tees ey oie a ee la 464 SCIENTIFIC INTELLIGENCE. Chemistry and Physics—Peculiarity of Nickel and Cobalt Sulphides, A. TuHieL and H. Gessner: Dimethyl Phosphates of the Rare Earths, J. C. Morean and C. Jamus, 475.—Qualitative Chemical Analysis, A. VORISEK: Introduction to Modern Inorganic Chemistry, J. W. Meiior, 476.—Prac- tical Methods of Organic Chemistry, L. GATTERMANN: Spectrography of Réntgen Rays, M. pe Brocuin, 477.—Removal of the Photoelectric Effect of Potassium, HatLwacus and WirpMann, 478.—Heat of Formation of Hydrogen from Hydrogen Atoms, I. Lanamurr: Thermal Expansion of Quartz Glass, K. ScHEEL and W. Herussz, 479.—Physics for Technical Stu- dents; Mechanics and Heat, W. B. ANDERSON: Natural Sources of Energy, A. H. Gipson, 480.—Effect on the Propagation of Electric Waves of the Total Eclipse of the Sun, 21st August, 1914, 481. Geology and Natural History—Manual of Petrographic Methods, A. JoHANN- SEN, 482.—History of Land Mammals in the Western Hemisphere: Nomen- clature of certain Starfishes, A. EK. Verriti, 483.—Birds of Connecticut, Bulletin No. 20, J. H. Sacz, L B. BisHop, and W. P. Burss, 484.—Letters and Recollections of Alexander Agassiz, with a Sketch of his Life and Work, G. R. Acasstz, 485. Obituary—G, W. Hinu: G. M. Mincuin: J. H. Poyntina: A. Saunas, 486. Vill CONTENTS. Number 222. Page Art. XL.—The Binary System MgO-Si0, ; by N. L. Bowrn and OLaF ANDERSEN — 2222255532 2252-2 2 XLI.—Application of the Displacement Interferometer to the Horizontal Pendulum ;. by C. Barus ._.._.._.--_- 501 XLII.—Hydrolysis of Esters of Substituted Aliphatic Acids ; by. Wi. -A: (DRUSHEL 222021202 3 ee ee 514 XLIII.—Thermochemistry and the Periodic Law. Heats of Combination ; Linear Functions of the Atomic Weights of Related Hlements); by W-G. MixnnR 25) =e 519 XLIV.—Picking Out and Mounting Diatoms; by J. M. BLAKE. o2).6. ce 555 a ee er 535 XLV.—Chemical Composition of Bornite and its Relation to Other Snlpho-Minerals; by E. H. Kraus and J. P. GOLDSBERRY 200. USS e Seo ee eS SCIENTIFIC INTELLIGENCE. Chemistry and Physics.—Relative Abundance of Several Metallic Elements, ~ F. W. CuarKker and G. STEIGER, 555.—Spitting of Silver, H. B. Baker: Laboratory Guide to the Study of Qualitative Analysis, KE. H. S. Batmny and H. P. Capy, 554.—Chemistry in America, EH. F. Smita: Annual Reports of the Progress of Chemistry for 1913, 555.—Progress of Scientific Chemistry in Our Own Times, W. A. TILDEN: Synthetic Use of Metals in Organic Chemistry, J. Hate: Electrical Analogue of the Zeeman Effect, J. StTaRK, 556.—Secondary Standards of Wave-Length in the Infra-Red, A. I@NATIEFF: Electrical Conductivity and Ionization Constants of Organic Compounds, H. ScuppEr, 558.—Theory of Heat Radiation, M. PLanck, 599. Geology—Virginia Geological Survey, 559.—Florida State Geological Survey : Illinois Geological Survey: Ohio and Mississippi Floods of 1912, H. C. FRANKENFIELD: New Zealand Geological Survey, 560.—Mountains; their Origin, Growth, and Decay, J. Grrkrm, 561.—La Geographié de Terre-Neuve, R. Perret: Fossilium Catalogus : Spitsbergen Map of Prince Charles Foreland from Surveys, W. 8. Bruce, J. Maraison, etc., 562. Miscellaneous Scientific Intelligence—National Academy of Sciences, 562.— Carnegie Foundation for the Advancement of Teaching; Highth Annual Report of the President, Henry S. Prircuett, 563.—Carnegie Foundation for the Advancement of Teaching; Education in Vermont, 564.—Mining World Index of Current Literature, G. E. Sistey : Les Prix Nobel en 1912: Manual of Bacteriology, H. S. Rrrp, 565. Obituary—N. H. WIncHELL: C.S. S. Petrce: E. Surss: J. Huser: R. K. GRAY. The American Journal of Science ESTABLISHED BY BENJAMIN SILLIMAN IN 1818. THE LEADING SCIENTIFIC JOURNAL IN THE UNITED STATES. Devoted to the Physical and Natural Sciences, with special refer- ence to, Physics and Chemistry on the one hand, and to Geology and Mineralogy on the other. Editor: EDWARD S. DANA. Associate Editors: Professor GEORGE L. GOODALE, JOHN TROWBRIDGE, W. G. FarRLow and Wm. M. Davis. of Cambridge; Professors A. E. VERRILL, H. L. WELLS. L. V. Pirsson, H. E. GreGory, and H. S. UHLER, of New Haven ; Professor H. 8. WILLIAMS, of Ithaca; Professor JOSEPH S. Amzs, of Baltimore: Mr. J. 8. DILLER, of Washington. Two volumes annually, in MONTHLY NUMBERS of about 80 pages each. This Journal ended its first series of 50 volumes as a quarterly in 1845; its second series of 50 volumes as a two-monthly in 1870; its third series as a monthly ended December, 1895. A FOURTH SERIES commenced in 1896. CONTRIBUTORS should send their Articles two months before the time of issuing the number for which they are intended. The title of communications and the names of authors must be fully given. Notice is always to be given when com- munications offered have been, or are to be, published also in other Journals. An author may be asked to pay for proof corrections if exceeding the usual amount. Thirty separate copies of each article will be furnished to the author free of cost and without previous notice from him. They will be provided with a plain eover (but with reference to volume and year). If the author orders separate copies, they will be understood to be in addition to the thirty mentioned above, and he will receive a bill for the extra expense involved, as also for that of a printed cover (with title, etc.), when this is specially ordered. These charges will conform to the following schedule; the rates will be somewhat increased if the article is accompanied by plates, or otherwise involves unusual expense. | No. Copies. eget 50 100 | 200 300 | 500 leesepapes 82 fee | $2.25 $2.75 $3.50 $1.95 | $5.50 NOMS ce A as 3.50 4.25 5.25 6.25 8.00 YO Hae Sea eee | 4.75 5.15 7.00 8.25 | 10.50 Cover ruin | $1.00 $1.25 | $1.05 $2.25 | $3.00 Subscription price $6 per year, or 50 cents a number, postage prepaid in the United States; $6.25 to Canada; $6 +40 to foreign subscribers of countries in the Postal Union. A few sets on sale of the early series. Ten-volume index numbers on hand for the second and third series. (@ Ten-volume Indexes, Vols. I-X, XI-XX, XXI-XXX, fourth series price one dollar each. Address, THE AMERICAN JOURNAL OF SCIENCE, New Haven, Conn. 9 vo MINERALOGICAL AND GEOLOGICAL GEN- ERAL CATALOGUE, Nr. 18. The SECOND edition of Part I. of our General Catalogue of teaching appliances, instruments and utensils for Mineralogy and Geology and their branches, has just come out. It has been extensively enlarged, the first part of this second edition alone covering 260 pages with 110 illustrations as compared with the first edition of 128 pages only. Compiled with the view of offering everything that is needed for the instruction of these sciences in middle- and high-schools, its completeness is such that a num- ber of the different collections of minerals, crystal-models, rocks, fossils, ete., will be found well suited for students’ practical work in the min- eralogical and geological institutes of the universities. AMONG THE NEWLY ARRANGED COLLECTIONS ARE THE FOLLOWING: “Practical Introduction into Petrology:’’ two small collec- tions of 25 specimens each of the MOST IMPORTANT (a) ROCK-FORM- ING MINERALS and (b) TYPES OF ROCKS with thin sections and lan- tern slides, and a description containing reproductions of microphotographs ot each reck-slide for the first introduction into the elements of petrology ; arranged by Professor Dr. Busz. Collection for the Demonstration of the Formation of Coal, arranged by Professor Dr. H. Potonie, according to his book on ‘“Wntstehung der Steinkohle.’’ 5th ed., Berlin, 1910. Collections for the Demonstration of the Decomposi- tion, Alteration, and New Formation of Rocks; 14 differ- ent collections of 8 specimens each, beginning with the unaltered rocks (eruptive and sedimentary) and following the different stages of decomposi- tion, etc. Collections to Demonstrate the Formation of the Moors according to the instructions of Professor Dr. Tacke, Director of the Moor- Institute at Bremen. For more extensive collections see our new catalogues Nr. | (Mineralogy, etc.); Nr. 4 (Petrography: rocks, thin sections and lantern slides); Nr. 17 (Practical Petrography: collection of 336 rocks and thin sections); Nr. 1b, 19 and 23 (Crystallography); Nr. 2a and 2b (Geology and Paleon- tology). Collections and Single Specimens of Minerals, Fossils, Meteorites, bought and exchanged. DR. FF. KRANTZ, RHENISH MINERAL OFFICE, BONN-ON-RHINE, GERMANY. ESTABLISHED 1833. ESTABLISHED 1833. ise e 1 | YMRAQWN, us YW. Se MUO, ' VOL. XXXVII. JANUARY, 1914. Established by BENJAMIN SILLIMAN in 1818. THE AMERICAN JOURNAL OF SCIENCE. Epirorn: EDWARD S. DANA. ASSOCIATE EDITORS Proressors GEORGE L. GOODALE, JOHN TROWBRIDGE, W. G. FARLOW ann WM. M. DAVIS, or CampBrwnce, Proressors ADDISON E. VERRILL, HORACE L. WELLS, LOUIS V. PIRSSON, HERBERT FE. GREGORY anp HORACE S. UHLER, or New Haven, Proressor HENRY S. WILLIAMS, oF ITHaca, Prorressor JOSEPH S. AMES, or Battimore, Mr. J. S. DILLER, or Wasuineton. { eee ae FOURTH SERIES VOL. XXXVII_[WHOLE NUMBER, CLXXXVII]. No. 217—JANUARY, 1914. WITH PLATES I-IV, NEW HAVEN, CONNECTICUT. POWs’: THE TUITLE, MOREHOUSE & TAYLOR CO., PRINTERS, 123 TEMPLE STREET. aN NE eM Pod EA BoP gaat Ore aegis MB ete A TNO ss 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). HODGKINSONITE, A NEW MINERAL. We have been fortunate enough to secure the best specimens of this very rare mineral. It is from the celebrated Franklin Furnace mines and is arare compound, the formula of which is Mn(ZnOH),SiO,. It crystallizes under the monoclinic system and is pink in color associated with barite. In a few specimens it is associated with the rare minerals Pyrochroite and Gageite. The whole makes a very pretty specimen. Prof. Charles Palache has ana- lyzed and will soon publish a description of it. The quantity found is scarcely enough to supply the scientific institutions who will want a speci- men. Prices range from $1.00 to $25.00. A NEW OCCURRENCE—Fluorescent Willemite with Rhodocrosite. The Willemite occurs in transparent crystals running in size from 2 milli- meters in diameter to almost a hair in thickness. They occur in cavities on Rhodocrosite, making a very beautiful specimen. Under a current they show a strong fluorescence. Prices range from $1.00 to $5.00. SYNTHETIC GEMS. It is remarkable the interest that has been taken by scientists in these wonderful scientific discoveries. The Corundums are now produced in Pigeon blood, Blue, Yellow, Pink and White. Also the new Indestructible Pearls in strings with gold clasps. These are identical in hardness and rival in color and lustre the real gems. They can be dropped and stepped on without injury and are not affected by acids. My collection of the above is unrivalled, and prices of the same are remarkably low. OTHER INTERESTING DISCOVERIES AND NEW FINDS Will be found in our new Catalogues. These consist of a Mineral Catalogue of 28 pages; a Catalogue of California Minerals with fine Colored Plates; a Gem Catalogue of 12 pages, with illustrations, and other pamphlets and lists. These will be sent free of charge on application. Do not delay in sending for these catalogues, which will enable you to secure minerals, gems, etc., at prices about one-half what they can be secured for elsewhere. REMOVAL NOTICE. Finding that at our recent location we were placed at a disadvantage in receiving and shipping our numerous consignments, we decided to return to our old location, where we are surrounded by the greatest gem and financial district in the world, and are near the vast shipping interests, Custom. House, etc., ete. Our old offices have been entirely renovated in a beautiful manner, new cases have been fitted and all who take the trouble to call on us will find themselves well repaid by the beautiful display of minerals, gems, curios, etc., ete. ALBERT H. PETEREIT 81-83 Fulton St. New York City. THE AMERICAN JOURNALOF SCIENCE [FOURTH SERIES.] —— Art. 1—The Determination of Mineral and Rock Densities at High Temperatures ; by Arruur L. Day, R. B. Sosman, and J. C. Hosrrrrer. 1. existing Data. BrFore discussing briefly the existing data on the change of specific volume with temperature, we shall mention several misinterpretations and misunderstandings which have occasion- ally appeared. In the first place, the sharp distinction between a glassy and a crystalline substance has not always been kept in mind. A glass is essentially a liquid which has cooled without erystalli- zation to a point where its viscosity has become so great as to give it the mechanical rigidity of a solid without the addition of any new property. The cooling has resulted in no release of the heat of fusion, nor has it developed a discontinuity in any physical or chemical property. There has therefore been no change of state (freezing point) and no one temperature which possessed more physical significance than another in the entire process of cooling to the temperature of the surround- ings. The same is true of the reverse operation ; inasmuch as the glass acquired no new property on cooling, it has none to give up on reheating, and there is no melting and no melting temperature. We must, therefore, associate glasses with “liquids,” not with “solids.” By definition, the temperature at which crystals of a sub- stance (the “solid”? phase) will exist indefinitely in contact with their ‘‘liquid,” unless heat is added or withdrawn, is the true melting temperature of the substance. Above it the liquid alone is stable, and below it only the erystals are stable. But if the substance be very inert it may exist for very long periods in the temperature region in which it is not stable (as, for ex- Am. Jour. Sct.—Fourtu Series, VoL. XX XVII, No. 217.—January, 1914. 1 2 Day, ete.— Determination of Mineral ample, in the form of a cold glass), and the properties charac- teristic of the substance in the form stable at this temperature may thereby be completely veiled. In passing from the amorphous or glassy condition over to the crystalline condition a compound usually undergoes a sud- den change of specifie volume. It is obvious that with rising temperature a glass might increase in specific volume continu- ously, and perhaps at the same time irregularly, in softening gradually to its normal liquid state; yet the fact would indicate nothing as to the amount or even the direction of its specific volume change in passing from the amorphous (liquid) state to the cry stalline state. In the second place, it should be noted that a rock, which is nearly always a mixture or solid solution, behaves very differ- ently from a single chemical compound. The single compound melts at a definite temperature, whether its liquid be viscous or not, in accord with the above definition ; the rock liquefies, continuously or in distinet steps according to the thermal char. acteristics of its individual components, and through a range of temperature which may be very wide. We cannot speak, then, of the “volume change on melting” of a rock as though it were a specific property, but only of the “ volume changes between the temperatures ¢, and ¢,.” With rising temperature, some of these changes of specific volume may be positive, some negative; some will occur at one temperature, some at another; and the total between temperatures 7, and ¢, may be an increase or a decrease. Indirect Estimates.—The most of the estimates of the vol- ume changes in rocks solidifying from an amorphous state are drawn indirectly from the specific Eales at 20° of the glassy and the crystalline rock respectively.* In general, the specific volumes of silicate glasses at 20° have been found to be greater than those of crystalline solids of the same composition. Hence if the dilatation coefticients of the solid and the glass are nearly equal, the rock may be assumed to have undergone a net ex- pansion in liquefying. Estimates made from this standpoint are subject to several large uncertainties. The unknown rate of volume dilatation with increasing temperature, and the frequent changes in this rate through inversions in the solid state, are large unknown factors. Then, too, the glass sometimes cools with the inelo- sure of fine gas bubbles which make its volume appear too * A. Harker, Natural History of Igneous Rocks, p. 158, 1909. F. Zirkel, Lehrb. d. Petrogr., i, p. 680-685, 1898. A. Delesse, Recherches sur les Verres provenant de la Fusion des Roches, Bull. Soc. Geol. France (2), iv, 1380, 1847. J. A. Douglas, On changes of physical eonstants which take place in certain minerals and rocks on the passage from the crystalline to the glassy state, Quart. Jour. Geol. Soc., lxiii, 145-161, 1907. and Rock Densities at Ligh Temperatures. 3 large, while on the other hand the process of fusion frequently changes the original composition by driving off water or other volatile constituents, a change which may lead to an apparently smaller volume for the glass than for the erystal. There are a few cases known of natural glasses which are denser than the corresponding rocks. It is not altogether cer- tain, however, that this is not due to small differences in com- position. Thus, Delesse* found that the glassy portion of certain basic dikes in Scotland was from 1:5 to 4 per cent denser than the crystalline portion. In general, it may be said that if the glass has a considerably greater volume than the crystalline solid (10 per cent or more), it is fairly certain that the solid will expand in liquefying. But as a basis for comparing the volume at high temperatures of one type of rock in the crystalline state with another type of rock in the molten state, the volumes at ordinary tempera- tures are quite unreliable. Evidence that is even more indirect than the comparison of volumes at ordinary temperature has been offered by other authors. Some of these have supported a supposed general expansion of rocks in erystallizing. For instance, Langt in 1875 concluded, from microscopic observation of basalt in com- parison with gypsum, that jointing of rocks is due not to con- traction but to the pressure developed by expansion during solidification. Stiibelt has based a theory of voleanicity upon such a supposed expansion. This same theory was put forth by Bornemann in 1887.§ He found experimentally that mel- anite crystals inclosed in leucite can be fused without disturbing the leucite, and that the fused silicate cools to a glass containing bubbles or vacuoles, indicating that the glass occupies a smaller volume than the original erystal.|| The amount of the contrac- _ tion was estimated to be 4 per cent. On the other hand, Gilbert, in 1877,4] estimated that the tra- chyte of the Henry Mountains had contracted by about one- tenth of its volume from the liquid to the cold crystalline state. This estimate was based on the densities of the sedimentaries and on the hypothesis that the horizon of the laccolitic intru- sions was determined solely by the relative densities of the liquid magma and the invaded rocks. * Delesse, Ann. Mines (5), xili, 3869, 1858. + Jabresh. d. Ver. f. Vaterliind. Naturk. in Wiirttemburg, xxxi, 336, 1875. t A. Stubel: Die Vulkanberge von Ecuador, p. 367-376, Berlin, 1897. §J. C. Bornemann: Ueber Schlackenkegel und Laven. Koenigl. Preuss. geol. Landesanst, Jahrb., 279, 1887. | Ibid., p. 252-258. §| Geology of the Henry Mountains, p. 72-80. (U.S. Geographical and Geological Survey Rocky Mt. Region, 1877.) 4 Day, ete.—Determination of Mineral Direct Measurements.—In addition to the estimates obtained by these indirect methods, direct measurements of the volume change accompanying the liquefaction of rocks and minerals have been attempted. ‘The most of these have depended: (1) upon the floating or sinking of a solid body in the molten rock ; (2) upon direct measurement of the volume changes in a given mass of molten material. Observations are on record of the floating or sinking of lava erusts in liquid lava and of the floating of solid bodies thrown into lava lakes.* The data are contradictory and unsatisfac- tory. The floating of lava crusts is usually due to their spongy, porous structure; in the lava lake of Kilauea, for instance, these crusts sink rapidly when freed from their entrapped or supporting gases. - Flotation experiments on a small scale, with molten rocks, artificial silicates, or metallurgical slags, have also yielded contra- dictory results. Whitley, + i in 1878, concluded from such exper- iments that there is contraction in the liguefying of basalt and of iron-furnace slags. Doelter,{ in 1901, from the sinking or floating of solid bodies of various densities, concluded that there is a net expansion in the liquefying of melanite, augite, and several rocks, and estimated its amount. As no account seems to have been taken of the expansion of the test substances, the results have little quantitative value. Doelter later improved the method by using a platinum sphere;§ he estimates from measurements made with the new apparatus that the density of liquid diopside near its melting point is 2°8. Fleischer| takes the opposite view, that lavas expand in solidifying; his argu- ments, drawn from a variety of laboratory data of a rather indirect kind, are unconvincing. Nies and Winkelmann{ obtained no satisfactory results by the flotation method. The earlier literature is well summarized in their articles, which, however, are concerned chiefly with metals. Measurements of the actual expansion or contraction of mol- ten slags and lavas began with Bischof** about 1838. He esti- mated a total contraction of from 10 to 25 per cent in passing from the molten to the cold crystalline condition, in basalt, trachyte, and granite. Similar measurements were made by * See F. Zirkel, Lehrb. d. Petrogr., i, 684, 1893. + Nature, xviii, 397, 1878. t Neues Jahrb. Min.. 1901, II, 141-157. § Doelter and Sirk, Sitzb. Wien. Akad., exx, i, 659-670, 1911. | A. Fleischer, Zeitschr. Deut Geol. Ges., lv, 56-68, 1903; lvii, 201-214, 1905; lxix, 122-181, 317-321, 1907; lx, 254-258, 1908. qF. Nies and A. Winkelmann, Ann. Phys. (2), xiii, 48-83, 1881. F. Nies, Uber d. Verhalten d. Silicate beim Ubergang aus dem gluthfltissigen i in den festen Aggregatzustand, Stuttgart, 1889. ** Neues Jahrb. Min., 565, 1841 ; 1-54, 1848. and Rock Densities at High Temperatures. 5 Forbes,* Mallet,+ and Barus,t all of whom observed contraction in passing from liquid to solid. Joly used a novel method§ in observing the expansion of spherical beads of different mineral glasses with rising temperature, by projecting the magnified image of a bead on ascreen. All of these measurements have been open to the serious criticism mentioned in a preceding par- agraph, that the substances cooled partly or wholly to glass; the measurements therefore did not include (or only partly in- cluded) the change of state liquid-solid concerning which infor- mation was sought. The measurements by Barus were probably the best up to that time, and have been so frequently quoted that they deserve further description. His method consisted in completely filling a platinum tube with melted diabase,| and following the down- ward movement of the surface, as the melt cooled, by means of a platinum contact point carried by a cathetometer outside of the furnace. By preliminary fusion in the open air, Barus found that the melted diabase cooled to a glass without crystallizing. He states, therefore, that ‘“‘ throughout this paper the molten rock solid- ifies into an obsidian.” On the basis of his own statements - it has therefore been assumed that he really did not observe the volume change accompanying the passage from liquid to erystalline, and the irregularity which he undoubtedly did observe has remained an unexplained anomaly. But, as we shall see later, the rock really crystallized in part under the conditions of slow cooling within his platinum tube, and the sudden volume change which he observed at about 1095° repre- sents a part of the difference in volume between crystalline and glassy rock at that temperature. According to Barus’ figures, the difference in specific volume between diabase and diabase glass at 20° is 11 per cent of the volume of the crystalline rock. At abont 1100° he found a contraction of 3-9 per cent in the glass.** Barus also observed that on reversing the measurements and heating the rock until fluid, the volume change was more grad- ual and came at a higher temperature than the contraction on solidification. The reason for this will become clear from our own experiments. *D. Forbes, Chem. News, p. 6, 1868. + R. Mallet, Phil. Trans., clxiii, 147-227, 1873. ~C. Barus, U.S. Geol. Survey Bull., ciii, 25-44, 1898. 8 J. Joly, Trans. Roy. Soc. Dublin (2), vi, 283- 304, 1847. | The rock is described as ‘‘a typical diabase obtained from Mr. Clarence King” (then Director of the U.S. Geological Survey). No information as to its origin was published. It appears to have come from the Palisade diabase along the Hudson River. “| Loc. cit. (Bull. ciii), p. 26; also pp. 25, 36. ** ioc, cit., p, 41. 6 Day, etc.—Determination of Mineral Method and Apparatus. After consideration of the experimental difficulties and the opportunities for error in the few methods available for dens- ity determinations at high temperatures, we adopted the method of Archimedes, namely, weighing the displacement of a liquid of known density. The principal diticulty, as with all investigations at high temperatures, is with materials of construction. Those used finally were graphite, which in a reducing atmosphere retains its form indefinitely, Marquardt porcelain, which does not soften appreciably below 1600°, and alundum (A1,O,), which melts at about 2050°, but is very por- ous. Pure magnesia is also available, but lacks mechanical strength and permanence. The liquid to be used should have the following properties : (1) high boiling point, (2) low melting point, (8) low volatility between these “temperatures, (4) high density, (5) no chemical reaction with graphite, (6) no action on the minerals to be examined. Silver and tin best fulfil these conditions. Tin has the advantage over silver of a very low melting point (232°), and the disadvantage of lower sensitiveness because of its lower density. The substance to be investigated, being always lighter than the metal, was contained in an inverted graphite crucible. The measurement consisted in determining the weight necessary to immerse the crucible and material to a given depth on the stem of the crucible. Because of the gases given off by natu- ral rocks and minerals when heated, both the liquid metal and the containing crucible must be such as to release these eases (or imprisoned air) at once. Imprisoned gas bubbles would destroy the significance of the measurements. The success of the particular form of apparatus used therefore depends upon the permeability of graphite to these gases. As noted above, the use of a graphite crucible requires the maintenance of a reducing atmosphere within the furnace. In the form of apparatus first tried, the weight was applied at the top of a long porcelain stem projecting upward from the crucible through the top of the furnace. But the center of gravity of this system was too high, with the result that to prevent tilting cuides were required, which introduced too much friction. The weight was therefore applied below (fig. 1), by means of a cage surrounding the large crucible contain- ing the metal. The fixed level to which the float-crucible was immersed was found by electrical contact. A circuit from a storage cell through a galvanometer entered the bottom of the furnace and made connection with the graphite crucible containing the and Rock Densities at Thigh Temperatures. a metal by means of a graphite rod (Ri in fig. 2), held always in contact with the bottom of the crucible by a spring at its lower end. The other terminal from the cell connected, through ‘the pan on which the weights were placed, with the graphite cage, which carried a pointed pin (A in fig. 1) on its upper eross-bar. The circuit was complete through the galvanometer when the graphite point touched the sur- face of the metal. This arrangement re- quired that the inverted float-crucible should be insulated from the cage, and also that the cage should not touch the large crucible containing the metal. Fig. 1 is a section of the apparatus as used, “showing how these requirements were met. G is the inverted graphite float-cruci- ble, containing a block (S) of the substance to be examined. F is the graphite cage or frame. The platinum wire P hangs ‘from the lower cross-bar of this frame through a small hole (H in fig. 2) in the bottom cover ot the furnace, and on its lower end is hung a copper* or aluminum pan, W, for the weights. The stem (B) on the upper cross- bar of the cage allows cage and crucible to be lifted free of the liquid metal before its solidification at the end of the measure- ments, by means of a hook let down from above. The float-crucible is insulated from the cage by the ring of magnesia (E), and the whole is pinned together by two graph- ite pins about a millimeter in diameter, so that the float can be lifted up with the cage. Maenesia was here used instead of porcelain, because the latter yielded slowly at high temperatures to the constant bend- Fie. 1. Graphite ing stress exerted by the crucible stem. apparatus for deter- The crucible, C, containing the metal, Par ai ae meee quid sili- was held upon a column of tubes and screens Gates, of Marquardt porcelain or alundum, locked together so as not to get out of alignment. The lowest tube stood upon three set screws (D, fig. 2) by which the column could be adjusted vertically until the float and cage swung free. Similar screens above the crucible prevented loss of heat by radiation upward. The temperature was measured by a platinum platin-rhodium thermo-element inclosed in a glazed Marquardt tube, which was immersed in the metal close to the side of the crucible in Fire. 1. Day, etc.—Determination of Mineral lige, 2, THERMO-ELEMENT WATER Wee {> WATER tbe risadesaesscssesesresesrsessnsside ZZ SSS As “- U7 BZ a 7, PNG} RANE RSIS Rvty 7 me ROR Pat wes Sy {fT Z L ccdvccciites YA = D SCALE ; xe pe WATER WATER ns) () GALVANOMETER CARBON _f MONOXIDE 2a TO PAN W FIG 1 Fie. 2. Carbon-tube electric furnace for apparatus of fig. and Rock Densities at High Temperatures. 9 order not to interfere with the float. The e.m.f. of the ele- ment was measured in terms of a standard cadmium cell with the help of a Wolff potentiometer. The readings were con- verted to degrees on the standard temperature scale, in the manner described in recent publications from this laboratory.* The opening H (fig. 2) was kept closed by a plug of asbes- tos, except for a few seconds while making the measurements, in order to maintain the reducing atmosphere within. The admission of a small amount of air at this point cannot be avoided, however, and to prevent its oxidizing the lower cross- bar of the cage, this was covered with a thin sheet of plati- num. The loss of weight from the cage during a series was then less than 0-1 gram. This was corrected for in the caleu- lations. The furnace required for this apparatus must give a uni- form temperature up to 1600° in a space about 80™™ in diam- eter and 200™" long. For this purpose the carbon tube furnace shown in fig. 2 was employed. It was built on the same plan as our small tube-furnace used for the melting point. of platinum,t except that only the middle portion of the tube was of carbon, ground on the lathe to a uniform thickness and diameter. This was set into end pieces of graphite which fit the brass water-cooled terminals and require no copper plat- ing or soldering to give a good contact. The carbon tube was surrounded by a fireclay tube, and the intervening space filled with carborundum. The outside packing was of petroleum coke. This furnace took about 750 amperes at 10-15 volts to reach 1600°. After the earlier measurements the apparatus was enlarged to obtain greater sensitiveness. The second furnace was of the granular carbon resistance type, developed at the Koenig- liche Porzellan-Manufaktur in Charlottenburg.t The heated zone in this furnace was about 130"™ in diameter and 250™™ long. The graphite apparatus was protected from oxidation by a mixture of nitrogen and car bon monoxide, introduced through a tube at the bottom. This. gas mixture was made by passing air through a porcelain or iron tube filled with coke and standing vertically in a_nickel wire resistance furnace kept at about 1100°. Only 18 grams of coke are required to furnish 100 liters of gas containing two parts nitrogen and one part carbon monoxide. The gas was passed through suitable * Day and Sosman, this Journal, xxix, 93-161, 1910. R. B. Sosman, ibid.. fani 1-15, 1910. Day and Sosman, Carnegie Institution Publ., No. 197, + Sosman, this Journal, xxx, 1-15, 1910. ¢ J. Bronn, Der Elektrische Ofen, p. 94, 1910. 10 Day, etc.— Determination of Mineral absorbing agents to remove traces of carbon dioxide, hydrogen sulphide and water vapor. 3. Preliminary Measurements. The first datum necessary is the density or volume curve of the metal up to the highest temperature to be used. As an absolute basis for determining this volume, we must have the measured linear or volumetric expansion of some solid sub- stance up to the same temperature. The only one available seemed to be amorphous silica (quartz glass), whose expansion has been determined by several investigators, and is, further- more, so small as to be almost negligible. We sought to employ this known datum by making two series of measure- ments: first, the displacement of metal by a solid graphite float ; second, the displacement of inetal by a similar float in which as much as possible of the graphite was replaced by some silica glass disks, cut from a block of glass made by Day and Shepherd in this laboratory. A series of simultaneous equations was thus obtained, containing the density of silver and the expansion coefficient of oraphite as unknown quantities. Although this series gave the density of silver within 0°5 per cent, the results were not entirely satisfactory, first because the silica glass crystallized to cristobalite at about 1200°, cut- ting off the measurements at that temperature; and second, because too much uncertainty was left in the expansion coef- ficient of the graphite, which was required to be used in later work as a correction. Therefore, direct determinations were made of the expansion of the graphite used, and these were taken as the basis for the volume determinations on the metals. 4. Expansion of Graphite. The data on the expansion coefficient of Acheson artificial graphite, together with a comparison and discussion of results, have already been published elsewhere.* The measurements were made on barsabout 700™™ long, in two types of electrically heated furnaces. The mean coefficient of expansion from 0° is expressed by the formula 10°B = 0°55 + 0°00168. The actual determinations are plotted in fig. 3. To obtain an accuracy of 0-1 per cent in the specific volume of tin, the expansion coefficient of graphite needs to be known _ with an accuracy of only 7:5 per cent at 1500°, and of 40 per * AL. Day and R. B. Sosman, Jour. Ind. Eng. Chem., iv, 490-493, 1912; Jour. Washington Acad. Sci., ii, 284-289, 1912. and Rock Densities at High Temperatures. 11 cent at 500°. The effect of errors in the expansion coefficient upon the measured volume of the metals was therefore extremely small. The effect of a similar error upon the measured volume of minerals was even smaller, because only a erucible or shell of graphite of small volume was employed. Fig. 3. [ay SR TT Cen aT Ih OU ANT [ears oa RU Mee [leaned ol eee nea ACHESON GRAPHITE MEAN COEFFICIENT eae ee OF LINEAR EXPANSION FROM O° | : ae g ° 100 200 300 400 500 600 700 800 900° =-41000 1100) 1200. 1300 1400 1500 1600 TEMPERATURE 5. | Specific Volume of Metals Used. It is hardly worth while to record the large number of observations made on silver, tin, lead, the eutectic of tin and lead, and other lead-tin alloys. The data on silver and the alloys are not complete, and as the values do not enter into the volumes of any substances discussed in this paper, they will be reserved for later publication. The experimental results on tin, lead, and the lead-tin eutectic are plotted in figs. 4 and 5. These include three series on tin (11, 26, and 28 October 1910), three on lead (30 June 1911, 138 March 1912, 31 January 1913), and one on the eutectic (16 June 1911). The eutectic contains 63 weight per cent of tin, 387 of lead.* The volumes and densities at round temperatures are sum- marized in Table I. The curves of figs. 4 and 5 have been completed between 0° and our lowest temperatures of measurement by inserting the data of Vicentini and Omodei. In the ease of lead their value for the density at the melting point is apparently too low, and we have put in Table I the value obtained by extrapolating our own curve over the very short interval of 11° between our measurements and the melting point. The volume curve for tin is very nearly a straight line from * Bornemann, Metallurgie, viii, 271-275, 1911. SaYuNLVYsdWNaL OoZzl oogl oosi oor OOEl ood ool ooo! O06 oos OOL 009 OOS OOv OOE foley4 ool (e) La) Ss ~ Se) SET g iS €9Q/ uospybisaa PUD S104 OFF sea) | a en aa rS 98g) !apowd puD TERY eh ee _~ Ort Ss = < S O 3 = Srl S = < “Ss : m ~ o} ~ = O ost SaaS 7 ~S oh D Q XR | | @ ser SS Zé, D NIL Qala ~ JO AWNIOA ODlslO3dS G ~~ S VLVG IWLNAWIYSdxg Sol 12 and Rock Densities at High Temperatures. 13 ATA OF VICENTIN VOLUME OF 1KG IN CC < b-----------,- TEMPERATURE 14 Day, etc.— Determination of Mineral the melting point up to 1600°, but is slightly concave upward. The graph for lead is perfectly straight from the melting point to 825°, and is represented by the formula: Kilogram-volume = 90°05 + 0°0108 4. The curve for the eutectic is slightly concave downward. At 650° the kilogram-volume of the eutectic, calculated on the assumption that the volume is proportional to the weight- percentage, should be 129°6. The value actually found, 129-88, is only 0 0-2 per cent higher than this. The differ ence might easily be brought about by the combined errors of com- pounding the alloy and making the three separate determina- tions of volume. In other words, the mixing of the two liquid metals in eutectic proportion is not accompanied by any contraction or expansion in the total volume. TABLE I. Density and Volume of liquid tin, lead, and lead-tin eutectic. Density Kilogram-Volume Temp. |- Tin Lead Kutectic Tin Lead Eutectic 181 (8:082)* (123-7) 200 (8-063) (124:0) 232 (6-988)* (143-1) 250 (6°982) 8-011 (143-2) 124°83 300 6°948 7-965 144.03 125°55 327 10°686 93°58 350 6908 10°658 7-921 144°76 93°83 126°25 400 6°875 10°597 7879 145°45 94°37 126-92 450 6-844 10°536 7838 146°11 94°91 127-58 500 6°814 10-477 7800 146°76 95°45 128 20 550 6.785 10°418 7-764 147°38 95:99 128:°79 600 6°755 10°859 77381 148-04 96°58 129°34 650 6°725 10-302 7-699 148-70 97°07 129-88 700 |! 6695 10°245 149°37 97°61 750 | 6°666 10°188 150-02 28°15 800 | 6:637 10°182 15067 | 98-69 850 | 6607 10:078 es | 99:28 900 | 6578 | 2-02 950 | 6:548 132°7 72 1000 6°518 153-42 1100 6°459 154°82 1200 6°399 156°27 1300 6°340 157-78 1400 6-280 159-24 1500 | 6-221 160°75 1600 | 6-162 162°29 | | *Vicentini and Omodei. Atti R. Acc. Lincei. iii, pp. 235, 294, 321, 1887. and Rock Densities at High Temperatures. 15 6. Quartz. A considerable amount of data already exists on the expan- sion of quar tz. Fizeau* and Benoitt have determined its linear expansion, both parallel and perpendicular to the optic axis, with great accuracy below 100°. Reimerdest has extended the parallel coefficient to 220°, and Randall§ to 500°. Le Chatelier| made approximate measurements on both axes up to 1060°, at a comparatively small number of tempera- tures. The quartz which we used was all from Minas Geraes, Brazil, and was in the form of clean transparent blocks weigh- ing from 35 to 85 grams. Measurements were made up to 1602° under metallic tin, in the inverted crucible apparatus described on page 7. Check measurements were also made in another form of apparatus designed for lower temperatures and for solid blocks only. This apparatus consisted of a cage similar in form to the graphite cage, but made of monel metal (a nickel-copper alloy). The graphite float-crucible was replaced by a four- pronged metal clasp which held the block of quartz. The crucible containing the metal was of steel. The furnace was wound with platinum wire. Measurements were made with this apparatus, using the molten eutectic of tin and lead, the density of which had been determined by means of a graphite block. The individual determinations are too numerous to be repro- duced here. Table II, giving the results of one of the series, will suftice to show the order of magnitude of the quantities measured, and the precision. The first column gives the temperature, and the second the total displacement, which is equal to the sum of the weights of the sample, the graphite crucible and cage, the pan, and the added weights. The third column gives the density of liquid tin at the temperature in question, and the fourth column the resulting volume of tin displaced. Subtracting the volume of the graphite crucible in column 5 gives the volume of the sample, and dividing this by the mass of the sample gives the specific volume in the last column. For convenience this is multiplied by 1000, so that the figures represent the “ kilogram- volume” or volume of one kilogram in cubie centimeters. *Pogs, Ann., cxxvili, 564-589, 1866. + Trav. Mem Bur. Int., vol. vi, 1888. t Inaug. Diss. Jena, 1896. § H. M. Randall, Phys. Rev., xx, 10-37, 1905. || H. LeChatelier, Compt. rend., eviii, 1046-1049, 1889. S| Steel, which was used at first, is too magnetic, and a small error is introduced on account of the force exerted by the furnace current. 16 Day, etc.—Determination of Mineral TaBLeE II. Example of Determinations of Specific Volume of Quartz. Datessc- 2th ees) sss ee tee eee 21 January, 1911 Wioirhtiof specimen ess] = eens eee 34°12 g. Weight.of (cage; ete) sis. os See ee eee ae eee 94°17 g. Volume graphite crucible at 0°______-_-_---___- 8°21 ce. Metal! used? 282): oh Se a ee ee Tin a | 2a eae Kilo- . otal | Volume | Volume | volume Temp-s | hotly Dens Loa a oluile | graphite. | quartz quartz | | | 1000 v. 20 Ree pee aay Mee tee ee emer ITT! | BARNS 450 146°6 6°844 RADE a Oreo ale ealiesl 386°6 514 TAG ee | MGSSOG Mee elcoD eral Spear mel aoe 390°5 550 1465 = 6-785 20-09) 38:23.) 138536 391°6 561 TAG SD A aGSINS el neol Ollie alin Oreo al eel ors 392°2 585 147°3 5-764 | 21:7 [p= (S728) ealdso0 397°2 600 LAL onl | MOLTO) salem eles Onn Him Se cannes | memel one 397-2 650 146-4 6°725 PV Oy 18:24 eel Sro8 396°6 701 145°6 C694 2 Or al S024 eld oil 396°0 705 145:0 i663 ee 2leGy Se Sre4) ee elasoe 396°3 800 144°3 6-637 | 21-74 8-25 | 18°49 895-4 850 VAS a OsG0 mel eel 8°25 13°50 395°7 900 143-0 6578 | 21°74 8°25 18°49 395°4 950 143°1* 6548 | (21°85) | 8:26 (13°59) (898-4) 1091 143-8+ 6-464 | (22°3) | 8-27 (14:0) (410°) 1167 142°3+ GALOIS A (2272) onl nOneo. a eM(bono) (407°) 1175 141°3+ 6-414 | (220) | 8:28 (13°7) (403°) 1204 140°6+ 6°397 (21°98) | 8:28 (13-7) (402°) 1250 138 8 6-370 | 21-79 8°29 13°50 395°7 1300 158°3 6-340 | 21°81 S329) Fie diss02 3963 1350 138°0 6-310 | 21°87 8:30 | 18:57 397°8 1392 138-2 6-285 | 21°99 8:31 | 13°68 401-0 1400 138°3 6°280 | 22:02 8°31 13°71 401°9 1450 139-4 6:250 | 22-30 8°31 13°99 410-1 1500 139°9 6221 | 2249 Cran if aldloilyy 415-4 1550 142°0 6-192 | 22°93 8°33 | 14:60 428°0 1594 144°2 6-166 | 23°39 8:34 | 15°05 441°2 1602 146-4 6-161 | 23°76 8°34 | 15°42 452°0 * Gas evolution beginning. +Gas escaping rapidly. + Gas evolution ceased. The volumes of the float-crucible and of the quartz at 20° were determined by replacing the molten tin by mereury. The volume of the sample was also checked by weighing under water in the usual manner. All of the experimental data on quartz are plotted in fig. 6, and the six series of measurements are listed in Table III. As will be seen from the curve, the voiume increases more and more rapidly as the inversion temperature (575°) is approached. Beyond that temperature the volume seems to decrease slightly. The escape of gases beyond 950° interrupts the measurements, and the points shown in the figure represent merely chance and Rock Densities at High Temperatures. 17 Fic. 6. ESCAPE OF GASES 1100 500 600 400 300 TEMPERATURE EXPERIMENTAL DATA SPECIFICS VOEUME Op QUARTZ 200 100 DD NI ONt AO ANNIOA {eo} ° fe) ° fo} Q Oo o i © 0 ” n ” ts) a) 9 fe) ° fo} fe) eo) © 9) 4 9) N = 7 v vt v vt NJ 400 Am. Jour. Sc1.—FourtH Series, VoL. XX XVII, No. 217.—January, 1914. 2 A 18 Day, ete.—Determination of Mineral values of apparent volume due to the fact that the gas escapes from the quartz more rapidly than it can get out through the graphite float-crucible. But above 1250°, when the evolution of gas had practically ceased, there are a few points which indi- cate that the volume either continues constant or goes on dimin- ishing over the range from 950° to 1250°. About this time, however, the formation of cristobalite becomes rapid and the volume begins to increase greatly. The form of curve below 575° is not due to a time lag. No effect due to the rate of heating was noticed, and the volumes obtained on lowering the temperature were practically the same a 8 with rising temperature, except for a small increase (about 2 per cent) due to the formation of a few cracks at the inver- sion temperature. Data with both rising and falling temper- atures are included in fig. 6. TaBLe III. Measurements made on Minas Geraes Quartz. Date Metal. | Wt. of specimen Mark in Fig. 6 9 January, 1911 Tin 43°98 Circle 18 | 6 same Square 21 aS ie ee | 34°12 Triangle 13 February a ts 39°02 Cross 21 sf ud | 39°96 Inverted triangle 20 June ‘¢ | Tin-lead eutectic, 85°18 Diamond The form of curve which we have found for the volume increase up to 575° is somewhat different from that calculated and published by LeChatelier.* His determinations below 575° were at three temperatures only, and the volumes calculated from his linear expansions at those temperatures fall almost exactly on our curve. Above 575° his values for the volume are too low, probably because his measurements were made (photographically) on bars 100™" long. With the ends exposed, as they must have been, a considerable percentage of the length may never have reached the transition temperature with its accompanying large increase in length. The increasing dilatation of quartz as the temperature approaches 575° is especially striking when compared with the curves for various other properties. Among those properties on which accurate measurements to 500° or higher have been made are the following: linear expansion parallel to the axis, * LeChatelier, Compt. rend., cviii, 1046-1049, 1889. Bull. ‘Soe. Fr. Min., xiii, 112-118, 1890. and Rock Densities at High Temperatures. 19 erystal angles (F. E. Wright, unpublished), double refraction, refractive index parallel and perpendicular to the axis, and specific heat. While some of these properties are interdepend- ent, nevertheless it is a striking fact that in every case the effect of the change at 575° begins to be felt 200° or more below the temperature of complete inversion. All of these various properties agree in showing a sudden break in the curve at 575°. In other words, although each property shows a rapidly increasing effect due to the approach- ing inversion, nevertheless there is apparently a discontinuity in the low-temperature branch of the curve. The final stage of the transition occurs very rapidly, probably within 0-1°, a range of temperature too small to be controlled and measured with our present forms of apparatus. This final part of the curve should therefore be drawn as a vertical straight line. The optical data show this point better than do the volume data. The form of the quartz curve is not that which would result from the presence of two molecular species in equilibrium, of which the one is increasing in concentration at the expense of the other. It is rather the form of curve which would result from the action of opposing mechanical forces. Various mechanical models of such a system will suggest themselves. or instance, the curve showing the relation be- tween the angular velocity and the moment of the forces on two co-axial cylinders rotating in water has a discontinuity like that in the quartz curve, since there is a limiting value of the veloc- ity below which the moment is proportional to the velocity, and above which the square of the velocity enters. The quartz curve therefore tends to confirm the view, well stated by Fenner in his article on the forms of silica,* that the alpha-beta inversions of quartz and tridymite represent only a rearrangement of the molecules in the crystal, whereas the change from one of these forms to another represents a real change in the constitution of the molecule itself. The unusually small variation in the properties of the high temperature quartz with, rising temperature is also striking. It is pertinent to inquire whether there may not be some con- nection between this fact and the unusually slow time-rate of melting of quartz. In both of these respects it differs from almost all other known solids. Orthoclase shows an interesting parallelism in being, on the one hand, one of the slowest- melting of the silicates, and, on the other, in having one of * This Journal, xxxvi, 365, 1913. + Since the above was written, F. E. Wright has published a comparison and discussion of the quartz curves. Jour. Washington Acad. Sci., iii, 480- 494, 1913. 20 Day, ete.—Determination of Mineral the lowest dilatation coetticients, actually contracting in two directions while expanding in the third.* A great deal of additional work on the properties of matter in the solid state is necessary before these relationships can be understood. Escape of gases from quariz.—As the data of fig. 6 show, there was a rapid escape of gases from the quartz, beginning at about 950° and continuing for a varying period of time depending upon the rate of heating. We made no attempt to find the lowest temperature at which this gas evolution would begin, nor did we study the effect upon athe gas evolution of holding the quartz for a long time at a fixed temperature. Chamberlint has found that the average amount of gas obtained from six specimens of quartz was 0°35 of the volume of the erystal, and that. the gases consisted chiefly of carbon dioxide and hydrogen. The amount of gas obtained by simply erushing the mineral to open the microscopic cavities was insig- nificant, even from quartz having comparatively large cavities. The amount obtainable by chemical reactions of water and impurities in the quartz was also small. The greater part of the gas seemed therefore to be either in solution in the solid erystal, or held in invisible cracks. Our experience supports this view. Quartz heated to 1300° under tin, after giving off a considerable quantity of gas above 1000°, was brought back to room temperature with relatively few cracks and with unimpaired transparency. The cracking seems usually due to the rapid volume change as the tempera- ture approaches 575°. The volume of gas escaping from our quartz specimens was, however, much greater than the volumes stated by Cham- berlin. It should be noted that the gases collected by Cham- berlin were all passed over calcium chloride before any measurements were made, so that he has no data on the amount of water vapor evolved nor on the temperature at which it escaped. A. W. Wright found that the smoky quartz of Branchville, Connecticut, contained, on the average, 0-062 per cent water which escaped below red heat, in addition to 1°32 volumes of gas chiefly carbon dioxide.g This quartz contained unusually large inclusions filled with water-and liquid carbon dioxide. Koenigsberger and Mueller] found that 80 to 90 per cent of the included material in three Alpine quartzes was water. * Fizeau, Poge. Ann., exxxv, 372-395, 1868. ‘cline T. Chamberlin : The Gases in Rocks, Carnegie Institution Pub. 106, ¢ This Journal (8), xxi, 209-216, 1881. Wright’s figure has been quoted erroneously as ‘‘0:07 volume” in Chamberlin’s table (p. 24 of Publication 106). | Centralblatt Min, 72-77, 1906. and Rock Densities at High Temperatures. 21 The Brazilian quartz used in our experiments showed under the microscope no trace of gas bubbles or inclusions. 8°37 grams (in large fragments) heated for a few minutes at 620° lost 1:8" or 0°022 per cent. Heated further for 24 hours at 1105°, the quartz lost 0°3 milligram additional, or 0-025 per cent in all. If this were all water vapor at 1000°, it would have a volume about 3:7 times that of the quartz. In contact with graphite this volume might be doubled by the formation of hydrogen and carbon monoxide. Such a’ volume of gas would be sufficient to account for the phenomena observed. This question was not pursued further as it would lead us too far astray from our present purpose, and as it presents a large problem for further work. Formation of cristobalite—Fenner* has shown in this laboratory that quartz is the stable form of silica under atmos- pheric pressure only up to 870°. From 870° to 1470° tridy- mite is stable, and above 1470°, cristobalite. Quartz in the dry state will not pass directly into tridymite between 870° and 1470°, but above 1000° it will be converted directly, though slowly, into cristobalite. At 1300° this conversion is fairly rapid. As will be seen from the curves of fig. 6, the volume begins to increase noticeably above 1800°. In the series of 21 January, 1911, with a maximum temperature of 1602°, cristobalite formed a surface layer over the block about half a millimeter thick, and had also grown into the larger cracks, causing con- siderable cracking and distortion of the block and increasing its apparent volume about 15 per cent. This block was above 1400° for about 40 minutes. In the series of 13 February, 1911, considerably more cristobalite had formed, and the block was easily broken to pieces; this had been above 1400° for 65 minutes. In the series of 21 February the block was held at 1500° for about 34 hours. The conversion to cristobalite was still quite incomplete, although the chalky white growth of erystals had penetrated all through the block. In every ease the cristobalite formed first on the exposed surfaces, and grew inward, so that even in the most completely converted block there were clear unaltered fragments weighing a gram or more in the center of the block, and microscopic unaltered angular quartz scattered all through the converted portions.t’ The apparent volume increase in this last case amounted to 52 per cent. On account of the friable and porous character of the cristobalite formed, however, this is considerably larger than *C. N. Fenner, Jour. Washington Acad. Sci., ii, 471-480, 1912; this Journal, xxxvi, 331-384, 1913. + This dependence of the growth of cristobalite upon the surface exposed has been noted by Endell and Rieke, Zs. anorg. Chem., Ixxix, 239-259, 1912. ? 22 Day, ete.— Determination of Mineral the true increase. At 20° the true specific volume of eristo- balite is 13-4 per cent greater than that of quartz. kb 7. Granite. Three specimens of granite were used: (1) a hornblende granite from Copper Mt., Alaska (No. 128); (2) a pink coarse- orained quartzose eranite from Kasaan Peninsula, Alaska (C. W. Wright No. 39); (8) a gray biotite- muscovite granite from Stone ‘Mountain, Georgia. The latter was furnished us Fie 450 ih EXPERIMENTAL DATA # at SPECIFIC VOLUME OF Ps GRANITE sg 430 = 420 410 O O Z 400 O % a 390 O WW = 380 =) | O > 370 Bee5 100 200 300 400 Soo 600 700 800 900 1000. in the form of two large blocks by the Stone Mt. Granite Co. of Atlanta. The first specimen was used only for a preliminary measure- ment. The second was tested under tin to 618° in the metal- frame apparatus (page 15). The third was tested to 935° in the graphite apparatus. The results are plotted in fig. 7. In both of the granites that were heated above 575°, the rapid volume increase on approaching that temperature iS notable. This increase is to be expected from the presence of iree quartz in the granite, but its amount is unexpectedly and Rock Densities at High Temperatures. 28 large. The volume increase of the two granites between 0° and 575° is 5-07 and 5:82 per cent, respectively, giving a mean of 5-45 per cent. The corresponding increase for pure quartz is 5°17 per cent. The most probable cause of this abnormally large expansion is the action of the pure quartz crystals in pushing the other mineral fragments apart, increasing the apparent volume of the rock. Evidence for this explanation is seen in the curve giving the data of a second heating of the Alaska granite. In this series the volumes are all lar ger, although ‘the curve is parallel to the first. The final kilogram-volume of this rock after the two heatings to 600° was 393-7, a permanent increase of 2°20 per cent. In the Stone Mt. granite, which was heated to 935°, a very considerable permanent expansion occurred. This is probably due largely to the unequal expansion of the different minerals, which causes them to push each other apart, fracturing the rock and rendering it porous and friable. The relative dimen- sions of the block, after heating, remained the same, but it had increased in volume by 9°3 per cent. The heating above 575° caused also the evolution of con- siderable gas between the temperatures 750° and 900°. During a part of the time the gas was evolved quite regularly, causing the floating system to rise and fall rhythmically, this rise and fall being caused by gas accumulating, lifting the float crucible, and escaping through the graphite top. Chamberlin found in the Stone Mt. granite 0°76 of its volume of gas, consisting chiefly of hydrogen. Water, which he did not determine, would add considerably to this, just as in the case of quartz already discussed (see p. 21). The escape of these gases probably contributes to the shattering and permanent dila- tation. This permanent dilatation in granite and other rocks has frequently been noted by other observers, and is indeed to be expected in any heterogeneous material. Any change of temperature will cause the constituent parts to expand unequally in accordance with their differing dilatation coeffi- cients, and if the temperature range is large enough so that the stresses thus set up exceed the breaking strength of the parts, there will be internal fracturing and opening of irregular cavities. N. E. Wheeler* has determined the linear expansion of several rocks used by Adams and Coker in their high pressure studies. In all cases a considerable permanent expansion was produced by the first heating. We have calculated from his data the volumes of the Westerly (Rhode Island) granite * Trans. Royal Soc., Canada, iv, III, 19-44, 1910. 24 Day, etce.—Determination of Mineral which he used, on the basis of a specific gravity of 2°627 at room temperature.* The results are plotted in fig. 8 These expansion measurements by Wheeler were meal on a 20-centimeter bar of rock the ends of which were viewed through side openings in the furnace tube. This method would tend to cover up any irregularities in the expansion, since the portion near the ends of a heated bar may lag in temperature considerably behind the remainder of the speci- men, even under conditions of stationary temperature. We Fic. 8. GRANITE BASED ON DATA BY WHEELER VOLUME OF 1KG IN CC TEMPERATURE have found this to be true in a similar apparatus in which measurements were made on platinum alloys.t On account of this temperature gradient, the effect of a transtormation which is complete at a definite temperature may be spread over a consider able range of temperature. The break in eurva- ture near 575° is therefore much less marked than in our own curves. It ‘a noticeable, however, especially on Wheeler’s cool- ing and reheating curves for the same block. Under the conditions above noted, there would have been no purpose in carrying on the experiments to a temperature suf- ficient to liquefy the granite, since the true dilatation of the erystalline rock can not be determined. Under the heavy pressure of overlying rocks the dilatation due to fracturing and differential expansion would be less than that obtained in ex- * Datum from F. D. Adams, private letter, 1913. + Day and Sosman, this Journal (4), xxvi, 425-441, 1908 ; xxix, 114, 1910; Carnegie Inst. Publ. 157, pp. 35, 63. and Fock Densities at High Temperatures. 25 periments at atmospheric pressure, but it would become zero only at depths where the minerals would flow or redissolve and fill all pore spaces. 8. Diabase. The diabase used was from the quarry of the Fairview Stone Crushing Company, near Granton, N. J. The diabase exposed at this locality appears to be an offshoot of the great Palisades sheet which outcrops along the Hudson River and southwestward through New Jersey. The particular specimen in question, in the form of a 100 lb. block, was from the Jower contact of the diabase against the underlying Newark (Triassic) shale and sandstone. Prof. J. Volney Lewis, who collected this material for us, has described the rock in his report on the “ Petrography of the Newark Igneous Rocks of New Jersey.”* At the locality in question there was, until quarrying operations followed it back into the underlying strata, an excellent example of a type of inclusion frequently occurring in the Palisade diabase, namely, slabs of the underlying shale or arkosic sandstone which have been “ floated” up into the igneous rock. These occur in various parts of the sheet and stand at various angles between horizontal and vertical.t+ It was therefore of interest to compare the volume of the molten diabase with that of the solid arkose and shale at the Same temperature. Our experience with granite had already convinced us that the expansion of a complex rock can not be accurately deter- mined by direct measurement above a few hundred degrees, because of the shattering due to the unequal expansion of the different minerals. Furthermore, a trial series on a pyroxenite showed that when the rock was heated up directly through its fusion range, its volume change in liquefying was concealed by the escape of a large amount of gas, which swelled the half- melted pasty rock into a sponge so voluminous as to cause the tin to overflow and thus spoil the series of measurements. We therefore made a glass of the rock by fusion in she open air, thus permitting the escape of all bubbles, followed by rapid cooling. The rock was fused in a cylindrical crucible made of platinum foil 0:05™ thick ; the foil was then stripped from the glass when cold, leaving a block of the proper dimen- sions. Each melt was held for a few minutes between 1300° and 1350° until all bubbles had escaped from the liquid. When *Geol. Surv. New Jersey, Ann. Rept. 1907, 97-168. See in particular, plate XVI; analyses V and XII, p. 121, and p. 127. t Ibid., p. 135 and plate XXVII, fig. 2. 26 Day, etc.—Determination of Mineral eold it formed a clear dark brown glass, with a refractive index varying from 1°59 to 1°61. No bubbles were visible, even under the microscope, and no crystalline material except at a few points on the surface of the block. These were cut away by grinding. It was to be expected that this fusion would not only drive off the water but would also cause some change in the state of oxidation of the iron, so that the glass would not retain exactly the composition of the original rock. The amount of moisture was found to be very small (see Table IV). The amount of change in the iron was determined analytically, and is shown in Table LV. For the present approximate determinations of the volume change, these small changes may be neglected, for they affect but a small percentage of the rock. ~ TABLE IV. Changes in State of Oxidation of the Iron in Diabase. © FeO; FeO Ferrous| Ferric | Total H.0 4 % 4 Tron % | Iron Z | Iron % Original Rock------ 1:51 8°74 6°79 1:06 7-85 0°61 Spesioagaetet aoe | ow | em | axe | oo | Recrystallized block: after volume deter-| minations under, 2:90 | 7:59 5°90 2°08 tin. (From glass of | al Jume 1 912) Nee | 7:93) |e The volumes of the rock and its glass at room temperature are given in Table V, in comparison with similar constants determined by Barus for his sample of diabase (see p. 5). The experimental data are plotted in fig. 9. For the sake of clearness the curves of rising and of falling temperature are plotted separately. No satisfactory measurements were obtained on the second block below 900°, as internal cracking in the glass block caused its volume to appear considerably too large until these cracks were opened by the shrinkage during erys- tallization. It will be seen from the curves that the glass expands slightly up toabout 900°, when a sudden shrinkage takes place. This is due to the erystallization of the glass, which is of and Rock Densities at High Temperatures. 27 TABLE V. Kilogram-volumes of diabase and diabase-glass. Date Temp. We ‘Dens. Kilo-vol. specimen Diabase 10 June 1912 22°7 90:017 2°982 835-4 11 Mar. 1913 19°45 104°553 2968 336°9 lle io us 21°3 40°516 2974 336°3 Mean 2975 =336°2 Diabase-glass 10 June, 1912 2°1807 2°778 360°0 aya 66241 2°752 368°4 128 :05 2-767 =361-4 80°34 2-770 361:0 85-802 2-747 364:°0 22 June, 1912 94. 3 11 Sept., 1912 W V9 WW DP HE WS VO WD DW NBDW Hn Mean : 2°7638 362-0 Diabase (Barus) 1891 20 22°895 3016 331°6 21 45°365 3°018 331°3 21 54°721 3:014 93318 21 69-494 3024 =330°7 Mean 3°018 381°3 Diabase-glass (Barus) 1891 21 60°933 2°702 370-1 19 33°766 2°745 364°3 19 29°978 2-705 = 869-7 Mean 2-717 368-0 course nothing more than an undercooled liquid which has failed to crystallize during rapid cooling, but finds a second opportunity to do so with the increased mobility accompanying slow reheating. The heat given off during crystallization, spreading thr ough the graphite float-crucible into the surround- ing molten metal, was sufficient to cause a sudden rise of 15° in the temperature indicated by the thermoelement. The volume to which the block shrinks is not its true crys- talline volume. Not only is there insufficient time for close adjustment of the crystals to one another, but the same forces of differential expansion that operate to shatter a complex solid rock operate here to leave the crystalline aggregate in a rather porous condition, with innumerable cracks of appreci- able aggregate volume between the fine crystals. The con- traction of the glass block on crystallization makes it certain, however, that thie erystalline rock has a smaller volume anew the glass at this temperature. If there had been no apparent volume change, or if the volume had appeared to increase, the question of the true relation between glass and crystal would have been left in doubt. neral t tion of M Mena Day, ete.—Deter Fie. 9. oos! oor! OoOE! _00z! ool fefereyl 006 008 (efey2 009 oos oor OOE (eler4 (efey! fe) SYNLVYsAdMN3at O9€ SYyYNnLVeadWN4al ONINIVS ose 00v Syl sO AWNIOA OOE ee ASVaVIC sO SINIMIOA DeEil@zlelS VIVG IWLNAWIYSdxS oze Z Q O o— A® (e asvavida xe aavsiivd Bef P9E % : Sssvi9 asvavid 3avsiivd SYNLVYAdWSAL ONISIY ose OOv and Rock Densities at High Temperatures. 29 Above the rather accidental temperature at which the glass erystallizes, the crystal aggregate again begins to melt. The melting does not occur at a single temperature, but is spread over a range of temperature, in “accord with the fact that this is a complex system of eight or more components. The remelting begins at about 1150° and is practically complete somewhat below 1300°. The rock flows readily at a tempera- ture of 1225°, at which temperature all is fused excepting about a third of the feldspar. < 0 eal * Trans. Roy. Soc. Canada, iv, 19-44, 1910. +F. D. Adams, private communication, 1913. + Adams and Coker, Carnegie Institution, Pub. No. 46, p. 57. and Rock Densities at High Temperatures. 33 good agreement with this is the coefficient of 25°5 x 107° for the contraction of the expanded rock from 1000° to 0°. Comparing the volume of the crystalline rock with that of the glass, and taking the volume of the glass as the basis of comparison, we find that the difference of 7-1 per cent at 0° has inereased to a value between 9:1 per cent as a minimum and 10°9 per cent as a maximum, the true value being much nearer to 10-9 than 9:1. 9. Shale and Sandstone. One series of measurements was made on a specimen of the shale which underlies the Palisade diabase at Granton. The block weighed 68°35 grams after drying four hours at 106°. Measurements were made with tin as the immersion liquid. The volume curve showed the characteristic break at 575° due to the inversion of the quartz which was present in the rock (see curves for granite, figs. 7 and 8). At 675° considerable quantities of gas began to come off, interfering with the vol- ume measurements, and the gas evolution continued up to 1090°. The rock was held between 1090° and 1113° for 80 minutes: at the end of this time its volume had increased to 563, an expansion of 46 per cent. This was found to be due to the formation of a spongy mass through the entrapping of innumerable bubbles within the partly fused rock. Subse- quent microscopic examination of the original shale showed the presence of datolite in the interstitial spaces: this accounts in Jarge part for the behavior of the shale on heating. Datolite was also found in the arkose sandstone. Its presence renders these rocks unsuitable for volume measurements to a high tem- perature. Measurements were also attempted on natural microcline, and a satisfactory curve was obtained as far as 600°. At higher temperatures the gas evolution again proved our un- doing, causing shattering of the mineral in the solid state and yielding a bubbly glass at the melting point, the liquid being too viscous to allow the bubbles to escape. 10. Geological Applications. Quartz and granite.—The volume change in quartz between 0° and 500° is 2°8 per cent, and between 500° and 575° it is 2-4 per cent, of the volume at 0°. In other words, the increase during the last 75° before the transition point from low to high temperature quartz is nearly as great as the increase dur- ing the preceding 500°. A volume change as rapid as this must be of considerable significance in the crystallization of the siliceous rocks. As Fenner has pointed out in his recent Am. Jour. Sct.—FourtTH SERIES, VoL. XX XVII, No. 217.—January, 1914. v 34 Day, ete.—Determination of Mineral work on the silica minerals,* the fact that quartz in pegmatites is frequently found to have er ystallized in the neighborhood of 575° may be connected with the opening of fissures in the cool- ing granitic rock by the sudden volume change taking place in quartz which has already crystallized. But the significance of this point is not alone in the sudden- ness or the amount of the volume change in the mineral, but also in the small heat effect which accompanies it. The erys- tallization of a magma may be accompanied by a considerable volume change, yet the amount of latent heat given out dur- ing crystallization might delay its progress to such an extent that the volume change of the mass as a whole would proceed relatively slowly. But the amount of heat accompanying the quartz inversion is too small to seriously affect the rate of cooling of the rock, so that the volume contraction, at this temperature (575°), of a rock containing much quartz would be sudden and considerable. In his papers on “The mechanics of igneous intrusion,’’+ Daly suggests that an important cause for the cracking off of blocks of rock in contact with a liquid intrusive mass is to be found in the unequal expansion due to sudden heating. The experiments on quartz and granite in the preceding pages emphasize rather the much greater importance which must be ascribed to the sudden volume change accompanying the transi- tion of quartz at 575°, or to other similar transitions. Their effect in shattering the rock is much greater than the effect of ordinary thermal expansion even in a region of steep tempera- ture gradient. The result, however, would not be the break- ing off of large blocks which, according to Daly’s “stoping hypothesis,” must then sink in order to be assimilated at greater depths, but rather the formation of a shattered porous material which would be in the best possible condition for immediate assimilation én s7tu. Moreover, the capillary penetrating power of liquid diabase into a porous mass of this character is very considerable. Dia- base melted in a crucible of alundum (granular fused alumina) came through the sides of the crucible, and rose by capillarity in the walls like water in filter paper. We have even found numerous tiny drops of diopside on the outside walls of the inverted graphite float-crucible of the apparatus of fig. 1, the liquid silicate apparently having been forced through the fine pores of the graphite by the small pressure of the surrounding molten tin. Daly has assembled in his paper on ‘ the secondary origin *C. N. Fenner, this Journal, xxxvi, 379, 1913. +R. A. Daly, this Journal, xv, 269-298, 1903 ; xvi, 107-126, 1908 ; xxvi, 17-50, 1908. and Rock Densities at High Temperatures. 35 of certain granites ”* a large number of instances in which it is believed that a basic intrusive has assimilated acid rocks along its contacts. The shattering of a siliceous rock by the quartz inversion, and the penetrating power of liquid diabase and diopside, which we have noted above, lend additional plausi- bility to Daly’s view of the phenomena. Diabase and sandstone.—Returning now to the question of the “floated” sandstone and shale slabs described by Lewis (see p. 25) our experimental data again find application. A Fie. 11. Fig. 11. Diagrammatic sketch of the contact of the trap and underlying shale at the base of the cliff at Linwood, on the Palisades. (After Kimmel.) fragment of pure quartzite without pore spaces would have at 1150° a kilogram-volume of about 395, supposing it not to have inverted to tridymite or cristobalite. Diabase glass at 1150° would have a volume of 384. The quartzite would be, therefore, 2°9 per cent lighter than the fused diabase, and would float up in it, if free to move. The arkosic sandstone which forms the majority of the inclusions in the Palisade diabase consists chiefly of quartz and orthoclase. Extrapolation of our curve for orthoclase (see p. 33) would give it a volume at 1150° of about 404. The aver- age volume of the arkose (about one-third quartz and two- thirds feldspar), would therefore be 400 at 1150°, or 4:2 per cent lighter than the liquid diabase. Loose slabs of the arkose would therefore tend to float up into the diabase sheet. Such floating due to normal gravitative adjustment would seem to offer the simplest explanation of the frequent occur- rence of slabs of the underlying rock found in various positions in the Palisade sheet. A good example is seen in the sketch *R, A. Daly, this Journal, xx, 185-216, 1905. 36 Day, etc.—Determination of Mineral by Kiimmel (fig. 11), made at the base of the Palisades above Jersey City. Fig. 12 is from a photographs at the Belle Mountain quarry mentioned by Lewis.* A slab of variable thickness here stands nearly vertical in the diabase sheet, and appears conspicuously at the center of the picture. The Hie. 12. Fie. 12. Slab of arkosic sandstone in the diabase sheet at the Workhouse Quarry in Belle Mt., east bank of the Delaware River, above Moore, New Jersey. (Photograph by R. B. Sosman, 24 July, 1913.) * Loe. cit., p. 135 (6). f and Rock Densities at High Temperatures. BT quarrying operations of the past several years have been fol- lowing this slab back through the diabase. Lewis" explanauion of these occurrences is found in the fol- lowing quotation :* ‘ The first step in such a process is seen in Plate XXII, fig. 1, where a thin sheet of the diabase has followed a beddine- plane about a foot below the base of the main sill. If any portion of the intervening sedimentary bed had broken or parted along a joint-plane and the edge had tilted up somewhat against the How of the intruding magma, it would have been raised by the current to a more steeply inclined or even vertical position.” but this can hardly be a complete explanation, as no force is suggested which could start the slab tilting up against the flow. The intruding dia- base can not be compar ed to a stream flowing freely over a flat, irregular bed, and lifting up blocks by its own momen- tum. On the contrary, we must assume that it flowed rela- tively slowly, under considerable pressure from overlying rock strata. Under these circumstances a thin sheet of liquid diabase, branching off from the main mass and following a bedding plane, would indeed exert a lifting pressure on the under surface of the slab of sandstone; but this pressure would exceed that on the upper surface by only a small amount, namely, an amount depending on the difference in density between the sandstone and the diabase. In other words, the sandstone would be lifted by the same force that lifts a piece of wood which is immersed in water. It is possible that the moving current of diabase might carry the slab along after flotation had lifted it from the floor. But this assumption is not necessary in order to explain the fre- quent vertical position of the slabs, since this is the position which any flat object tends to assume when floating up through a liquid. Temperature of intrusion of the diabase.-—A comparison of the fusing temperatures of the arkose and the diabase yields the interesting fact that the arkose fuses at a lower tempera- ture than the diabase. Nevertheless, in the inclusions actually found there is no sign of fusion. This discrepancy led us to make a more thorough investigation, with the assistance of Dr. Merwin, of the fusion temperatures of the two rocks. This was done by means of quenching tests and microscopic exami- nation, of which the results have already been published else- where.t Briefly summarized, the conclusions were as follows : (1) The “ basaltic” facies of the Palisade diabase begins to fuse at about 1150°, and enough of it is fused at 1225° to permit the rock to flow readily. (2) The arkose now found in * Loc, cit, p, 1384, 1907. +R. B. Sosman and H. E. Merwin, J. Wash. Acad. Sci., iii, 3889-395, 1918. 38 Day, ete.—Determination of Mineral the diabase in the form of inclusions is more than half fused at 1150°, but shows no fusion at 1025°. (8) The inclusions as actually found show no indication of fusion or flow. As ind?- cated by their present properties under atmospheric pressure, there is therefore a gap of at least 100° between the maximum temperature to which the arkose inclusions could have been subjected, and the minimum temperature at which the diabase will flow. The most probable agent which possesses the power to bridge this gap is the water content of the diabase magma, of which the surrounding rocks show abundant evidence. At the time of the intrusion water was undoubtedly present and served to lower considerably the fusion temperature of the diabase. It is evident, then, that we are by no means at the end of the geological problem of the relative volumes of crystalline and molten rocks when we have determined their volume under the conditions of pressure and composition in which we now find them at the surface of the earth. The effect of the volatile components, now present only in traces or perhaps entirely absent from the rock, is still to be taken into account. The effect of dissolved water, for instance, on the volume of liquid diabase as given in the preceding pages, is unknown, although an amount sufficient to produce large changes in melting temperature and viscosity will probably be found to produce relatively small changes in volume. Data such as we have presented are absolutely necessary, however, as a basis upon which to build further knowledge of the volatile compo- nents, when methods are worked out for including them. SuMMary. After reviewing existing and rather conflicting data on the change in volume on fusion of various rocks, we have described apparatus by which the volume of metals and of solid and liquid silicates can be determined from 250° to 1600°. . The sensitiveness of the method is from 0:03 to 0-2 per cent. The accuracy varies with the metal used, the temperature, and the size of the sample, but is usually between 0:2 and 0°5 per cent. The fundamental constant is the expansion coefticient of arti- ficial graphite, which is given by the formula 10° = 0:55 + 0:0016¢. In Table I (p. 14) are given the volumes of tin, lead, and the tin-lead eutectic in the liquid state at various tem- peratures. Volume measurements on quartz up to 1600° show a dilatation which increases rapidly as the inversion point of 575° is approached. Above this temperature quartz contracts slightly. At about 1300° there begins a second and (under these and Rock Densities at High Temperatures. 39 conditions) irreversible dilatation due to the formation of eristobalite. Granite has a eurve of similar form. Above 575° granite undergoes a permanent dilatation caused by the shattering effect due to the different expansion coefticients of its minerals and to the escape of gases. The volume of diabase glass was determined up to 1250° and compared with Barus’ well-known values. A simple explanation of some of the conflicting features of his observations has been found. Diabase expands considerably in liquefying. The difference in volume between crystalline and glassy (liquid) diabase at 1150°, at which temperature the crystalline rock begins to fuse, isabout 10°9 per cent of the volume of the glass, or 12 per cent of the volume of the crystalline rock at that temperature. At 0° this difference is 7°7 per cent of the volume of the rock. A consideration of the comparative specific volumes of quartzite or feldspathic sandstone on the one hand, and liquid diabase at the same temperature on the other hand, shows that the diabase is denser, and slabs or fragments of the sandstones would therefore tend to float init. This seems the best explana- tion of the peculiar sandstone inclusions in the Palisade diabase described by Lewis. Volatile components, now absent from the diabase as we find it, seem to have greatly modified its temperature of intrusion, and make it impossible at present to solve completely the problem of relative volumes under the original conditions of intrusion. The foregoing measurements, however, furnish a necessary basis, on which the variation due to volatile components ean be established. Natural rocks are unsatisfactory for measurements of this kind for three reasons: (1) from the chemical standpoint they are very complex systems, between whose properties it is difficult to find a simple relation: (2) the unequal expansion of the various minerals fractures the rock so that it is impossible to determine its true volume change over any considerable range of temperature ; (3) the natural rocks give off considerable quantities of gas. The best plan for future work, therefore, would seem to be to determine the volumes at high temperatures of simple silicates, prepared in the laboratory and free from gases. Their other properties can then be controlled, and it will become possible to find definite relationships between composition and dilatation. Such measurements have been begun, and will be reported on at a later date. Geophysical Laboratory, October, 1913. 40 FE. R. Van Horn—Pisanite and Arsenopyrite. Art. I].—Wotes on a new occurrence. of Pisanite and Arsenopyrite, and some large Staurolite Crystals from the Ducktown District, Tennessee ; by Frank Rogpertson V An Horn. Introduction. Durine the month of June, 1912, the writer spent about three weeks with a class of Case mining students in the vicin- ity of Ducktown, Tennessee. While visiting an abandoned open cut of the Ducktown Copper, Sulphur, and Iron Com- pany at Isabella, considerable amounts of a sulphate deposit of secondary origin were observed. These minerals were col- ored green in some places, but blue was the prevailing color for most of the formation. It was at first thought that the green mineral was melanterite (FeSO,+7H,O), and that the blue substance was chaleanthite (Cuso, +5H,O). However, the green deposit was rubbed with a knife blade which was immediately colored by the precipitation of copper, which proved that the mineral could not be melanterite. Subsequent tests and analyses showed that neither melanterite or chalean- thite were present, but that both the green and blue minerals Were pisanite with varying proportions of iron and copper. Since pisanite has never been reported from the region, as far as the author is aware, the new occurrence has been thought worthy of record. Pisanite. As has been previously stated, the mineral was found in two varieties, green and blue, on the walls of an abandoned open cut at Isabella, Tenn. This locality is situated about one and one-half miles east-southeast of Ducktown. A view of the open cut workings is shown in fig. 1, in which the light-col- ored deposit along the right hand wall consists of pisanite. The blue variety is found in greater abundance than the green mineral. Both substances are transparent and are found as crusts, botryoidal coatings, and stalactites on the walls of the cut, which ‘consist chiefly of pyrite with some pyrrbotite and chalcopyrite. Fig. 2 is a nearer view of the occurrence shown in fig, 1. In fig. 2 some small stalactites are visible, and the botr yoidal appearance can also be noticed. Limonite has pre- cipitated in large amounts from the water to be seen in the lower right hand corner of the picture, and the water is quite highly charged with copper. In some places the deposits are more or less hollow, and the cavities show distinct crystals. Many specimens were collected, and no change was noticed in their appearance until after six months when the surface of practically all pieces became white and opaque. I. Rk. Van Horn—Pisanite and Arsenopyrite. 41 Fie. 1. Fie. 1. Open cut workings at Isabella, Tennessee, showing light-colored deposit of Pisanite above water in the lower right hand corner, Fic. 2. Nearer view of Pisanite deposits showing stalactitie and botry- oidal formations. 49 I. R. Van Horn—Pisanite and Arsenopyrite. Chemical Composition.—Some five months after the min- eral was collected, transparent specimens of both blue and green varieties were submitted for analysis to Dr. W. R. Veazey, Assistant Professor of Chemistry at Case School of Applied Science, and the writer is under great obligations to Professor Veazey "for the analyses. Although the specimens had been in a dry place so long after being “collected, nevertheless some moisture was noticeable when the mineral was broken. It was, however, not deemed advisable to dry the specimens by artificial means on account of the possible oxidation of the iron, and also for fear of driving off some of the water of erys- tallization which has been stated by Schaller* probably to take place between 100°-110° C, Binue Pisanite.—Below is given under (1) the analysis of the blue pisanite from Isabella, Tenn. Under (2) and (3) are to be found analyses of pisanite by Dr. Schallert from the Alma mine, Leona Heights, Calif. Under (4) is to be seen the original analysis of the pisanite from Turkey by Pisani. (1) (2) (3) (4) SO, ee OOS 29°28 29°25 29°90 KheOneesa les 16°47 5°46 10°98 CuOy so 28250 9°22 17°95 15°56 Mic Ones Os4i7 Nance 2°82 Pere H.O 2.) 46:47 45°74 45-21 43°56 100°49 100°61 100°69 100°00 An inspection of the various analyses shows a very great similarity between analyses 1 and 2. The water is evidently too high in 1, and was probably due to a small amount of moisture in addition to the water of crystallization. The water in 4, on the other hand, is certainly too low, and was probably determined by Pisani only by difference as ‘the anal- ysis totals 100 per cent, and this kind of analysis practically always totals higher than 100. One interesting feature of the Isabella mineral is the small percentage of magnesia. This indicates the presence of a small amount of the very rare cupromagnesite$ molecule, (Cu,Mg)SO,+7H,O, which is iso- morphous with melanterite. Analyses of mine waters| from the region show the presence * Minerals from Leona Heights, Alameda Co., California, by W. T. Schaller, Bull. Univ. of California, vol. lii, pp. 204-206, 1903. + Op. cit. , pp. 206-207 {Comptes Rendus, xlviii, 807, 1859. S$ System of Mineralogy, Dana, p. 944. || Preliminary Report on the Mineral Deposits of Ducktown, enines08, by W.H. Emmonsand F. B, Laney, Bull. 470, U. S. Geol. Survey, p. 172 , 1910. F. R. Van Horn—Pisanite and Arsenopyrite. 48 of magnesium in small quantities. In analysis 8 by Schaller much more magnesia is found than in the Isabella mineral. In order to calculate the formula from the analysis, since the magnesia replaces ferrous iron, it has been recalculated as that substance with the following results : Recaleulated Combining Molecular analysis weights ratio SOR Pees ee a. 27°87 °348 3°25 (FeO,MgO) aes soe 18°02 "251 2°34 CuO: ere. 8°50 107 1:00 Ja Os eh a neers ce 46°47 2°585 24°15 100°86 The molecular ratios obtained from analysis 2 by Schaller are 2FeO.Cu0.38O,+21H0, and the ratios of the blue pisanite - from Isabella conform very closely to this formula, with the exception of the water, which is 3 molecules too high and must be ascribed to moisture in the original substance. The theo- retical percentage required by the formula 2FeO.CuO.3SO0,+ 21H,O would be as follows : SOM ee ine 28°53 MeO elem rele? Sane 17:08 CuO ear ee cae Men 9°45 EIS Op ere eee eee ee 44°94 100°00 Green Pisanite.—The results of the analysis of the green mineral were obtained by Dr. Veazey as follows: IS Oise beech iat pa D879 GO) cep te: care pies ee be 21°45 CuO ei eee eee 3°83 Mic, Oper Oe eae 0°39 EEO) 285 ee ae ee ee 45°58 99:97 When this analysis is compared with that of the blue vari- ety, the most noticeable differences are the increase of iron and decrease of copper which was to be expected from the green color. In other words, the green mineral is a pisanite nearer melanterite, whereas the blue substance is a pisanite approaching the rare mineral boothite (CuSO,+7H,O), which 44 F. R. Van Horn—Pisanite and Arsenopyrite. Schaller* has shown to be isomorphous with melanterite form- ing pisanite. Of ae interest is the fact that the forma- tion of both green and blue pisanite have caused a secondary enrichment of the copper by oxidation, since the sulphide ores of the district contain an amount which aver ages slightly under two per cent. The small percentage of magnesia is seen to be fairly constant in both varieties. When we calculate this MgO as FeO, the analysis becomes as follows: Recalculated Combining Molecular analysis weights ratio S Opa ee ee oe 28°72 359 7°48 (FeO, Nig ©) eee 22°14 308 6°42 CoORe renee 3°83 ‘048 1°00 1 RO ie ec ee 45°58 2°530 52°71 — 100°27 About the nearest logical formula for the green pisanite would seem to be seven times that of the melanterite formula, or 6FeO.CuO.78O0,+49H,O. In this analysis it would seem that the water is even larger than in the blue mineral although the same is nearly a per cent lower. In all published analyses of the pisanite the size of the molecule seems to vary, although all occur in multiples which adhere closely to the general formula (RSO,+7H,O),. It is very clear from the analyses of the Isabella varieties as well as those from other localities that iron and copper have no fixed relation to each other, but that they may replace each other in any proportion. Crystallography.—Although the analyses of the mineral showed quite conclusively that the substance was pisanite, and not a mechanical mixture of melanterite and chaleanthite, as might be claimed, nevertheless, it was thought advisable to see if any measurable crystals could be obtained. The writer, after breaking up several specimens of the blue pisanite, was able to find some cavities lined with well-detined crystals. Because of his previous experience with pisanite from Cali- fornia, Dr. W. T. Schallert of the U. 8S. Geol. Survey, Washington, D. C., was asked to see if he could obtain any crystallographic results on the material selected. I wish to express my thanks to Dr. Schaller for his data which follow, and also especially for his drawings, figs. 8, 4, and 5. Dr. Schaller has observed seven forms which are fairly common for pisanite with the exception of o (121), which was first dis- covered by him on the mineral from the Alma mine, Leona Heights, California. Dr. Schaller’s results follow: * Op. cit., p. 207. + Op. cit., pp. 199-207. FR. Van Horn—Pisanite and Arsenopyrite. 45 Fie. 3. Fia. 4. Fia. 5. @ GE oe Ww D b-| 772 77t 77 772 77U 770 Fras. 3, 4, and 5, Pisanite crystals from Isabella, Tennessee, drawn by Dr. W. T. Schaller. Forms determined : 6 (010), very narrow, figs. 4 and 5. e (001), large figs. 3, 4,5. Cleavage. m (110), large, dominant form, figs. 3, 4, 5. o (011), small, figs. 4 and 5. w (103), minute, fig. 4. t (101), minute but larger than w figs. 3, 4, and 5. o (121), minute, fig. 5. (Average) Measured and Calculated angles: Measured Calculated “ p v) p b (010) OO i=3290 00: 0°00’ 90° 00’ e (001) 89 23 15 54 90 00 oyelat m (110) 4] 34 90 00 41 36 90 00 0 (011) 9 35 57 19 10 07 57 04 w (103) 88 27 36 29 90 00 35 48 t (101) 89 21 47 06 90 OO 47 09 (on (121) Ite} 3333 72 06 19 32 72 46 The results of the crystallographic observations given above prove conclusively that the blue mineral is pisanite and not a mechanical mixture of melanterite and chalcanthite as some might claim. Arsenopyrite. One of my students, Mr. P. A. Fruehauf, found in the schis- tose walls of the ore body of the London mine, Ducktown district, two crystals which he brought to the writer for deter- 46 LF. R. Van Horn—Pisanite and Arsenopyrite. mination. One crystal measured 5x7™", and the other was 9x<14™". The general habit of both resembled that of danaite or glaucodot rather than that of most arsenopyrite. Blowpipe tests revealed the presence of iron, arsenic and sulphur with just a trace of cobalt, so that the mineral is undoubtedly arsenopyrite rather than danaite. The specific gravity is 6°05. Mr. Walter F. Hunt, Instructor in Mineralogy at the Uni- versity of Michigan, very kindly offered to measure and draw the crystals, and I wish to Fic. 6. thank him for his assistance. The following forms were determined m (110), g (011), n (O12). Angles measured. m 2m 68° 39' 2 Op 100° =e nmin’ Gl o7 The drawing is shown in fig. 6 where it is readily seen that the largest face is the Fic. 6. Arsenopyrite from London brachydome q (O11), and that mine, Ducktown, Tennessee. Drawn the brachydome (012) is by Mr. W. F. Hunt. very small. The habit is rather unusual for arseno- pyrite, and as far as the writer can learn, that mineral has also never been previously reported from the Ducktown region. Staurolite. This mineral has long been known from the general district of eastern Tennessee, western North Carolina, and northern Georgia. Emmons and Laney* have carefully traced certain staurolite layers in the schists of the Ducktown region on account of their probable relation to former sedimentary rocks which had presumably been replaced, with the consequent formation of the iron and copper sulphide deposits. It is the present purpose of the writer only to call attention to a fact of mineralogical interest which has apparently been overlooked, namely the extremely large size of some of the staurolite crys- tals. The largest were found in Mill Creek Valley not far from the Copperhill smelter of the Tennessee Copper Com- pany. The locality was visited three times, each with a dif- ferent section of students, and the writer is not sure but that the students may have collected larger crystals than are shown in fig. 7 which are approximately half the size of the natural * Op. cit., p. 158. FR. Van Horn—Pisanite and Arsenopyrite. 47 erystals. The only faces observed were m (110) and 4.(010). No terminal planes were developed and all faces were ea The single crystal on the left in fig. 7 measures about 5 wide, IKOes long, and 4° thick. The other two are twins after 2 (232). The central crystal is 5™ wide, 8™ long, and 6™ thick. The largest specimen is the twin on the right which is tabular and measures 8™ wide, 9™ long, and 34 thick. The writer has Fic. 7 Fic. 7. Large staurolite crystals (one-half natural size) from Copper- hill, Tennessee. The one on the left is a single crystal, and the other two are twins after z (232). never seen any staurolites in this country or Europe which approach the size of these crystals. All crystals were either single individuals, or twinned after z(232). and none twinned after «(032) were found. It may possibly be a matter of interest to know that out of several tiundred small staurolite crystals found near Mineral Bluff, Fannin Co., Georgia, only 11 were found twinned after (032). The remainder were about equally divided between single crystals and those twinned after 2(232). The large staurolites of the Ducktown district are all altered to an irregular aggregate of chlorite, muscovite, and garnet, with a little ‘magnetite. In some places the erys- tals were evidently formed after the schistose structure of the surrounding rocks was developed. Geological-Mineralogical Laboratory, Case School of Applied Science, August, 1913. 48 Perret—fepresentation of Volcanic Phenomena. Arr. II].—The Diagrammatic Representation of Volcanic Phenomena ; by Frank A. Perret. With Plates I to lV. As an aid to comprehension in the presentation of a subject the visual sense is now appealed to in a constantly increasing degree and a description without illustrations, a report without a drawing, or a lecture without projections has become alimost an exception. Exquisitely sensitive to infinitesimal variations of form and of angle, the eye is capable of conveying to the mind, not only the reproduction of a scene, but even the summed- -up impression of a series of measurements, or the net results of a laborious analysis, as presented in a diagram or curve. And, in such a case, it is precisely this property of aS so to speak,—of eliminating distracting intervals and showing, almost at a glance, those collected values which must otherwise have been considered one by one—which constitutes the principal advantage of this method, and gives to it the quality of being at once analytic and synthetic. We may see, therefore, the gradual cooling down of a molten mass plotted out in the form of a curve, of which beginning, pro- gression and end are visible and comparable; the details of plant growth analyzed and totalized in a diagram ; the composi- tion of rocks exhibited for comparison in a series of geometri- cal designs. But it would be superfluous to insist further upon the value of the principle—it is a theorem not needing demonstration. Can a diagrammatic method be devised for presenting the characteristics of voleanic action ¢ It may be said, first of all, that the process of photographic illustration has alt eady rendered immense service to this branch of science whose phenomena are unfamiliar to the mass of mankind and are of a grandeur and beauty which baffle description. The earlier sketches were often but caricatures, in many instances reducing the average volcanic cone to the apparent dimensions—and almost to the angles—of a chureh steeple, and it is by photography alone that the reality can be adequately portray ed and thus conserved as a precious docu- ment for all time. But even this process is circumscribed and limited in its practical application, not merely by the frequent elimination of the sine yua non by clouds of eru pted material but also by reason of the fact that a large proportion of voleanie mani- festations are invisible and, finally, because the photographie record is fragmentary and lacking i in the above-mentioned ele- ment of totality. It is evident, therefore, that a further means Perret— Representation of Volcanic Phenomena. 49 of expression is needed—a method of graphically setting forth all the phenomena in their just proportions and thus to ‘reveal, in a composite figure, the true character of the volcanic activ. ity. The complete presentation of any volcanic event will then be of a threefold nature, consisting of the written deserip- tion, the photographic illustrations and the diagrammatic inter pretation of the eruptive characteristics. Long convinced of this need, the writer has devoted himself to the “problem of devising a method which should be simple and thoroughly practical and, at the same time, so comprehen- sive as to fully attain the scope of the undertaking, and this is believed to have been finally achieved in the form herewith presented. The Diagram Chart. This is furnished in the form of a simple postcard which is easily handled, readily carried in the pocket and mailable at lowest rates. One side is spaced for the address and for notes, while the other is devoted to the diagram proper (see Plates LIV, giving twelve examples, each reduced one-third) which is constructed as follows: Eight radiants diverge from a center and terminate in a cor- respondingly divided circle which, with its extensions, forms the key of the diagram. The circle divisions are marked to represent the various phenomena, which are here expressed in the most general terms—explosive, effusive, seismic, etc.—and are then developed and extended into subdivisions indicated by letters and numbers. Each radiant is a line of ten small, blank circles destined to be blackened or filled in progressively from the center out- ward, according to the value assigned to the phenomena which it represents. Tie lines are then to be drawn between the radiants intercepting these at the marked extremities, thus enclosing a space and constituting a figure, to the form and angles of which the eye is peculiarly sensitive. This construction of the radiants has been adopted, not only as forming a convenient decimal scale of values, but because it is easier to blacken a small circle than to draw a straight line, and also for the reason that a more salient appearance is thus given to the radiant proper—which indicates the real value assigned to the characteristic—than to the rest of the figure as formed by the intercepts. The subdivision is to be indicated in the diagram by its letter and number, and in the case of there being more than one—and of lesser value—its circle may be left blank, thus contrasting with the others and marking the value of that par- ticular detail, all of which is clearly shown in the examples accompanying this article. Am. Jour. Sci1.—Fourtu Serigs, VoL. XXXVII, No. 217.—January, 1914. 4 50 Perret—Representation of Volcanie Phenomena. Before taking up the consideration of the various phenom- ena it may be well to realize the scope and limitations of the method. The volcanic forces and manifestations not being generally measurable in exact terms, it is obvious that the dia- gram will not be of mathematical precision. It is not intended to indicate, for example, the explosive pressure in any given outbreak nor to compare this rigorously with that of another volcano, but the object is rather to present the various phe- nomena of an eruptive phase in their relative proportions, and thus to reveal its character. The maximum values once adjudi- cated, it is easy to estimate the others in relation thereto, and this is far simpler than might, at first, be supposed. At Etna, for instance, in 1910, the great stream of lava, issuing at tor- rent velocity and extending nearly ten kilometers, easily formed the distinguishing characteristic and, constituting as it did, even for this volcano, a very notable outflow, it is indi- cated on the diagram with the full ten points. At Teneriffe, instead, although the lava-flow constituted the principal external manifestation, its magnitude was mediocre, not merely per se but also more in comparison with former outflows in the same general locality, and the value assigned to it on the dia- gram is seven points. It will be evident, also, that the difference of a point either way will not, as a rule, materially affect the integrity of the diagram. At Stromboli, 1912, the electrical phenomena are indicated by a radiant of six points. A second observer might have argued against a higher value than five on the ground of the brief duration of the phenomena, while a third might have assigned seven points in order to give prominence to a manifestation extremely rare at this voleano—all would have been agreed, however, as to a characteristic of very notable, but not paramount, importance and thus, in any case, the radi- ant would have a value conveying this impression. In the case of certain outbreaks of brief duration—and espe- cially those in which the character of the action does not greatly change during the course of the eruption—a single dia- gram may be made to represent the entire event. This has here been done with Teneriffe 1909, Stromboli 1907 and 1912, and, somewhat reluctantly, with Etna 1910. But it is evi- dently preferable that the diagram should show the eruptive conditions at a given time, different phases being represented on separate charts and forming a series of documents reveal- ing the progression of the eruption, and it is here that the beauty and value of this graphic method is most clearly seen. At the great Vesuvian outbreak of 1906 the character of the eruptive phenomena swung through a succession of inter- merging but widely differing phases, the more salient of which Perret—fepresentation of Volcanic Phenomena. 51 are here represented in six diagrams showing the development and progressive variations of that important event. It may be of interest to note that any voleanologist could, from these diagrams alone, mentally construct and vividly and correctly describe the eruption. This is evidently working in the wrong direction and is not to be recommended, but the fact is mentioned as showing the effectiveness of the method. Considering now the characteristics constituting the key of the diagram, the reader will at once perceive their emplace- ment to be that ee is most in accord with the nature of the phenomena. Thus, the upwardly diverging radiants represent the explosive phases, the vertical being devoted to volatile and aérial products and phenomena, with the fragmental ejecta adjacent on either side. The lower perpendicular indicates degradation and deformation of the volcanic edifice by the general phenomena of descent. The effusive line is horizontal, on the right, while the left-hand corresponding radiant repre- sents electrical effects, and is adjacent to the ash and detritus which are essential to their development. Seismic phenomena are placed at the lower right, while the so-called solfatariec, i. e. fumarolic, phase of activity is represented at the lower left. Speaking in the most general terms, the right-hand divisions present the more direct products and phenomena, while those at the left are of an accompanying or resultant nature. These will be seen to correspond in some degree—especially as regards the ejectamenta—with certain terms which have been proposed, and in part adopted, for distinguishing the principal classes of phenomena and products, but which, however convenient for general descriptions, either fail in consistency under the test of the diagrammatic method or are inapplicable because of the diffi- culty of observational classification. They invariably constitute an imperfect and limiting terminology, and are here best avoided. We shall now examine in detail the various subdivisions of the analytical key, beginning with the “‘ Explosive” phase. The gases issuing from active lava on its first coming to the surface —and even under other conditions—will frequently burn with the production of true flame, which is sometimes lost to view in the glare of the lava. At Etna in 1910 the flames were ten meters in length, and at Kilauea in 1911 each lava-fountain jet emitted a flash of burning gas. Apparently a product of this combustion, but possibly also formed without it, the transparent bluish vapor constitutes one of the commonest of volcanic exhalations. It is the bane of the photographer in inter-crateric work, where its transparency deceives the eye and its actinic color veils the photographic image. The white “panache” is a characteristic feature of 52 Perret—lepresentation of Volcanic Phenomena. modern activity, being divided, when the emission is strongly intermittent, into puffs, rings, vortices, ete. Stronger explosions form waves of aérial concussion, often destructive, as at Stromboli in 1907, when most of the win- dows on the island were shattered to fragments. At Krakatoa the aérial coneussion was of enormous power. When these waves result from explosions, which are very sharp and sudden as well as powerful, they may be visible near their source as “flashing ares,” springing upward and outward from the crater.* Subterranean detonation is frequently so powerful as to be heard at enormous distances. The magnitude of these deep- seated manifestations is sometimes greater than that of the surface eruption, as was the case at Teneriffe in 1909. ‘i Catastrophic explosion is that of extraordinary violence, gen- erally involving considerable deformation of the voleano itself, and extending its destructive effects to the surrounding regions (Vesuvius 1906, Pelée 1902, Krakatoa 1883, etc.). The “ Fraomental ie division, on the right, represents those products issuing most directly from the active lava, commenc- ing with that of the most liquid consistency. Lava Fountains, with the constant product of their scattering jets—the filamentary lava, or “ Pele’s hair.” Glass foam—the so-called thread-lace scori, or ‘‘limu,” and the interesting little lava drops—‘“ Pele’s tears.” + Section b represents ejecta from a somewhat more viscous quality of lava, the vitreous sand often formed from the surface material in oreat quantities by the expanding gases, the well-known lapilli and scorize and the figured projectiles and bombs. In section ¢ we have ejected pumice and that marvelous result of rapidly exploding viscous lava which is the nwée ardente. A digression must here be made in order to point out that the diagram is not, and cannot be, concerned with the origin or cause of the phenomena, nor with the chemical nature of the products. For example, explosion may be due to juvenile gases or to those of purely “phreatic” origin, but this is a matter for ulterior inv estigation and description —the diagram records the physical phenomenon. Similarly, the water of a mud-flow may be volcanic or wholly meteoric, or it may result from the melting of snow by hot volcanic owases—this i is matter for the observer’s notes. And—more important still—ejected ash may be due to the trituration of “old material,” or it may, on the contrary, consist of finely divided magma from the lower parts of the conduit and, therefore, be a direct ejection of co- * This Journal, xxxiv, 329, October, 1912. + Volcanological Review, No. 1. ¢ This Journal, xxxv, 611, June, 1913. Perret—Representation of Volcanic Phenomena. 53 eval lava.* But it is often exceedingly difficult to distinguish between the two varieties and, as regards the diagram proper, an emission of ash is to be represented as such, and ulterior information given in writing. As for the nuée ardente, its ¢ complex constitution of explod- ing lava, avalanche form, and cauliflower ash-cloud development, makes it somewhat difficult of emplacement, but this has here been done on the basis of its fundamental nature, viz., that of a direct, explosive-fragmental product. This rare phenomenon should not be confounded with the hot ash avalanches men- tioned below. A study of the chart will also show that, in some cases, a certain subdivision will not need to be indicated on the dia- gram, as it will be understood from the other phenomena. For example, in the case of strong ejection of bombs and scoriz directly from the active lava—especially at a volcano of the liquid lava type—it would be superfluous to indicate the “ crater-glow.” Another point requiring explanation, and which some may be disposed to criticise, is the omission of a number of forma- tions resulting from the various activities. Avalanches and mud-flows will form “bréches” and ‘“‘conglomerates,”’ cones may be erected by fragmental ejecta, ete., but all these involve the time element and are not, strictly speaking, contained within the category of the phenomena and immediate products of eruption, and should not, therefore, be made to encumber the diagram. We now return to the diagram key, at the left hand “ Frag- mental” division. The products here are non-luminous, or only partially so, according to the conditions of their ejection and the temperature of the gases, and consist of debris pro- jected in rapid jets or majestically unfolding volutes, or quickly darting spear-head formations. These, in case of a powerful emission, will form the ash “ino,” from which descend the showers of ash. When the explosive power is sufficient, angu- lar blocks and bowlders are added to the ejected debris. The presence of water, either volcanic or meteoric, may result. in the concentration of the air-borne ash into pea-like spheres— the pisolites,—or, if in great quantity, precipitate the ash cloud in veritable showers of mud. The “ Effusive ” division begins with inter-crateric flows of lava which, with the ejecta, constitute the usual process of crater filling. Lava lakes also are included in this division, *See Johnston Lavis: ‘‘ The Mechanism of Volcanic Action,” Geological Magazine, October, 1909, p. 441. See also paper by the present writer in this Journal, xxxiv, 405, Nov ember, 1912, written in ignorance of having been gaulcipated by Dr. Johnston Lavis in this matter of direct ash production. 54. = Perret—lepresentation of Volcanic Phenomena. which is not an inconsistency, as might be supposed, for their material is in constant circulation and therefore truly fluent. One of the most characteristic of volcanic phenomena is the “crater glow ”’—that searchlight projection of the Java’s lumi- nosity whose reflection by thes vapor clouds so closely resembles true flame. It is here included as constituting a precise indi- eation of the presence of active lava in acrater to which a near approach may be inconvenient. Terminal overflow—a com- mon occurrence at some volcanoes, especially during their period of growth,—and the more frequent out-flow from lateral vents, are “provided with subdivision into the two main types of lava—the smoothly undulating or ropy ‘ pahoehoe” and the bristling, disunited surface ‘“a-a.” Finally we have the phenomenon of viscous extrusion in the form of domes, spines, ete. (Pelée, Bogoslof, etc.) “Electric” effects are inevitable whenever a powerful gase- ous emission is accompanied by abundant detritus, the voleanic lightning appearing in linear flashes (Vesuvius 1906) or as lobes of fire (Stromboli 1907) or in stellar form (Pelée 1902.) hen the “pino,” as a whole, acquires high potential, the silent discharge from pointed objects—St. Elmo’s Fire—is fre- quently seen (Vesuvius 1906). In the “Seismic” division, volcanic earthquakes are sub- divided—for convenience in the case of representing an erup- tion with a single diagram—into premonitory, concomitant and consecutive. Bradyseismic phenomena involve local ele- vation or depression which may be transient (Vesuvius 1906) or persistent (Usu-san 1910). Lastly, voleanic “ Tsunami,” or sea waves, may be of great magnitude and destructiveness. (Krakatoa 1883). The so-called ‘Solfataric ” phase of volcanism is of greater importance than is generally supposed, especially as regards the primary fumaroles which develop in communication with the interior. These may continue in operation through long periods of external repose, during which time they serve as indicators of the internal conditions. Those of the second- ary type, forming on erupted lava or upon accumulations of elastic ejecta, have, necessarily, a more ephemeral existence although they may, for a time, exhibit the same temperatures and characteristics as those of the primary type. Flame-emit- ting fumeroles have been, perhaps, the least studied of all the phenomena relating to this interesting phase of volcanic action. ‘‘ Degradation ” and deformation of the voleanic edifice may occur through demolition, subsidence or collapse, either ter- minal, (Vesuvius 1906) lateral, (Vesuvius, Etna, Santa Maria, etc.) or subsurface (Bogoslof, Krakatoa, etc). By processes of translocation, the materials left or deposited upon the upper Perret— Representation of Voleanie Phenomena. 55 parts of the mountain are gravitationally transported to lower and more peripheral resting places, forming hot ash flows, if of freshly erupted substances and, in any case, dry avalanches, of which a variation is constituted by the equally imposing, but often less directly observable, inter-crateric slides. If these last are of considerable magnitude, the result may be an external ash cloud greatly resembling those due to explosion, and it is important that the two should not be confounded. With the advent of water the translocation is effected by the mud lavas which so frequently form a highly destructive adjunet toan eruption. The diagram key terminates with the item of true landslides, of volcanic causation, which have been frequent in Japan. In illustration of this diagrammatic method, twelve exam- ples are reproduced on Plates I to LV (two-thirds size) a glance at which will be more explanatory than many pages of descrip- tion. Attention is called to the comparison of the consecutive outbreaks of Stromboli and their contrast with the effusive eruption of Etna; to the still greater difference between the extreme types Kilauea and Krakatoa ; and to the phase varia- tions in the progress of the great Vesuvian event of 1906, for the full analysis of which even more diagrams might profit- _ ably be employed. It is, in fact, highly desirable that the charts be used freely and that, at every visit to a vantage point of observation, a diagram should be plotted. It will scarcely be necessary to state that the charts need only be marked in the field and the final diagram filled in at leisure and after mature consideration. It is recommended that the intercept lines be drawn before the insertion of the letters and numbers. These last may be done in red ink, if desired, which gives a lesser interference with the figures, or the circles may be done in red, which is very effective. The circles can easily be filled with pen and ink—an operation which the employment of a lens will facilitate and expedite. ‘Very elegant diagrams for lantern projection may be made by coloring, on the slide, those circles which would otherwise be blackened. The diagram, in this case, is to be formed on the card by the intercept lines and other markings but without blackening the circles. This is then photographed, the slide is made, and then the diagram circles are filled in with trans- parent colors. Suitable tints are the following, beginning at the upper radiant and proceeding clock-wise: violet, orange, red, green, purple or black, yellow, blue, brown. It is, perhaps, inevitable that each living investigator would have worked ont a somewhat different diagrammatic method, as personal experience is always vivid and necessarily affects the perspective of a world-wide voleanism. It is only by trial 56 Perret—Lepresentation of Volcanie Phenomena. ona large scale under actual working conditions that the value of a new process can be determined, and, to this end, the writer is prepared to furnish these charts ‘oratis, in several lan guages, to investigators, and will endeavor to keep a supply on hand at all observatories or stations at voleanic centers, or with the local authorities, for free distribution to visiting scientists. Believing that a special punishment is reserved for those who lightly or unworthily bring forward any new thing, the writer has labored to produce a “method which should include all essentials without undue complication and without the sacrifice of workability, bearing constantly in mind the not infrequent moments of his own observations and seeking the counsel of practical fellow workers, for whose kindly aid he is duly grateful. Posillipo, Naples, October, 1913. Am. Jour. Sci., Vol. XXXVII, Jan., 1914. Plate |. VOLATILE AND AERIAL a. 1. Burning gases — true flames. 2. Transparent bluish vapor. b ft. Opaque vapor cloud - the white « panache». *\2. Light explosive effects - puffs, rings, vortices. ft. Strong explosive effects — aerial concussion. ¢-\2, Subterranean detonation. 3 Catastrophic explosion. VOLCANO STROMBoLI. 1, Lava fountain jets. 2. Pele’s Hair. 3. Thread-lace «limu». 4, Pele's Tear-drops. &. The ash «pino». Showers of ash. f 1. Vitreous sand. 2. Lapilli & scoriae. 1. Large angular blocks & bowlders. *\ 3. Bombs & figured projectiles. c. 1. Pisolites. 2. Mud showers. c. 1. Pumice. 2. Nuées ardentes. a. Dark jets, volutes, spear-head projections. f 5 . a. (chaotic ejectamenta — cone debris). \ a.Inter-crater !ava-flows. 1.Lavalakes. 6.The « Crater-glow » — pseudo flame. c.Terminal overflow jf 1. pahoehoe, d.Qutflow, lateral vents! 2. a-a type. e. Viscous extrusion — Spines, domes, etc. fi. Linear flashes \2, Globe 3. Stellar J 1. Sharp crackle \ 2. Heavy thunder. c.Silent discharge —St Elno’s Fire. a.Volcanic lightning 4.Detonation a. Fumaroles — primary type (devcloped on volcanic channels). 6. Fumaroles — secondary type _ (on erupled lava or clastic ejecta) c. Flame-emitting fumaroles. | 1, Premontlory. 2. Concomitant. 3. Consecutive. 6. Bradyseisins. 1. Transient. 2. Persistent. c. « Tsunami» of volcanic causation. a. Volcanic earthquakes a. Demolition, subsidence, collapse. 1. Terminal 2. Lateral, 3. Subsurface f 1. Hot ash flows & dry avalanches. \ 2. Mnler-crateric slides. 3. Mud lavas. ¢ Landslides of volcanic causation. DIAGRAM CHART (Perret system) of Volcanic Activity OBSERVER PERRET. 6. Translocation (Physical) VOLCANO VOLATILE AND AERIAL DATE . a. 1. Burning gases — true flames. 2. Transparent bluish ‘vapor. TuLyY — Aue 5 § TR OM B OL! 55 b Ae Opaque vapor cloud - the white « panache». : sees seeee *\2. Light explosive effects - puffs, rings, vortices. /91 a2 . fr. Strong explosive effects — aerial concussion. \2. Subterranean detonation. 3. Catastrophic explosion. a. Dark jets, volutes, spear-head projections. (chaotic ejectamenta — cone debris). 6. The ash « 19 og, 1. Opaque vapor cloud ~ the white « panache ». ves u vi US : a(t: Light explosive effects - puffs, rings, vortices. A.M. 1. Strong explosive effects — aerial concussion. i" *\2. Subterranean detonation. 3 Catastrophic explosion. ne j 1. Lava fountain jets. 2. Pele’s Hair. \ 3. Thread-lace «limu». 4. Pele’s Tear-drops. (1. Vitreous sand, 2. Lapilli & scoriae. \ 3. Bombs & figured projectiles. c. 1. Pumice. 2. Nuées ardentes. a. Dark jets, volutes, spear-head projections. (chaotic ejectamenta — cone debris). 6. The ash «pino». Showers of ash. 1. Large angular blocks &'bowlders. c. 1. Pisolites. 2. Mud showers. 6. % a.Inter-craterlava-flows. 1.Lavalakes. 6. The «Crater-glow » — pseudo flame. d c.Terminal overflow f 1. pahoehoe. d.Outflow, lateral vents\ 2. a-a type. e. Viscous extrusion — Spines, domes, etc. 1.Linear flashes 2.Globe 3.Stellar 1. Sharp crackle. 2. Heavy thunder. c.Silent discharge — St. Elzno’s Fire. a.Volcanjalightningy 5 2 6.Detonation { 0000000 ifs Premonitory. a, Volcanic earthquakes ; 2. Concomitant. 3. Consecutive, b. Bradyseisms. 1. 7vansient. 2. Persistent. c. «Tsunami» of volcanic causation. a. Fumaroles — primary type (developed on volcanic channels). 4. Fumaroles — secondary type _ (on exupled lava or clastic ejecta) c. Flame-emitting fumaroles. a. Demolition, subsidence, collapse. 1. Terminal 2. Lateral. 3. Subsurface DIAGRAM CHART \ iB Trnnclocation: 1. Hot ash flows G dey avalanches. OBSERVER (Perret system) \\ 2. Inter-crateric slides. 3. Mud lavas. of Volcanic Activity c. Landslides of volcanic causation. Pe RRET (Paysical) VOLCANO ~— VOLATILE AND AERIAL DATE a. 1. Burning gases — true flames, 2. Transparent bluish vapor. A PR. is 2 19 06, aft Opaque vapor cloud — the white « panache». *\2. Light explosive effects - puffs, rings, vortices. 2.00 P.M. 1. Strong explosive effects — aerial concussion. *\2. Subterranean detonation. 3 Catastrophic explosion, a, Dark jets, volutes, spear-head projections. : a f 1. Lava fountain jets. 2. Pele’s Hair. (chaotic ejectamenta — cone debris). Z> “\ 3. Thread-lace «limu». 4. Pele’s Tear-drops. &. The ash « cme Whence Vi Wea xm wolts: The limiting sensitiveness is thus of the same order as in the cylindrical case for the same V,, as might have been antici- pated. It is probable, however, that the present conditions may be more nearly realized in practice, as the apparatus is essentially simpler. If V,= V,=0, the idiostatic case is identical with the absolute electrometer. 6. The closed field electrometer (disk and cylinder com- bined). If the two cases of the preceding paragraphs be com- bined, with the symbols of potential and distance the same as before, the energy of the system is, in electrostatic units, PY he Rivr aaa +r\ ) w=t 1(v.-V) (+t R)+(% Mgt Pa) and bence the forces become a qe Ee A is V,+V,\ Mg x=1(p+ 7p) (y- v,] (7- 5 ‘Ve ] AN ti Ve — 0. a ae 4 Mg AN ae v (/d*?+(R+r)/(R—r)) 1 after commutation, where Z is the length of the cylinder. The sensitiveness of this apparatus depends very largely on d, that is upon the disk, in any practical case; but it should be more nearly absolute. Experiments were not made. 7. Corrections. The most important consideration to be made here is the effect attributable to the want of symmetry in the orientation of the disk or movable charge, supposing the latter and the plates are quite parallel. Let & be the axial dis- tance of the uncharged disk from the plane midway between the condenser plates (guard ring). In such a case, if the elec- tric displacement is A/V, the quantity A= +A must be added on one side and subtracted on the other, so that d= D/2+ A,d'= D/2—A_ Thus the displacing force is, after reduction, pers ae Vea Bays eee DEYN EIDEED 2) (GS DYN ILD which may be further reduced. if nothing is neglected, and under the assumption of a uni- form field, for a positive charge, V, corresponding to a dis- placement AN' with V,= OF and for a negative charge, — V, 2 Barus—Displacement Interferometer. 13 corresponding to a displacement — AV", the mean of the equa- tions, AV = 4AN' + AN"), after reduction becomes MgAN D (14+2A/D/) l iP V 1 the corrected equation required, reducing to the above case when A= 0. The correction factor is thus 1//’= (1 + 2A/D)’ and was computed in an auxiliary table. It is now easy to discuss the conditions of equilibrium, for the forces X are given by the equations (second member referring to the pendulum and third member to the electrical forces) =V, , Mga N sin Me 1 Ae as a (ee A/D) Hence, if AV = A, & = 0, the condition under which the disk just moves, without interruption, from the guard ring to the condenser plates, 2. ¢., the limiting value ot the potential products, (V,V,) for stable positions of the disk, since a ae ee GaN ae ae 4c 2.) (HROLAY/D) are (in electrostatic units) (y, V. eer =a (1+2 A/D)" 8 44 9 A/D Assuming V, = 250 volts, the forces Y for the pendulum and the electric field were also computed. If now, we insert the value of JV’, from the above equation and reduce —6 A/D=1.0r —A=D/6 Hence the value of V, which corresponds to tangency is ee oy (: nn es VAT IN Go) Ny mee Vee or V, = 145 volts, above which charge the disk passes contin- uously from guard ring to plate. If the suspension is provided with a horizontal micrometer by which it can be shifted as a whole from & to k’ taking the needle with it, 4 may be eliminated. But the expression is not simple. The idiostatic method needs a corresponding correction, and if A’ is neglected in comparison with D, 2 Mg D? br? As the new factor is practically & or constant, V,’ is linear with AY. If, as above indicated, the suspension is provided AN=V? (1 +4 4/D) 74 Barus—Displacement Interferometer. with a horizontal micrometer by which it can be shifted as a whole from # to &' taking the needle with it, & may be elimi- nated. The equation may be written in terms of the two forces X (gravitational and electric) MgANN 47 ie: A/D CUBR Atay V? if peg a A= 2 (ANAC?) A table was computed for 4” and the two forces XY may be found at once. Again, the condition may be practically stated, inasmuch as the values dX/dX& may be easily derived. They are for the pendulum M/qg/Z and for the electric field dG WA Vatire vle 2 VAO/D IN DY. NIN Oe Equating them, we derive the voltage of transition to be char- acterized by ( V,), as pg Dp (v,)= ay aia since A = 0 at the origin. Hence above (V,) the disk would move on any slight disturbance without interruption from the guard ring, A = 0, to either plate of the condenser ; but within (V,) the apparatus is available for accurate measurement of V,, if & is equal to zero by trial. 8. Experiments. Cylindrical electrometer ; movable cylin- der within.—Tests of the apparatus described were begun by using a rather heavy cylinder of aluminum, damped with a vane submerged i in acup of oil. The cylinder itself weighed but 6546 grams; but air damping was quite inadequate, though the cylinder might easily have been etched to a more appropriate degree of thinness in acid. It is therefore necessary to make use of paper cylinders, even if the possibility of securing more nearly accurate values of the diameter of cylinder is the advantage of the metallic appurtenance. No trouble was encountered in using a light plane mirror about 1/4 inch in diameter, with axes for adjustment in a bit of cork. To sharpen the ellipses the (yellow) image at the movable mirror should be screened off, leaving the blue image and but two superposed spectra. Light gilt paper cylinders, weighing with mirror, etc., only about a gram, were therefore installed. An example (Table if - and fig. 3 a and b) of the many results obtained can only be given here, JZ being the mass and 7 the radius of the movable cylinder, and # the radius of the fixed cylinder. Barus—Displacement Interferometer. (® TABLE I, Cylindrical electrometer ; movable cylinder within. Ji NU TOS 8 Ie = OO Gung he ley eines YN Kee 8 Ved “(GART V3 Voltmeter °08 15 15 ONS 28 29 “60 43 4] Wey 60 61 1°87 7 81 IT,.— M = 1:189.¢. 8 Ih = POM Git, 2 PS NO Ging b= Mill Cans V,=0; V, = 250 volts. 10°AN Vs Voltmeter 1-74 6°1 5 27 7°6 1 3°31 11°6 11:5 4°53 16°0 15 5°85 20°6 19 eal OF EO AO 60S SSO HOR 0 Fig. 3. Electrometer data compared with actual voltages. The observations with the needle charged at 250 volts show the maximum sensitiveness obtained with short suspensions. These results and those by the idiostatic method are quite satisfactory and the differences are more liable to be sought for in the voltmeter with which the comparison was made than in the electrometer. It indicates the sufficiency of the equation assumed, apart from correction for the end of the cylinder, so that the latter cannot be menacing. In addition to these experiments, a long suspension (2 = 150 centimeters) in which the sensitiveness would have been increased 7 times or equivalent to 510% centimeter per volt or ‘006 volt per vanishing ring was installed. The ellipses were easily found but they were continuaily in motion, owing to the friction of air currents moving across the wires. To make the adjustment available it would have been necessary to build a closed case around the bifilar suspension over 1°5 76 Barus Displacement Interferometer. meters in height. It was not thought worth while to do this, and further experiments were abandoned. 9. Cylindrical electrometer; movable cylinder within. Summary.—The above experiments were made’ in the midst of the turmoil of a large city and near the engineering labora- tories of a University. “Tt is rather remar kable, therefore, that the ellipses were so easily found throughout and so easily made use of ; but it was quite out of question to use the eva- nescence of rings by which the sensitiveness could have been increased over twenty fold. Though the work was carefully done, it is intended merely to exhibit the general character of the method, inasmuch as the storage battery which was drawn upon was unavoidably in use elsewhere in the laboratory and the potential may have fluctuated. The design of the apparatus with a movable inside cylinder is probably the least interesting of those used. Thus it is difficult to keep a paper cylinder quite smooth or to give it a rigorous cylindrical shape and the correction for the ends can scarcely be estimated. For high charges auy asymmetry of the movable cylinder is lable to place it in contact with the fixed cylinders. Thus there is a limit of sensitiveness from a purely instrumental point of view, not contemplated in theory. In the above experiments with short suspension (21 centimeters) ‘710% centimeter per volt was the largest double displace- ment practically obtained, which would mean about ‘04 volt per vanishing ring. With the long suspension -006 volt per ring may be estimated. It is far short of the theoretical datum of §4. 10. Cylindrical electrometer ; movable cylinder without.— After a few preliminary experiments with the metal cylinder of aluminum, the gilt paper cylinder was tried and was at once successful. The mirror was attached to the paper wing needed for damping by the aid of a bent piece of thin steel wire, cemented on so as to give a horizontal axis on passing through the perforation of the bit of cork holding the mirror. Fora vertical axis the whole apparatus is rotated (cf. figs. 1 and 2). The results are given in Table 2 and fig. 4a. TABLE II. Cylindrical electrometer ; movable cylinder (gilt paper) without. I.—M = 2075 g.; foe 1:7oem. 5) == 1°59 zem4 ia) iemee V, = 250 volts. Drum 10°AN V3 Voltmeter Ibex hei 18°7 18 22°9 ashe 38°2 38 SONS 8°83 SOP 58 27°6 6°90 46:1 48 Barus—Displacement Interferometer. LYS IT,—_M = 1:0930 grams. Drum 10°AN Vz Voltmeter 24°9 6°22 21°9 ON ALES} 11°82 41°6 41 Fic. 4. v 0 0 40 6 0 2% 40 Fic. 4. Electrometer data compared with actual voltages. The differences in this table are more lable to be of the voltmeter than of the electrometer. The needle easily carried a potential of V,= 250 volts. With V,=100 volts, the needle became asymmetric and observations could no longer be taken. The sensitiveness is about -0003 centimeter of double displacement per volt for the short suspension, so that ‘1 volt per ring is the equivalent datum. The present form is thus easily made as sensitive as the preceding form, besides being much simpler in general design and installation. The experiment was now pushed a step further in the direc- tion of sensitiveness by using an even lighter paper cylinder with paper wings, the needle having the constants given in the second series of Table 2 and fig. 46. The sensitiveness is now about ‘0006 centimeter of double displacement per volt, twice that of the preceding case as it should be. This apparatus did not, however, behave as satisfactorily as the other, there being greater tendency of the movable cylinder to cling to the stable parts, whenever V, was high. In case of smaller potentials or of the idiostatic method there would be no difficulty in this respect. No doubt, if greater time and patience were spent on the work (a cylinder of celluloid suggests itself) the final result could be somewhat improved ; but the limit found (6x 10* cen- timeter or 20 rings per volt) will not easily be exceeded in practice. 11. Disk electrometer. Apparatus and experiments.—The disk electrometer deserves special attention because it is simpler in design and practically more sensitive than the cylinder types. At the same time, however, it is more treach- erous and without special precautions there is danger of short- circuiting the highly charged disk, while the sensitiveness is enormously variable in response to any unsymmetrical position 78 Barus—Displacement L. ntenperometer. of the disk. The instrument used is shown in figs. 5 and 6, in sectional side and front elevation. J, is the char ged disk on the bifilar suspension of very thin copper wire yy’, -007 centimeter in diameter. The disk is made of thin mica, silver-plated, carried by the horizontal axial steel rod dd, and surrounded by the guard ring of very thin copper V,’. Parallel to the disk and equidistant (d) from it are the plates V, and J, of the condenser at a distance apart, so that D is ‘twice d. The plate V, is earthed and firmly held on the arm of the inter- ferometer. The plates V, and V, are spaced at three points MiG. oO: HiGisG: RG Aan 2 | aS KG Fie. 5. Disk electrometer, vertical section. Front view. Fie. 6. Disk electrometer, vertical section. Side view. by hard rubber gaskets, cc, through the holes of which hard rubber screws are passed, the plates being ultimately secured by nuts a, a’, and a", on the outside of the plates. The nuts a terminate in clamp screws. The guard ring JV," and the suspension wires yz’ are in metallic connection at the tops of the suspension, the latter being the same already shown in figs. 1 and 2. The small mirror, m, is attached to a small plate of cork, n, which is slotted parallel to the rod dd, the latter being clutched by the jaws of cork which make up the sides of the slot. In this way m may be rotated around a horizontal _ axis, while the apparatus as a whole may be revolved about a vertical axis. If the grating is capable of being raised or lowered, it is not difficult to adjust the apparatus and find the ellipses. The damping of the disk is naturally good, though it may be improved by surrounding it with the case shown in fig. 1. Barus— Displacement Interferometer. , 12. Hxperiments with the disk electrometer.—The experi- ments were begun with but a small distance, D, between the plates of the electrometer. In such a case the disk can not carry very high potential before it is drawn across to the condenser plates fand short circuited. In fact, though the mica disks may be made very light, the annoyance of short- -circuiting is correspondingly increased and the great advantage of sensitive- ness cannot for this reason be realized unless the idiostatic method is used. The disk was charged with the storage battery to 101 volts and a number of small potentials measured at the plates. The computed and uncorrected results were about 3 times as large as the true voltages. The sensitiveness is here surprisingly large, almost 8 drum parts or 2X 10~* centimeter per volt, 7.e., 70 rings per volt. The distribution of results is, nevertheless, linear and proportional to the true voltages. To account for this result by aid of the equation MgAN D [eatin (1-2 a/D) is not impossible, if the asymmetry is such as to make the terms in A/D essentially negative. Yet in the case of a large value of A the results are so lacking in probability that a want of uniformity or variation in the field is more liable to be in question. On spacing the plates further, D—1:04 centimeters, and carefully adjusting the disk, it carried a charge equivalent to V,=250 volts easily, so long as V, did not exceed about 20 volts. A new and lighter disk was thereupon introduced, weighing but 607 gram. As it failed to carry a charge quite as large as V,=250 volts, the only available smaller one, V,=104°5 volts, was given to it. The uncorrected results for V, are here much nearer the true values, z.e., about 1:2 times too large, than were those of the preceding case, so that the correction for asymmetry is probably applicable. The mean trend of the locus is through zero. Sensitiveness is naturally much lower than above, showing AV =5 x 10~“ centimeter per volt. The experiment indicates well how markedly dependent the con- stants of the instrument are on the position of the disk. The same disk was now weighted with a rider, making the total mass 1°086 grams, with the object of charging it to 250 volts. The uncorrected results, while maintaining proportion- ality, were about 2 times too lar oe. The data obtained idiostatically, however, where the deflec- tions are essentially small, fig. 7a, show an excellent agree- ment with the true voltages. Discrepancies are as liable to be 80 Barus—-Lisplacement Interferometer. in the voltmeter as in the electrometer. The values of & given for this table show how nearly a true condition of symmetry was attained, as its mean value from the equation 2 Mg D A \ 72 v == Vie pe aw (1-2 ) is but —:0085 centimeter. The data, as a whole, illustrate the peculiarities of the disk method very well, giving evidence both of its relatively great Fic. 7. i a 95 1 100 aE : J ae ea 7 80 | Vine a 60 Bie DA \35 40 VARIG ¢ | Gigs) Gh OG 20 ae [ as f 0 We 40 o|__ votis— foeaeiia eal hacenl Pe] 0 0 40 60 80 100 Fic. 7. Electrometer data compared with actual voltages. sensitiveness (the double displacements being even 2 x 107° centi- meter per volt) on the one hand, and the variation of sensitive- ness as the result of a more or less unsymimetric¢ position of the disk on the other. The ratio of observed and actual volt- ages is fairly constant. The question then arises as to the degree to which the latter difference may be corrected. There is, IN many eases, a peculiar shift of the zero of displacement for the uncharged apparatus which is not easily accounted for; though, from another point of view, it is truly astonishing that a suspended disk should adjust itself to a given position with an accuracy comparable with the wave length of light. In the experiments detailed, such difficulties were eliminated by taking mean results, but usually the position in question was actually stable. If the shortcomings can be overcome, the practical limits of the apparatus as here constructed should, for short suspension (23 centimeters), be about 015 volt per vanish- ing ring. Barus—Displacement Interferometer. 81 13. TheSame. Further experiments. The interesting feature of the preceding result is the large range of variability in the sensitiveness of the apparatus. It was therefore thought worth while to throw further light upon the investigation by purposely tipping the apparatus, in order that the disk might le on one side or the other of the guard ring. Mg D> Ve a applied to the experiments did not give results quite as consistent as was anticipated, inasmuch as the same value of k did not reappear when the apparatus was tipped back apparently to its original position. Moreover, the central position did not clearly correspond to k=0. Hence it seems that k=0 is determined by the geometry of a non-uniform field. Neither did the values of V, correspond very closely to the values of &. The apparatus is more sensitive when the disk V, is on the side of the plate V, and about equally so when central or when on the side V, = 0. In other experiments the disk was weighted with a rider and a similar series of data was taken, although V, = 250 volts was carried by the disk only in cases where its position was nearly central. This was even the case on adding heavier riders. The computed values of V, were in all cases 2 or 3 times the true values, though the proportionality is maintained. The conditions are thus too complicated for consideration here. 14. Case of the enclosed disk. With the object of obtaining a more uniform field, the electrometer was now modified by surrounding the disk with a short cylindric or drum-shaped tube within the guard ring. The plates and the drum guard ring are spaced, as before, by three hard rubber rings, with a hard rubber screw passed through the perforations. Experiments were begun with the light disk, J/ = -607 gram, which was charged to 110 volts, as it would not carry an appreciably higher voltage. Consequently the rider was thereatter added, making J/ = 1-086 grams, and the disk charged as far as 250 volts. The results are given in fig. 7 0. These voltages, though obtained under very different condi- tions, are practically coincident and the discrepancies are as liable to be in one instrument (voltmeter) as the other. The remaining experiments were of the same nature. To summarize ; the enormous variation of the sensitiveness of the disk electrometers as depending upon the position of the disk, &, has, therefore, failed of interpretation; 2. e., the cor- rection for V, does not seem to follow the above equations. The field is, therefore, probably far from uniform, possibly consisting of a conical tube of force between plate and disk on the side toward which the disk leans and of a more nearly Am. Jour. Sct.—FourtH SErRiss, VoL. XXXVII, No. 217.—January, 1914. 6 The approximate equation for k/A V = 82 Barus—Displacement Interferometer. cylindric tube on the opposite side, where the tube must pass through the circular perforation in the guard ring. The result is an empiric instrument in which the deflections are propor- gonsl) to the voltages to be measured, of increased sensitiveness, even 2 X 10-* centimeter per volt, but not yet sensitive enough to be immediately valuable for refined practical purposes. The idiostatic instrument, for voltages above 20 volts (after the disk has been adjusted fork = 0 by comparing the posi- tions of the charged and uncharged disk) may in some eases be useful. 15. Case of the unsymmetric disk. The cause of the departure from linearity in the preceding experiments is to be referred to a slight asymmetry in the disk, whereby an effect varying as the square of voltage of the disk is superimposed on an effect varying as its first power. To eliminate the non- linear term, it should be sufficient to obtain no deflection in the case of earthed plates, when the needle is successively charged and uncharged. In fact, the charged non-symmetrical needle between earthed plates introduces an interesting method. of electrometry as follows : Let & be the amount of non-symmetry for a disk whose mean distance from the plates D centimeters apart would be d = D/2. UHence for any displacement AW the distance of the disk from one condenser plate will be d= D/2+ (k +A) and from the other plate, D — d = D/2 —(k + AW). Hence the displacement force is (after reduction) DV St ett ONG 1 Mg A N 7p DIM DYES CURRIN) = As (k+ AW)’ may usually be neglected in comparison with D*/4, the equation becomes Mig NON Vil mae kta N If & is large as compared with A/V, which will usually be the ease, V,’ will vary linearly with AZ. To use this method, #& inust, therefore, be known, and it may be determined with the aid of a given voltage preliminarily. Many experiments were made in this way, an example of results being given in fig. 7¢. The asymmetry of the disk was,about °5 ) millimeter, and the plates were over 6 millimeters apart. The equation employed answers the requirements as closely as the observations could be made. At high voltages (100 volts) there is liable to be divergence, while at low volt- ages the displacement AY is too small for accurate work, see- ing that the underlying equation is quadratic. An apparatus like the present, intended for actual measure- ments, should be provided with a micrometer suspension for Barus—Displacement Interferometer. 83 shifting the disk as a whole in the direction of its axis and with more elaborate means for sighting with a view to hori- zontal and vertical parallelism of ‘the disk with the plates of the condenser than was the case with the improvised apparatus here treated. It has been stated that it is an inversion of the present displacement which is superimposed on the displace- ment in case of a permanently charged needle, and which thus demands an apparatus tested for absence of non-symmetry, if the voltages are to be proportional to the displacements. 16. Displacement Interferometry Applied to the Quadrant Llectrometer.—The method of measuring small angles given elsewhere may be made use of in measuring very smal! volt- ages or small increments of potential, by attaching a pair of light mirrors, symmetrically and parallel, to the needle of a quadrant electrometer. In the present experiments this was an improvised instrument, constructed by myself, the quad- rants being of sheet copper fitted and soldered together and supported on cylinders of hard rubber. The bottom of the stem of the needle was submerged in sulphuric acid as in Kelvin’s instrument and the suspension was bifilar. The insulation was throughout excellent. The needle was kept charged to about 150 volts with a Zamboni pile, any variation of charge being indicated by Elster and Geitel’s electroscope. In fic. 8, gq shows a pair of quadrants in vertical section, E the needle on the stem ss’, the lower end of which is platinum, bent as shown, thus making a clip to hold the light mica vane v (if necessary) submerged in sulphurie acid of the vessel ¢ charged by the Zamboni pile. fi’ are the hard rub- ber supports of the quadrants. At a suitable distance below them the light parallel mirrors 7 and n (less than 1 centimeter in diameter each) are supported by the light cross piece of hard rubber vr attached to the stem ss’ of the needle. The axial line of the needle #'is parallel to the line 77 between the mirrors and the latter are placed at a horizontal angle of about 45° to rr. To adjust m and n to adequate parallelism, each is supported by an attached fine needle, fitting snugly in a vertical groove in the ends of rv. The needle, as a whole, is to be clamped at rr in a suitable support and sunlight i is to be used. When the horizontal beam reflected from m to m and thence to a distant white screen falls within the direct shadow of n, the needles are fixed in place by resinous cement. The mirrors should be equally high. The quadrant electrometer with the needle in position was now placed, with the aid of three long foot screws, on a cireular platform just below the iron arm holding the fixed inter- ferometer mirror J, and the reflection took place as shown in fig. 9. In effect the arc light from a collimator reflected from 84 Barus— Displacement Interferometer. the front face of the grating (blue image due to scattering) is next reflected at n, thence to m, W henee it goes to WV and returns by the same route, passing however from m through the grating to the observer’s telescope. The case surrounding the electrometer must, therefore, be provided with a front and rear window. With somany reflections (altogether ten, inelud- ing the one at the grating and the effect of the two windows) this one is considerably weakened, and it is advisable to use black glass and not a silver mirror at the micrometer I. It is prefer able, moreover, to use the glass side of the mirrors at m and n, as the silvering is liable to be brighter in the rear and the effect of thin glass plate is of no consequence. This, Fic. 8. Fie. 9. Fie. 8. Quadrant electrometer with parallel mirrors. Fic. 9. Diagram showing direction of rays, quadrant electrometer. in fact, was the greatest difficulty encountered, as the mirrors m and n were not at first adequately silvered and polished. Even when the direct image of the slit is clear the spectrum is apt to be dull and the ellipses are hard to find. The adjustments, when so many conditions have to be met, were not at first easy of attainment. The electrical installation should first be completed and the needle in place between the quadrants, the electrometer being placed so that the beam from the interferometer strikes m (see figs. 8 and 9). A white screen behind n, catching the light passing beyond the edges and showing the shadow of 1, facilitates this adjustment, the \ Barus— Displacement Interferometer. 85 slit being opened wide. The electrometer is now rotated as a whole around the vertical, until the light reflected from n strikes m and a similar screen behind the latter in the line mn is necessary here. Next, the mirror JV is adjusted to normality, a white screen behind m in the direction Vm being essential, whereupon the light is reflected from n again towar d the grating. As the pencil must pass through the grating G again, this is mounted to be capable of being raised or lowered and rotated around the vertical and horizontal (adjustment screws); but these are the usual adjustments on the inter- ferometer. It follows that the electrometer must also be capable of being raised and lowered on its long foot screws, as already indicated. Moreover, the spectrum should be so placed by raising and ae the source of light and including the grating together with WV that the higher orders of direct spectra are not in the same plane with the spectra superposed for interference. When the adjustments described are well made, there is no further dithiculty m ‘finding the solitary ellipses. As there is compensation owing to.the two glass plate windows in the case of the electrometer, the ellipses are lable to be enormous, practically vertical straight lines, which are displaced with cor- respondingly great rapidity by the micrometer screw, and are therefore hard to find. Hence some counter compensation at the micrometer mirror is desirable, in order that centers may appear, and the displacement may be slower. A compensating plate about 1 centimeter or less in ‘thickness, with the vertvcal foeus of the light from the slit on the stationary mirror, pro- duces very clear and sharp ellipses, admirably adapted for measurement, 17. Observations.—The stems of the mirrors m and n on 77 were about @ = 4°5 centimeters apart. Hence, since 7 = 45°, a = 4:5/+/% and if 6 is the path difference sre = ‘0175 X 2a sin = ‘079 cm. per degree. Supposing that 1 volt gave a deflection of 45° and that 10+ centimeters are measurable at the micrometer, the micrometer sensitiveness should be LOs AS 079; or about 3 : volt per vanishing interference reaver” 16% would thus seem probable that with a lighter needle and a more delicate suspension the possibility of measuring 10~* volt would not be out of the question. The following observations show that these surmises are correct so far as the method is concerned. The great hardship encountered in the present work was the unavoidable agitation of the laboratory, and this unfortunately is insuperable, After the ellipses were found they were always in motion, so that the displacement work was bound to 86 Barus— Displacement Interferometer. be rough and the use of interference rings out of the question. As the rings, however, may easily be obtained and used under suitable conditions , and as the purpose of the present paper is merely to test the method, the annoyances in question are of less consequence. To obtain small potentials, a thin bare German silver wire about 1 meter long was stretched and insulated on a board and the ends supplied with a constant potential difference of one volt. Two points of the wire, one or more centimeters apart, were then used as a source of potential, the wires from these points leading to a Mascart key, which suitably earths and commutates the charged wires leading to the electrometer. A thermo-couple might have been used; but the long wire is preferable because of its simplicity. In the first experiments made, the suspension was a bifilar about 10 centimeters long (each strand of several silk fibers) and about °5 to 1 millimeter apart. No doubt the torsional stress of the fiber was here of an order commensurate with the bifilar force. The needle, being damped in concentrated sulphurie acid, moved very slowly and about two minutes were allowed for each deflection. It was thus possible to follow the ellipses on commutation, from one to the other extreme elongation, by moving the micrometer screw proportionally to the displace- ment of ellipses. As a rule, the ellipses were quite clear all the way, showing that the adjustment for parallelism of mirrors on the needle by the aid of sunlight is adequate. The potential of the charge on the needle was of the order of 150 volts. In the first experiments, A/V/A V = -57 centimeter was the displacement on the micrometer per volt, so that AVY = 10~ centimeter corresponds to ‘000170 volt, or the sensitiveness is about -000960 volt per vanishing interference ring. The instrument was now improved by inserting a Jonger bifilar suspension, consisting of a single fiber of silk. The dam per and other details were retained. The potentials were tapped from points of the long Wire, respectively 5 and 10 centimeters apart. Owing to the very large deflection, the ellipses were not equally clear thronghout the whole displacement from elongation to elongation. The instrument behaved much better in the present case and the sensitiveness had been much increased, particularly in the later series of observations. Since A/V/A V was of the order of 2-9 centimeters per volt, 10~ centimeter corresponds to -000035 volt, z.é., to about -000010 volt per ring, so that under favor- able circumstances a few millionths of a volt should be dis- cernible. Brown University, Providence, R. I. J. Barrell— Upper Devonian Delta. 87 Arr. VII.— The Upper Devonian Delta of the Appala- chian Geosyneline ;* by JosnpH BaARRELL. Parr II. Facrors ConTroLumnGc THE PRESENT Limits OF THE STRATA. CONTENTS. Introduction mandy sumaniaty ss ee sy eee ese ae 87 Present extent and thickness of the Upper Devonian... .__ 88 Comparison of this restoration with other maps --------- 91 IbAIHEINGS OE CROSIOIN CyCles eae lose aoes bake Geese se oo ed 93 Present limits dependent upon intervening baselevels__ 98 Relation of post-Ordovician erosion to original limits of De VOmTema te 1S eerie aera (SUN ey a debe ae Miyy eb N Ld 95 Influence of late Paleozoic cycles of erosion _______---- 96 Degree of destruction by the pre-Newark cycle -.-.__-_- 97 The Newark erosion cycle and the quartzite conglom- TATE Sree OU MMR Eme NO oak ee eal eee Coe ea ue oe OO The Jurassic erosion cycle AY SRA aie Se ts ee oO Nd es 102 The post-Jurassic erosion cycles ._______-------------- 108 Absence of structural relations between present and orig- Thal IkewaAs 5 oo ve, a eas Saat IVS Epa tee nye aaa DLs 105 INTRODUCTION AND SUMMARY. Towarp the west and southwest the Upper Devonian delta faced the sea. The study of its strata, guided by criteria for distinguishing the modes of origin of the several formations, has led in the first part to certain conclusions. But facing the east the marginal outcrop is near the region of maximum thick- ness. The beveled edges of the strata, measured in thousands of feet, bear witness to a former extension toward the ancient land of Appalachia. The nature of the surface and the relations of the delta to the sea are but. one half of the problem. The former extent and nature of the strata now destroyed and the relations of the delta to the old-land constitute the other half. Although Appalachia and its marginal waste plains no longer exist, we may attack these questions by inferences from the nature of the remaining strata, from the structure of the rocks beyond, and by the evidences as to the amounts of erosion which they have suffered. Merely because the study is infer- ential, the results are not necessarily less secure. The com- monest conclusions in geology are in fact inferential. Therefore it is not the absence of inference, rather is it the definiteness and convergence of independent lines of evidence and the soundness of the principles according to which the conclusions are inferred which determine the security of the answers. It is seen then that the principles of interpretation are as impor- * Continued from this Journal, (4), xxxvi, pp. 429-472, 1913. 88 J. Barrell— Upper Devonian Delta of the tant as the facts interpreted ; but those criteria which are used by geologists to determine the original nature and limits of eroded formations are less well demonstrated and are much more vaguely employed than are those used for determining the modes of origin of strata. In the study of the existing sediments observation and demonstration are able to follow fairly close to hypothesis, but in the more difficult study of that which is no longer existent unchecked hy pothesis still largely prevails. This second part of this subject is taken up, there- fore, in discussing the causes which have determined the present limits of the Upper Devonian formations and the lack of relationship between the present outcrops and the former limits. It has been customary, on paleogeographic maps, to draw the original limits of formations at no great distance beyond their present outcrops. On the other hand, areas of ancient rocks tend to become regarded, unconsciously to the thinker, as land areas through all the younger ages. To a certain degr ee both these principles as rules of guidance point in the direction ot the truth, and it is only by following them in a general way that paleogeographic maps may be constructed. “Yet in any particular problem these principles may lead to large errors. Sediments may originally have existed hundreds or thousands of feet thick and reached hundreds of miles beyond their present boundaries and now leave no trace. An extended discussion of these subjects is required in connection with the Upper Devonian, as is seen on comparing the present conclusions with those of previous writers. This part, on the relation of present outcrops to causes other than the original limits, is preparatory for the more definite evidence which follows in the third part. The preliminary discussion shows that there is no reason why the Upper Devonian sediments may not have extended north- ward to beyond Lake Ontario and eastward to the margin of the present coastal plain. In the third part, to be published in a following number, the indications given by the strata as to their former extension are taken up and several independent lines of evidence converge to the conclusion that the Upper Devonian did extend originally to the neighborhood of these limits. PRESENT Extent AND THICKNESS OF THE UPPER DEVONIAN. The map of the Appalachian geosyncline, fig. 1,* shows the limits of the Upper Devonian outcrops and the contours give approximately the total thickness of sediments which were deposited during the Upper Devonian. * Figs. 1 to 4 inclusive are published in Part I, this Journal (4), xxxvi, pp. 429-472, 1913. Fig. 1, for convenience of reference is republished in this part also. 89 Appalachian Geosyncline. Ines AL, TH TAT -| eT o a io) 500 j i) TTT Contour interval = 1000 Feekdepth strata The Appalachian geosyncline at the close of the Chemung—Catskill Fie. 1. epoch, 90 J. Barrell— Upper Devonian Delta of the In southern New York the strata dip gently southeast and outerop in a broad hilly upland from 1500 to 2000 feet in ele- vation; the general surface however slopes north, contrary to the slope of the strata, and overlooks the central lowland of New York, lying at a veneral level of 400 to 500 feet, devel- oped on the soft Ordovician and Silurian strata. The old upland surface is now trenched by deep transverse valleys leading to the north or south and giving a sinuous outline to the limiting outcrop. In eastern New York the strata are more resistant, dipping gently southwestward and extending northeastward in the Catskill Mountains until the higher peaks reach elevations in the neighborhood of 4000 feet. Close in front lie the Mohawk and Hudson Rivers near the level of the sea. The thickness of the Upper Devonian at its margin is about 2000 feet in western, about 4000 feet in eastern New York. The strata must therefore have extended at one time much farther to the north. Facing the Appalachians the Upper Devonian as far south as Lat. 41° forms a mountainous tableland of gently flexed strata, the elevations descending from. the heights of the Cats- kills to about 2000 feet. Immediately to the east lie the folded structures of the Appalachian system over which the Upper Devonian strata, here from 4000 to 6000 feet thick, must once have extended, participating in the folded zone. South of Lat. 41° to 41° 15’ the folded structures advance westward and involve a greater breadth of the Appalachian geosyncline. The Upper Devonian strata plunge steeply beneath the synclinorium of the anthracite coal basins, the eastern outcrop showing here maximum thicknesses of between 7000 and 9000 feet. West of this synclinorium is a broad cross anticlinorium running north and south between Long. ° and 78°. This upwarp crosses a series of great folds, pro- ducing very zigzag outcrops in the strata pitching away from it. The Upper Devonian is completely removed across this transverse anticlinorium and exposes in its flanking outcrops a thickness decreasing northwestward to 6000 feet. On the west the isolated Broad Top Coal basin shows the Upper Devonian on all sides and offers good opportunities for meas- urement. The last exposure of the full thickness is on the Allegheny front, at the limit of the strongly folded structures, but the upper part outcrops in three gentle anticlines to the west and bore holes show the continuation of the strata. Tn the region of folding the measurements of thickness can be made in a limited distance and with small error but transi- tions at base and top prevent a high degree of accuracy. Some changes of thickness due to folding are also probable. In the other regions the thicknesses are not so easy to obtain. Appalachian Geosyncline. 91 Over the western part of Pennsylvania the bottom of the Upper Devonian is nowhere exposed. In southern New York, on the other hand, the upper beds have been removed and detailed geologic mapping has not crossed that wide belt between the middle Devonian and the Mississippian outcrops. In northern New Jersey also the isolated Green Pond Moun- tain axis preserves merely a remnant whose original thickness is not known. In the neighborhood of Lake Erie, however, more exact figures are determinable. But although these difficulties prevent high accuracy, the errors are small in com- parison with the total thicknesses of thousands of feet. A striking fact in regard to the distribution of thicknesses is that they are greatest near the southeastern margin. It is to be noticed further that the boundary from Harrisburg to Lake Erie is everywhere oblique to the lines of equal thick- ness, decreasing from about 9000 feet in the southeast to 2000 feet in the northwest. In the region of the folded Appala- chians the thicknesses range from 6000 to a little over 9000 feet. It appears therefore as if ‘there were preserved to our geologic period but a portion of the original basin. The eastern side is almost wholly gone and on the north erosion has etched back the margin for an unknown distance. The restoration of this basin showing its original extent and the thickness and character of the strata is the task which we have set. CoMPARISON OF THIS RESTORATION WITH OTHER Maps. Before entering upon a detailed analysis of the problem it is desirable to see how the present restoration of the Upper Devonian as shown in fig. 1 compares with others which have been made. J. D. Dana in 1895 published a map of North America at the commencement of the Carboniferous era.* On it the limits of the sea, and of the sediments, are shown as approximately following the north boundary of Pennsylvania, Lat. 42°, and lying close to the limiting outcrops on the southeast. The most detailed map was published by Willis in 1899,+ and is the map with which the present one should be most closely compared. The two differ materially. In Willis’ map the whole area of sedimentation is indicated as a shallow sea and the north limit of the basin is drawn at Lat. 48° 15’ where the present map indicates a former thickness of between 2000 and 3000 feet of Upper Devonian strata. On the east the shore line as shown by Willis follows closely the marginal outerops. The thicknesses taken by Willis include the Hamilton, which * Manual of Geology, p. 688. + Maryland Geol. Sury., vol. iv, pl. vii. Geography of the Eastern United States During and at the Close of the Devonian Period. 92 J. Barrell— Upper Devonian Delta of the is here excluded, and the present map rests in part upon publi- cations by the United States and New York Surveys issued since the map by Willis was published. A large difference ne however, owing to another cause. In Lycoming County, in North Central Pennsylvania, Willis takes data from Andee Sherwood* for the thicknesses of the Upper Devonian. But these are measurements of partial and not complete sections. Their use gave an apparent lack of sedimentation and conse- quent constriction ot the Beene ne The present map in that region has been controlled by the very complete section measured by H. M. Chance.+ But it is in the location of the margins of the basin that the two maps differ most widely ; the differences being the result of the application of different principles for the location of the original margins of eroded formations. Willis has crystallized these more definitely and consciously than others of the living geologists and their dis- cussion is taken up in detail in later sections of this article. The two maps are important in showing the wide difference in results which may flow from different interpretations of the same set of facts. In 1906 Chamberlin and Salisbury published a set of paleo- geographic maps which showed distinct advances over those preceding.t The map for the Upper Devonian (II, p. 431) shows the northern limits of water as possibly existing to the line located by Willis. A possible extension of water on the east is shown about to the limit of sediments indicated on the present map. Continental deposits are not discriminated. A probable water body is shown as existing south of Connecticut and extending possibly northward up the Connecticut valley. In 1910 Schuchert published his Paleogeography of North America§ which gave a wealth of detail beyond anything previously published. In the construction of his maps he followed the principle that the shores should be shown as near to the limiting outcrops as is reasonable: that is, the extension of the seas to the limits set by Schuchert is practically certain, but, as is noted by Schuchert (p. 446), this method may err on the side of too much restriction of the continental seas. Fol- lowing this method in his map of the Upper Devonian (pl. 77) the shore is shown on the north where it is drawn by Willis. The Catskill Mountain area is shown as an estuary (see p. 545) separated from the sea by a barrier. The necessity of a sea- way for the migration of faunas causes him to locate a strait across northern New Jersey, making the Skunnemunk con- * Second Geol. Surv., Pa., vol. G-2, 1880. +Second Geol. Surv., Pa., vol. F, Appendix B, 1878. t Geology, Earth History, IJ, III. $ Bull. Geol. Soc. Am., xx, 427-606. Appalachian Geosyncline. 93 glomerate marine. The beds in Pennsylvania are shown as ‘wholly marine and advancing east somewhat farther than indicated on previous maps. The present map, fig. 1, shows the shore line at the close of the Devonian farther west than any previous map but the margin of the sediments farther east, except for the New Jersey strait of Schuchert which is here eliminated. Between the two limits is shown, not a shallow sea or estuary but a subaérial delta plain. The comparison with the previous maps crystallizes the differences between them and introduces the need for a discussion of the principles which have controlled the past and the present restoration. INFLUENCE OF ERosion CYCLES. Present Limits Dependent upon Intervening Baselevels. The original development of a formation was dependent upon the baselevel of the time, the initial baselevel in its his- tory. Ata distance from the shore the inland baselevel deter- mining the relations of erosion and continental sedimentation are controlled by the river grades and may depart appreciably from sea level. The limits of deposition, however, are limited in the initial cycle to the parts below baselevel even if some- what above sea level. In the drawing of the earlier paleogeographic maps the views prevalent in the middle of the last century determined the principles of construction. It was originally assumed that the present limits of a formation are controlled by but two base- levels of erosion, the initial one, giving the original limits of sedimentation, thought of as necessarily a shore Tine ; and the present cycle, ‘giving the existing relief to the land. The idea of clearly separated baselevels was not however present in theory. Uplift was looked upon as progressive, the successive shore lines forming on the whole a descending series. It apparently was not thought that any part of the continental surface save the anticlines of folded regions had lost very much in elevation. No quantitative comparison had been made between the great volumes of the sediments and the enormous denudation which they imply. Consequently the old shore lines were located by filling up in imagination the river valleys and extending the strata to the higher upland levels of the region. It was not believed that extensive overlaps or outliers could have existed, for it was thought such could not have been wholly eroded. This blended two-cycle principle, inherited from the past, still unconsciously controls the work of many geologists whose specialty does not lie in the application of present physiographic knowledge to the distant past. 94 J. Barrell— Upper Devonian Delta of the In the past two decades has grown up an increasing apprecia- tion of the complexity of erosion cycles. Davis early recog- nized one baselevel as the floor of the Triassic formation and another as the surface truncating the higher ridges of Penn- sylvania. The latter he correlated with the floor at the base of the Potomae group and therefore it was placed by him as of Cretaceous date. Another erosion level, existing in the val- leys and beveling the softer formations, he placed as of Ter- tiary age. Since “then still other minor cycles have been added. What effect should this development of the erosion theory have upon the problem of Paleozoic geography? Its influence may be considered by introducing first the effect of the Cre- taceous baseleveling. In that Jurassic-Cretaceous cycle it has been held that all of eastern North America except certain residual mountain groups was reduced to a peneplain. Con- sequently if ancient strata, for example those of the Devonian age, were warped upward at the close of the Paleozoic above the level of the Cretaceous peneplain, no matter how far they originally extended, so long as they did not enter into the residuals still remaining upon that plain, they would have become completely destroyed. Even if they had extended outwards for hundreds or thousands of miles at elevations above that level, now no vestige would remain. Consequently the extension of the strata to the present level of the uplands would only restore the boundaries as they existed in the Cre- taceous, not in their time of origin. This may be called an in- terpretation of paleogeography with respect to a three-cycle basis, the Triassic and Tertiary cycles being looked upon by most writers as relatively partial and local. Chamberlin and Salisbury in their Geology and Willis in his Outlines of Geologic History have recognized the pertineney of this principle by. showing a probable extension of the early Paleozoic seas upon the Canadian shield where now only pre- Cambrian rocks exist below the Cretaceous level. Such rela- tions of intervening erosion cycles to the problem of the ancient boundaries does not however appear to have been clearly formulated. A paper in which is stated distinctly the relation of the Cretaceous cycle to an earlier probiem, here the problem of Permian folding, is that by R.T. Chamberlin on the Appalachian Folds “of Pennsylvania.* Here the Cretaceous baselevel is detinitely assumed to be essentially the same as the post-Permian baselevel.+ It is the conclusion of the present writer, however, that many erosion cycles, developed with respect to successive base- levels, have intervened between the Devonian and the present and that the Cretaceous cycle is a broad term for a number of * Jour. Geol., xviii, 228-251, 1910. + Loe. ecit., 237-241 Or Appalachian Geosyncline. 9 partial cycles extending from the close of the Jurassic into the early Tertiary.* If this be so, it still further destroys any real relationship between the limits of the present outcrops and the original limits of sedimentation, even in cases where residuals still rise above the upwarped Cretaceous baselevel. Relation of Post-Ordovician Erosion to Original Limits of Devonian. A more or less widespread crust movement took place at the end of the Ordovician. Uplift and folding occurred over the Appalachians as far west as the eastern side of the Great Valley, as is shown by the unconformity at the base of the Silurian and the fact that fragments of Ordovician formations enter into the Silurian basal conglomerate. This has been called the Taconic revolution from the region where the signif- icance of the unconformity was first recognized. The Taconic and Green Mountains still stand in Vermont at elevations in the neighborhood of 4000 feet. The question arises,—how much of the metamorphism and folding of that region should be ascribed to the close of the Ordovician, how much to later movements? Their elevations, high above the Cretaceous baselevel, have been commonly looked upon as an inheritance from the Silurian. Did they contribute waste to the Devonian sediments and form a barrier in Devonian times, or were the Silurian mountains already leveled before the Middle Devonian so that the Devonian waste had to come from farther regions and may possibly have mantled the beveled folds ? The platean, outher of the Rensselaer Grit, shown in fig. 1, goes far toward supplying an answer. In the first part of this article this formation has been described and reasons advanced for holding that it is more probably of Middle Devonian age, though it was assigned by the older geologists to the Silurian and by J. M. Clarke to the Upper Devonian. The grit rests unconformably upon the Upper Ordovician schist and lies immediately west of the Taconic range. The latter is com- posed of these same schists and rises a thousand feet above the surface of the piateau. The Rensselaer forination contains beds of slate, others of grit, and still others of conglomerate. Feldspar, gneiss, and quartzite pebbles are abundant. The material of the plateau shows therefore that it was not largely derived from the schists of the Taconic range which now over- shadows it. Furthermore the Rensselaer grit is itself folded and metamorphosed, indicating that it has once been somewhat deeply buried and while so buried subjected to powerful earth *See Abstract, ‘‘The Piedmont Terraces of the northern Appalachians,”’ Bull. Geol. Soc. Am., 1913. Paper read Dec. 30, 1912. 96 J. Barrell— Upper Devonian Delta of the pressures and high temperatures. It is, however, on the western margin of these orogenic activities. These relations of attitude, of composition, and of internal structure show that the present Taconic region was lowlying in the Devonian, no longer a mountain barrier, and that a large part of the strong folding thrusting, and metamorphism which now mark the Taconic and Green Mountain region was imposed after the Rensselaer grit was deposited, probably in the closing revolution of the Paleozoic. In New Jersey also the Green Pond Conglomerate of Silurian age rests unconformably upon Ordovician limestone and pre- Cambrian gneiss. Its pebbles, except at the immediate base, are water-worn pebbles of white quartz and not local accumulations from an unreduced region. Much erosion had taken place between the close of the Ordovician and the beginning of Silurian deposition. Further, Lower Devonian quartzites and fossiliferous lime- stones are found immediately east of the Green Mountains in northern Massachusetts over a region doubtless affected by the Taconic folding. Clearly then we must conclude me the nee of the Taconic Revolution did not endure to supply the Upper Devonian sediments. They no longer rose as in- superable barriers and the sediments were more or less free to transeress their eroded structures. Influence of Late Paleozoic Cycles of Erosion. Here will be considered from the standpoint of theory the possible effects of the erosion cycles which followed the deposi- tion of the Upper Devonian but preceded the Permian fold- ing. The original marginal parts of a formation are the least down warped. They may, in fact, if terrestial in origin, have accumulated at considerable elevations above sea level. Even if not originally above sea level a slight upwarp of the margins of the basin might completely destroy them when the central and thicker parts would still remain. While a deposit exists as an unconsolidated surface formation it is particularly sensi- tive to even minor changes of level. The most widely extended formations are the ones least protected by overlying materials and most constantly subjected to erosion through later time, so that, as Gilbert has pointed out, the evidence “of the oreatest past transgressions of the sea becomes most completely “obliter- ated.* After such an erosion of marginal strata let the sea again invade the land and a disconformity is the result, in which the time interval marked by the hiatus represents some- * Continental Problems, Geol. Soc. Am. Bull., iv, 187-190, 1893. Appalachian Geosyncline. 97 thing more than the time in which sedimentation has been absent. For the Upper Devonian marginal sediments the problem 1 ig somewhat different. A great mantle of terrestrial deposits extended toward the mountains. The river grades which determined whether erosion or deposition should take place must have been subject often to minor disturbances of a climatic or diastrophic nature. The disconformities produced have been destroyed along with the marginal strata, but each pulse of erosion on the margin of the plain must have been recorded by a pulse of deposition of coarser waste toward the center of the basin. Climatic and diastrophic movements dur- ing and following the Upper Devonian would thus tend to affect the margin and the center of the basin in different ways. The relation between the two will, however, be treated more fully in a later part of the paper. Degree of Destruction by the pre- Newark Cycle. The next related problem turning upon ancient baselevels deals with the erosion of the early Mesozoic. Sedimentation began about the middle of the Triassic in certain tracts of the Appalachians; giving rise to the red conglomerates, sandstones and shales of the Newark group. These sediments rest upon floors which had been intensely folded and metamorphosed near the close of the Paleozoic, as shown by the Carboniferous strata which are involved in the Maritime Provinces of Canada and in Eastern New England. Further, near the Susquehanna river in Pennsylvania the Triassic approaches within ten miles of. intensely folded Upper Paleozoic rocks. Except in Nova Scotia the Triassic nowhere, so far as the base is visible, rests upon Devonian rocks. Does that mean that the Upper Devo- nian never extended over the region now occupied by the Triassic of the United States, or does it mean that sufficient erosion could have taken place after the folding of the Permian but before the deposition of the basal Newark of the mid-Tri- assic to have removed thousands of feet of resistant Devonian and Silurian formations ¢ Light on this question may be gained by an examination of the floor of the Newark rocks. Davis long since called atten- tion to the fact that the western contact in Connecticut is such an old floor resurrected by Tertiary erosion. He notes that it is approximately a tilted plain, since the outcrops extend in a nearly straight line, except where broken by faults. It cuts across greenstone and schist and granite without notable deflec- tion. The basal beds of the Newark are moderately fine-grained and are not the coarse agglomerates which skirt a rugged, moun- Am. Jour. Sci.—FourtH SEeRies, Vou. XX XVII, No. 217.—Janvuary, 1914. 98 J. Barrell— Upper Devonian Delta of the tainous upland. The conglomerates lie mostly on the oaeen side in Connecticut, prevailing especially in the higher beds, and are mostly of cobbles under a foot in diameter. They indi- cate the repeated rejuvenation of hills not far distant from the present eastern margin and appear to mark a zone of faulting and uplift on the east during the progress of sedimentation to the west. Thus, in a region of resistant metamorphic rocks, erosion toward an early Triassic baselevel had become largely completed. This does not mean, however, that a smooth plain was necessarily, or even probably, developed across hard rocks. Minor movements of the crust and the long time consumed by the last stages of the erosion cycle may per petuate a hilly relief long after mountains have vanished. Irregularities in the Triassic boundary in Massachnsetts suggest that a mature relief, measured by hundreds of feet, existed at the time of burial, but this is a minor feature in comparison with the ero- sion of thousands of feet which the metamorphic and crumpled structures of the floor imply. The floor can be studied further in Pennsylvania, on the southeast side of the Triassic. Here it is tilted 15 to 20 degrees northwestward and erosion has cut across both the Newark and the underlying formations. The floor crosses pre-Cambrian gneiss, Cambrian quartzite, and Cambro-Ordovician limestone. Although the margin is highly irregular in detail, it is seen to intersect hard and soft forma- tions without large response to the erosive resistance of the formations. It is difficult to eliminate completely the influence of faults and determine how much of the irregularity was due to the original surface. In the Germantown quadrangle of the Philadelphia folio* the irregularities in the margin of the tilted floor suggest hills of gneiss rising to 1300 or perhaps 1700 feet above the pre-Newark valleys, but on the other hand the contact crosses the broad anticline of pre-Cambrian in the Phoenixville quadrangle for 18 miles without showing any greater deflections than this from the limestone at each end. The relief, therefore, was not due to the pre-Newark folding, but was related to the hardness of rocks largely at least inde- pendent of structure. It seems clear, then, that the folded structures of Permian date had suffered profoundly from erosion even by the begin- ning of Newark deposition, and the geology of the Newark floor cannot be used as an argument to prove the absence of the Upper Devonian from those formations which were involved in the Permian folding. If they were high above baselevel they would have suffered destruction before the be- ginning of Newark time. Although such seems to be the conclusion, it assists in the conception of the process of rapid * U.S. Geol. Survey. Appalachian Geosyncline. 99 erosion to note the fatal weakness in the structures of the Appalachian geosyncline. The three notably resistant forma- tions, the Kittatinny, Pocono, and Pottsville formations, are separated and underlain by oreat thicknesses of non-resistant rocks. Wherever this series is exposed to erosion the rivers rapidly sink toward baselevel upon the soft formations and the mountain-makers become separated by deep valleys. Each is sapped from all sides. It is the old proverb—united we stand, divided we fall—applied to geology. It is not thought, from other considerations, that the Upper Devonian ever extended over the Connecticut Valley. It is concluded, however, in the following parts of this article, that, as shown in fig. 1, it once thinned out over the region now occupied by the Triassic in New Jersey and Pennsylvania, but was destroyed before the beginning of the Newark deposition. The conclusion is made more impressive by noting that on the Susquehanna River the Tri- assic rests on Cambro-Ordovician limestone. Only eight miles north lies the ridge of Silurian quartzite. Beginning with this formation, four miles of strata, including Devonian, Mississip- pian, and Pennsylvanian formations, rise vertically in the somewhat overturned arch of the syncline. The great eleva- tions which they imply on the east, even though they rapidly thinned out in that direction, had been removed before the beginning of Newark time. The Newark Erosion Cycle and the Quartzite Conglomerates. The Newark group of red shales and sandstones occur in a number of isolated areas from Nova Scotia to North Carolina. The one which enters into*the present discussion is the New York-Virginia area. This shows a monoclinal structure, dip- ping northwest at average inclinations of 15 to 20 degrees, but broken by faults. As interpreted by the present writer, the structure across the basin during sedimentation appears to have been analogous to that which exists at the present time across the Great Valley and Sierra Nevada of California. There, uplift on the east, downsinking on the west during the later Tertiary has given rise to profound erosion of the uplifted side and transfer of sediment on to the sinking floor. sss ee 6 10. Sands, fine, thin-bedded, with small amount of clay2ayets 183°3 Section IL. Kast branch Calvo brook. 100 feet from junction. Feet 1. Soil,—brown, mixture of gravel and sand __.-..-_--- 15 2, Sand and fine gravel, containing numerous carbonized fragments: Of plantsig ees ae seo eee eee 3 3. Peat, brown, porous, mixed with mud; plants too frag- mentany, for determination! yaq- ss) ssee see ae Oe Unconformity. 4, Gravel, composed of angular pebbles one-half inch to four Inchesmmediameter ees seer See ate ee 1:5 Unconformity. 5. Sand, with small amount of thin-bedded clay, yellow, traversed by carbonized roots and root-tubes; prob- ablysanvoldisoils ots 2 oe trey ee ee eee 3 6. Clay, light brown; and sand, fine, clear, in alternate, very thin strata. Clay and sand about equal in AMOUNU See = Soe eS Se it oes Neve pert eee 30 52°75 bo Section IIf. West branch of East branch of Calvo brook. Dip siNyZ22 Feet 1 aGravel;andisand;\ surface wash 22 52 soe oe soe Gravel and fine sand, grey; loosely cemented by iron. 2 Unconformity. Clay, very thin, even layers, forming strata three inches to one foot thick, which alternate with strata of fine sand. A few decomposed bones ---- 2-25 -252522= 10 Volcanic ash, white, microscopically fine-grained, thin- bedded, porous. ‘Locally called“ chalk” 22222 = aaa Clay and sand, very thin layers grouped in alternate Strata some Soe Saeco ee ae eee ene 8 Ayusbamba (Peru) Fossil Beds. 139 Notes on measured sections.—The individual clay beds are nowhere thick and at all places are interbedded with fine sand. At the north edge of the basin, where 135 feet of lacustrine deposits are exposed, the section consists of alternating layers, one-tenth of an inch to four inches thick, of fine sand and banded, dark red clay in the midst of which occur three lenses of coarse sand. The layers of sand and clay interleave as lenses fifty to two hundred feet long. Minute folds and faults further interrupt the regularity of bedding. The clays are red, pink, or brown in tone, due to their content of ferric iron in hydrated form, and are believed to owe their origin to the decomposition of the surrounding pre-Cretaceous sandstones. No “clay dogs” or concretions were observed, but a few small, light-colored patches in the clays are highly calcareous in marked contrast to the clay in general. Apparently the process of segregation is in operation at the present time. The regularly banded white ash, interstratified with the clays, decomposes under slight pressure to an impalpable powder which has a gritty “feel.” Microscopical study of this rock revealed the presence of glass arranged as threads and cusps and hooks, and constituting about 90 per cent of the mass; laths of plagioclase and frayed ribbons of biotite complete the list of component minerals, and determine the classification of the deposit as dacitic ash. No true peat or coal was found among the beds, but at sev- eral localities thin bands of an earthy mixture of sand, clay, and vegetable fragments were noted. This mixture was found by microscopic examination to consist of minute shreds of glass, tiny feldspar, quartz, biotite and muscovite fragments, together with broken bits of voleanic ash and portions of cal- cite crystals. The plant remains present were too fragmentary for determination. The remains of vertebrates are found among the upper grav- els, on floors of tiny ravines, and embedded in clay or sand layers. The content of the calcareous sand lenses (No. 9, Sec- tion I), commonly a mixture of sand, lime, and clay, is in places fully one half bone fragments. Bones are widely scattered horizontally, and in a given locality may be found unrelated parts of a skeleton mingled with bones of animals belonging to entirely different species. No entire skeletons were found in place, and it is probable that several feet of strata intervene between portions of the same carcass. The species represented by parts of skeletons are listed and described by Dr. Eaton (pp. 144-154 following). Field evideuce justifies the conclusion that bones of animals which died on shore or floundered in bogs or guicksand were redistributed by surface wash and running water. 140 HH. EF. Gregory—Ayusbamba Fossil Beds. Origin and character of Lake Ayushamba.—From the sec- tions and descriptions given above it is evident that we are dealing with deposits which in part are truly lJacustrine,— material laid down in a body of quiet water which existed for a relatively long period of time. On the assumption that each layer of sand and of clay represents the amount deposited during a single rainy season, approximately 100,000 years would be required for the accumulation of the materials ex- posed in the present fragmentary sections. It is probable that the lake beds had oreater thickness. That they formerly extended much farther northward is shown by an unprotected, truncated section including 185 feet of strata perched high on the valley side overlooking - the Chipura river. The containing wall of rock is complete except on the north side, where it has been entirely removed. Unlike the remaining portions of the rim, the north wall probably consisted of fluvial and glacial débris washed from the highlands. Moraines extend to the edge of the present lake deposits and may have formed the original barrier. During the life of the lake, fans from the high ridge at the south encroached upon its waters and sepa- rated the original sheet into more or less detached bogs. This process, combined with the development of a channel through the unconsolidated northern barrier, led to the extinction of the water body. Judging from the physical data at hand, this mountain tarn may have beautified the landscape of the late Pliocene or any portion of the early Pleistocene epoch,—a con- clusion which is in harmony with the paleontological evidence.* Ancient Lake Ayusbamba is not an isolated case of extinet water bodies in Peru. Duefiast speaks of similar deposits near the pueblo of Paruro, and the unpublished reports of engi- neers and travelers indicate the existence of unexplored Tertiary and Pleistocene deposits scattered over the Andean highlands. * See the following article by Dr. George F. Haton, pp. 141 to 154. + Loe. cit. Eaton —~ Vertebrate Fossils from Ayusbamba, Peru. 141 Art. [X.— Vertebrate Fossils from Ayusbamba, Peru; by GrorcE F. Eatron.* With Plates V, VI, VII. WHILE engaged upon the field-work of the Peruvian Expedi- tion of 1912, I had the pleasure of accompanying Dr. Albert Giesecke, President of the University of Cuzco, on a hasty visit to a locality near Ayusbamba, among the mountains about thirty miles south of Cuzco, where he had previously obtained some fragmentary vertebrate fossils. On this occasion a few hours only could be spent in the field; but as it seemed prob- able that by going over the ground carefully, further material might be secured, Professor Bingham, the Director of the Expedition, gave his consent to my inaking another visit to the locality, this time in company with Professor Gregory, the geologist, Mr. K. C. Heald, assistant topographer, and Mr. C. Duque. Although other important work caused the post- ponement of this trip until the middle of November, when the rainy season, unfavorable to fossil-hunting in the moun- tains, had set in, we were able to obtain material of consider- able value during the brief time our party was in the field. To Sr. Emeterio Calvo, the master of the delightful hacienda Ayusbamba, I owe my sincere thanks, not only for permission to collect upon his land, but also for the generous hospitality extended to our party. The fossils that form the subject of these notes occurred at an altitude of about 12,400 feet, in gravel and clay beds and in surface-wash along the ponthenn margin of a small lake, the original contours “of which are partially indicated in the map (fig. 8) accompanying the preceding article by Professor Gregory. An excellent view, looking southerly across the fossil grounds, appears as fig. 6 of Professor Gregory’s report. Very nearly the reverse view (N. 5° E. Mag.) taken’ from the south rim of the basin, is shown in text-figure 1 of the present article. After taking ‘this photograph the camera was turned a little to the right (N. 30° E.) and the view shown in text- figure 2 was taken. These two views together cover practically all of the fossil grounds. Almost at the exact center of Professor Gregory's fie. 6, a mastodon’s blade bone was found. The nearer view of this spot (text-figure 3) is equally charac- teristic of several other places where fossils occurred. In this photograph, taken while waiting for the protective jacket of burlap and plaster to dry, appears one of our faithful arrieros whose interest in collecting fossils made him very helpful. With few exceptions, fractured and dissociated material only * Osteologist of the Peruvian Expedition of 1912. Am. Jour. Sci.—FourtH Series, Vout. XX XVII, No. 218.—Frpruary, 1914. 11 142 Luton— Vertebrate Fossils from Ayusbamba, Peru. was found. This indicates clearly that, in most cases, the bones of animals, perishing near the borders of this ancient lake, were widely scattered before being finally covered by alluvium from the neighboring heights. It is possible that Fig. 1. Fic. 1. View N. 5° E. over Ayusbamba fossil beds from the southern margin of the basin. some of this vertebrate material may have been originally embedded at a slightly higher level. I see no reason, however, to question the contemporaneity of the extinct species of animals represented in this collection. No vertebrate fossils had been previously described from this part of the Peruvian Andes. Therefore every recognizable specimen, that might add to our Eaton— Vertebrate Fossils from Ayusbamba, Peru. 148 knowledge of the extinct fauna of the region, was collected, without regard to its perfection and availability for exhibition. Nearly all of this material has now been identified, and it is found that specimens can be referred to five mammalian Fig. 2. Fie. 2. View N. 30° E. over Ayusbamba fossil beds from the southern margin of the basin. 5; families, the Camelidee, Cervidee, Equidee, Elephantide, and Mylodontide. No fossil remains of Rodents and Carnivores were observed. The material may be assigned to the follow- ing genera: 144 Haton— Vertebrate Fossils from Ayusbamba, Peru. Lama sp. The fifth lumbar vertebra and portions of the right ilium and ischium. The specimens compare closely with the corre- sponding skeletal parts of a medium-sized animal of the recent species, Lama huanachus. It is, of course, impossible to assign the present specimens definitely to any one of the several species of this genus that have been described from the Phocene and Pleistocene of Argentina and Brazil. Fic. 3. Fic. 3. A mastodon’s scapula ready for removal, Ayusbamba, Odocoileus brachyceros. The material from Ayusbamba referable to this genus and © species includes a fifth cervical vertebra belonging to an animal considerably smaller than a full-grown Virginia deer, and also a number of fragments of antlers, picked up on the surface of the lake-beds. Fortunately the basal portion of one of the antlers has been preserved. No upright snag rises from the base of the inner side of the beam, and the antlers fork near the burr, the basal portion being extremely short. These characters, together with the texture of the surface, serve to identify the Ayusbamba specimens with Odocodleus Eaton— Vertebrate Fossils from Ayusbamba, Peru. 145 brachyceros as defined by Professor Lydekker,* who states: “ It does not appear that this deer comes close to any existing species.” Philippi has caused some confusion by giving the name Cervus brachyceros to three individuals of the Venado de Cajamarca, which he proposed to separate, under this name, from the recent species of Andean deer, Odocoileus antisiensis = Cervus antisiensis.t. It is apparently far from Philippi’s intention to convey the idea that the Pleistocene Odocozleus brachyceros has persisted until the present time. The antlers of this fossil species are quite different from those of the recent deer to which a similar specific name has been thus unfortunately assigned. Dibelodon bolivianus. Remains of Dibelodon were the most abundant fossils at Ayusbamba, bones and teeth, usually dissociated and incom- plete, occurring at various places in the beds of clay and gravel, and also superficially. It is significant of the history of these deposits that only small and compact bones should have been preserved entire. The individuals of the species, whose remains were found here, differed considerably in size, . and it appears from the dentition and from the condition of the epiphyses that many of these animals had not attained their full growth. An accurate comparison of their mature size with that of other mastodons cannot therefore be made. The maximum stature indicated by the largest bones is not more than three-quarters as great as that of the medium-sized example of Mammut americanum, whose mounted skeleton is exhibited in the Peabody Museum‘ of Yale University. The height of this animal, taken at the shoulder, is about 8 feet and 3 inches. Of the six South American species of Mastodon described by Ameghino,t the two most generally recognized are M.andium, Cuv.,and M.humboldti, uv. According to Pom- peckj§ there are two other valid species, namely, J/. bolive- anus, Philippi(emend. Pompeckj), and JZ. chilensis, Philippi. While Pompeckj does not adopt Cope’s separation of Dzbelodon and Tetrabelodon from Mastodon, it should be understood that both J. bolivianus and M. chilensis, as wellas I. andium and M. humboldti, belong to the Dibelodont division of the original genus. * Paleontologia Argentina, II, p. 79. + Anales del Museo Nacional de Chili, 1894, Entr. 7, Primera Seccion, p. 5. ¢ Mamiferos Fosiles de la Republica Argentina. §$ Mastodon-Reste aus dem inter-andinen Hochland von Bolivia, Paleon- tographica, vol. lii, 1905. 146 LHaton— Vertebrate Fossils from Ayushamba, Peru. Most of the mastodon bones found at Ayusbamba present no characters that can be utilized in the identification of the species. As might be expected, it is best to rely principally upon the form and structure of the teeth. Two specimens have therefore been selected to illustrate the type of dentition. From the piece of a right (upper) tusk shown in Plate V, figures 1 and 2, the following measurements have been taken: Circumference of proximal end...--.-- 354mm Breadth of ‘enamel band==_-_2--2--2-- 90 Maximum diameter at proximal end__-- 117 Minimum diameter at proximal end.__- 104 Ane lezO fs tOrslOM see eee ee The last of these measurements requires a few words of explanation. A mastodon’s tusk upon which the enamel band, or other line of growth, describes a helix, presents a certain analogy to a screw; and the torsion of ‘such a tusk may properly be regarded as a pitch angle. The angle of torsion, used here, is defined as the angle that a tangent to the spiral enamel band, at any point, makes with the adjacent axial line of the tusk. This measurement is easily taken. It has the further advantage of being independent of length, diameter and circumference; and may prove to be of considerable taxonomic value. The mere statement that an enamel band makes one revolution, or one-half revolution, about the axis, without any record of length and diameter, does not adequately describe the torsion. The form of the tusk of J/. bolivianus is described by Pom- peck]* as follows: ‘* Der Querschnitt ist vollkommen elliptisch ; die Durchmesser sind an der hinteren Bruchfliche 8-9 :6-2™, an der vordere 6:2: 4°7°™. Das z. T. abgefallene Schmelzband hat vorne eine Breite von 5°65.” In another placet he com- pares this tusk with that of M. andium: “ Die schlanken, sehr wenig gekriimmten, in Schraubenspirale von ungefihr 1/2 Windung gedrehten Stossziihne haben bei Mastodon bolivianus stark komprimier ten elliptischen Querschnitt, nicht den kreis- runden der Zihne von Mastodon Andium.” That the enamel band is very broad in J/. bolivianus is evi- dent from Pompeckj’s measurements of the anterior portion of a tusk and also from his illustration of a basal fragment, where the enamel appears to cover one-third of the circumference. The band is a trifle narrower in the fragment of tusk from Ayusbamba, which differs also from J/. bolivianus in the ratio of the maximum and minimum diameters. To comply with tine diametrie ratio (8°9 : 6-2) obtained from JZ. bolivianus, the * Op. cit., p. 38. t Op. cit., p. 44. + Op. cit., Taf. IV, Fig. 56. Eaton— Vertebrate Fossils from Ayushamba, Peru. 147 Ayusbamba specimen having a maximum diameter of 117", would require a minimum diameter of only 82", instead of 104"™, as recorded. Yet its section, as shown in fig. 5, is still very far from circular, and cannot be described by the same terms that Pompeckj uses in reference to the tusks of J/. andium. The tusk of the latter species, figured by Ameghino,* greatly exceeds the Ayusbamba specimen in diameter, and possesses an enamel band relatively much narrower. Its curva- ture and the angle of torsion of the enamel band are approxi- mately the same as those of the Ayusbamba specimen. The torsion of the tusk of JZ. bolzvianus is not recorded in such a way as to be available for comparison. Views of the grinding and labial surfaces of an unworn molar from Ayusbamba are shown in Plate V, figures 3 and 4. This tooth, imperfect anteriorly, is presumably a lower molar of the left side. Large accessory tubercles rest like buttresses against the outer columns, but the tubercles that flank the inner columns, and those that lie in the depths of the inner valleys, are so small as to be almost hidden beneath the thin layer of cement. As a result of this arrangement, the tooth would be characterized, during early stages of wear, by the presence of trefoils on the outer columns only, while the inner valleys would remain open. Although this tooth agrees, in the foregoing characters, with the specific definitions of J/. andium given by Lydekkert and by Ameghino in his work already cited, the correspondence fails entirely when the con- tours of the inner columns are considered. Lydekker states that ‘The form of the dentine-disk on the columns which do not bear tretoils is pear-shaped, with the apex directed toward the adjacent column.” The inner columns of the Ayusbamba molar are as stout where they approach the median cleft as at their lingual borders, and would not, at any stage of wear, assume a pear-shaped outline. In view of this essential differ- ence, I do not think that the Ayusbamba molar should be referred to J/. andium. It compares much more clorcly with the lower molar of J/. bolivianus figured by Pompeckj.t The small accessory tubercles that flank the inner columns and line the inner valleys of this tooth find their counterparts half-con- cealed beneath the cement layer of the Ayusbamba molar. The transverse measurements of the tooth of WV. bolivianus, taken at the ae of the crown, viz., for the Ist cross cusp 74°", for the 2nd 7-75, for the 3rd S-0e, for the 4th 7:8, for the 5th 7-4", borenoed very nearly with the following transverse measurements taken in the same way from the Ayusbamba * Mam. Fos, de la Repub. Argent., p. 640. + Cat. Foss. Mam. Brit. Mus., 1886. POp: city, Dat Til Bigs Ja. 148 Eaton— Vertebrate Fossils from Ayusbamba, Peru. molar: for the first cross cusp 73°, for the 2nd 7:4°™, for the 3rd 7:2". Taking into consideration the characters of this molar tooth and also those of the fragment of tusk, previously described, I am confident that both of these specimens should be referred to Jlastodon bolivianus, or Dibelodon bolivianus, if it is best to use the generic name proposed by Cope. It is not surprising that remains of this species should be found at Ayusbamba, for Ulloma, where the type material of D. bolivianus was obtained, is less than 350 miles distant, being situated at an altitude of about 3,800 meters, in the high Bolivian table-land south of Lake Titicaca. There seems to be no reason to doubt that the same climatic conditions pre- vailed in these two localities during the early Pleistocene. Mylodon sp. The Gravigrade Edentates hold such an important place in the Pleistocene fauna of South America, and have been so closely associated, in their distribution, with the Elephantide and Equidee, that one might expect to find the group repre- sented in the collection from Ayusbamba. Vertebree, pieces of the ribs, including an ossified sternal section with its charac- teristic double articulation, and an ungual phalanx, from this locality, should be referred provisionally to the genus Aylodon, so closely do they resemble the corresponding parts of the mounted skeletons of JZ. robustus and IM. myloides in the American Museum of Natural History. I have no means of determining the species. In this connection may be men- tioned a humerus of Mylodon that was found, not at Ayus- bamba, but within two miles of Cuzco, where the road to Ayusbamba ascends from the Huanearo Valley. The location of the bone beneath well-defined beds of fine and coarse allu- vium, about 28 feet in thickness, is marked in text-figure 4 by the point of an arrow. This humerus, shown in Plate V, figures 5 and 6, differs from the humeri of J/. robustus and M. myloides at the American Museum in not having so prominent a deltoidal tract. The outer tuberosity is also smaller, and the posterior outline, immediately below the head, is more deeply concave. These differences, while slight, are sufficient to indicate that some species of J/ylodon other than M. robustus and M. myloides occurred in this part of the Cordillera. Potsherds and bones of llamas were observed at a depth of 24 feet in the surface stratum at the top of the bank where this humerus was found; but no objects were seen in the lower strata that could, in any way, associate the J/ylodon bone with the period of human inhabitation. Am. Jour. Sci., Vol. XXXVII, Feb., 1914. Plate V. i) 6 Fie. 1. Dibelodon bolivianus, fragment of right tusk. x 0°26. Fic. 2. The same fragment, proximal end. x 0:29. 3. Dibelodon bolivianus, left lower molar. x 0°29. Fic. 4. The same tooth, labial surface. x 0°26. 5. Mylodon sp., right humerus, anterior view. x 0°25. 6, The same specimen, external view. x 0°25, Am. Jour. Sci., Vol. XXXVII, Feb., 1914. Plate VI. Fie. 1. Skull of Parahipparion sp., from Ayusbamba, lateral view. x 0°18. Fie. 2. Skull of Onohippidium compressidens in the Museum of La Plata, after Lydekker. x 0°15. Fie. 8. Skull of Parahipparion sp., from Ayusbamba, posterior view. x 0°25. Am Jour. Sci., Vol. XXXVII, Feb., 1914. Plate VII. 2 Fig. 1. Skull of Parahipparion sp., from Ayusbamba, superior view. x 0°30. Fic. 2. Skull of Parahipparion sp., from Ayusbamba, inferior view. x 0°28. : ‘ Eaton— Vertebrate Fossils from Ayusbamba, Peru. 149 Parahipparion. Equine remains collected at Ayusbamba include several teeth, a fragmentary skull and a portion of the distal end of a femur. The teeth were all surface specimens, and there was nothing about the manner of their occurrence to prove that any of them belonged to the same individual as the skull. Two of Fia. 4. Fie. 4. Location of Mylodon humerus (indicated by arrow-point) by the side of the road from Cuzco to Ayusbamba. them, however, may be from one individual, not necessarily because they were picked up within a few rods of each other, but because they show equivalent structure, size and stage of wear. These two teeth, first and second functional lower 150 Haton— Vertebrate Fossils from Ayusbamba, Peru. cheek teeth (p, and p,) are accordingly represented together in text-figure 5. Owing to their imperfection exact measure- ments cannot be taken; but the length and width of their crowns appear to have been approximately as follows: p,, length 29™™, width 14:2"; p,, length 27", width 16™™. These dimensions and the pattern of the saamel foldings show that the teeth cannot be referred either to Hgwus curvidens or to any of the three species of //ippzdiwm whose teeth have been minutely described and figured by Sefve.* Their affinity lies rather with the genus Parahipparion. In view of the difti- culties that beset the study of isolated equine teeth, especially those of the lower series, it would perhaps be unwise "to attempt the specitic identification of these two lower premolars ; but it is significant that among all the teeth with which they have been” compared, they most nearly find their counterparts in the first and second lower deciduous molars, and the first molar, of Parahipparion peruanum from Tirapata, Peru, fig- ured by Sefve.t The united length of the first two milk molars of this last mentioned specimen amounts to approxi- mately 56™; and this measurement exceeds the restored united lengtht of the two teeth under discussion by 2"", which is very nearly the difference in length that one night expect to find between the first and second lower deciduous teeth and their permanent successors. So many variable and uncertain factors enter into these measurements that the comparison should not be carried too far. In another lower cheek-tooth (p,) from Ayusbamba the forms of the inner valleys (text-figure 6) are simpler than in the teeth last described, the contrast in this respect being almost too marked to be readily accounted for by the fact ‘that the teeth present different stages of wear. It should be noted that the teeth shown in text- fioures 5 and 6 possess, in common, a very deep outer valley that penetrates between the inner val- leys and nearly traverses the crown. The deepening of the outer valley, according to Sefve’s observation, characterizes P. devillec and P. peruanum, and distinguishes them from P. saldiusi. The premolar of text. tizure 6 measures, on the crown, 29°™x14™™_ In its enamel pattern, as well as in its size, it is remarkably like the p, of P. devillez,t which has a length of 28™™ and a width of 13™™. Although the species of co} the. Ayusbamba specimen is not determined by the above com- * Die Fossilen Pferde StidAmerikas ; Kungl. Svenska Vet. Hand., 1912, Band 48, No. 6. + Hyperhippidium ; Kungl. Svenska Vet. Hand., 1910, Band 46, No. 2, Taf. 2, Fig. 4. t The united length of overlapping teeth is less than the sum of their indi- vidual lengths. t Sefve, 1912; Taf. 2, Fig. 21. Eaton— Vertebrate Fossils from Ayusbamba, Peru. 151 parison, there is no doubt that it belongs under the genus Parahipparion. Examination of an upper molar (text-figure 7) leads to about the same conclusions as were obtained from the study of the lower teeth already described. This tooth appears to be either the first or the second upper molar of the right side. Probably it is the first. The length is 21°5™™ and the width 29m". The depth of the crown is S greatly reduced, only about 30™™ of the prism being left. In fact, it is so low as to present an almost senile stage of wear. Under these circumstances not only would the pattern of the enamel foldings be simplified, Hie. 5d. Fig, 6. Fic. 7. CEG TY NC 05 Fic. 5. Pm, and Pm;, Parahipparion sp., Ayusbamba. x 0°80. Fic. 6. Pm», Parahipparion sp., Ayusbamba. x 0°80. Fic. 7. -M!, Parahipparion sp., Ayusbamba. x 0°80. as compared with the condition at four or five years of age, but the antero-posterior diameter, or length, of the crown would be considerably reduced. The width of the crown has probably been a little reduced also, but judging from the very slight convergence of the labial and lingual sides of the prism, this transverse reduction is very small, probably not exceeding 1™™. Making due allowance for advanced stage of wear and for individual variation, it is conceivable that the enamel pat- tern of this tooth might have been derived from the pattern exhibited by the first upper molar of any one of the species of Parahipparion in which the true molars have been observed. These teeth have not been observed in P. peruanum, the species that from its occurrence at Tirapata would seem most 152 Eaton— Vertebrute Fossils from Ayusbamba, Peru. likely to be found in Pleistocene formations at Ayusbamba. On the whole it appears that very little progress toward the specific identification of the upper molar of text-figure 7 is to be made through the study of its enamel pattern alone. Much better results can be obtained by comparing its dimensions with those of the various species of Parahipparion. Accordingly the length and width of the upper true molars of P. saldiasi, P. bolivianum, and P. devillec and the same dimensions of the upper milk molars of P. peruvanum are quoted from Sefve and arranged in the following table. Dimensions of Molars of Parahipparion in millimeters. M! M? M3 Pe eee 4 ice | joa es ayaa an |e | | oO LS ete S| Se Parahipparion burmeisteri- ---- | 29-5 | 30: 29° 29°D || 27 24° IP ssaldiasiasce mem at seers ne. || 29: 29° 26° 26° 82°09 | 22° IPS bolivianume see sae eee 25 | 27: 2 of ou de Pe aevillelea staan = elie eer Po NPs is ae = os ‘Pe Gevalll Giesss ay ee eee 24: | 25° 23 23 24 21 D! D? D3 PM eruan mies eee ene 35:7) | 23: 28 22°5 || 29°2 | 22 The width of the crown of the upper molar from Ayusbamba, viz., 22™™, is less than that recorded, in the table, for m* of any species of Parahipparion. P. devillec comes nearest, with a width of 25™ for m', while P. saldiast and P. bolivianum are much farther removed. | FO ee (1) Ges 0) where F denotes a force, while XY, Y, and Z denote the magni- tudes of its components. (2) Equilibrium of a rigid body. It is shown that a rigid body is capable of having two distinct and independent types of motion, namely, a motion of translation and a motion of rotation. Consequently, a rigid body is capable of being sub- jected to two distinct types of action, one of which relates to translation, while the other is connected with rotation. These * Letters in black type denote vector magnitudes. + In the Newtonian system of Mechanics the term action is used in a sense different from that in advanced Dynamics, therefore there is no danger of confounding this principle with the principle of least action. 162 Dadourian—Progresswe Development of Mechanics. are called linear action and angular action, respectively. Therefore, the action principle states: Sy sb \) =O, (A’) where A, denotes a linear action and A, an angular action. But since A, and A, are independent of each other, the prin- ciple may be stated in two parts: I. The sum of all the external linear actions to which a body or a part of a body is subject at any instant vanishes : SA, =0. (A,) Il. The sum of all the external angular actions to which a body or a part of a body is subject at any instant vanishes ; SA. — 0. (A,) In ease of equilibrium, forces are the only type of linear action which come into play; on the other hand, torques (defined as that action of one body upon another which tends to pro- duce a motion of rotation) are the only type of angular action. Therefore, the following well-known conditions of equilibrium follow, immediately, from the two sections of the action principle: >) GEE Ne 3F=0, 3Y=0, (1) 710) SiGei—01, SG —)0r > G.— 0. (II) > C= (0) 5 where @ denotes a torque, while G,, G,, and @, denote the magnitudes of its components. (8) Motion of translation. Here the concept of linear kinetic reaction is introduced as a linear action which is the counterpart of force and which appears whenever a body is given a linear acceleration. The quantitative definition, or the measure, of this new form of action is illustrated by means of the two following ideal experiments. (a) An apparatus consisting of a spring balance S (fig. 1), an extensible string of great length, and a block B is set up on a perfectly smooth horizontal table 7. The string connects the block with one end of the spring balance, the other end of which is fixed on the table. Suppose, now, two persons to perform the following experiment. One of the experimenters Dadourian— Progressive Development of Mechanics. 163 pulls the block away from the balance, thereby stretching the string; while the other observes the readings of the balance. Since the system is in equilibrium, two equal and opposite forces must act at the two ends of the string. In other words, the pull exerted by the block upon the string equals that regis- tered by the balance. The pull exerted by the block is obvi- ously due to the pull exerted by the person who holds it. When the block is released it is observed that the reading of Bie the balance does not drop to zero at once. In fact the balance registers a pull so long as the length of the string is greater than its natural length. In other words, the block exerts a pull upon the string in spite of the fact that there is no force which tends to pull it to the right, nor a force, like friction, which resists its motion towards the left. In order to account for this action of the block we may take one of the following two points of view. We may attribute it. to the block itself, and using the common mode of expression state, “the block resists the accelerating force of the string and ther eby acts upon the latter because it has inertia.” Evi- dently this amounts to stating, “the block offers a resistance because it has the property of resisting.” In this “explanation” the term imertza acts as a sop to the mind. We can, on the other hand, take the position that the block is just as helpless against acceleration as a person falling in space is against fall- ing: that bodies have the property of interaction but not one of resistance to action. If we adopt this point of view we must suppose that the inertia of the block and its action upon the string is the result of an interaction between the block and the ether (or whatever may take its place in future physical the- ories). In other words, we must suppose that the ether acts upon a body whenever the latter is given an acceleration.* This new type of action, which we will call linear kinetic reac- * Some physicists may consider it a mistake to bring the ether into Dynam- ics. These men should remember that the ether has already entered this field, and that our acceptance of the electromagnetic origin of mass is a very important concession to the claims of the ether for a place in the Science of Dynamics. 164 Dadourian—Progressive Development of Mechanies. tion, differs from forces and torques, which represent interac- tions between matter and matter and not between ether and matter. In order to obtain a quantitative definition of the kinetic reaction, suppose the experiment to be continued in the follow- ing manner. After the block is released one of the experi- menters observes the readings of the balance, while the other records the position of the block at different instants. Then a comparison of the readings of the balance with the correspond- ing accelerations of the block (obtained from the observations of the second experimenter) gives the following result : UR! JE of PA ele fie ele as era (1) where f# denotes the acceleration of the block at the instant when the balance registered the pull #2, while m is a constant which is called the mass of the block. Thus the kinetic reac- tion of a body is proportional to its acceleration. If the blocks be replaced by other blocks and the experiment repeated the same result will be obtained, with the exception that the constant m will, in general, be different for different blocks. If the readings of the balance are compared when different blocks have equal accelerations it will be found that es eee Shia Mey Al (2) 7 Mm, mM, where m,, m,, ete., denote the masses of the blocks, determined in the foregoing manner, while 7 denotes the common acceler- ation. Therefore, the kinetic reaction of different bodies hav- ing equal accelerations are proportional to their masses. It follows from (1) and (2) that the measure of the linear kinetic reaction is given by the relation R, = — m_F, ’ (3) where the subscripts are introduced to emphasize the fact that all three of the magnitudes involved in (3) relate to a motion in which the acceleration is tangential, or, longitudinal, while the negative sign is introduced to take into account the fact that R; and f, are oppositely directed. (b) In the preceding experiment the acceleration was longi- tudinal: we will now consider a case in which it is transverse. Let P (fig. 2) be a particle placed upon a perfectly smooth horizontal table 7, and connected to a spring balance S by means of an inextensible string, which passes through a smooth hole in the center of the table. If the particle be set in motion Dadourian— Progressive Development of Mechanics. 165 so that it will describe a circular path it will be observed that its speed remains constant and that the spring balance registers a pull. It follows from the second observation that the particle exerts a force upon the string and in turn is pulled by the string. Considering the particle alone, we see that it is pulled towards the hole but does not approach it, and this in spite of the fact that no force is acting upon the particle to counter- balance that exerted by the string. In order to account for this fact we must introduce, again, the concept of the kinetic reaction. The measure of this reaction may be obtained in a manner similar to that employed in (a). When different transverse accelerations are given to the particle by changing its speed, the radius of its path, or both: and then these accelerations are compared with the corresponding readings of the spring balance, the following relations will be found : ne Ve inn es i re where /™ is the reading of the balance which corresponds to the acceleration #®, while 7 is a constant which may be called transverse mass. ‘Therefore the transverse kinetic reaction of a particle is proportional to its transverse acceleration. If, on the other hand, the particle is replaced by others and the experiment repeated, the following relations are obtained : R IR Ie a ay (5) where m,, m,, etc., denote the transverse masses of the particles, while #,, /,, etc., denote the readings of the spring balance 166 Dadourian—Progressive Development of Mechanics. which corresponds to a common transverse acceleration //. Thus the transverse kinetic reactions of different particles having equal transverse accelerations are proportional to. their transverse masses. It follows, therefore, that the quantitative definition of the transverse kinetic reaction is given by the relation rn (6) where the subscripts are introduced in order to emphasize the fact that the magnitudes involved in (6) relate to the normal, or transverse, direction. When the same system of units are used to measure R:;, R,, f and f,, it will be found, that m: and m, are equal. Therefore, when a particle has a transverse as well as a longitudinal acceleration, the total linear kinetic reaction is the sum of R and R*. Thus, R, — R- + R.. = — mh, ta) (s | = — M\V-—- — p = — Mv, (7) where v denotes velocity, vr; the tangential component of the acceleration, and v the total acceleration of the particle; while p denotes the radius of curvature of the path, measured from the center of curvature. The linear kinetic reaction and forces are the only types of linear action which a body experiences. Therefore, the first section of the action principle leads to the relations : SF; = R, = (0) 9 or F = R, ==, (IIT) where F denotes the resultant force. The last relation, which is called force equation, is then resolved into two component equations which correspond to the directions parallel to the tangent and to the normal to the path of the particle. Thus and 1) = Se oe te p°? Dadourian—Progresswe Development of Mechanics. 167 where /, #7, and /, denote, respectively, the magnitudes of the resultant force and of its tangential and normal components. (4) Motion of rotation. Here the concept of angular kin- etie reaction is introduced as the counterpart of torques, which come into play whenever a body is given an angular accelera- tion. The quantitative definition of this new type of action is obtained in a way quite analogous to the manner in which the measure of the linear kinetic reaction was obtained. Consider a flywheel, which is perfectly balanced and _ per- fectly free to rotate about a horizontal axis. In order to start the flywheel in motion or to stop it a torque must be applied. Therefore the flywheel experiences an angular kinetic reac- tion when it is given an angular acceleration. If the angular acceleration of the flywheel is observed at different instants and simultaneous observations are made of an arrangement which measures the kinetic reaction it will be found that the following relations hold : IBD IRR We : — pS Se Sd, (8) . where w denotes the angular acceleration, /2, the correspond- ing kinetic reaction, and Z a constant which is called moment of mertia or angular inertia and which plays a role analogous to that played by the mass of a body in motion of translation. Thus the angular kinetic reaction of a body rotating about a given axis is proportional to the angular acceleration. If, on the other hand, equal angular accelerations are given to different bodies or to the same body about different axes and the moments of inertia determined in the foregoing manner the following relations will be found : Tae ‘Ri! ei Ti a ee S11 eM, (9) where denotes the common angular acceleration. There- fore, the angular kinetic reactions of different bodies, or of the same body relative to different axes, are proportional to the corresponding moments of inertia. It follows therefore that the following relation defines the angular kinetic reaction : Ds Ie, (10) where the negative sign is introduced to take into account the fact that R,, and @ are oppositely directed. The angular kinetic reaction and torques are the only types of angular actions which come into play; therefore the second section of the action principle states : 168 Dadourian—Progressive Development of Mechanics. 3G: +R.=0, or G =—R. LR (IV) where G denotes the resultant torque. By comparing the expression for the moments of the kinetic reactions of the elements of a rotating body with equation (IV) it may be shown that the foregoing experimental definition of moment of inertia is equivalent to the common analytical definition. In this treatment of the science of mechanics the principles of the conservation of dynamical energy, of linear momentum and of angular momentum are derived from the action princi- ple and are used as supplementary Poe The form of the statement of the action principle and the way in which its significance is developed do away with the ditiiculties which arise from confounding the linear kinetic reaction with a force and the angular kinetic reaction with a torque. A clear-cut distinction is made between force and torque on the one hand and the two forms of kinetic reaction on the other. Zhe former represent interactions between matter and matter while the latter represent interactions between ether and matter. On this view inertia is not a resistance which bodies offer to an accelerating force. The conception of two gravitating particles attracting each other and at the same time resisting attraction, imparting accele- ration yet offering resistance to acceleration, is obtained by analogy from tug-of-war. But the analogy does not hold good unless something is introduced into the conception of the eravitating particles which will play the same roll as the com- mon ground on which the men playing tug-of-war stand. Without such a common ground there can be no tug-of-war. We can not “ without destroying the clearness of our concep- tions take the effect in inertia twice into account—first as mass and secondly as force.” The point of view outlined here com- pletes the analogy by introducing the necessary “ common ground.” In addition to clarifying the fundamental principle, the development outlined in this paper unifies the presentation of the science of Mechanics to a great extent, makes it progressive and graded. A tremendous gain is thus made in economy of effort. The following derivation of Newton’s third law of motion illustrates the simplicity with which dynamical laws and theorems can be derived from the action principle. Consider two interacting particles: Let m, and m, be the masses, and v, and v, the accelerations of the particles. Fur- Dadourian— Progressive Development of Mechanics. 169 ther let F,, denote the force on m, due to m, and F,, the force on m, due to m,. Then if the two particles are considered as one system the only external actions are their kinetic reactions, therefore the action principle gives MN, + MN, — 9, (1) V ; or wi =— —, (2) which states that the mutual acceleration of two particles are oppositely directed and are inversely proportional to their masses. When the particles are regarded as separate systems, then the action principle gives E an mV, = 0 (3) and F., + mV, = 0. (4) From the last two equations and equation (1) we obtain F,, aaa FE ’ (5) which is Newton’s third law of motion. Sloane Physical Laboratory, Yale University, New Haven, Conn. 170 T. L. Walker—Temiskamite. Arr. XI].—TZemiskamite, a new nickel arsenide from Ontario, by T. L. Watxer. A FEw weeks ago the Royal Ontario Museum of Mineralogy received some specimens of niccolite from the Moose Horn Mine, Elk Lake, Ontario. On examining some of the mate- rial it was observed that much of the supposed niccolite was much paler in color than normal niccolite. The examinations thus initiated are detailed in the following paragraphs. Paragenesis.—This mineral occurs in calcite veins carrying niccolite and smaller amounts of native bismuth and silver. Rr Gales 4 The new mineral appears to be one of the first to form. It is bordered and fringed by niccolite and bismuth, while the cal- cite appears to be later, as it fills in all the space between the branching masses of the other minerals. Physical properties —Color, silver-white with a touch of red, tarnishing very quickly to madder gray, and after pro- longed exposure to tints resembling those of bornite. Hard- ness, 5°5; specific gravity, 7-901; fusibility, 2; luster, metal- lic; opaque; not magnetic ; streak, brownish black. Crystallogra aphy. —This mineral occurs in radiating fibrous masses suggesting rammelsbergite in structure. No erystal- lized material has yet been obtained. Traces of cleavage are seen upon examining polished surfaces with the microscope; T. L. Walker—Temiskamite. Wel The radiating fibrous masses are built up into arborescent botryoidal forms, the outer surface sometimes covered with niccolite or bismuth or by a very thin layer of some undeter- mined black mineral resembling native arsenic (figure 1). Chemical properties.—In strong nitrie acid the mineral is attacked with violence accompanied by the evolution of red fumes and the separation of a very small amount of sulphur, which later oxidizes so that complete solution results. In sul- phurie acid the decomposition is slower while in hydrochloric acid the mineral is only very slowly dissolved. Closed tube, a very slight deposit of crystallized arsenious oxide; open tube, abundant deposit of the same oxide, the oxidized residue oreenish in color; plaster cast, coating for arsenic and in smaller degree bismuth ; charcoal alone, the easily fusible mineral melts to a br ight bead and oxidizes, covering the char- coal with arsenious acid, the bead after long treatment being magnetic. Chemical analysis.—Several grams of this mineral appar- ently free from the associated minerals was selected, ground finely, dried at 110° C. and analyzed with the following result : Nickel sy 49-07% - 58°7 = ‘8359 | oa.g @obalt tenet 1:73% + 59° = 0293 ( °°°* Ones se es trace PAT SenIC so oie ee 46°34% + 75° = °6179 t +6500 Sulplury ses s2 os - ORGS Se Be a ROL ANTULNOM SoA not determined Bismuth sam 55% 98-724 °8652 + 2163 = 4 ratio of Ni + Co, 6500 + 2163 = 3:005 ratio of As + S. From the above analysis and calculation it seems very prob- able that this is a new mineral whose chemical composition may be expressed by the formula (Ni,Co),.(As,S), or, apart from very small proportions of isomorphous elements, Ni, As,. The small amount of bismuth, it will be observed, is not included in the above calculation, as it seems probable that it occurs in the form of small particles of native bismuth included in the mineral. Although this mineral has not yet been found crystallized it seems reasonable to regard it as a new species, first, because of the exact ratio resulting from the analysis and calculation, and second, because we do not know any minerals which could be mixed so as to produce an aggregate of the above composition. Millerite contains as high a percentage of nickel but in it the 172 T. L. Walker— Temiskamite. sulphur percentage is very high. In the classification this new mineral belongs to the dyser asite group and to the basic division of the “sulphides. * The mine from which the material was obtained lies some distance to the west of Cobalt, but as the mineral associations are the same in both camps it seems probable that this new ~ mineral should be found in the Cobalt mines if carefully looked for. To indicate the region from which the first specimens were obtained I suggest that this mineral be called Temiskamite from the district of Temiskaming in northern Ontario. I wish to thank my assistant, Mr. H. V. Ellsworth, for assist- ance in connection with the chemical analyses. Royal Ontario Museum of Mineralogy, Toronto, Ontario, December, 1913. * Since writing the above, my attention has been ealled to the new mineral Maucherite, recently described. This mineral has the chemical composition represented by Nis;As2 and is apparently closely related to Temiskamite, ! Thornton—Separation of Titanium from Iron. 173 Arr. XIIL.—The Use of the Ammonium Salt of Nitrosophe- nylhydroxylamine (“Cupferron” ) in the Quantitative Separation of Titanium from Iron; by Wit11am M. THORNTON, JR. [Contributions from the Kent Chemical Laboratory of Yale Univ.—ccliv. ] NIrRosOPHENYLHYDROXYLAMINE was first synthesized by Wohl.* The ammonium salt of nitrosophenylhydroxylamine was brought into service in analytical chemistry by Baudischt for the estimation of either copper or iron, the separation of these two metals from various others, and indirectly for the separation of the former from the latter. Owing to these properties, the trivial name of ‘ Cupferron” has been applied to the ammonium salt of nitrosophenylhydroxylamine. The analytical data given by Baudisch are few and not absolutely confirmatory. Since then, however, various other workers (notably Biltz and Hodtke,t Hanus and Soukup,§ and Frese- nius|) have thoroughly demonstrated the value of this reagent for the quantitative precipitation of either copper or iron and their separation from various other bodies. In connection with other work, Schroeder has made the statement** that titanium and zirconium could be quantitatively precipitated from their acid solutions by the “cupferron” reagent, and that experiments were in progress for the estimation of these two elements. Schroeder, however, gave no experimental data and has not published further upon the subject. Bellucci and Grassitt+ have shown that from solutions, moderately aciditied with either sulphuric or hydrochloric acid, titanium could be quantitatively precipitated by the “cupferron” reagent, and that titanium could also, under these conditions, be quantita- tively separated from aluminum. Under these circumstances the titanium comes down as a very bulky, readily filterable, precipitate of canary yellow color. In the opinion of the afore- said authors the precipitate, after having been crystallized from ethyl alcohol, is the titanic salt of nitrosophenylhydroxylamine, CH,-N=N-—O corresponding to the formula | ° lI Ti: O * Ber. chem. Ges., xxvii, 1435, 1894. + Chem. Zeitung, xxxiii, 1298, 1909. ¢Zeitschr. anorg. Chem., lxvi, 426, 1910. § Ibid., Ixviii, 52, 1910. || Zeitschr. anal. Chem., 1, 35, 1911. “| Zeitschr. anorg. Chem., Ixxii, 89, 1911. ** Loc. cit., page 95. ++ Gazetta Chimica Italiana, Anno xliii, parte I, 570. 1913. Am. Jour. Scr.—Fourts Series, Vout. XX XVII, No. 218.—Ferpruary, 1914. 13 174 Thornton—Separation of Titanium from Tron. It has been known fora long time that certain organic acids, containing both hydroxyl and carboxyl groups, such as tartaric acid and citric acid, have the power to prevent the precipitation of certain metals when their solutions are made alkaline with sodium or potassium or ammonium hydroxide. This principle - was made use of by Gooch* for the separation of titanium from iron; for, if the solution contain sufficient tartaric acid, the iron can be precipitated by ammonium sulphide as ferrous sulphideand the titanium will then be found entirely in the iron free filtrate. The next step is to oxidize the tartaric acid; for titanium is not precip- itated in its presence by any of the reagents previously used for its gravimetric estimation. This was accomplished by Gooch,t+ after strongly acidifying with sulphuric acid by adding potas- sium permanganate to the boiling aqueous solution. This process is open to the objection that a great deal of manganese is thus introduced into the solution and is coprecipitated in some measure when the titanium is subsequently thrown down by hydrolysis of titanic acetate. A second precipitation is therefore necessary, which must be preceded by fusion with an appropriate flux and solution of the melt in acid. The author’s experiments show that, after acidifying the filtrate from the ferrous sulphide, the titanium can be quantitatively precipitated by the “cupferron” reagent, notwithstanding the presence of tartaric acid. Two solutions of titanic sulphate were employed for these experiments ; which were prepared by warming potassium fluotitanate with concentrated sulphuric acid until all the hydrofluoric acid had been volatilized, pouring into cold water, and making up to known volume. The quantity of sulphuric acid used was such that the resulting solution contained about 10 per cent of absolute acid. In the case of the second solu- tion, the trace of platinum was removed by saturating the solution with hydrogen sulphide, filtering off the platinic sul- phide, boiling out the hydrogen sulphide, filtermg again, and making up the solution to definite volume. The first solution was standardized by taking weighed portions of 25°™* and pre- cipitating the titanium by hydrolysis of the acetate. The solu- tion was made nearly neutral with redistilled ammonium hydroxide—until a faint permanent turbidity appeared. One em* of a strong solution of ammonium hydrogen sulphite was added, followed by 15 grams of ammonium acetate and 20 grams of glacial acetic acid, and the solution made up to 400™*. This solution was brought rapidly to boiling and maintained in ebul- lition for one minute. The precipitate was washed twenty times—first with boiling 5 per cent acetic acid, and finally * Proc. Am. Acad. Arts and Sci., n.s., vol. xii, p. 455; Chem. News, lii, 59, 68. ft Loe. cit., p. 440. Thornton—Separation of Titanium from Iron. 175 with boiling water. In the usual manner the precipitate was ignited to titanic oxide and brought to constant weight over the Meker burner. Duplicate determinations gave the follow- ing results: Titanic oxide Titanic sulphate soln. Baas Cre ()) POS SBOE) aks Bae eae ee ee ae 0: 1427 grm. 0°5226% AD) te oc es pollOLonms: aa) 2 22m es) Ocl4 28 orm 0522 1% Since these two values agreed so closely, (b) was arbitrarily taken as correct. The second solution was standardized by taking two weighed portions of 25° and 24°™* respectively and determining the titanium in one (a) by the acetate method given above and in the other (b) by the “cupferron” method of Bellucci and Grassi* (the exact technique of which will be given presently). Duplicate determinations gave the following results : Titanic sulphate soln. Uitectitlo Onsatels (a) 25°"? c= 97-814 grms, 9.2... -..- '0°1066 grm. 038324 (b) 24°™* ——— 96°667 orms. ..-..__.---- 9°1022 grm. 0:3882% Since these two determinations agreed exactly, the value here obtained was taken as correct. A solution of ferric sulphate was prepared by dissolving Baker’s analyzed ferric ammonium sulphate in water, adding 25°™* of concentrated sulphuric acid per liter to prevent the formation of basic salt, filtering, and making up to definite volume. In one portion (a) of 25°" the iron was determined by titration with potassium permanganate after reduction by zinc —the potassium permanganate having been previously stand- ardized against sodium oxalate.t In another portion (b) of 25°" the iron was determined by precipitation with redis- tilled ammonium hydroxide in a platinum basin and ignition of the precipitate to ferric oxide. Parallel determination gave the following results : Ferric salphate soln. Ferric oxide CNY. Ieee ee a 5d Os ae ee ee 0:2267 grm. (CLD) kee oS Aes a ee Re ara 2's oh | een OR Lg Sn a 0°2269 orm. The value obtained in (a) by the volumetric method was arbi- trarily taken as correct. * Loc. cit. + The sodium oxalate was obtained from the Bureau of Standards, Wash- ington, D.C. 176 Thornton—Separation of Titanium from Tron. The supply of ‘“‘cupferron” for these experiments was pre- pared in this laboratory according to the directions given by Baudisch.* An approximately 6 per cent solution of the salt was made by dissolving it in cold water and filtering from any insoluble residue that remained. The first series of experiments was carried out with a view to ascertaining whether or not titanium could be completely precipitated and accurately determined in the presence of tar- taric acid. To a solution containing a known quantity of titan- ium a little more than three times the weight of the titanic oxide present was added of tartaric acid. The solution was made neutral to litmus with ammonium hydroxide, then acid again with 5°™* of sulphuric acid (made by diluting acid of sp.g. = 1:84 with an equal volume of water), and the volume made up to 200°", A little more than the calculated amount of “cupferron” solution was added and the beaker set aside for the precipitate to settle. The supernatent liquid was tested by adding a few drops of the reagent, which were made to run down the wall of the beaker. The formation of a white pre- cipitate of nitrosophenylhydroxylamine indicates that the re- agent has been added in excess, while the formation of a yellow turbidity shows that the titanium had not been completely thrown out. It is well also to test the filtrate. The precipi- tate was filtered on paper using very gentle suction and washed twenty times with cold water. * During the washing the suction should be almost stopped to prevent the wash water from run- ning through too fast to accomplish much solvent work. The precipitate is very prone to develop mud cracks and should therefore be agitated with the stream of water from the wash bottle as much as possible. After having been sucked free from drainage water, the precipitate along with the filter was placed in a tared platinum crucible, dried at 110° C., very carefully heated until the volatile products of destructive distil- lation had escaped, the inclined open crucible ignited till all earbon had been consumed, and finally brought to constant weight over the Meker burner. If it is desired to save time, the ‘precipitate can be dried by inclining the crucible, support- ing the lid tongue downward on the triangle and edge of the crucible, and applying a small flame beneath the lid. ‘The heat is thus deflected downward and the precipitate gradually dried from above with little danger of spattering.t It is unadvis- able to dry the precipitate in the funnel; for at a low tempera- ture the substance melts, or at least assumes a plastic condition, and penetrates the pores of the filter; moreover, the dried * Chem. Zeitung, xxxv, 913, 1911. +See Analytical Chemistry, Treadwell [translated by Hall], vol. ii, p. 29, 1910. Thornton—Separation of Titanium from Lron. 177 material is very brittle and on attempting to fold the paper and introduce it into the crucible particles are likely to fly away and be lost. In experiments (1) and (4) the filtration was earried out on asbestos in the perforated platinum crucible. Due to the above mentioned property, a little titanic oxide was found on the outer surface of the crucible after ignition. This method of filtration was therefore abandoned. Table I con- tains the results of three experiments. TABLE I. TiO, TiO. Tartaric No. taken found Error acid grm. germ, grm. grm, Ik 0°1498 0°1426 —0:°0002 ae De 0°1428 0°1429 + 0°0001 0°5 Bi 0°1066 0°1063 —0:0003 0°4 In the second series of experiments actual separation of titan- ium from iron was carried out. To facilitate the reduction of the iron the solution, to which tartaric acid equal to three times the weight of the titanic oxide and ferric oxide present had been added, was neutralized by ammonium hydroxide, 1 to 2° of sulphuric acid (1:1) added, and the volume made up to about 100%. Hydrogen sulphide was then introduced till the solution appeared colorless. If the iron is not thus reduced before its precipitation, titanium will be thrown down in part also.* The solution was then made ammoniacal and more hydrogen sulphide introduced, until the iron had been com- pletely precipitated as ferrous sulphide—leaving the solution, however, alkaline to test-paper. The ferrous sulphide was filtered off and washed ten times with very dilute colorless ammonium sulphide. The filtrate was acidified with 25™ of sulphuric acid (1:1), the hydrogen sulphide boiled out, the acid partially neutralized with ammonium hydroxide so as to leave about 2°5°™* of sulphuric acid (1:1) for every 100°™° of the solution, and the ‘“‘cupferron” added in the cold. Table IT contains the results of four experiments. TABLE ITI. TiO, FeO; TiO, No. taken taken found Error grm. grm. grm. erm, 4, 0°1428 0'2267 0°1424 —0°0004 5: 0°1428 0°2267 0°1430 + 0:0002 6 0°1066 0°2267 0°1068 + 0:0002 7 0°1063 0°2267 0°1061 —0:0002 * Cathrein, A., Zeitschr. Kryst., vi, 248, 1882; vii, 250, 1883. 178 Thornton—Separation of Titanium from Tron. In the third series of experiments the solution containing the titanium and iron was divided into two aliquot parts by weight. In one part the titanium was determined by the method already outlined. In the other part the iron was deter- mined by the method of Gooch and Newton.* Tabie III con- tains the results of two experiments. TaBLe III. TiO. Fe.Os3 TiO. Fe.0O3 Error Error No. taken taken found found TiO, Fe. Os; grm. erm, grm. germ. grm. grm. 8. 0°0716 0°11380 0:0719 0°11380 + 0:00038 0°0000 9. 0:0717 0°1129 0:0716 0°11385 —1°0001 +0:0006 From the work of Bellucci and Grassi,+ Fresenius,t and the author, it would seem that there should be no great difficulty attending the separation of titanium from iron, aluminum, and phosphoric acid, which are the substances with which titanium is commonly associated in nature. Experiments are now in progress by the writer for accomplishing this separation with the aid of the “ cupferron” reagent, and the results will appears in a later issue of this journal. 6 January, 1914. * This Journal, xxiii, 365, 1907. + Loe. cit. t Loe. cit. W. E&. Ford—Optical Study of the Amphiboles. 179 Arr. XIV.—A Contribution to the Optical Study of the Amphiboles ; by W. E. Forp. Tue late Prof. S. L. Penfield was engaged at the time of his death upon an investigation of the chemical composition of the minerals of the Amphibole Group. A series of eleven analyses for this work had been made under his guidance, by Dr. F. C. Stanley, of amphiboles which ranged in composition from trem- olite to hornblende. The specimens selected for analysis afforded unexcelled material for chemical investigation, and it is safe to say that few such series of authoritative analyses of the amphi- bole minerals can be found elsewhere. The article, as far as the discussion of the composition of the amphiboles was con- cerned, was practically complete at the time of Prof. Penfield’s death and was subsequently published in this Journal.* It had been the intention of Prof. Penfield to supplement the discussion of the chemical composition of these amphiboles by a study of their optical properties. Unfortunately he was not able to do this, but it has seemed of considerable importance that this work should be done. In the majority of cases the identical specimens from which material for the analyses was taken were preserved in the Brush Mineral Collection. In the few cases where it was impossible to positively identify the specimen which had furnished the material, a specimen in the Collection and undoubtedly of the same suite, as the one ana- lyzed, was used. In studying the literature of the last twenty-five years it is surprising to find in how few cases both an analysis and an optical description of the same amphibole have been recorded. There are plenty of instances where one or the other is given, but unfortunately the two are seldom found combined in the same description. Therefore, if only as a matter of record, it seemed advisable to complete the investigation of these amphi- boles by a determination of their optical characters. This work was started a number of years ago in the laboratory of Prof. Rosenbusch at Heidelberg, but through various causes it has only recently been possible to carry it to completion. As far as possible the following optical facts were deter- mined in each case: (1) the indices of refraction and the aver- age index of refraction, (2) the angle of extinction on 6 (010) measured with the trace of the prismatic cleavage, and (8) the pleochroism. The value of the optical angle, 2V, was deter- mined whenever possible by calculation, but no direct meas- urements of it were made. The indices of refraction were measured upon a total refractometer, at least two differently *Vol. xxiii, p. 23, 1907. 180 W. E. Ford-—Optical Study of the Amphiboles. orientated plates being used in each case. In a few instances, where the character of the material made the use of the total refractometer impossible, the mean index was determined under the microscope by the immersion of small fragments of the mineral in liquids having known indices of refraction. The extinction angle on 6 (010) was determined on carefully orien- tated plates, the measurements being made in sodium light and repeated a large number of times in order to give a cor- rect average. In the majority of cases these plates were made from well-developed crystals, so that the direction of the inclination of the clino-axis was known. In all these cases the extinction direction ¢ was found to lie, as is usual with amphi- boles, in the obtuse angle 8. In the case of the pleochroic varieties a second plate was prepared parallel to a@(100) in order to give the vibration direction b.. Below are summarized the results obtained, the corresponding analysis being given in the table on p. 181. 1. Tremolite from Richville, near Gouverneur, New York. a = 15992, B = 1°6132, y = 16246. Mean index = 1-612. Vy — a — 0204. INE a= OEY CoN eS OO? 2. Tremolite from Lee, Massachusetts. a = 16022, 8B = 1°6192, y = 1°6347. Mean index = 1-618. y — a = ‘0325. OV =186 2 20 cA Ce 216 r38e 3. Actinolite from Russell, St. Lawrence Co., New York. a = 1°6162, B = 1°6304, y = 1°6412. Mean index = 1°629. y — a = 0250. ONE 285, 30. oO A Cipla aie 4. Actinolite from Greiner in Tyrol. a = 1:6173, B = 1°6330, y = 1°6412. Mean index = 1°631. y—a= 0277. 2N = 381/38 Cun Cl l4 a 5oe 5. Actinolite from Pierrepont, St. Lawrence Co., New York. a = 1°6237, 8 = 1°6382, y = 1°6503. Mean index = 1°637. y —a = 0236. 2V = 84° 8’. c ~Acimpossible to measure because of wavy character of extinction. Pleochroism, a = almost colorless, 6 = faint yellow-green, c = faint bluish green. Absorption, c > b > a. 6. Actinolite from the mines of Krageré, Norway. a =1°6280, B = 1°6442, y = 1°6547. Mean index = 1°641. y —a = 0267. INK iG SIG “CN 6 Sse Bs" Pleochroism, a and 6 = pale brown, c = light green. Absorption, c> b= a. W. E. Ford— Optical Study of the Amphiboles. 181 <7 . Hornblende from Edenville, Orange Co., New York. aj= 16583.) B= 16701) y = 16789. Mean index — 1:668- VSO = "0206. 2N) =| Bl Aol ici— 23.48), Pleochroism, a and 6 = light brownish green, ¢ = dark green. Absorption, c> b= a. oa) . Hornblende from Renfrew, Ontario, Canada. Meanvindexs—1-G7e) Cen ¢=/330 20. Pleochroism, a and b = olive-green, c = dark blue-green. Absorption, c> 6 > a. 9. Hornblende from Mte. Somma, Italy. Mean index = 1°68. Pleochroism, a and 6 = olive-green, c = dark blue-green. Absorption,c b=a. 10. Hornblende from Cornwall, Orange Co., New York. Mean index =1:-71). ca c= 9°. Pleochroism, a = light olive-green, b = yellowish green, c = bluish-green. Absorption, ¢ > b> a. From the above facts it is possible to make the following general statements concerning the optical character of what ~ may be called normal and typical amphiboles. The mean index of refraction ranges from about 1°61 to about 1°71. The mean index of refraction of typical tremolite lies between 161 and 1°62; that of typical actinolite ranges from 1:63 to 1 2 3 4 5 6 " 8 9 SiO. 57°45 57:69 54°80 56:25 52°31 51°85 41°99 48°76 39°48 TiO, Unseae 0°14 0:10 gee 0:28 1:26 1:46 0°78 0°30 Al,0; 1°30 1°80 2°58 1°24 2°69 4°36 11°62 8:33 12°99 Fe.O, 0-18 0:00 2°50 0:78 3:09 2°58 2°67 6:90 7°25 FeO 0°22 0°55 4:75 5°50 6°68 5:46 14:32 10°47 10°73 MnO 0:07 es Gr 0°48 0:70 0°35 0:25 0°50 1:00 MgO PAL8o 2412-20230) P21 19-27 O48 e > TA 1263) = a7 CaO 12°89) 13519) 5 12:08) 1208 10-88) 10:60) 11752 9°84 12-01 ' KO 0:54 0:22 0:24 0:28 0:50 0°35 0:98 1:28 2°39 Na.O 0:67 0°48 0°82 0.19 0:78 2°15 2°49 3°43 1°70 H,O 1:16 1°56 1°60 1°81 1°42 1:21 0°61 0°65 0°76 F, 0-77 0°37 0:77 0:04 0:93 0°46 0°80 1°82 0:05 Loss at 110° 0:09 0:10 0-11 0:08 0:13 0:08 0°10 0°12 100719 100°22 100°65 99°84 100°59 100°24 99°96 100°49 100°25 O=F, 0°32 0-15 0:32 0°39 0°22 0°33 0°76 0:02 99°87 100°07 100:33 100°20 100:02 99°68 99°78 100:23 All analyses by Stanley except No. 10, which is by J. L. Nelson. 1. Tremolite from Richville. 6. Actinolite from Krageré. 2. Tremolite from Lee. 7. Hornblende from Edenville. 3. Actinolite from Russell. 8. Hornblende from Renfrew. 4. Actinolite from Greiner. 9. Hornblende from Mte. Somma. 5. Actinolite from Pierrepont. 10. Hornblende from Cornwall. 182 WW. E. Ford—Optical Study of the Amphiboles. 1:64; that of typical hornblende between 1°66 and 1°71. The birefringence of the whole series is about the same, the extremes observed being ‘0206 in the hornblende from Edenville and °0325 in the tremolite from Lee, the average of the series being ‘0259. Theaxial angles as far as they could be calculated range from about 77° to 86° 30’... The extinction between ¢ and c measured on 6 (010) ranges from 16° to 20° in tremolite, from 18° 30’ to 15° in actinolite and from 9° to over 33° in the typical hornblendes. Strong pleochroism was found only in the horn- blendes, the actinolite members showing it but faintly, if at all. The absorption was always ¢c >b6 >a, although frequently little difference could be distinguished in the color of the light vibrating parallel toa and. The plane of the optical axes in these minerals was always in the symmetry plane and the bisectrix c lay in the obtuse angle 8. The optical character, where it could be proven, was found to be uniformly negative. With our present meager knowledge concerning the relations existing between chemical composition and optical characters, such general statements, as given above, are perhaps all that should be attempted. The amphiboles, because of the wide variation in their composition and the large number of variable radicals which they contain, present extraordinary difficulties to any exact correlation between the chemical and optical properties. Such a correlation it is quite possible can never be made. Yet it is thought, that within somewhat wide limits and with the recognition of numerous exceptions, a tentative correlation based on the work with this series of amphiboles can be given. It was found on studying the above optical determinations in respect to the corresponding analyses that the variation in the mean index of refraction came nearer to showing a correlation with the variation in composition than any of the other optical characters. Consequently the analyses have been chiefly studied in the endeavor to discover, if possible, how they varied with increase in the value of the mean index of refraction. The analvses, as given on page 181, are arranged in order of the rise in the value of their mean indices of refraction. With the purpose of making these relationships clearer, figs. 1 to 9, which show them in graphical form, are given. In these figures the mean indices of refraction are shown as the abscissas while the percentages of the different radicals present in the minerals form the ordinates. The numbers given to the different points located on the figures correspond to those of the analyses on page 181. In each figure a straight line has been drawn which falls as closely as possible to the different points. Very likely a curve would have better expressed the relations between the indices of refraction and W. E. Ford—Optical Study of the Amphiboles. 183 OFS os SESS SRS SSeS 28 Ssesaea 28 2S S878 Ie Se ee Sn eS ES = 68S 1690 16 6/5 1620 162 1630 /65 1640 164. 1650 /b5e SESS SSS SS) SS ss ~s ~~ ss Sere So est espe ee the percentage composition, but, with the data available, it seemed best to assume that this line was straight. It will be noted that the first six analyses form a series, the members of which lie close together. Then there comes quite a break, with analyses 7, 8 and 9 falling again in a group by themselves. There is another break and analysis 10 lies at the end by itself. It is unfortunate that other data were not available by means of which these gaps could have been closed. 184 =o W. E. Ford—Optical Study of the Amphiboles. Figure 1 shows the decrease in percentages of SiO, with the rise in value of the mean index of refraction. The relations here are reasonably well expressed by the straight line given inthe figure. Of the first nine analyses, the one which departs farthest from this theoretical line is No. 4 and this is only about 2°5 per cent out of the way. In figure 2 the increase in the percentages of Al,O, are shown corresponding to the increase in value of the mean refractive index. The points determined in this case show more general variation from the straight line than in the case of the silica, yet here the greatest variation among the first nine analyses is only a little over 2 percent: Figure 3, which shows the variation in the percent- ages of Fe,O,, is much the same, the greatest variation being about 2°7 per cent in the case of analysis 7. Figure 4 shows the variation in the sum of Al,O, and Fe,O,. As might be expected, the position of the points in this tigure in general approximate much closer to the straight line than in the two previous figures, analyses 4, 9 and 10 being the only ones that depart from it in any notable dgree. Figure 5 shows the variation in amount of FeO with the increase in value of the index of refraction. Eight of the analyses give points that le reasonably close to the line; analyses 8 and 9 however depart considerably from it, 9 being nearly 6 per cent out of the way. Strikingly different is the agreement shown in figure 6, in which the amounts of total iron are given. Here analysis number 9 is the only one that shows any notable discrepancy and the amount of variation in this case is only about 2 per cent. Figure 7 shows that the decrease in percentage of MgO follows closely the rise of the index of refraction, the greatest variation from the line being 2 per cent in the case of analysis number 9. Figure 8 gives the percentages of CaO in relation to the indices. There is a slight decrease in the amount of CaO as the index rises but the extreme variation is not much over 3 per cent. Consequently the influence of the lime upon the index of refraction must be fairly uniform throughout the series. Figure 9 gives the percentages of alkalies present, K,O+Na, O. They show a slight increase in amount with the rise of the indices, but as the largest amount of the alkalic oxides is under 5 per cent they cannot have much influence upon the optical characters. From the above it can be seen that with this series of amphiboles it is the silica, the total iron and the magnesia that follow in their varying percentages most closely the variation in the mean index of refraction. It must be to them that this variation is chiefly due and in all probability the percentage of total iron present is the factor possessing the greatest influence. With full realization of the incompleteness upon which these pen) CO I boles. 0 cal Study of the Amph 7) W. £. Lord—Opt 5-7. Figs. 186 }=W. E. Ford—Optical Study of the Amphiboles. figures are based and that there will be many exceptions found, yet, it is thought that by their use the composition of a normal and typical amphibole can be predicted within reasonable limits when its mean index of refraction is known. It will be of interest to next consider two other amphiboles analyzed by Stanley, together with several described by other authors, in order to see how far they agree with those given above. As already stated, the literature shows surprisingly few cases of recent analyses with which a complete description of the optical properties are given. Unfortunately also for use in the present discussion, such amphiboles, as have been described, are for the most part exceptional in their nature Fies. 8, 9. SSS Se SoS SS SS SOS a See S SS © SS S08 SSS SoS SS SS SSS 5S. 5°8 8 = &§ ST RSS) Ss SiS SSeS SS SS ~ DSRS Ra SS SON SS 1655 SS wea as) SS Se SE aS SS SESS NEES S S ~ ~S “~“ ~ N N ~ ~~ N SS ~ ~ since the interest in them, which led to their description, was naturally caused by their unusual character. Therefore, many of these analyses cannot be taken as belonging to amphiboles that could be considered normal and _ ty ‘pical in character. Nevertheless it will be of value to consider them and endeavor to account for their divergence from the series given above. Analysis 11 is of a tremolite from Switzerland. Its optical characters are given by Kreutz as follows: a = 16000, 6 = 1:-6155, y=—1°6272. Mean index=—1°61386. y—a=-0272. 2V = 85° 30’ (calculated by present writer). This analysis will be found to answer closely to what would be expected, from the above study, to be the composition of an amphibole having its mean index of refraction. Analysis 12is of the unusual hornblende from Grenville town- ship, Quebec, Canada, analyzed by Stanley. Its optical proper- ties are as follows: a=1°6128, 8 = 1°6180, y=1°6328. Mean index = 1:623. y—a=0-020. 2V = 56° 8’. cn ¢= 29° 44", For a hornblende with its mean index of refraction, it is char- acterized by a very low percentage of silica, by a high amount of alumina, by low percentages of the oxides of iron and by high W. E. Ford—Optical Study of the Amplhaboles. 18% percentages of the alkalies. It is especially distinguished by the unusual amount of fluorine that it contains (2°76 per cent). Another factor to be considered is the 1:2 per cent of TiO, which is present. The effect of an increase in the percentages of fluorine and the alkalies is generally thought to lower the indices of refraction and that of an increase of TiO, to be opposite in character. With all these constituents present in unusual amounts it is difficult to come to any reasonable con- clusion concerning the correlation of the chemical and optical properties of this amphibole. The axial angle, diverging as it does so markedly from the usual value, shows clearly that this Grenville mineral is abnormal in character. Analysis 13 is of the hornblende known as pargasite from Pargas, Finland. Its optical characters as given by Kreutz are as follows: a =1°6158, 6 =1°6205, y =1°6353. Mean index——1-6955. y—@ = 0195) IVI 59730" (ene 272 13", As Kreutz has pointed out, this mineral is closely analogous to the hornblende from Grenville described above. Like it, it is characterized by a low amount of silica, high percentages of alumina and alkalies and by the presence of a notable amount of fluorine (1°82 per cent). The extinction angle and the axial angle are also closely similar. Like the Grenville mineral it cannot be considered as normal in character. Analysis No. 14 is of a hornblende from Chester, Massachu- setts, described by Dupare and Pearce. Its optical properties are given as follows: a =1°6598, 8 = 1:°6729, y = 1°6798. Mean index = 1°6698. y—a = ‘0200. 2V = 75° 44’ (calculated by present writer). cac=14° 40’. This analysis agrees rea- sonably well with what might be predicted for an amphibole having its mean index of refraction. The amounts of silica and magnesia agree closely with the corresponding percentages to be derived from figures 1 and 7. The amount of sesqui- oxides and the total iron are somewhat higher than would be expected. Analysis No. 15 is of a variety of amphibole called soretite, coming from Koswinsky in the northern Urals and described by Dupare and Pearce. Its optical character is as follows: oC N662i, 0 — oOo. vi 16856.) Mean index — 1-644 y—a = 0228. 2V = 82° 30’ (observed). cac=17°. Pleochro- ism, a = pale greenish yellow; b = green, c= deep green. The analysis agrees duite closely with the theoretical one to be derived from consideration of the figures 1 to 9. The amounts of silica, alumina and magnesia are within less than two per cent of what would be expected. The percentage of ferric oxide is high and that of ferrous oxide correspondingly low, but the amount of total iron present is almost exactly that which would have been predicted from the study of figure 6. 188s W. E. Ford—Optical Study of the Amphiboles. Analysis 16 is of the basaltic hornblende from Bilin, ana- lyzed by Stanley, the optical properties of which follow: Mean index = 1- 692. Ccxaé=1° 12’ in the acute angle £. Pleochroism, a= light yellow-green, 6 = orange, c = dark reddish brown. Absorption, c>b>a. The percentage of silica present is about what would be expected from a horn- blende with a mean index of refraction of 1°69. The sesqui- oxides are low, the amount of iron protoxide is very deficient and the magnesia is about double what would be expected. This hornblende contains 1-68 per cent of TiO,, which undoubt- edly has a marked influence upon its optical properties, although it would hardly seem sufficient to account for all its peculiari- ties. It is characterized optically by its unusual extinction angle and by the deep reddish brown color of the light vibrat- ing parallel toc. This hornblende shows little correspondence with the series given above. _ Analysis 17 is of the variety of amphibole known as kaer- sutite, from Kaersut, Greenland, as it is given by Washington. Its optical pr operties follow a —=A:6((65 8 — 1-694 iy lene Mean index = 1°692. y-—a= :0382. OV = 82° 6’. Whis aa) an unusual amphibole in that it contains over 10 per cent of TiO,. As might be expected, the amount of iron present is small for an amphibole having this index of refraction. It obviously cannot be classed as a normal amphibole. Analysis 18 is of the similar mineral from Linosa, also given by Washington. Its optical properties are given as follows: G—INeOO2s “B= = 1°730, y=1-760. Mean index = 1°726. y—a = 068. 2V = 79° 54’. This, like the Kaersut mineral, has a large percentage of TiO, (8°47 per cent). It differs in having higher percentages of the iron oxides and lower percentages of the alkalies. Optically it is distinguished by having the high- est mean index of refraction and strongest birefringence of any of the amphiboles described in this paper. From the foregoing study the following tentative conclu- sions may be drawn. Given a normal and typical member of the Amphibole Group, its composition can be predicted within reasonable limits from knowing its mean index of refraction. Of the various constituent oxides, the silica, lime and magnesia follow most closely in their variations the change i in the index of refraction. The alumina varies more widely but does not seem to have a great effect upon the index of refraction. The iron oxides in a less degree also frequently show a variation from what might be expected, but on the other hand the total iron present shows a close approximation to the theory. The introduction of notable amounts of titanium, alkalies and fluorine produces marked changes in the optical character. The presence of titanium serves apparently to raise the index of refraction while the alkalies and fluorine tend to lower it. W. E. Ford— Optical Study of the Amphiboles. 189 alu 12 13 14 15 16 We 18 SiO, 58°22 45°79 48°38 42°74 40°52 39°95 39°52 40°85 TiO. Rees 1:20 0:05 1:08 Ue 1:68 10°31 8°47 Al,0; Seca 0;83 5:48 10°99 17:58 11:22 9-89 Fe.0; 0-04 0-42 0-76 §=©11°92 9°64 7°20 1°22 8°85 FeO 0°61 0:42 1°56 =11°46 9°83 2°18 8°81 3°96 MnO 0:04 0°39 0:04 0:06 tr. tr. 0-06 0-12 MgO 2a Digalettan 20°78) | 11-60). 11:82) 14:15. 13-31 12-47, CaO 12D el eile ote ee eee) eso 2h 96 4) 10:93" a2 K.0 0-04 1°69 1°38 0:56 0°68 1:98 1:07 0°63 Na,O 0-24 2°51 2°69 2°25 2°38 3°16 2°95 2°01 H,0 cP = OLOGY 0°91 0:50 0-41 0:59 0°19 F 0-17 2°76 1°82 0:08 0:28: NiO 0°18 0°10 99°82 101:06 101°44 99°69 100-40 100-46 100:00 99:98 O=F 0-07 1:16 0-7 ‘01 99°75 99:90 100-68 100°45 11. Tremolite from Switzerland. Kreutz, Min. Mitt., xxvii, 251, 1908. 12. Hornblende from Grenville, Penfield and Stanley, loc. cit. 13. Pargasite from Pargas, Finland. Kreutz, Min. Mitt., xxvii, 249, 1908. 14, Hornblende from Chester, Mass., Dupare and Pearce, Bull. Soc. Min., xxxi, 119, 1908. 15. Soretite from Koswinsky in northern Urals. Duparc and Pearce, Bull. Soc. Min., xxvi, 131, 1908. 16. Hornblende from Bilin, Penfield and Stanley, loc. cit. 17. Kaersutite from Kaersut, Greenland. Washington, this Journal, xxvi, 198, 1908. 18. Kaersutite from Linosa, Washington, ibid., p. 192. It was felt that with the data at hand it would be of value to discover, if possible, any regularity in the chemical changes that accompany the variation in the angle of extinction, cAc, measured on 6 (010). Some sixteen analyses of normal amphi- boles were used for this purpose. These analyses were arranged in a series with decreasing angles of extinction. It was at once seen that they could be divided into two distinct groups, each of which covered much the same range of varia- tion in the extinction angles but were quite distinct from each other chemically. The first series showed extinction angles ranging from 20° to 8° with percentages of silica varying from about 58 per cent to51 per cent. The second series showed a variation iu extinction angles from about 38° to 9° with silica percentages from about 44 per cent to 86 per cent. That is, we may have two amphiboles of distinctly different composi- tion, one belonging to the tremolite-actinolite group and the other to the hornblende group, which have nevertheless the same angles of extinction; compare, for instance, analyses LV and XIV, p.190. ‘These two series of analyses were separated from each other and each studied by itself. The analyses of the tremolite-actinolite series are given on p- 190, arranged with decreasing angles of extinction. In Ay. Jour, Sct.—FourtTH SERIES, Vou. XX XVII, No. 218.—Frsruary, 1914. 14 190 =W. &. Ford—Optical Study of the Amphiboles. Tremolites and Actinolites. I II III IV V VI VII VIII SiO, o747 «8605769 «= 02) 56-25) = :d54°80 «= 55-210 185 Ss«4-D TiO; Sf W 0-14 Dae Ra 0-10 Le 1:26 0°39 Al,Os 1:28 1:80 4°58 1°24 2°58 3°45 4°36 9°25 Fe,0; 0:18 Bee 1:04 0°78 2°50 eee 2°58 4:44 FeO 0°22 0°55 3°28 5°50 4-75 7:49 5°46 9°81 MnO 0:07 tr. ae 0:48 tr Sah 0°35 0°46 MgO 24°87 2412 20°36 21°19 20°30 18:97 19:48 10°38 CaO 12°84 13:19 8:00 12°08 12°08 10°50 10°60 1:98 K,0 0-49 0:22 1°52 0°28 0:24 Bee 0°35 0:16 Na.O 0°68 0:48 6°71 0719 0°82 2°45 2°15 7°56 H,O0 1:30 1°56 0°51 1°81 1°60 1°75 1°21 ier F 0:77 0:37 pair 0-04 0-77 uae 0°46 aes 100°19 100°22 100:97 99°84 10065 99°82 100°24 100°68. Total Fe 0-29 0:42 3°27 4°81 5°44 5°82 6:24 10°84 Extinction angle 20°1' 16°38’ 16°16’ 14°59’ 14°47” 14°34’ 13°35’ 8° Mean index 1°612 1°618 Soe 2 abteeh ae) atseB) alsaeel ee I. Richville, see p. 181. II. Lee, see p. 181. IIl. Széchenyite, see Krenner, Zs. Kr., xxxi, 502, 1899. IV. Greiner, see p. 181. V. Russell, see p. 181. VI. Berkeley, Blasdale, Bull. Univ. Calif., ii, 338,1901. VII. Kragero, see p. 181. VIII. Blasdale, Bull. Univ. Calif., ii, 338, 1901. Hornblendes. IX aXe XI XII XG eae XV XVI SiO» 43°7 41:99 43°35 40°52 45°20 39°58 42°74 36°86 TiO, 0:78 1°46 aie iP ral 0°84 tr. 1:08 1:64 Al.Os 8°33 =11°62 8:11 10°99 eae a9 5°48 12°10 Fe.0; 6°90 2°67 7°91 9°64 755 4°01 11:92 7°41 FeO 10°47 14°32 10:11 9°88 15°80 10°67 11:46 23:35 MnO 0°50 0°25 pee tr. 1°52 tr 0:06 0-77 MgO 122637 1 ale 3 3) 82 8:40 18:06 11°60 1:90 CaO 9:84 ddAch ot 1S 21 Ae 338)5) 12:80 = 76s lec eOa9 K.O 1:2 0:98 ERS 0°68 0°37 0°62 0:56 3°20 Na.O 3°48 2°49 2°18 2°38 0:80 2°87 2°25 1:20 H.O 0°65 0°61 0:91 0:50 0:70 2:79 aie 1.30 F 1°82 0°80 Rea apie vee age Sais 0:27 100°49 99°86 100°98 100-40 100°82 100°27 99°69 99-99 Total Fe 12°98 12°99 1338 14:38 17°35 11:09 17:24 23°33 Extinction angle Sasl 0mm co ns See lio Oe muller Tiss? OF aa? 14°407 g° Mean index 1°67 1:668 Bae he SO Atari pee toa 1°67 ike 7l TX. Renfrew, see p. 181. X. Edenville, see p. 181. XI. Artificial amphibole, Chrustschoff, Jb. Min., ii, 86, 1891. XII. Soretite, Dupare and Pearce, Bull. Soc. Min., xxvi, 126, 1903. XIII. Philipstadite, Daly, Proc. Am. Acad. Sci., xxxiv, 483, 1899. XIV. Piedmont, Van Horn, Am. Geol. xxi, 370, 1898. XV. Chester, Dupare and Pearce, Bull. Soc. Min., xxxi, 118, 1908. XVI. Cornwall, see p. 181. Qn iS & W. E. Ford—Optical Study of the Amphiboles. 191 general these analyses show, together with the decrease in the angle of extinction, a decrease in the percentages of SiO,, MgO and OaO and an increase in the percentages of Al,O,, Fe,O,, FeO and the alkalies. This variation in composition is by no means regular and the relations existing between the com- position and the angle of extinction are not at allclear. The only relationship that seems at all close in character is that between the angle of extinction and the total amount of iron present, there being an increase in the percentage of total iron Fies. 10, 11. hotol te = S that follows closely the deer ease in the extinction angle. The correlation between these two is shown oraphieally i in figure 10. That the alkalies have little effect upon the angle of extinction is seen by comparison of analyses I and V and III and VIII. It will be noted further that the decrease in the value of the extinction angle is accompanied by a fairly uni- form increase in the value of the mean index of refraction ; see figure 11. sine hornblende analyses are given on p. 190, also arranged in the order of decreasing extinction angles. These analyses have been studied in various ways but without finding any reasonable correlation between their percentage composition and the angle of extinction. The variations in the percentages 192 W. &. Ford—Optical Study of the Amphiboles. of the different oxides are quite irregular. The amount of total iron present increases in general with the decrease in the angle of extinction, yet this increase is irregular (note the low percentage of total iron shown in analysis XIV) and does not even begin to show the same degree of uniformity as in the case of the tremolite-actinolite series. The fact that we can have two members of the Amphibole Group differing decidedly in their percentage composition but still showing the same angles of extinction proves at once that the value of the extinction angle does not depend directly at least upon the percentage composition. It seems probable that it must depend ultimately in a large measure upon the kind and amount of elements present in the mineral, but it is likely that it is also influenced largely by the character of the physical molecule into which they are grouped. The relation- ships here are certainly complex and the available data are far too few to allow us to hope that the problem can be solved as et. z Various writers have previously attempted to find a correla- tion between the chemical and optical properties of the amphi- boles. Murgoci in a study* which was concerned chiefly with the characters and classification of the glaucophane-like amphi- boles makes, however, some general statements. On p. 362 he says: “ The chemical constituent with the most influence on the physical properties seems to be Fe,O, (viz. Fe,Si,O,) and that not only in the glaucophane series, but in general in the amphibole family . . . Further, the size of the angle of the optic axes, the position of the axial plane and even the angle of extinction, up to a certain point, are in relation with the amount of F e,0,. ” This may be true in respect to the alkali- rich amphibole-like glaucophane, but the facts as given above hardly bear this statement out in respect to the ordinary amphi- boles. The amount of FeO shows a closer relation to the opti- cal properties than the Fe,O, and the amount of total iron present conforms more closely yet. Further, the same author states on p. 373 that “ A comparison of the whole group of AlFe’” amphiboles has convinced me that in general the size of the angle of extinction is related neither to the amount of Al,O, nor to the amount of Fe,O,, but to the proportion of their molecular coefficients of ‘combination in the amphibole constitution.” A stndy of the analyses given above has failed to confirm this theory. In fact it only needs a hasty inspec- — tion of the relative percentages of Al,O, and Fe,O, in the analyses quoted on p. 190 to show that the assumption does not hold true with them. * Contribution to the Classification of the Amphiboles, Bull. of the Geol. Dept. Uni. of Calif., iv, 359, 1906. W. E. Ford— Optical Study of the Amphiboles. 193 The fact that the iron content of an amphibole has a large influence upon the optical characters was stated as long ago as 1871 by Tschermak* and has recently been emphasized by Kreutz.t The results of the present study, as already given, bear out his statement. Wiik+t endeavored to show that the size of the extinction angle, cA¢, increased as the percentages of AI,O, increased. This theory has already been shown by others to be untenable and the results of the present study point to the same conclu- sion. Mineralogical Laboratory of the Sheffield Scientific School, Yale University, New Haven, Conn., July 2, 1913. * Min. Mitt., p. 17, 1871. + Min. Mitt., xxvii, 250, 1908. t Zs. Kr., vii, 79, 1883. SCIENTIFIC INTELLIGENCE. I. Cyermistry Anp Puysics. 1. The Action of Aluminium Carbide upon Solutions of Metallic Salts.—Hitrert and Dirmar have found that aluminium carbide, which has the formula Al,C, and which reacts with water and with acids giving methane, CH,, is capable of methylat- ing certain metals when it acts upon solutions of their salts. For instance, when aluminium carbide is placed in a hydrochloric acid solution of mercuric chloride, mercuric methyl chloride, HgCH,Cl, is formed, and when the solution is neutral or slightly acid, mercury dimethyl, Hg(CH,),, is produced. In the same manner bismuth trimethyl, Bi(CH,),, which is otherwise difficult to obtain, can be synthesized. Tin salts also are methylated by aluminium carbide, and in this case the tin-methyl compounds, on account of their intense and characteristic odor, may be used as a qualitative reaction for tin, for even as small a quantity as 0-1 mg. of stannous or stannic chloride in 2 ccm. of dilute hydro- chloric acid, upon boiling with a small quantity of aluminium carbide, gives a strong odor. Preliminary experiments have shown that not only arsenic and antimony can thus be methy- lated, but also other metals with which such compounds have not been previously obtained. For example, there is formed in the reaction between copper sulphate and aluminium carbide a com- pound with a disagreeable odor resembling mercury methyl. The exact mechanism of these interesting reactions is uncertain, and it is remarkable that the authors have not been able to methylate 194 Scientific Intelligence. any pure organic halide by means of this reaction. For instance, ehloracetic acid could not be changed to propionic acid, nor benzyl chloride into phenylethane. It does not appear, however, that electrolytic dissociation is a requirement for the reaction, for mercuric cyanide reacts like the chloride. — Berichte, xlvi, 3738. H. L. W. 2. Solubilities at the Critical Temperatures of Solvents.— Fritz Frrepricus has made experiments with a large number of substances, mostly inorganic salts, by placing them in sealed glass tubes with liquid anhydrous ammonia and also with liquid sulphur dioxide, and then heating the tubes to the critical tem- peratures of the liquids. Although the continuity of the liquid and gaseous condition of solutions has been observed previously, Friedrichs has made an interesting contribution to our knowledge of the solubility of salts in the two liquids used for his experi- ments. For example, the fact that potassium iodide dissolves readily in liquid ammonia and passes into the gaseous state with the ammonia at the critical temperature of the latter is striking, as is also the similar behavior of dilute solutions of ‘silver iodide in the same solvent. It would be interesting to find to what extent such gaseous solutions of solids could be expanded, and still hold non-volatile solids in solution. H. L. W. 3. Organic Chemistry for Advanced Students ; by Juxtus B. Couzn. Vol. II, 8vo, pp. 427. New York, 1913 (Longmans, Green & Co. Price, $4.50 net).—The first volume of this very useful text-book was published in 1907. The present volume fills up gaps left in the first one according to the author’s present plan of presenting the subject in his lectures. The course as modified gives more attention to the physical side of organic chemistry than was the case with the first volume. The chapters of the book deal with the valency of carbon, the nature of organic reac- tions, dynamics of organic reactions, physical properties and structure, and color and structure. The book makes an excellent addition to the first volume. H. L. W. 4, Industrial Chemistry for Engineering Students ; by HENRY K. Benson. 12mo, pp. 431. New York, 1913 (The Macmillan Company. Price, $1.90).—The purpose of this text-book is to describe from the standpoint of chemistry the more common materials used in the various branches of engineering. ‘The topics of greatest interest in engineering, such as fuels and com- bustion, clay products and cement, are treated more fully than others. There are short chapters on the manufacture of pig iron, the commercial forms of iron and steel and industrial alloys, but no other metallurgical subjects are treated. ‘lhe references to the literature of the subjects that are included are very exten- sive and valuable, and the book appears to give a very good pre- sentation of the subjects for the purpose in view. As the book does not touch upon many important chemical industries it is hardly suitable for students of chemistry and chemical engineer- ing. The book appears to be lacking in the critical discussion of Chemistry and Physics. 195 many of the materials described. For instance, little is said about the relative merits or the prices of the various pigments, oils, resins and solvents used in the manufacture of paints and varnishes, although many of them are described. H. L. W. 5. Outlines of Theoretical Chemistry ; by FRrpERIcK H. GETMAN. S8vo, pp. 467. New York, 1913 (John Wiley & Sons). — This text-book is designed for classes beginning the study of theoretical or physical chemistry. A working knowledge of elementary chemistry and physics is presupposed in the presenta- tion of the subject, although an introductory chapter of 19 pages is devoted to a brief review of the principles with which the stu- dent is assumed to be fairly familiar. These elementary principles include the laws of definite, multiple, and combining proportions, the atomic theory, Avogadro’s hypothesis, atomic heats, isomorph- ism, valence, and atomic weights, while the main part of the book is advanced and physical in its character. The presentation is mathematical in its aspect, but the use of calculus has been restricted as far as possible. A good feature is the introduction of many numerical problems for solution by the student. There is a chapter on the electron theory, but the author, unfortunately it appears, has decided against the incorporation of any account of radiochemistry. H. L. W. 6. Oxygen in the Sun.—In the year 1896 Runge and Paschen showed that the first triplet of the principal series of the oxygen spectrum agreed in wave-length and relative intensity with three Fraunhofer lines in the solar spectrum. Since there are good reasons why the remaining oxygen lines would not be susceptible of detection in the solar spectrum, the coincidence and character of the members of the triplet have been accepted as sufficient proof that oxygen is present in the solar atmosphere. On the other hand, in 1912 Geiger obtained three lines in the are spec- trum of iron which agreed exactly with the triplet and which he ascribed to iron. If the lines observed by Geiger really owe their origin to iron then the argument of Runge and Paschen is not tenable because experience shows that, when sufficiently high dispersion and spectroscopic resolving power are used, the sepa- rate lines of different elements do not coincide exactly, much less in groups. In order to settle the question as to the origin of this triplet the experimental side of the problem has been taken up again by Runes and Pascuzen. These investigators employed a Rowland concave grating of 3 meters radius of curvature in conjunction with a concave reflector which sent plane waves upon the grating. By working near the principal normal to the grating a non- astigmatic image of the slit was formed on the photographic plate. Furthermore, by using a photographic objective a non- astigmatic image of the source of light was thrown on the slit of the spectrograph. The first source used was a vertical arc between iron electrodes, the lower rod being the anode. The condensing lens was moved by hand in such a manner as to keep 196 Scientific Inteiligence. the image of the intensely white spot on the oxide globule, from which the are starts, focused exactly on the slit. The spectro- gram shows that the members of the triplet in question extend only a relatively short distance from the globule, whereas the remaining iron lines recorded by Geiger extend across the entire field of view and are most intense in the region corresponding to the arc proper. No difference in the appearance of the triplet was observed when the voltage was changed from 220 to 60. These lines also were present near the anode when the are was burned in air at a pressure of 1°" of mercury. On the other hand, when a small piece of iron wire was laid on the positive electrode of a carbon arc, in air at atmospheric pressure, the triplet could not be found, notwithstanding the fact that all the other lines recorded by Geiger were photographed. When the processes of oxidation and reduction which occur in the are between iron electrodes are taken into account together with the conditions which hinder oxidation in the carbon are, little doubt can remain as to the oxygen origin of the triplet. The wave- lengths of the lines are given as 777201, 7774°21 and 7775°50 A. U. The proof seems to be made final and conclusive by the follow- ing argument of the authors. The ‘“ frequencies” of the lines are 12866°68, 12863°04, and 12860°91 so that the first differences become 3°64 and 2°13. The mean frequency differences of the remaining triplets of the oxygen spectrum are 3°70 and 2:08. Corrections to the frequencies of the lines of the first triplet would only have to be +0°02, —0°04, and +0:01 respectively in order to give the mean values just quoted. The corresponding corrections to the wave-lengths would be —0-012, +0°024, and —0:006 A. U., all of which lie within the limits of experimental error. Therefore, it is clear that the triplet belongs to oxygen and that the evidence for the presence of this element in the solar atmosphere has been strengthened as a consequence of the apparent mistake of Geiger.— Physik. Zeitschr, No. 25, p. 1267, Dec. 1913. H.-S. Us 7. Fluorescence of the Vapors of Sulphur, Selenium and Tellurium.—Although the number of fluorescent compounds is comparatively great, nevertheless the number of chemical ele- ments which have been shown to possess this property, in the vapor state, is relatively small. In fact, prior to the latest work of W. Sreupine, the following list was probably complete, namely: Sodium, potassium, rubidium, mercury, thallium, oxygen, bromine and iodine. In the case of oxygen it was found by Steubing that the fluorescent light, as well as the exciting radia- tion, had wave-lengths less than 2000 A. U., that is, they fell within the Schumann region so-called. By analogy with the behavior of the vapors of the elements of the first group of the Mendelejeff table one would expect the heavier elements of the sixth, or oxygen group, to exhibit fluorescence having wave- Jengths greater than 2000 A. U., and hence to be susceptible of investigation with ordinary spectroscopie apparatus. Further- Chemistry and Physies. HOT, more, the maximum of fluorescence should shift toward the red as the atomic weight increases. These expectations have been realized by the recent work of Steubing. The details of the apparatus used will not be dis- cussed in this place; suffice it to say that the elements were vaporized in specially designed quartz vessels which were heated by means of either a Bunsen flame or a helix of wire carrying an electric current. For sulphur the exciting light was derived from an iron arc of about 5 amperes and 220 volts, while for selenium and tellurium an ordinary carbon arc sufficed. For selenium, and especially for sulphur, it was necessary to have the element in as pure a condition as possible, and hence great care had to be taken to exclude foreign gases from the quartz cylinders. On the other hand, the fluorescence of tellurium vapor was not very sensitive to the presence of impurities. The most important remaining facts discovered by Steubing may be summarized in the following words : (1) The four elements of the sixth column of the periodic system exhibit fluorescence in the vapor state. (2) As the atomic weight increases the fluorescence spectrum as a whole is displaced from the shorter towards the longer region of wave-lengths, (3) The region of excitation lies, (a) for oxygen, below 2000 A. U., (6) for sulphur, between 2500 and 3200 A, We (ce) for selenium, above 3000 A. U., and extends into the visible spectrum, (d) for tellurium, in the more refrangible part of the visible region. (4) In order to obtain fluorescence it is necessary for the vapor to have a definite density and a definite temperature. This is especially true in the case of sulphur, for which the vapor must be super- heated between 400° C. to 500° C. (5) The fluorescence spectra are discontinuous and consist of more or less ‘ washed-out ” groups of lines which appear to be bands.—Physik. Zeitschr., No. 18, p. 887, Sept., 1913. H. S. U. 8. The Series Lines of Neon.—In the year 1911 Watson pub- lished an account of his unsuccessful attempt to discover line series in the spectrum of neon. He only found, among the relatively strong lines, several groups which repeated them- selves with constant frequency difference. Recently R. Rossi has subjected Watson’s wave-lengths to careful scrutiny and dis- covered three series among the faint lines. ‘Two of the series lie entirely in the visible spectrum and are composed of doublets with constant frequency difference (167°5) between the members of the pairs. Since these two series also converge towards the same limit, 4119°8 A.U., they resemble subordinate series. Never- theless all efforts to find the associated principal series have failed. Furthermore, the neon series do not correspond to any of the subordinate series of helium, for, although the neon series are nearer to the red end of the spectrum than the helium series, as would be expected, the approximate relation between the sepa- ration of the doublets and the square of the atomic weight is not fulfilled. Rossi uses Rydberg’s formula because the number of lines found for each series is small, namely, seven doublets in the 198 Screntific L niellagentee: one series and four in the other. The third seriés calculated by Rossi consists of single lines all of which fall within the ultra- violet region. Finally, by using a quartz “end-on” spectrum- tube with neon at a pressure of about 4 mims., five new lines were discovered having respectively the following wave-lengths and intensities: 2352°0 (0), 2396-5 (2), 2464-0 (1), 2639°9 (1), and 2660°0 (0).— Phil. Mag., vol. xxvi, Dec., 1913, p. 981. 4.5. U. 9. The Mass of Rapidly Moving Hlectrons.—It is a matter of great theoretical importance to obtain experimental data which will enable a final decision to be reached concerning the relative merits of the theories of Abraham and Lorentz-Einstem. As is well-known, the experimental data of Bucherer and Wolz favored the longitudinal contraction of swift electrons as assumed in the theory of Lorentz, but Bestelmeyer has raised objections which cannot be disregarded. One of Bestelmeyer’s adverse criticisms of the work of Bucherer and Wolz is that the number of experimen- tal data obtained by these investigators is too small to justify the conclusion that Abraham’s theory of a rigid, spherical electron is untenable. This objection has been removed by the recent work of GUNTHER NEuMANN which was carried out at Breslau under the direction of CLEMENS SCHAEFER. The method of crossed fields was employed. The apparatus formerly used by Bucherer was borrowed and partly remodeled. Twenty-six negatives were selected for the final calculations. The speeds of “the electrons emitted by the radium salts varied from 0:4 to 0°8 of the speed of light. For each speed observed, the values of °— (for small speeds) were calculated by aid of the m formule deduced on the theories of Abraham and Lorentz-Ein- 5 A € Z stein. The resulting values of —— are plotted as ordinates on a Wb, diagram having as abscissze the ratio of the corresponding elec- tronic speed to the speed of light. The points derived from the formula of Lorentz fall close to a straight line parallel to the axis of abscissas, whereas the points pertaining to the hypothesis of Ape nae) show a systematic deviation from a constant value of ——as the speed increases. Also Bucherer’s data give points which lie almost exactly on the mean straight line representing Neu- mann’s experimental results. Between 0:4 and 0-7 of the speed of light the theory of Lorentz-Einstein is verified to about 1°5 parts in a thousand. Because of the ever increasing experimental difficulties as the speed of light is approached, the verification is not quite so good in the interval 0°7 to 0°8. On the other hand, the curve repr -esenting the hypothesis of Abraham consistently ap- proaches the axis of abscissas until, at the ppeer -ratio 0°8, a devi- ation of about 8 per cent from the value of = at low speeds is attained. It is therefore clear that the work of Neumann has Chemistry and Physics. 199 settled the question in favor of the theory of Lorentz. Finally, the weighted mean value of — deduced by Neumann from his experimental data is given as 1°765 E.M.U., which agrees re- markably well with the average value of the results obtained by other investigators and by other sources of radiation.—Physik. Zettschr., No. 22/23, Nov., 1913, p. 1117. H. S. U. II. Grotogy AnD MINERALOGY. 1. Changes in. Level in the Earth's Crust; by OsMonp FIsHER. (Communicated.)—In the addendum to a paper by me, on “a suggested cause of changes of level in the earth,” published in the March number, 1906, of this Journal (vol. xxi, p. 216), I referred to a paper by Col. Burrard, Surveyor General of India, published in the Transactions of the Royal Society. In that paper Col. Burrard called attention to the “astonishing ” differ- ence of 0:103°" between the value of gravity at Dehra Dun as determined by Basevi and Heaviside and that about 30 years later by Capt. Lenox-Conyngham. From this difference I argued that there had been a change of density beneath that station in the interval. However, in a note to a reply to some criticisms of mine upon his theory of mountains, Col. Burrard now says, that the change referred to is apparent only, and was due to the vibration of the brick pillars on which Basevi’s pendulum was swung. Mr. Oldham, late of the Geological Survey of India, has now examined this question, and on referring to Gen. Walker’s account of the “Operations” of the Survey, 1879, he finds that Basevi made two sets of observations at Dehra Dun. In the first, his second’s pendulum was swung on a wooden support, standing on a concrete floor. The result of this set of observations agrees fairly well with Lenox-Conyngham’s. Subsequently Basevi made a second set of observations at Dehra Dun, in which his pendulum was swung on brick pillars and the result obtained differed from the former, and he seems to have considered the value so obtained as the more trustworthy. It was to this latter result that Col. Burrard referred in his paper at the Royal Society, which I quoted. It now appears that Lenox-Conyngham, who swings his half- second’s pendulum on brick pillars, has found that a correction is necessary to allow for the vibration of the pillars, and there can be little doubt but that Basevi’s second set of observations was vitiated by a like cause. Consequently, as Col. Burrard now says, the change of gravity, deduced by comparing Basevi’s value quoted with that lately obtained, is apparent only. Neverthe- less, there does seem to be some reason to think that small changes in the value of gravity are at present going on at Dehra Dun, as reported in ‘ Nature,” in a note, vol. xci, p. 143, 1913. Graveley, Huntingdon, Noy. 20, 1913. 200 Scientific Intelligence. 2. Meteorites vs. The Earth; by Oxtver C. Farrineton. (Communicated.)—The criticism by Merrill* of the writer’s com- parison of the composition of meteorites and the earth suggests to the writer some further observations. With Merrill’s view, that the average compiled by him, by omitting the metallic iron and associated metals and sulphides from the sum of meteorite compo- sitions, is “the most acid phase conceivable” for a meteoritic magma, the writer can hardly agree. The quantity of unknown matter of the earth so greatly ex- ceeds that of the known, that in comparing meteorites with the known matter of the earth only a few meteorites need be taken. Assuming a knowledge of the composition of the earth to a depth of ten miles,t and regarding the mean diameter of the earth as 7913 miles, the unknown substance of the earth exceeds the known in the ratio of 131:1. Hence only 4 of the analyses, numbering 443 in all, which were used by the writer to obtain an expression for the composition of the earth,{ need be used to represent the composition of the earth’s crust. Such a sum, moreover, would have the advantage of being derived from actual analyses rather than from a secondary treatment of them. For these four, in order that only well-known and representative meteorites may be taken, those of Juvinas, Frankfort, Petersburg and Stannern, as given in the writer’s list, may be used. The average of these calculated to 100 is shown below and, for com- parison, Washington’s average for terrestrial rocks and Merrill’s Table V. I II Iil Washington’s Four average of Merrill’s Meteorites _ terrestrial rocks Table V SiQi ae seas ae 49°85 58°24 45°46 ell © aie Sone ee 11°14 15°80 3°21 Fe,O. } cake fs ae FeO ( AIO DSR OSS 18°87 721 19°29 Mig Ops eae ea eee 9°83 3°84 26°86 CAQ ae 51s eeu 9°44 0°22 2°06 iNav OPS kere 0°63 3°91 era IKE Oa ee Seas 0-14 3°16 0°38 He On gee SG ae 1:79 MnO 065 Ti Oe pete Se. 0-03 1-04 Feo [0-98 PL On ieee 0:04 Sie ae 100°00 100°58 100°00 That a magma of the composition shown in Table I might differentiate into the present rocks of the earth’s crust seems to the writer entirely conceivable. Differentiations much greater in * This Journal, (4), xxxv, 324, 1913. +Clarke’s assumption, Bull. U.S. G.S., 491, p. 22. ¢ Publ. Field Museum, Geol. Ser., vol. iii, p. 213, 1911. Geology and Mineralogy. 201 kind occur among observed rocks of the crust, and it is not diffi- cult to conceive that differentiation on a profounder scale in the early periods of the earth’s history could have produced the. crust of the present time from a magma like the above. The view of Merrill that the excess of iron and magnesium in the older rocks is altogether the result of secondary accumulation seems also to the writer open to question. If time is the only factor necessary for the accumulation of metals and their com- pounds, it would seem that all metals and their salts should show preponderance in the early rocks. As is well known, this is not thecase. That the two metals most abundant in meteorites should also be most abundant in the early rocks is prima facie evidence in favor of the excess of these metals at that time. Even if the Lake Superior iron deposits can be shown to be the result of sec- ondary accumulation, it should be borne in mind that these are but a part of the iron deposits of early periods. The great bodies of magnetite of Laurentian age in the Adirondacks and Canada appear to be of primary origin. Important iron deposits in Norway are reported to be magmatic segregations of Archean granite rocks. The great iron deposits of Sweden are connected with early eruptive rocks, and so, too, are large ones of the Urals. It seems doubtful, therefore, whether the excess of iron in the older rocks can be altogether explained by the theory of secondary accumulation. With regard to the origin of dolomite, it is probably true that the view that it is formed by the gradual replacement of lime by magnesia through the downward percolation of magnesian waters is a generally accepted one, but it is hoped that this will not prevent acceptance of another theory if satisfactory evi- dence can be presented in behalf of it. Several students of the subject have of late given excellent reasons for dissatisfaction with the replacement theory, and other methods of origin are being sought. One of the most complete of the recent summaries is by Steidtmann,* who after an extensive study expresses the view that “the occurrence of dolomites of vast thickness and extent cannot find a ready explanation in the mutative agency of underground waters.” A similar view is expressed by Daly.t+ That the sea, and therefore the limestones formed in it, are steadily becoming more calcic and less magnesian on account of the greater solubility of lime salts, is clearly shown by Steidtmann in the article mentioned. The highly magnesian nature of crinoid tests, as shown by analyses,{ seems to the writer a further indi- cation of this change. In a more magnesian ocean such forms could flourish in great abundance and become important rock- forming agents, but as the amount of magnesia diminished they gradually became extinct. * Jour. Geol., xix, 342, 1911. + This Journal, (4), xxiii, 109, 1907. tH. W. Nichols, Pub. Field Col. Mus., Geol. Ser., iii, 49, 1906; Clarke, Bull. 491, U.S. G.S., p. 540. 202 Scientisie Intelligence. 3. The San Franciscan Volcanic Field ; by Henry HouuistER Rosinson ; Professional Paper No. 76, U. 8. Geol. Surv., 8vo, pp. 2138, maps and figs., 1913.—The San Francisco Mountains, so- called, in Arizona have in recent years become well known as they are situated beside the tourist route to the Grand Canyon of the Colorado. In addition to the prominent central mass of San Francisco Mountain there are a number of other peaks situated in its vicinity or at no great distance from it, the whole constituting a distinct field of volcanic activity on this portion of the great plateau. It is of this voleanic group that Dr. Robinson here pre- sents the results of extensive studies in the field and of the mate- rial collected by him in the laboratory. The result forms one of the most detailed investigations of an extinct volcanic group which has yet been undertaken in this country. After the introduction and a general consideration of this region, the author gives an account of the sedimentary platform upon which the volcanoes have been built, and then takes up the general geology of the volcanoes and lava fields. He shows that there were three general periods of volcanic activity, beginning with outflows of basalts, followed by the formation of the volcanic cones, and ending with another outbreak of basalt, forming a great number of small cones and of flows. The second period presents the most complex stage of volcanic activity, as the cones in general are built up of various kinds of lava. Thus at San Francisco Mountain five stages of eruption, each yielding a different type of lava, have been worked out. The geology of each of the main peaks is described in detail, first with San Francisco Mountain, the main cone of the district, fol- lowed by Kendrick Peak, Bill Williams’ Mountain, O’Leary Peak, Sitgreaves Peak, and Mormon Mountain. In addition to these cones the author shows that in the second period there were also intrusions of peculiar laccoliths, which were not only intrusive, but also partly extrusive in nature. The succeeding portion of the work is devoted to presenting the results of the petrographic study of the lavas and of the gen- eral petrology of the district. The different kinds of rocks, rhyolites, dacites, latites, andesites, and basalts, have been thor- oughly investigated, and their description is accompanied by numerous chemical analyses made by the author. Under the heading of the petrology there is given a series of statistical studies of the results afforded by the chemical analyses, affecting the origin and differentiation of the lavas, from various points of view. The author here comes to some very interesting conclu- sions which have a bearing on the general petrology of igneous rocks, but it would be beyond the scope of this review to present and discuss them. The whoie work forms an important contri- bution not only to our knowledge of the geology of this part of the country, but also to the subject of theoretic petrology in gen- eral, and the final conclusions which Dr. Robinson has drawn from the results of his studies should be read by all petrologists. INS 12 Geology and Mineralogy. 203 4. The Devonian of Maryland. Geological Survey of Mary- land, 3 vols., with 1280 pages and 171 plates.—This great work treating of the entire Devonian of Maryland was begun about fifteen years ago, since which time it has grown to unfore- seen proportions. The Lower Devonian (pp. 1-560, 98 plates) is the work of C. K. Swartz, Charles Schuchert, C.S. Prosser, E. O. Ulrich, R. S. Bassler, T. P. Maynard, D. W. Ohern, and R. B. Rowe. The Middle Devonian (pp. 1-338, 40 plates) is by C. 8. Prosser, E. M. Kindle, E. O. Ulrich, R. 8. Bassler, and C. K. Swartz; while the Upper Devonian (pp. 339-701, 33 plates) is by C. S. Prosser, C. K. Swartz, and John M. Clarke. Of the fossil forms discussed, there are 785, of which 397 (142 new) are from the Lower Devonian, 185 (21 new) are from the Middle Devonian, and 203 (60 new) are from the Upper Devo- nian. These fossils occur in the following formations : Lower Devonian— Helderberg formation, thickness --.-_-.--,- 290— 350 fee Oriskany formation, COOP IAS WA Moat emai teary 50—- 417 “ Middle Devonian— Romney formation, Ever auicreaat tesa 600— 1650 “ Upper Devonian— Jennings formation, UM Reef ae eagh 3400-— 4750 “ Catskill formation, Ci, WAP pace ANB 1200— 3800 << Total thickness, 5540-10,967 feet This storehouse of information, together with the Paleontol- ogy of New York, will be the mecca to which all students of the paleontology and stratigraphy of the Devonian in eastern America must go. We extend our hearty congratulations to State Geol- ogist William Bullock Clark upon the completion of this model report on the Devonian of Maryland. C!s: 5. Stratigraphy and Paleontology of the Alexandrian Series in Illinois and Missouri, Pt. I; by T. KE. Savace. Illinois Geol. Surv., Bull. 23, pp. 1-124, pls. I-VII, 1913.—A careful, detailed, and well-illustrated study of the basal Silurian formations of southern Illinois, with a description of the faunas of the two low- ermost members, the Girardeau (28 species, 12 new), and Edge- wood limestones (59 species, 36 new, and the new coral genus Calvinia). In northern Illinois occurs the Channahon limestone, which is correlated with the Edgewood; of the 23 species (12 new) found in the former series, 8 are also found in the Edge- wood; but the general aspects of the two faunas are more harmo- nious than the figures indicate. c. 6. 6. On the important part played by calcareous Algce at certain geological horizons, with special reference to the Paleozote rocks, by E. J. Garwoop. Geol. Mag., dec. v, vol. x, pp. 440-446, 490- 498, 545-553, 1913.—Because of the marked interest that is now being manifested the world over in the important réle played by 204 Scientific Intelligence. the calcareous alge in the formation of limestones and other cal- careous strata, attention is directed to this important paper, delivered by Professor Garwood as his Presidential Address be- fore the Geological Section of the British Association for the Advancement of Science, at the Birmingham meeting. The more important literature on the subject is referred to in the paper. C..8. i. The Geological Survey of Oklahoma.—At a meeting of the State Geological Commission of Oklahoma, late in December, the resignation of D. W. Ohern as Director of the Oklahoma Geological Survey was accepted. L. C. Snider, the assistant direc- tor, declined to consider the directorship and C. W. Shannon, field geologist, was appointed director. The personnel of the scientific staff of the Survey as now constituted is as follows: C. W. Shannon, director; L. C. Snider, assistant director; L. E. Trout, field geologist; Wm. A. Buttram, chemist. 8. Minéralogie de la France et de ses Colonies; per A. Lacrorx. Tome cinquiéme ; Deuxiéme Supplément et Index Géographique. Pp. 501. Paris 1913 (Librairie Polytechnique). —This is the second supplement to the exhaustive work by Pro- fessor Lacroix. It is chiefly given to a geographical index of localities with the species occurring at each. The places are arranged alphabetically under the countries, France, Alsace, Bel- gium, etc.; also Algeria and other French colonies, the most important of which from the mineralogical standpoint is Mada- gascar. ‘The interesting new species from this remarkable island with notes on some other minerals are described briefly in the opening pages. 9. Handbuch der Mineralogic; von Dr. Cart HInTze. Erster Band, Lieferung 16; pp. 2401-2560. Leipzig, 1913 (von Veit & Comp.).—Begun in 1899, the Mineralogy of Hintze has now reached its twenty-eighth part. The minute, careful labors which the author has devoted to this great work are beyond praise. The present part is given to the fluorides, the species fluorite occupying much of the space. Ill. Muscetnangrovus Sctentiric INTELLIGENCE. 1. The Semi- Centennial Anniversary of the National Academy of Sciences, 18638-1913. Pyp.vii, 108, with illustrations. Washing- ton, 1913.—A memorial volume, giving a history of the first fifty years of the National Academy of Sciences, was published some six months since and noticed in this Journal at the time (see vol. xxxvi, p. 185). The semi-centennial anniversary meeting, held in Wash- ington in April, was an occasion of great interest not only to the Academy itself but also as regards American science in general. We have now an account of this April meeting prepared by the Home Secretary, Dr. Arthur L. Day, which gives in detail the Miscellaneous Intelligence. 205 exercises of the three days concerned. The volume opens with a copy of the official program, following which is the introductory address of President Remsen. Other formal addresses by distin- guished men especially invited for the occasion follow. These include President A. T. Hadley, who spoke on the relation of sci- ence to the higher education in America; Professor Arthur Schuster, on international codperation in research ; Dr. George E. Hale, on the earth and sun as magnets ; and Professor J. C. Kap- teyn, on the structure of the universe. The remarks of Dr. Wood- ward and President Wilson on the occasion of the distribution of medals, are given in full. The meeting closed with a dinner for the members of the Academy and invited guests, and the secre- tary is able to give here the complete speeches by the various gen- tlemen called upon by Dr. Woodward, the toastmaster. These were Vice President Thomas R. Marshall, Hon. James Bryce, the late Dr. S. Weir Mitchell, Dr. W. W. Keen, and Hon. Theodore E. Burton. These speeches, with the remarks of the toastmaster himself, form one of the most interesting features of the present volume, which closes with the register of those in attendance. 2. Annual Tubles of Constants and Numerical Data, chena- ical, physical, and technological, published by the International Commission of the VII and VILLI International Congresses of Applied Chemistry.—Volume III of these tables is now in press and will be issued in the first half of 1914. A descriptive circu- lar with references to reviews of previous volumes may be secured on application to the University of Chicago Press. The subscrip- tion to volume III, now opened, will be closed March 31, 1914. The names of subscribers should be sent to the University of Chicago Press, to which subscriptions are payable at the time of publication. The price of volume III (as for volume II) is $6 unbound, $6.80 bound (carriage free). Members of contributing societies (the National Academy of Sciences, the American Acad- emy of Arts and Sciences, the American Chemical Society, the American Electrochemical Society, the Society of Chemical Indus- try) and of contributing manufacturing establishments are entitled to a discount of 20 per cent (but not on the binding) provided their subscriptions are received by March 31, 1914. After March 31 the price will be $6.40 (unbound) and $7.20 (bound), with a charge for carriage ; no discounts will be allowed. The Commissioners for the United States are: Julius Stieglitz, University of Chicago; Edward C. Franklin, Leland Stanford University ; Henry G. Gale and Albert P. Mathews, University of Chicago. 3. The Fungi which cause Plant Disease ; by F. L. STEvVENs. 8vo, pp. 754, 449 figures. (The Macmillan Company, New York, 1913.)—The title of this work might naturally convey the impres- sion of a popular general treatise on pathogenic fungi, perhaps after the order of H. Marshall Ward’s book, “ Disease in Plants.” Although, of course, such books have their peculiar value, the present work is, as a matter of fact, not at all of this type. Dr. Am. Jour. Sci1.—FourtH Serigs, Vou, XX XVII, No. 218.—FrEsrvuary, 1914. 15 206 Scientifie Intelligence. Stevens’ earlier book, “ Diseases of Economic Plants,” lays espe- cial emphasis on the host in its treatment of the various diseases, dealing with the symptoms of the disease, prevalence, damage caused, methods of control, ete.; the present work is more or less complementary to this, in that it treats especially of the causal organisms. After a brief introduction, the first part is devoted to the Myxo- mycetes and Bacteria, and Phycomycetes, Ascomycetes, Basidio- mycetes and Fungi Imperfecti follow on in the order named. All genera and species of economic importance in this country, and a few so far known only in Europe, are included. Each genus of pathological importance is represented by at least one figure. Practical keys for the various groups of fungi are supplied throughout the book. One of the most commendable points is the comprehensive bib- liography appended after each part. Such references to the litera- ture are particularly helpful to the student engaged in research, and as one looks through the book, the large number of fungi which are said to be “ probably,” or “ perhaps” parasitic, or of which little seems to be known save the bare fact of their occurrence on a given host, is certainly convincing as to the great need of such research. In addition to the bibliographies, lists of the more useful gen- eral works and periodicals, and a glossary and index are added. A few typographical errors appear here and there, and also some inconsistencies in the headings and bibliographical references; yet the book represents such a vast amount of labor and is so exceedingly valuable to students of plant pathology that we are loath to mention these imperfections at all. A. H. G. 4, Measures of Proper Motion Stars made with the 40-inch Refractor of the Yerkes Observatory in the Years 1907 to 1912, by S. W. Burnuamu. (Carnegie Institution of Washington, Publi- cation 168.)—As one examines this record of a great contribution to an important branch of astronomical research, it is natural to contrast it with the earliest work of the same observer, done with- out any previous training with the simplest instruments in such fragments of time as could be saved from the business of a reporter for the courts of Chicago, but stamped, then as now, with the mark of an unequaled ability for observations of this kind. It need hardly be said that with the best telescope in the world, the best atmospheric conditions, and all the appliances of a great observa- tory at command, Mr. Burnham’s work cannot. be rivalled else- where. The present volume, No. II, records about 9500 measurements, following vol. I, which covered the period from 1897 to 1906, published in the general Catalogue of Double Stars of the Carne- gie Institution. The greater number of stars given in this volume are taken from the same volume. Stars of appreciable proper motion being in general our nearest group of neighbors, astro- nomically speaking, as well as fairly bright for the same cause, Miscellaneous Intelligence. 207 the reason for selecting such double stars for measurement before exploring the field further is evident. “For many years,” the author remarks, “I have been hunting for a faint star with some certain proper motion, but so far without success.” Considerable attention has been given to an interesting group of 39 bright stars, principally in Taurus, first noted by Boss, which appear to have a proper motion in common of about 0°10” in the same direction. Mr. Burnham has added three more, and most of the 42 have been located by position angle and distance from one or more neighboring faint stars. Ww. B. 5. Astronomy: A Populur Handbook ; by Haroxip Jacosy, Rutherford Professor of Astronomy in Columbia University. Pp. 435. New York, 1913 (The Macmillan Co.).—The accumu- lation of new material in astronomy warrants the publication of popular handbooks at short intervals, and this work brings the record up to date most satisfactorily, both in the choice of mate- rial and the form of presentation. If the ordinary reader cannot digest all that it offers, it is due to the subject rather than the author whose faculty both for elucidating and condensing is unusual, as is his independence of thought. An example of the latter is his treatment of the question of the Martian Canals, a subject on which the public possesses a large amount of misinformation. Dr. Jacoby dissents from the theo- ries of Percival Lowell whom he thinks to be the author rather than the discoverer of most of these canals, and those who wish to know the argument for the negative should not fail to read this book. Many points of interest are excluded from this notice for want of space, but something should be said of the plan of the author to make the book serve the double purpose of a popular treatise and a satisfactory text-book for high schools and colleges. This he aims to do by keeping the text for 360 pages free from formal mathematics, condensing this into an appendix of 60 pages. The attempt does not seem to the present writer likely to succeed without much hard labor on the part of the teacher ; except for the author, whose method is no doubt best suited to his own class room. W. B. 6. The American Chemical Journal.—The fiftieth volume of the American Chemical Journal, completed in 1913, closes the independent career of this important periodical. Started in 1879 at the Johns Hopkins University under the editorship of Profes- sor Remsen, it has occupied from the start a prominent place in chemical literature and has done a great work in stimulating the ever-increasing activity of chemical research in this country. In future the papers, which would have come to it, will be cared for by the Journal of the American Chemical Society. An index to the fifty volumes of the American Chemical Journal is to be issued and will be sent, to those ordering it, by the Johns Hop- kins Press, Baltimore ; price one dollar and fifty cents, 208 Scientific Intelligence. OBITUARY. Dr. Siras Weir Mircuett, gifted alike as investigator, phy- sician, novelist, and poet, died on January 4 in his eighty-fifth year. Kew men of any nation or time have been able to accom- plish so rare a work in physiology and medicine and at the same time to take a place in the first rank of men of letters. His inter- ests were as varied as his vigor was unlimited, and his activity continued till near the close of his long life. He will be mourned alike by a circle of friends as broad as the nation and by a host of grateful patients whom he helped to overcome their nervous ailments. Dr. Bensamrin Oscoop Peirce, Hollis professor of mathematics and natural philosophy in Harvard University, died on January 14 in his sixtieth year. He was born on Feb. 11, 1854, and edu- cated at Harvard, Leipzig, and Berlin. His original scientific work was varied and important, extending over both mathematics and physics; the phenomena of magnetism were of particular interest to him. Dr. Serum CarLo CHAanpDieER, the distinguished astronomer, died at his home in Wellesley, Mass., on Dec. 31, at the age of sixty- seven years. Proressor Winstow Upton, head of the astronomical depart- ment of Brown University, Providence, died on January 8, aged sixty-one years. Dr. Cuartes Bupp Roginson, economic botanist of the Bureau of Science of the Philippine Islands, recently met his death at the hands of the natives of the Amboyna Islands; he was in his forty-third year, Str Ropert Staweiyt Barz, the eminent astronomer, died in London on November 25, at the age of seventy-three years. Born in Dublin in 1840, he was professor of astronomy in the University of Dublin and Astronomer Royal of Ireland from 1874 to 1892. Later he became Lowndean professor of astronomy and geometry at Cambridge University and director of the Cam- bridge Observatory. He was knighted in 1886. Sir Trevor Lawrence, the English botanist and president of the Royal Horticultural Society, died on December 22, 1913, in the eighty-second year of his age. Dr. Penry Vaueuan Brvan, the English physicist, professor in the Royal Holloway College, died on December 15, 1913, at the early age of thirty-eight years. Dr. W. PoprpLewELt Bioxam, formerly professor of chemistry at Madras, died on December 26, at the age of fifty-three years. Warns Naturat Science EstasisHMent A Supply-House for Scientific Material. Founded 1862. Incorporated 1890. DEPARTMENTS: Geology, including Phenomenal and Physiographic. Mineralogy, including also Rocks, Meteorites, etc. Palaeontology. Archaeology and Ethnology. Invertebrates, including Biology, Conchology, ete. Zoology, including Osteology and Taxidermy. Human Anatomy, including Craniology, Odontology, ete. Models, Plaster Casts and Wall-Charts in all departments. Circulars in any department free on request; address Wards Natural Science Establishment, 76-104 College Ave., Rochester, New York, U.S. A. Complete Laboratory Furnishers Chemical Apparatus, Balances, etc. C. P. and T. P, Chemicals and Reagents Platinum Ware, Best Hammered Blowpipe Outfits and Assay Goods WE CARRY A LARCE STOCK OF MINERALS FOR BLOWPIPE WORK, ETC. EST’SB - 1851 203 -211- THIRD -AVE ae NEW -YORK: CITY CONTENTS. Page Arr. VIIL—Geologic Reconnaissance of the Ayusbamba (Peru) Fossil Beds ; by H.R GREcory =. cosas 125 IX.—Vertebrate Boeale from Ayusbamba, Peru; by G. F. Eaton. ee Plates V, Vi; Vil) -2.. 2. ae X.—The “Dam” at Cheshire, Connecticut ; by F. W sco 155 XI.—Progressive Development of Mechanics based upon a New Form of Fundamentai Principle of the Science ; by HOM. DapouRIAN. 222. eo ee XII.—Temiskamite, a new nickel arsenide from Ontario ; by Ss Ta, WAL RRR GEC oe BE A a Shame 170 XIII.—Use of the Ammonium Salt of Nitrosophenylhydro- xylamine (‘‘ Cupferron’”’) in the Quantitative Separation of Titanium from Iron; by W. M. Tuornton, JR, ----- 173 XIV.—Contribution to the Optical Study of the Amphiboles ; bys W.oE) Monpess (22 ae eg ee ee 179 SCIENTIFIC INTELLIGENCE. Chemistry and Physics—Action of Aluminium Carbide upon Solutions of Metallic Salts, Hruperr and Dirmar, 193.—Solubilities at the Critical Temperatures of Solvents, F. FRrEpRICHS: Organic Chemistry for Advanced Students, J. B. Conmn: Industrial Chemistry for Engineering Students, H. K. Benson, 194.—Outlines of Theoretical Chemistry, F. H. Gmrman: Oxygen in the Sun, RunGE and PascueEn, 195.— Fluorescence of the Vapors ef Sulphur, Selenium and Tellurium, W. Steusine, 196.—The Series Lines of Neon, R. Rosst, 197.—Mass of Rapidly Moving Hlectrons, G. NEUMANN, 198. Geology and Mineralogy—Changes in Level in the Earth’s Crust, O. FisHEr, 199.—Meteorites vs. The Harth, O. C. Farrineton, 200.—San Franciscan Voleanic Field, 202.—Devonian of Maryland : Stratigraphy and Paleontol- ogy of the Alexandrian Series in Illinois and Missouri, Part I, T. EH. SAvAGE: Important part played by calcareous Algz at certain geological horizons, EK. J. Garwoop, 203.—Geological Survey of Oklahoma: Minéralogie de la France et de ses Colonies, A. Lacroix: Handbuch-der Mineralogie, C. Hintze, 204. Miscellaneous Scientific Intelligence—Semi-Centennial Anniversary of the National Academy of Sciences, 1865-1918, 204.—Annual Tables of Con- stants and Numerival Data, etc. : Fungi which cause Plant disease, F. L. STEVENS, 205.—Measures of Proper Motion Stars, S. W. BuRNHAm, 206,— Astronomy, H. Jacopy: The American Chemical Journal, 207. Obituary—S. W. MircHeL.: B. O. Perrce: S. C. CHANDLER: W. Upton: C. B. Ropinson: R. §. Batu: T, LAwREncE: P. V. Bevan: W. P. BLoxam, 208. & z ee, _ Smithsonian Institution. CU. 3. Vigne, yw.) ae VOL. XX XV IT. MARCH, 1914. | Established by BENJAMIN SILLIMAN in 1818. THE | AVR R TC AN JOURNAL OF SCIENCE. Epirorn: EDWARD S. DANA. ASSOCIATE EDITORS Proressors GEORGE L. GOODALE, JOHN TROWBRIDGE, W. G. FARLOW anp WM. M. DAVIS, or CameBrince, Proressors ADDISON E. VERRILL, HORACE L. WELLS, LOUIS V. PIRSSON, HERBERT E. GREGORY anpD HORACE 8S. UHLER, or New Haven, | | | Proressor HENRY S. WILLIAMS, or Irwaca, | Proressorn JOSEPH S. AMES, or Battimore, | Mr. J. S. DILLER, or Wasuineton. | : | | | | | pr FOURTH SERIES VOL. XXXViI—_[WHOLE NUMBER, CLXXXVIT). No. 219—MARCH. 1914. WITH PLATE VIII. NEW HAVEN, CONNECTICUT. LOLA . / | é i ! | THE TUTTLE, MOREHOUSE & TAYLOR ©O., PRINTERS, 123 TEMPLE STREET. a | 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). HODGKINSONITE, A NEW MINERAL. We have been fortunate enough to secure the best specimens of this very rare mineral. It is from the celebrated Franklin Furnace mines and is arare compound, the formula of which is Mn(ZnOH),SiO,. It crystallizes under the monoclinic system and is pink in color associated with barite. Ina few specimens itis associated with the rare minerals Pyrochroite and Gageite. The whole makes a very pretty specimen. Prof. Charles Palache has ana- lyzed and will soon publish a description of it. The quantity found is scarcely enough to supply the scientific institutions who will want a speci- men. Prices range from $1.00 to $25.00. A NEW OCCURRENCE—Fluorescent Willemite with Rhodocrosite. The Willemite occurs in transparent crystals running in size from 2 milli- meters in diameter to almost a hair in thickness. They occur in cavities on Rhodocrosite, making a very beautiful specimen. Under a current they show a strong fluorescence. Prices range from $1.00 to $5.00. SYNTHETIC GEMS. It is remarkable the interest that has been taken by scientists in these wonderful scientific discoveries. The Corundums are now produced in Pigeon blood, Blue, Yellow, Pink and White. Also the new Indestructible Pearls in strings with gold clasps. These are identical in hardness and rival in color and lustre the real gems. They can be dropped and stepped on without injury and are not affected by acids. My collection of the above is unrivalled, and prices of the same are remarkably low. OTHER INTERESTING DISCOVERIES AND NEW FINDS Will be found in our new Catalogues. These consist of a Mineral Catalogue of 28 pages; a Catalogue of California Minerals with fine Colored Plates; a Gem Catalogue of 12 pages, with illustrations, and other pamphlets and lists. These will be sent free of charge on application. Do not delay in sending for these catalogues, which will enable you to secure minerals, gems, etc., at prices about one-half what they can be secured for elsewhere. REMOVAL NOTICE. Finding that at our recent location we were placed at a disadvantage in receiving and shipping our numerous consignments, we decided to return to our old location, where we are surrounded by the greatest gem and financial district in the world, and are near-the vast shipping interests, Custom House, ete., ete. Our old offices have been entirely renovated in a beautiful manner, hew cases have been fitted and all who take the trouble to call on us will find themselves well repaid by the beautiful display of minerals, gems, curios, etc., etc. ALBERT H. PETEREIT 81-83 Fulton St, New York City. pale ease Plate VIII. Am. Jour. Sci., Vol. XXXVII, March, 1914. aug WjIM ‘sMel pue [[nys eq} Jo yoodse yy, eyy SuLMoYs ‘ eaAOqB UO IJ POMOLA ‘aspe zaddn oy} piemoy Surutod ofpped 4ysia ezIs “yeu g/{ ‘“[ISsOJ oltyue Jo Ydeisoyoyg THE AMERICAN JOURNAL OF SCIENCE [FOURTH SERIES.] +o Art. XV.—Fossil Dolphin from California; by Ricrarp Swann: Lurz. With Plate VIII. [Contributions from the Paleontological Laboratory, Peabody Museum, Yale University, New Haven, Conn., U.S. A.] Introduction. Extent and preservation of the specimen. Dedication of the specimen. Morphology. The axial skeleton. Skull. Jaw. Dentition. Vertebral column. Ribs. Sternum, The appendicular skeleton. Scapula. Humerus. Forearm. y Hand. Summary. Introduction. Slit eos Tuere has recently been presented to the Peabody Museum at Yale University a slab of sandstone containing the impres- sion of a small fossil whale, apparently one of the Delphinide, but otherwise new to science. It was found by Mr. Edwin White Newhall, Jr., of the Newhall Land and Farming Com- pany, San Francisco, and presented by him to the Peabody Museum through the intermediation of the writer. In view of the scarcity of cetacean remains from the Tertiary rocks of the Pacifie coast, and of the perfection of the relic, the specimen is eminently worthy of consideration. Am. Jour. Sct.—FourtH Serigs, Vou. XX XVII, No. 219.—Marcu, 1914, 16 210 R.S. Lull—Fossil Dolphin from California. The slab was found on the Suey ranch, owned by the New- hall Land and Farming Company, in San Louis Obispo county, California, in what is called “Sulphur Creek” canyon, about 700 feet above sea-level and 30 miles in a direct line from the ocean. According to the company’s maps, the place is in the southeast corner of Section 17, Range 32, West Township 11, north San Bernardino Base and Meridian. Sulphur creek empties into the Santa Maria river at about the southern extremities of Sections 16 and 17, and Mr. Newhall found the specimen about a quarter of a mile up the creek. The fossil is presumed to be of Miocene age, judging from the Geologic Map of North America published in 1911. The precise level, however, has not been determined, but is at present under investigation through the courtesy of Professor John ©. Merriam, of the University of California. Extent and preservation of the specimen. The specimen, which bears the catalogue number 10040, Yale Museum fossil vertebrate collection, consists of a single slab of sandstone, upon the surface of which the articulated skeleton of the dolphin was impressed (see Plate VIII). Most of the bones have been eroded away, so that their imprints alone remain, with but little of the osseous material clinging to their depths. Of the skull, however, a great deal was pre- served, for, while the exposed surface had suffered from weath- ering, removal of the matrix from the reverse side displayed most of the right aspect of the cranium and jaws. There remains of the animal, therefore, the entire skull and jaws with many of the teeth, some entire, some having but the root embedded in the alveolus, and others represented only by their imprint in the matrix. The vertebral column is indi- cated as far as the twelfth dorsal, though each bone cannot be clearly distinguished. There are twelve pairs of ribs and the impression of the sternum is distinct. Of the limbs, the left scapula, the humerus, and the ulna are represented, and at least four phalanges of the second or third digit, while of the right paddle, impressions of all the elements remain except the distal ends of the several digits. There is unfortunately no trace of lumbar or caudal vertebrae, nor have I recognized any portion of the pelvis, which would be of great interest in so old a form. I should judge the animal to be fully adult despite its rather small size, as the bones show scarcely a trace of the epiphyses. The only preparation for study necessary was the removal of the matrix from the right aspect of the skull, and, after a careful preliminary examination, it was deemed expedient to R.S. Luli—Fossil Dolphin from California. 211 eliminate the bone fragments clinging to the depths of certain of the impressions, especially those of the cervical series and of the right paddle, in order to secure by means of casts and otherwise the more exact contour of the bones themselves. One should fully realize that while the imprints are often deep, certain outlines become very vague and the extent of the vari- ous processes very difficult to determine. The entire specimen is in a.sense comparable to the deeply incised but partially effaced inscriptions on an Egyptian obe- lisk, and the interpretation of such an ancient hieroglyph is necessarily halting. Dedication. I propose for this form the name Delphinavus newhalli, n. gen., n. sp., the generic distinctions in contrast with Delphinus being outlined below. The species is dedicated to the family of Newhall, twice a contributor to the Yale collections, first through the late Henry G. Newhall of the class of 1874, Shef- field Scientific School, who accompanied Professor Marsh upon the Yale Expedition of 187 3, and again through his nephew, Edwin White Newhall, Jr., ‘Massachusetts Agricultural Col- lege 1906, the discoverer and giver of this important specimen. Morphology. For the purpose of comparison, I have, through the cour- tesy of Doctor George F. Eaton, used a mounted skeleton of Delphinus delphis, catalogue number 265, Peabody Museum osteological collection. This specimen is not quite mature, as the vertebral epiphyses are still free from their respective centra. The skeleton measures 5 feet 10 inches from tip of rostrum to that of the caudal series. The fossil, which as I have said, seems to have been fully grown, would have an estimated length of perhaps 5 feet, though the proportions of the various regions which serve to make up the total length differ in the two forms, as I shall show. Tuer AXIAL SKELETON. Skull.—In the proportions of cranium to rostrum the form under consideration agrees most nearly with the modern dol- phin, Delphinus, in contrast with those proportions seen in Phocena. The most apparent cranial contrast with Delphinus is the higher, more nearly vertical occiput which in the fossil makes a sharp angle with the roof of the skull at the vertex, instead of rounding forward as in the modern form. The pre- cise limits of the various bones are not to be determined with assurance in many cases, but they agree in general with those 212 R.S. Lull—Fossil Dolphin from California. of Delphinus, with the following exceptions. The exoccipital bones in the fossil have a much less lateral extent in proportion to their height. The squamosal bone is very robust and broadly expanded antero- posteriorly so as to be both actually and rela- tively larger and heavier. The orbit is large and well rounded and the superior limiting bones, frontal and jugal, are compar- Inaxey, al, Fic. 1. Skull seen from the right, about % nat. size. atively heavy. The frontal plate of the maxilla does not appear to have the full lateral or posterior expanse which is character- istic of the modern type, as its presence cannot be detected in the antorbital or orbital regions viewed laterally. What appear to be fragments of the zygomatic process of the jugal are at least twice the diameter of that of Delphinus. In the sharp- edged posterior limitation of the temporal fossa, the fossil agrees with Delphinus in contrast to Phocend. ” Jaw.—The two forms may be contrasted in the proportions of this bone, for whereas the entire length of the mandible is 278"" in the fossil to 370" in the recent form, their vertical expanse at the posterior end is essentially the same. This R.S. Lull—Fossil Dolphin from California. 213 coupled with a more robust angle in the fossil implies a greater proportionate jaw power in the older form. In the latter the symphysis is apparently longer, though the distal two-thirds of each jaw is about equal in slenderness. oe . 40 Dentition.—The dental formula of Delphinus is ee In the fossil the number of teeth in the upper jaw is exactly the same, Fie. 2. Fic. 2. Right lateral aspect of skull and jaws. 1, nat. size. Bo, basi- occipital; C, condyle of jaw; D, dentary; Evo, exoccipital; Fr, frontal ; J, jugal; Max, maxillary; Nas, nasal; Oc, occipital condyle ; Pa, parietal ; Pmx, premaxillary; Pn, ? posterior nares; Sq, squamosal; So, supraoccipital. but the imperfection of the mandible renders a count of the lower dentition impossible. There is no reason, however, to suppose that the number here was not also essentially similar. In each instance the premaxillary bone is toothless. The teeth themselves are also alike, being simple, slightly re- curved cones without carinee, and with smooth, polished enamel. They contrast sharply with the spatulate crowned teeth of Phocena. In so faras Iam able to judge, the skull and dentition of the fossil agree essentially with those of Delphinus, the exceptions being all in the direction of greater cetacean specialization on the part of the latter, such as the character of the occiput, the reduction of the zygomatic arch, and the lessened jaw power. 214 L. S. Lull—Fossil Dolphin from California. Measurements of the skull. Delphinavus Delphinus Cat. No. 10040 Cat. No. 265 Keneth including rostyomy. esse eee 299mm 44mm és anterior end of jugal to tip of rostrum 181 265 Height, vertex to lower margin of exoccipitals 117 150 Antero-posterior diameter of orbit -_--.-.-- 35 49 ene thvot mandibles sss ae seen ee eS 370 Height of mandible at posterior end _.-.--. 64 66°5 Space for 10 alveoli in rear of maxillary -. -- 3 47 Maximum length of crown of tooth -_--_--- 06 08 Sguamrosal hero tases a as ya eres 40 41°5 cs exposed: breadthi== 23-5" = 22555 3 21 Tympanic bone, greatest diameter ----- ---- 30 oa Height of mandible at mid-length ____----- 175 26°5 ene thvofsymphysis 2.2 ese A 37 Vertical diameter of mandibular condyle --_- — 20°5 23 Vertebral column.—Cervical vertebre. The cervicals are recorded only as deep impressions in the matrix, so that their Fic. 5. Impressions of cervical vertebre. 15 nat. size. J, atlas; fac, posterior facet of atlas; z, postzygapophysis of axis; d/, ?first dorsal; VJZ, seventh cervical. interpretation is accompanied with great difficulty and the apparent results may not be in harmony with the actual facts. The vertebree seem to be entirely free, the atlas presenting its posterior aspect, with well-developed, slightly convex facets for articulation with the axis, such as Delphinus might well exhibit were the bones not coalesced in their lower half. The right diapophysis is indicated. The axis seems to have been a rather RF. S. Lull—Fossil Dolphin from California. 215 massive bone, and exhibits again the posterior face of the cen- trum, which is concave with well-rounded margins. The axial arch and spine, the former exhibiting a well-developed post- zygapophysis upon the left side, are well preserved but appar- ently entirely detached from the centrum. I can discern with fair assurance the impressions of five or six cervicals in proper sequence, but what is apparently the sixth shows a diapophysis comparable in development to that of the seventh in Delphinus, while just beyond it there lies an impression very suggestive of the diapophysis which in the modern form articulates with the tuberculum of the first rib ; one cervical is therefore either missing or too obscurely defined to be recognized. In comparison, then, with Delphinus, the fossil again exhibits more generalized traits in the free cervicals, in their actually greater length in the smaller specimen, and their much greater relative size. They seem to represent more nearly the degree of evolution shown by Jnza, but display no feature which would debar the older type from direct ancestry to Delphinus. Measurements of the cervical series. Delphinavus Delphinus Cat. No. 10040 Cat. No. 265 Meng thvotentire seriesse eee ace = en ee ed Aa Wadthrotlatlas Sasi sl Sk oG eee 62°5 119 Height of axial spine measured from base of POstzyeapophysisea-m © ess eee eee 46 40 Dorsal vertebre.—At least twelve rib-bearing vertebra are preserved, comparable in number to those of Phocewna, and three fewer than in Delphinus delphis. However, this is of little moment, as the remainder of the body is lacking in the fossil so that the number of dorsals and ribs may readily have been greater. Fia.4. The vertebree differ markedly from those of either of the above mentioned genera in the much greater length of centrum and in the low neurapophysis in contrast with the very high spine of an approximately equivalent vertebra of the recent form. The proportions of the vertebral centra are more nearly ye. 4, Ninth dorsal those of nia again, though in that genus vertebra, left aspect. 14 the neutral spine is much higher. mab size.) Imp.) impres- s10n, * This is not a just measurement, for the vertebre are slightly displaced. 216 RS. Lull—Fossil Dolphin from California. Measurements of dorsal ? LX, Delphinacus Delphinus Cat. No. Cat. No. 10040 265 iene the oteacemtr umes senate 3gmm 23 3m Height of centrum, posterior end.__. 21:8 29°5 Total height, including neural spine.. 44:1 110 Libs.—Twelve pairs of ribs are preserved in the specimen, of which at least four pairs were double-headed, in comparison Fic. 5. Fic. 6. | % Fie. 5. Fourth rib. 1, nat. size. Fic. 6. Sternal impressions. 1¢ nat. size. with five in Didelphis and six in Phocena. It seems hardly possible that in this regard the ancient type is more specialized than Delphinus, while in every other respect in which they do not agree the reverse is true. The ribs are quite comparable to those of Delphinus in proportions and general curvature. Measurements of the ribs. Delphinavus Delphinus Cat. No. Cat. No. 10040 265 Length of first rib, measured along the chord 99™™ 10mm Lene th of tournthenios sesso eee 200 225 R. S. Lull—Fossil Dolphin from California. 2 The torso of the fossil form was evidently fully as long and deep as that of the dolphin used for comparison, though the skull of the latter is materially larger. Sternum.—There are two impressions not far removed from their normal position which I take to be those of two coalesced and one isolated sternebree. The outlines are very vague in each instance but the anterior one seems to differ very mark- edly from that of Delphinus and to resemble more nearly that ~ Fic. 7. Fic. 7. Right pectoral limb. 1, nat. size. ac, acromion process ; cor, coracoid process; cun, cuneiform ; h, humerus, anterior aspect ; ha, hume- rus, internal aspect ;/, lunar; m, magnum ; mcp, metacarpal ; p, phalanx ; r, radius; s, scapula; se?, ?scaphoid ; ¢?, ? trapezoid ; u, ulna; wne, unci- form. Drawn from the impression. of Platanista. The difference is mainly in the greater size of the element in the fossil and in its simplicity of ontline. The anterior margin has a simple semicircular notch about one-third the width of the entire bone, whereas in Delphinus the notch 218 R.S. Lull—fossil Dolphin from California. has become converted into a foramen through the closing together of the anterior margins. In Delphinus the lateral edges of the anterior sternal are prolonged into outward and backward projecting cornua, while in the other there are simple rounded angles in their place and the width of the sternal posterior to them is relatively much greater. Measurements of the sternum. Delphinavus Delphinus Cat. No. Cat. No. 10040 265 Length of sternals 1 and 2-.--. 120™™ ae Greatest breadth __-.--------- 70 84 I cannot positively identify sternal ribs in the fossil. THE APPENDICULAR SKELETON. Both paddles are preserved, that of the right side almost entire, the left more or less occluded by the skull but supply- ing certain details lacking in the other due to the orientation of certain of the bones, notably the humerus. The limb is essentially delphinoid, showing apparently the same reduction of digits, but differing mainly in the more primitive character of the individual elements. In so far as it is preserved, the limb is of approximately the same size in both modern and fossil forms. Scapula.—The scapula is less widely expanded than that of Delphinus, otherwise the two are very suggestive of one another. In the fossil the horizontal extent of both coracoid and acromion processes cannot be ascertained ; they are, how- ever, similar in the extent of their origin to those of Delphi- nus. The prescapular fossa seems to have been relatively greater in the ancient type. Humerus.—The ratio of humerus to radius is much greater in the fossil, as one would be led to infer in the case of the more primitive form, The humerus is also relatively more slender, but has much the same general outline as in Delpha- NUS, except that the tuberosity projects slightly above the head and has a lesser transverse extent. A faint indication of the capitular epiphysis may be seen in the fossil, practically the only indication of one in the entire skeleton. The right humerus presents the impression of its posterior aspect, while the left is viewed from without. The latter shows the charac- ter of the distal facets, each of which is planoconcave, the curved axis lying in the transverse plane. The radial facet is the greater, as in Delphinus. Forearm.—The forearm presents perhaps the greatest diverg- ence in the two forms of any part of the limb, for instead of R.S. Lull—Fossil Dolphin from California. 919 being nearly parallel as in Delphinus, the two bones are sepa- rated by a wide lenticular space, the curved line of the radius being particularly noticeable. _In the ulna the olecranon is better dev eloped but simple in outline. Hand.—The carpus is essentially alike in both genera, but here again the fossil is the more primitive in that the magnum is repr resented by a well-defined lenticular bone lying at the proximal end of metacar pal IIT; whereas in the specimen of Delphinus used for comparison, it is reduced to a tiny nodule closely adpressed against the unciform and visible only from the inner aspect of the wrist. Its position, however, is exactly the same in each form. In the fossil the bones of the anterior side of the carpus have been displaced, but their identity seems fairly certain. The metacarpals show much the same proportions in each genus, but the preserved phalanges are relatively longer and more slender in the fossil, no separate phalangeal epiphyses being discernible. The digital reduction has gone further in the modern type, as evidenced by the proximal phalanx of digit IV, which is a much better developed bone in the fossil. Measurements of the pectoral limb. Delphinavus Delphinus Cat. No. 10040 Cat. No. 265 Scapulavsherght qs es aoe a ee else LOR eE 104™™ es ne ad tht Sree eee eee 146 160 Fiuimenuss: Lenod ie eee ee eee ere GO) 60 gs proximal breadithmt a =e see 37 49 ss ant.-post. diameter, distal end_ 31°5 38°5 . transverse “ mid shaft. 19°5 24 vadiussslem ote = ee ee hee ae = OA 84 a diameter, proximal end._--.---- 25 28°5 es ss distale end aemiss sae 33 37°5 Wihnarlen Ghose naan ee eta See elie 72 diameter, proximal endy: 22252222 26°5 30 e distalende 222 eee 26 oe Metacarpal EN lene tives sas wie keene 20 32 <= breadth of shaft 2-22-22 tel 14 Digit III, 1st phalanx, leneun Se Neat eee OLD 21-5 (73 6c od 2 Para gate Ce Gn es te Rete Sine ee 11°5 1 7 ce 66 ‘3d ce SC teller epee mred eae 08 13} Length of digit, inc. metacarpals, estimated 80 100 “entire limb, inexscapula, — “ 320 360 Suminsary. The specimen under consideration is clearly a dolphin allied to Delphinus and hence of the family Delphinide. It com- * No epiphyses nor articular cartilages indicated, but bones are spaced. + Including epiphyses, bones in contact. 220 R.S. Lull—Fossil Dolphin from California. pares favorably with the dolphins of this genus in point of size, proportion of rostrum to cranium, in dental formula, char- acter of the dentition, in having approximately the same num- ber of double-headed ribs, and in the general character of the pectoral limb, more especially of the manus. The contrasts are in practically every instance due to the more primitive character of the fossil, and I fail to recognize one which would debar the more ancient type from a direct ancestry to Delphinus delphis itself. To,sum up these con- trasts, they are: In the skull, the straighter, more nearly ver- tical occiput, the minor proportions of the individual cranial elements, the more robust zygomatic arch, the less expanded maxillary plate, and the more broadly expanded and powerful rear portion of the lower jaw. The cervical vertebree, in so far as may be learned, differ mainly in the lack of fusion of atlas and axis in the fossil and the somewhat greater collective length of the entire series. The dorsals are much longer in centrum and lower in the height of the neural spine, while the number of double-headed ribs in the two specimens is appar- ently four for the fossil and five for the recent type, the only instance wherein the former seems to be more specialized, and this may be due to lack of perfection of the fossil. The sternum of the older form is simpler and more general- ized, but easily modified into the modern type. In the limbs, the contrasts lie in the less widely expanded scapula, relatively longer humerus, wide lenticular space separating the radius and ulna, and in the simpler form of the olecranon in the fossil. In the carpus the magnum is relatively large, while in the recent type it is vestigial. The phalanges are relatively longer in the ancient type. The degree of evolution of the two dolphins is doubly inter- esting when compared with the advance manifest in the same length of time in certain terrestrial families, notably the horses. It points to the early adaptation of the Cetacea to their aquatic habitat and their slow subsequent evolution, which amounts, in this instance, merely to the adjustment of certain minor details, some, as in the jaw power, along the line of degeneracy, others mainly for the perfection of speed. This evolution shows the impelling force of topographical changes on the land, which not only directly modify the animal form and motor organs, but also, through influencing climate and thereby vegetative life, necessitate new adaptations on the part of terrestrial animals or the elimination of those forms which cannot thus change. All of this passes harmlessly over the head of the cetacean, which can migrate when the waters become too cold, and upon which none of the other influences can make themselves felt. Pirsson and Schuchert—Oriskany Formation. — 221 Arr. XVI.—Wote on the Occurrence of the Oriskany For- mation on Parlin Stream, Maine; by L. V. Pirsson and CHARLES SCHUCHERT. | Tue material which is the subject of this communication was collected by one of the writers (L. V. P.) while traversing this portion of Maine with his guide, and as Oriskany fossils occur here it has seemed worth while to put this on record and sive a list of the species obtained, together with a mention of _ the exact locality and some additional observations which may be of future service in the study of the geology of this part of the state. Parlin Stream is the outlet of Lake Parlin, which lies 20 miles southwest of Moosehead Lake, where Oriskany fossils have long been known to occur. While the mere crossing of a country which is so heavily forested, as is this portion of Maine, affords little opportunity to gain much idea of its geology, the following facts were observed, in relation to this occurrence. The fossils were found in loose, angular blocks of a much indurated shale or shaly sandstone of a dark green to brown eolor, weathering with a blackish chocolate-brown crust. They are seen on exposed bedding planes of the firmer mate- rial or in more weathered, cavernous, sandy fragments and they constitute a great part of the rock mass, like the so-called shell limestones of other places. These blocks occur on the west side of the large pool immediately below the partly burned lumbering dam, about a quarter of a mile above the falls on Parlin Stream. The locality is about two and a half miles from the outlet of Parlin Lake. This same locality was collected from by Gilbert van Ingen in 1889 (Bull. U. S. Geol. Surv. 165, 1900, p. 89). The country in this locality is composed of heavily bedded, much indurated, or metamorphosed, strata which have been upturned. At Parlin Falls there is a good exposure of the beds. The falls are made by the stream passing downward over the inclined back of a heavy resistant bed. At this point the strata have a general east and west strike and dip north- ward at a considerable angle, perhaps 45°. Below the falls the stream has cut a very respectable canyon, whose exposures should afford a good opportunity for study of the formation. All the exposures weather with the blackish crust mentioned above. While the material containing the fossils was not observed in place, it is quite certain from considerations of the loeal geology seen that it belongs in its immediate neighborhood, 222 = Pirsson and Schuchert— Oriskany Formation. and had suffered little transport. This is shown by the abund- ance of blocks, their large size, angular shapes, unworn condi- tion and characteristic resemblance to the near by rocks in place. It is presumed that they represent the debris of broken down beds of the immediate vicinity. The supposition that the blocks are local in origin is further supported by the rather noticeable absence in this ‘neighborhood of the coarser material of glacial transport. Glacial sands and gravels containing peb- bles or ver y small bowlders are common, ‘and in morainal forms, as in drumlins, ete., are rather frequent topographic features, but the larger elacial erratics, which are so striking a feature of the region farther south, were not observed in the stream beds, on ‘the lake shor es, or in other places where erosion might have brought them to the surface. Toward the lower end of Lake Parlin there is a large intru- sive mass of trap rock. It is exposed in outcrops on the road ; large bowlders of it are seen along the lake shore, and massive outcrops at several places. It rises into Parlin Mountain, where, as it was found at the north end and on the trail to Long Pond as it winds about the mountain, it is inferred that it forms the main upper portion of this eminence. No con- tacts were seen and its intrusive nature is inferred from the size and position of the mass, the nature of its jointing, and the petrographic characters of the rock. Ona fresh surface of fracture it isa dark gray to green and of uniform and fine tex- ture. Examination in thin section shows that it is a greatly altered diabase. Originally it consisted of plagioclase and a brownish augite with accessory iron ore and apatite, with occa- sional very fine, delicate micrographic intergrowths of quartz and orthoclase filling interstices between the plagioclases. The structure was the characteristic ophitic one, but the proportion of plagioclase to augite was large, making the rock one of feldspathic type. From this condition the rock has been greatly altered ; the augite has been changed into chloritic substance, which pre- serves the outlines of the ophitic fillings between the feldspars and contains in a few places still unaltered pieces of the unchanged augite. While this chloritic substance, for the inost part, appears of the nature of ordinary green, pleochroie chlorite with low double refraction, it contains frequent minute bundles of radiant fibers of a more yellowish color and with rather high double refraction, which is suspected to be an iron- bearing serpentine. Some separated silica also is seen, but no formation of any carbonate was observed. The feldspars are also altered and the original albite twin- ning so much obscured or obliterated that their character can- not be accurately determined, but they are thought to be Pirsson and Schuchert— Oriskany Formation. 228 andesine-labradorite. They are also filled with infiltrated chloritic substance. The iron ore is changed to leucoxene and the apatite and micrographic intergrowths are the only unchanged constituents. The rock finally weathers with a brown crust of limonitic material. The mass is a large one and must be a prominent factor in the structure of the local geology. The fossils collected by Mr.and Mrs. Pirsson on Parlin Stream are from the Moose River sandstone, one of the Oriskanian formations. The first good geologic account of this region dates back to 1861, when Charles H. Hitchcock determined the Oriskany sandstone here. In 1889, Van Ingen collected fossils along Parlin Stream for H. 8. Williams, and in 1905 O. O. Nylander gathered other material for John M. Clarke. The latter has brought together all that is known about the Oriskanian of this area in his work, “ Early Devonic History of New York and Eastern North America” (Mem. 9, N. Y. State Mus., 1909, pp. 52-90, pls. 12-21). Below is given an annotated list of the Oriskanian fossils from Parlin Stream, just below the old dam above the falls, Somerset County, Maine, collected by Mr. and Mrs. Pirsson. BRACHIOPODA. Pholidops terminalis Hall. ‘Two specimens. Dalmanella drevermanni Clarke? 'Two small specimens. Chonostrophia dawsoni (Billings). One large and two young specimens. Leptocelia flabellites (Conrad). Common. Spirifer pirssone, n. sp. Clarke did not name this form, but has it from Telos Lake dam, and Moosehead Lake, 7 miles north of Kineo. He states that the form is common and that he thinks it “will prove unlike any known to us.” The interior resembles very much the common Oriskanian guide species, S. murchisoni, but as all of the specimens are mature, never attain to one-half the size of S. murchi- soni, and have a smaller number of plications, it is here proposed to name the new form S. pirssone, after Mrs. Pirsson, who collected and presented to the Pea- body Museum of Yale University an abundance of nat- ural moulds. Further, the six plications on each side of the fold and sinus never attain the prominence or angularity of those in 8. murchisont. The species is based on the cotypes figured by Clarke in the work cited on Plate 20, Figures 9-12. Spirifer, a. sp. This common species gives the impression of being 224 Pirsson and Schuchert—Oriskany Formation. closely related to S. pirssone, but although both are fimbriate Spirifers, they are clearly of different stocks. The two forms are not associated in the same layers. The species under discussion is more quadrate than the fore- going one, has a-narrower ventral cardinal area, and therefore is also less thick shelled and less rostrate in these parts. It differs chiefly, however, from 8S. pirs- sone in having a wider and more flabellate ventral muscle field. As the material is hardly good enough to figure, the form is left unnamed for the present. Meristella lata (Hall)? . aS 1400 got gh gb vqiod WbF 19l06 19/6] 19 0§ 19/94 ge. H.W. Fic. 1. Curve of the consecutive annual temperatures observed at Are- quipa. A,B, C,D: pleionian crests. The meionian depression 6. may have been accentuated by the effect of volcanic dust produced by the Mt. Pelé erup- tions or by other causes. The curves connecting the crests and the depres- sions demonstrate the existence of a long range variation different from the sunspot frequency variation. Rico, give a slightly retarded curve. This is a most interest- ing fact because the pleionian crests of Porto Rico could not be retarded if these temperature anomalies did not have a tendency to persist, combined with a tendency of displacement. Considering, for the same period of years, consecutive curves for a great number of stations of temperate or even arctic regions, I found that the phenomenon of pleionian vari- ations is less complicated than. would be expected, if we only admit that it is a dynamical phenomenon. The pleions displace themselves as waves. But, as the variations of Havana, Bulawayo, Mauritius, Batavia, ete., are nearly simultaneous, one would be inclined to think that the displacements of pleions go from the equatorial regions towards the poles. H. Arctowski—About Climatical Variations. 309 This is not the case. The variation in New York presents most striking coincidences with the Arequipa variation, and the same may be said about the curve of Kazan, in Russia, and North of Iceland, the observations made on Grimsey Isl., under the polar circle, show also all the crests and depressions of the Arequipa curve, slightly retarded. Therefore, the impulse which produces the pleionian vari- ations must be felt simultaneously all over the world, and Fic. 2. Fri a te een has aK + by 2 m I 1 ay ae Nee whe ab ron ve i exe i Y twa ci . IS EW ex ne 1 Fic. 2. Pendulations of the pleionian center : 1900-1909. then,—just as in the case of a handful of stones projected on a quiet surface of water,—from distant points where the impulse is felt directly, the so produced anomalies propagate themselves. We see now how the conception of thermopleions has to be enlarged. Pleions are not simply the areas occupied by posi- tive departures of temperature and antipleions the areas of negative values. They are thermal waves. It follows that, for practical purposes, it would be best to define a pleion as being the area where the consecutive annual means present a tendency to rise. But then the rise of temperature will be felt all over the area only in the case of stationary pleions. If the pleion displaces itself the rise will be observed on the wave front, whereas on the other side we shall notice a progressive 310 H. Arctowski—About Climatical Variations. decrease of temperature. This, of course, applies to consecu- tive means. In order to study the displacements of pleions, I made detailed annual departure maps, for the years 1900-1909, for Scandinavia and Central Europe, for North America and for the Atlantic. I hope to publish these maps as well as the discussion of the results obtained and, at present, I will simply mention some of the interesting points of my investigation. The main result gained by the study of the European maps is that, during the years 1900-1909, the pleions and antipleions did not move from the Atlantic towards Asia. On the con- trary: the displacement was from the NE. towards the SW., or from the E. towards the W. Moreover, these displacements did not cross the areas of maritime climate. The big pleion- ian variations of Europe appear to be a purely continental, and perhaps, arctic phenomenon. But, the areas where they are formed, in situ, are probably not always the same. For the United States I have utilized, besides the annual detailed departure maps and the consecutive curves of a selection of stations, also consecutive maps obtained from the district means published in the Monthly Weather Review. I will notice a few problems arising from a closer examination of these consecutive maps. First of all, during the years 1900-1909, the pleions as well as the antipleions displayed a tendency of persistence. No see-saw movement, between a pleion and an antipleion, leading to the gradual disappearance of both and then to the forma- tion in situ of a pleion on the place formerly occupied by the antipleion and vice versa, could be traced. Minor see-saw oscillations took place, but served simply to increase or de- crease the contrast between the pleions and antipleions with- out destroying them. Together with the tendency to persist goes a tendency of displacement. These displacements are generally gradual and continuous, but sometimes they may be very fast and in a striking contrast to the nearly stationary conditions which preceded or followed the rapid change of position. Another fact is the tendency of the pleions and antipleions to remain on the continent. In other words, the phenomenon of the variation in the distribution of the anomalies of yearly temperatures in North America, is a purely North American phenomenon. The pleions and antipleions seem to be cor- related or bound together. One depends on the other. If one moves the other moves. The area of the North American continent seems not to be wide enough for the simultaneous presence of many pleions and antipleions. In order to remain on the continent the motion of a pleion involves a displace- HH. Arctowski— About Climatical Variations. 311 ment of the antipleion in an opposite direction. A rotatory movement is the consequence. Jt ts a pendulation (fig. 2). The principal problem is, of course: what keeps the pendu- lation going? Without some exterior impulse the movement would die out or could not even originate. I think that it is the cause of the formation of pleions which, repeating itself more or less periodically, gives the impulse to the clockwork. The Russian pleions have shown some correlations with the equatorial variation of temperature, illustrated by the consecu- tive curve of Arequipa. The consecutive curve of New York also belongs to the Arequipa type. We see now how the tendency of the pleions to maintain their existence complicates the prob- lem of their mode of’formation or origin. Because, since for certain parts of the United States the con- secutive temperature curves belong to the direct type,—that is to say, are similar and coincide more or less in time with the equatorial curves,—the impulse producing these variations must be the same as that which produces the tropical varia- tions. This impulse is evidently extra-terrestrial. Therefore, where the variation is direct, the departures of temperature will not be due to abnormal conditions of atmospheric circula- tion, but will, on the contrary, produce such changes of atmos- pheric pressure, wind direction and velocity, etc., as may be characteristic for either pleions or antipleions. But on the maps the pleions do not disappear: they move away. Now the question is how—in a direct type of variation— the pleion corresponding to the second crest of the consecu- tive curve is renewed? Isit the same pleion coming back from the region it was pushed away from by the formation in situ of the direct antipleion, or is it a new pleion, and if so what became of the first one ? Let us call the pleionian crests of the Arequipa curve A, B, C and D (fig. 1). The consecutive maps show that the crest B of New York went northwest over Canada and then southwest towards California. The pleion came back nearly the same way during 1904-1906 (fig. 2). The crest C of New York is therefore the same as B. But if we try to follow this pendula- tion on the consecutive curves of individual stations we do not succeed very well. And this is because the amplitude of the departures changes independently of the pendulation. The pleions pendulate and surge at the same time. An old pleion may be reintensified. In “the case of the pleion ©, the surg- ing is nothing but the superposition of a new pleion upon an old one, so that C is the residual of B, plus a new impulse produced in situ under the influence of the direct solar action. In this way it is conceivable how the pleionian variations may be more important on the North American continent than identical variations in tropical regions. S12 H. Arctowski—About Climatical Variations. I have taken the differences between the highest and lowest departure for each of the 109 consecutive maps, in other words, the total amplitudes between pleions and antipleions. From 1900 until 1906 the reversed curve expressing these fig- ures graphically and the Arequipa curve are similar. Then, between 1906 and 1907 an interesting anomaly is noticed. Moreover, the Arequipa temperatures “decreased from 1900 to 1909. In the United States the differences between pleions and antipleions were also decreasing during that period of years. Independently of the pendulations, the Arequipa vari- ation affects therefore the entire system of North American leions and antipleions. I should like now to pass in review some results obtained from the close examination of European rainfall data, for the years 1851-1905, and to other researches | made on the varia- tions of atmospheric pressure and of sunshine records. For rainfall and the other meteorological elements the con- secutive curves also show perfectly characteristic crests and depressions. Moreover, the rhythms are absolutely similar to those of temperature. In consequence we shall have to speak of pleions of atmospheric pressure or baropleions, or, to use a simpler expression, of baros and antibaros. And likewise, in the case of rainfall, we shall have ombropleions or ombrons and antiombrons. And in the case of sunshine, we shall have to speak of helions and antihelions. It will be necessary to establish the details of all these varia- tions before it will be possibie to search for the mutual rela- tionship of thermopleions, baros, ombrons and helions, and only then will it be possible to examine the role of the centers of action of our atmosphere, and this will lead to the study of seasonal anomalies. It is an immense field of research work to be made, but, as will be seen immediately, all this is only one side of the ques- tion of climatic variations. We have indeed to consider also the macropleionian variations. 2. Macropleionian variations. In order to detect these variations it is necessary to form consecutive means of groups of years. It is impossible to take 4, 5, or 6 yearly means because then the chances are that the figures will be greatly influenced by the pleions. To eliminate the pleionian variation it is necessary to take 10 yearly means. But even then some predominant pleions may affect the means. On the other hand, it is impossible to take a longer period than ten years, first of all because it is only exceptionally that we have to deal with long temperature records which may be con- HI. Arctowski—About Climatical Variations. Silla sidered as being perfectly homogeneous, and then, because a variation of about 18 or 19 years seems to exist. Utilizing Bigelow’s tables of temperature data for the United States, for 1873-1905, I made consecutive means of decades of years and inscribed the departures on maps. To sum up the results obtained by the inspection of these maps, I will say that the long-range variations of temperature, of particular stations in the United States, are due to irregular pendulations of macropleions and macromeions, that these. pen- dulations are complicated by the existence of slight see-saw movements (or vibrations) which increase or decrease the departures, making the macropleions more or less accentuated, and that, finally, the entire system of macropleions and macro- meions moves up and down. ‘This last movement is shown on the maps by an increase in size of the macropleions and a decrease of the macromeions or vice versa. This is the real long range variation. The decade of 1873-1882 is a typical example of a widely spread out macropleion, the decade 1883-— 1892 shows a predominant macromeion. For Europe, I also made maps giving the distribution of the percentage of rainfall during the consecutive decades of 1851— 1905, and found that the macroombrons display phenomena very ‘similar to those of the thermopleionian variations. A priori, one may presume that the variations interesting longer periods of years willalso pendulate. I do not think that this hypothesis is in contradiction with the results Ellsworth Huntington has obtained from the discussion of historical data. Concerning geological records it would be absolutely prema- ture to express any opinion. But, for practical purposes, the brachypleionian variations present a greater interest. As yet I have only studied the brachypleions of Arequipa. 3. Brachypleions of Arequipa. For this study I had to take the daily values of temperature most kindly sent to me by Prof. Edw. C. Pickering. I made 5 day or pentade means, and then, consecutive monthly means,—the pentade and not the day being the unit. In this way I noticed that during certain years (1902, 1903, 1906 and 1907 in particular), a brachychronic variation of about 55 days duration, from one maximum to the following maximum, is the most striking feature of the curves (fig. 3). The amplitudes of these variations may be as important as the amplitudes of the pleionian variations, that is to say, equal the value of the annual variation of temperature. 314 H. Arctowski—About Climatical Variations. Conclusion. The main question is to know what produces these variations of short and of long duration 4 During some volcanic eruptions great quantities of pulver- IRSRCeBAAbeL ee Ate \Va ie iso lS | ae eee Peameie | BRE Re RB Be BB Po | Lele ce i a coe fe cay Be IM IZIN RUT ha LV |" calle rate ae! papel TP x. Af, Fie. 3. Brachypleionian variations at Arequipa. Vae = Ne (=) ° da ized ashes have been projected to altitudes higher than the mean elevation of the cirrus clouds. Being above the clouds, H. Arctowski—About Climatical Variations. 315 this voleanic dust could not be washed out from the atmos- phere, and remained therefore in suspension long enough to be spread out all around the world, by the western drift of the general circulation. Recently, C. G. Abbot has shown that the dust projected by the eruption of Mt. Katmai, on June 6 and 7, 1912, affected the solar constant determinations made by him and Mr. Fowle in Algeria and on Mt. Wilson. However, the occasional presence of volcanic dust, produc- ing a general haziness of the atmosphere, as in the case of the Krakatoa and Katmai eruptions, in particular, is a very inade- quate explanation of the formation of antipleionian depressions of temperature. The consecutive curves prove it very well. The antipleionian depressions, on the curves of tropical sta- tions, reoccur more often than the eruptions producing atmos- pheric haziness. Besides that, the sedepressions are not V- or U-shaped discontinuities, marking interruptions on a continu- ous curve, but are preceded and followed by crests. Now, the presence of such crests could not be explained by the volcanic dust hypothesis. Al! this, supposing of course striking coinci- dences between the depressions of the equatorial temperature curves and the volcanic eruptions producing dust veils in the higher layers of our atmosphere, which is only exceptionally the case. The brachypleionian, the pleionian and the macropleionian crests of the consecutive curves demonstrate therefore with evidence the fact that more or less periodical changes of the solar constant must be the real primary cause of the various climatic variations. Considering the means of the solar constants, observed by Abbot during the summer months of 1905, 1906, 1908, 1909 and 1910, and comparing the differences between the mean values with the corresponding differences of temperature in Arequipa, I arrived at the conclusion that a difference of 1° F. corresponds to a change of 0-01 of the solar constant. If this is the case, the lowering of temperature required to produce ice age conditions, on our earth surface, could easily be explained by variations of solar activity, in a measure which does not exceed the differences actually observed between the individual values of the solar constant. Differences of 0°15 to 0°20 have been observed every year. Such differences, if permanent during some centuries or thou- sands of years, would produce the required differences of tem- | perature. Hastings-on-Hudson, December 27, 1913. 316 Lahee—Late Paleozoie Glaciation. Art. XX V.—Late Paleozoie Glaciation in the Boston Busin, Massachusetts ; by Freperitc H. Laner. Tuer Boston Basin is essentially a down-faulted block of sedimentary rocks and associated lavas which were folded and locally metamorphosed at the close of the Paleozoic Era. They are thought to be of late Carboniferous or of Permian age. As regards the stratigraphic sequence, the “Roxbury conglom- erate” constitutes the lower 8000 feet,* and the “Somerville slates” constitute the upper 2300 feet,* of the formation. Where these two groups of strata pass into one another there is a considerable thickness (two or three hundred feet) of tran- sition beds—shales or slates, sandstones, and conglomeratic rocks. The main mass of the Roxbury conglomerate “displays charac- ters suggestive of piedmont fluviatile deposition, but within the transition beds are several features which point strongly to Paleozoic glaciation in this vicinity. In his “Squantum Tillite’t, Mr. R. W. Sayles has ably pre- sented the case for glacial action in the deposition of the upper part of the Roxbury y conglomerate. It is this upper portion, belonging to the transition beds, that he calls the Squantum tillite. While Mr. Sayles’ paper is written with a force and ~ clearness which should convince, nevertheless there may still be some of his readers who will remain somewhat skeptical. Having some familiarity: with the Boston Basin and, at least in the large, strongly favoring Mr. Sayles’ convictions, I there- fore take this opportunity of offering a few suggestions. Has not Mr. Sayles used the word crzterion inadvisedly ?. On page 144 of his article he lists fifteen characteristics of the Squantum tillite and subsequently refers to these as criteria. The presence of “Some rounded, water-worn pebbles or bould- ers” is not a criterion for the Squantum tillite nor for any other tillite or till. They do not assist in the determination of a deposit as till or tillite. The same may be said of a few of the other phenomena cited in the list. Criteria which indicate the action or association of ice in the deposition of any sedimentary series may be classified as fol- lows :— A. Criteria observed in ice-laid deposits (till and tillite).+ The deposit as a whole, (a) Is “an unstratified mass of miscellaneous and unsorted rock materials.” * Mansfield, G. R.: The Origin of the Roxbury Conglomerate. Mus. Comp. Zool., Bull., xlix, pp. 209-210, 1906. + Mus. Comp. Zool., Bull., lvi, pp. 141-175, 1914. {In this classification I have drawn freely from Mr. Sayles’ paper. Lahee—Late Paleozoic Glaciation. olw (6) May contain intercalated “nests” and layers which are often contorted. 2. The matrix consists usually of rock-flour (sometimes of sand). 3. The included fragments, (a) If of compacted (frozen?), but not consolidated, mate- rials (clay or sand) when handled by the ice, (1) Are usually angular, and (2) May show crumpling, caused by the pressure of superincumbent or moving ice. (4) If of consolidated rock when handled by the ice, (1) Are generally angular or subangular, showing the effects of bevelling on one or more sides and of blunt- ing on one or more corners. (2) Are often striated. (8) May show concave fractures. B. Criteria observed in aqueo-glacial deposits. 1. Isolated pebbles and bowlders, striated or not, may be found in beds which, by their fine texture, uniform bedding lamination, or lack of marked cross-bedding, indicate that they were deposited by relatively weak currents or in quiet water. Such pebbles and bowlders suggest ice-rafting. . Contemporaneous deformation, due to grounding of float- ing ice blocks.* bo ©. Criteria observed in connection with the older formations over which the ice may have passed. 1. Contortion and disruption of mud and sand beds, ete., suggesting that these were unconsolidated or too weak to resist the ice pressure. This phenomenon is generally accompanied by criteria cited under A, 3, (a). In this case the surface of unconformity above the older formation is very irregular and shows no striation or other evidence of strong abrasion of this kind. . The surface of unconformity is polished, striated, or grooved, and shows other characters typical of ice erosive work. These features suggest that the older formation was consolidated hard rock when the ice passed over it. bo Of the criteria noted above, the only ones which are rare or absent in the case of the Squantum tillite are the striated peb- bles and the striated subjacent rock floor. ; For those who would question the action of ice in the deposi- tion of part of the Boston Basin sediments, the following remarks may be added to what Mr. Sayles has already said :— (1) Absence of a striated pavement beneath the tillite is to be compared with the relations between the Kansan till and the * See Lahee, F. H., Contemporaneous Deformation: a Criterion for Aqueo- glacial Sedimentation, Journal of Geology, vol. xxii. 318 Lahee—Lute Paleozoic Glaciation. underlying Aftonian stratified drift, as figured by Chamberlin and Salisbury.* (2) Searcity of striated pebbles and bowlders in the tillite has been compared by Mr. Sayles to the similar case in the Pleistocene of southern Rhode Island. Last summer (1913), in company with Dr. A. Knopf of the U. 8. Geological Survey, I had an oppor- tunity of examining a number of well-defined lateral and frontal moraines which were built by the Pleistocene valley glaciers of the Sierra Nevada in California. We were much impressed by the great scarcity of striated rock fragments in these moraines. In the Boston Basin this feature may be codrdinated with the fact that here ice seems to have travelled over unconsolidated sediments instead of over a hard rock floor. (3) The large angular blocks of slate and shale included in the tillite} resemble the masses of Aftonian gravel inclosed in the Kansan till, as figured by Chamberlin and Salisbury. (4) Mr. Sayles observes that near the top of the tillite, “inter- calated stratified beds may be found.” I have frequently noted similar beds of sand, some contorted and some not so, in the Pleistocene till of the Boston Basin, both in irregular till sheets (ground moraine) and in drumlins. (5) Contemporaneous deformation was observed last autumn in Pleistocene sand deposits at Woodland, Mass.§ (6) The presence of isolated pebbles and bowlders was reported in the Cambridge brick clays which are of Pleistocene age.|| (7) The Somerville slates, at the base of which are found evi- dences for glacial action, are remarkable for the absence of fossils. It is interesting to note the similar absence of fossils in the Pleisto- cene brick clays of Cambridge and other localities. 4 * Geology, vol. iii, p. 385, 1906. + Sayles, op. cit., pp. 148, 149. t Geology, iii, p. 386. § Lahee, F. H., op. cit. | Shaler, N. S., Woodworth, J. B., and Marbut, C. F., U.S. G.S., Ann. Rept., 17, Pt. 1, pp. 970 and 990. “| Shaler, N. S., ete., op. cit., pp. 993-994. Massachusetts Institute of Technology, February 10, 1914. Faurwell— Wilson Tilted Hlectroscope. ° 319 Arr. XX VI.—Wote on the Use of the Wilson Tilted Hlectro- scope; by H. W. Farwe t. In using the tilted electroscope I have frequently been annoyed by the lack of steadiness of the zero of the leaf and have found various causes for the wandering. Of these I will mention one which I have not seen described, although very likely many who have used electroscopes adjusted to a position Fic. 1. a TV Cee SNE Rana eee ee Bs ea a Ns a ee aE a om Lt | | I A i A A PS a YT Ne ae | ae | PC a Ea ic a UR | 2A a TST i TA ab a fs ie Nee (AES a FAG a PS oe J DU 27 | PS a | fa 0 i A A | TR VR a a We I es 7 al | Ie a J Se a ah a | ESA Amn Me AS eee ee ee Aa | RS SS Hee NE a 1 PS ea] TI | i ce Pa a a | Jee See ee ae ae so s of electroscope Dak ° SIIGISE ears — Reading pe 10 Time im minutes ——> very near to unstable equilibrium may have had the same trouble. The wandering which had been ascribed to defective insula- tion began to show some relation to the deflections of a gal- vanometer connected with a thermocouple placed near the electroscope case, and on making observations with a 16 e. p. _ carbon lamp as a heater, the disturbance was soon shown to be due to convection currents inside the electroscope. The pos- sibility of such a source of disturbance had been considered far too remote, but the curve given above shows how distinct is the relation. The lamp in this case was placed a short distance above the thick lead box which contained the electroscope. 320 Farwell— Wilson Tilted Electroscope. The top of the box was fully an inch over the heavy brass case of the electroscope. The galvanometer, of low resistance, was fairly sensitive, about 10~° amperes, and a resistance of 200 ohms was used in series with the iron-constantin thermocouple. The electro- scope leaf was not at its most sensitive position in this par- ticular observation, the whole fluctuation being about the same as would be caused by changing the potential of the leaf by about °1 volt. By moving the position of the lamp to the other end of the electroscope case, the leaf was observed to move down as the temperature increased, showing that the direction of the con- vection currents had been reversed. It would appear to be most desirable to use suspensions adjusted as is the gold leaf in this case in a fair vacuum, since it is generally necessary to use some sort of lamp to illuminate the background for the leaf. ~ Columbia University, Jan. 29, 1914. C. Schuchert and R. S. Lull—Mammut Americanum. 321 Arr. XX VII.— Mammut Americanum in Connecticut; by Cuaries Scuucnerr. With a note on the Farmington specimen by Ricuarp 8S. Lutt. [Contributions from the Paleontological Laboratory, Peabody Museum, Yale University, New Haven, Conn., U.S. A.] Introduction —The unearthing of good mastodon bones anywhere is worth noting, and when a fine skeleton is found, and especially in Connecticut, the discovery is all the more important. In addition, the latest find can be somewhat directly connected with the vanishing of the Wisconsin ice sheet, the last glacial episode of the Pleistocene. During the past century mastodon bones have been discovered in Connect- icut but five times, the last one found preserving more of the skeleton than al! the previous ones combined. Sooner or later these occurrences will assist in the formulating of a more detailed history of the Glacial Period, a time as yet unravelled and unmarked by the successive faunas and floras that inhab- ited North America during this age of decidedly changing climatic conditions. As yet we know but little of this life, and consequently the chronology of the Pleistocene in Amer- ica thus far depends solely upon the physical characters of the deposits and the topography of the period. Previous jinds.—(1) The earliest record of the finding of mastodon bones in Connecticut dates from previous to 1828, when ‘‘some remains of the mammoth were found in Sharon,” Litchfield county, near the New York State line.* Nothing more is known of this discovery nor of what has become of the bones. (2) Silliman states that while excavating the Farmington canal in the summer of 1828 there were found near Cheshire “three or four large molar teeth of the mammoth” in gravel but a few feet under ground. Some of the teeth are said to have been much worn and therefore were of an old animal. It appears that all of them except one were broken up at the time of their excavation, and that the remaining tooth was pre- sented to Yale College by Edward Hitchcock (loe. cit.). This is an uncut second lower molar of Mammut americanum (Yale Museum catalogue number 11985). The three crowns are well preserved and but little mineralized, as the tooth was found in a more or less dry gravel. (3) In 1833, while a canal was being dug for the New Britain Knitting Company in New Britain, through a pond which then covered what is now the corner of Elm and Church * This Journal, xiv, 187, 1828; also xxvii, 166, 1835. Am. Jour. Sct.—Fourtu Serizs, Vou. XX XVII, No. 220.—Aprin, 1914. 9 23 B22, C. Schuchert and R. S. Lull— streets, there was found, according to Silliman, a nearly com- plete dorsal vertebra of a mastodon. The dorsal process was 17 inches long, the centra 5:5 inches in diameter and nearly 3 inches in thickness, while the neural canal was 3°5 by 2°75 inches in height and breadth respectively. The bone had a dark chocolate color and was ‘‘not mineralized in the least.” It was taken out of mnd or clay 3 feet beneath the surface. Associated with the bone were freshwater shells “of the genera Planorbis, Lymneea, Cyclas, ete., similar to those occupying the waters of the vicinity, to Doctor Walter L. Barrows, of Trinity College, directs atten- tion, in a letter to the writer, to the “Connecticut Courant” for June 2 1834, in which there is a statement that Silliman exhibited this bone at a lecture given at Hartford May 27, 1834. From this account is taken the following : “ From the perfect condition of the bone thus accidentally discovered, there is reason to believe that a complete skeleton may be recovered in the morass where the specinien in question was found.” This bone was presented to Yale College by Elijah H. Burritt, but is evidently no longer in existence. (4) A second mastodon was discovered in New Britain in September, 1852, and the bones were for many years on exhi- bition there at the State Normal School. What became of these bones after 1885 is not known. The present information is taken from an account by Hon. David N. Camp,t entitled “‘ Mastodons roamed Connecticut once.” While excavating in a soft swampy soil for a pond on land belonging to Mr. Wil- liam A. Churchill, the workmen came across a considerable portion of a mastodon skeleton, “the thigh bone, humerus, tibia, several of the ribs, and two or three teeth. Some of the bones, on being exposed to the air, crumbled.” This locality is now in the city of New Britain back of the Young Women’s Christian Association, or near the junction of School and College streets. (5) Professor Edward Hitchcock got in 1871 a mastodon molar that was taken “out of a muck bed on the farm of Elias Bardwell” in the town of Colerain, which is near the north line of Massachusetts. This was then the first known occur- rence of mastodon in that state. The Farmington or Pope mastodon.—Late in August, 1918, Italian workmen, while digging a trench to drain a small up- land swamp on the beautiful estate of the late A. A. Pope, near * This Journal, xxvii, 165, 1835. Here the locality is given as Berlin, but D. N. Camp, in the article cited, gives it as above described. + Hartford Times of January 3, 1914. The first statement regarding the 1852 specimen appeared in the same. paper, on September 21 of that year. ¢ This Journal, (8), iii, 146, 1872. Mammut Americanum in Connecticut. 323 Farmington, reported to the superintendent, Mr. Allen B. Cook, the finding in the bog of a “black devil.” Mr. Cook, a graduate of the Massachusetts Agricultural College, soon saw the significance of the find and reported the matter to Mrs. Pope, who was much interested in the remarkable discovery and through Attorney Charles T. Brooks brought the informa- tion to the writer. The trench at the time of the latter’s visit revealed a number of Jarge bones of one of the fore limbs and the back part of the skull, which had unfortunately been greatly damaged before the Italians became aware that they were removing bone and not a prostrate tree. The exhumation of the skeleton was undertaken by Mr. Hugh Gibb, assisted by three other preparators from Pea- body Museum at Yale, and by five Italians. So careful were these “bone diggers” that all the clay immediately around the skeleton was dug out with their hands, as they felt their way through the sticky clay down to and around the bones. The greater part of the skeleton was taken out in two weeks’ time. Subsequently it was decided to enlarge the shallow hole and to make of it a water reservoir for the estate, and during this excavation the workmen late in November came upon one of the great tusks, lying alone, 23 feet away from the skull, and in perfect preservation. The skeleton of the Farmington mastodon consists of all of the essential parts, minus most of the small bones of the feet, a few of the smaller leg bones, most of the caudal vertebra, and one of the tusks. The greater part of the animal lay together with the bones more or less jumbled, but in the main there was still considerable natural skeletal alignment,— the head at one end and the pelvis less than 10 feet away. Scattered about and often many feet away from the central mass lay single bones of the feet, tail, vertebrae, ribs, and one scapula. The recovered tusk lay farthest away and on a level about 2 feet higher than the main mass of bones, which were in the lowest part of the swamp. Fig. 1, drawn to scale by Professor Lull, shows the general disposition of the bones of the mastodon as they were found. Geologic position of the Farmington mastodon.—The Farm- ington skeleton lay in a shallow trough directly on bowlder- clay, a thin covering of ground moraine that originally mantled the adjacent hills and valleys alike and was deposited by con- tinental glaciers during Wisconsin time. (See contour map, fig. 2.) The bowlder-clay is light blue in color and consists of a decidedly sandy fine mud, with some coarse sand, an abun- dance of muscovite, and many striated bowlders, large and small, mainly of trap with some quartzite and crystalline rocks. That the till was largely derived from the adjacent Triassic 324 C. Schuchert and R. S. Lull— area is attested not only by the sandy blue clay but more espe- cially by the abundance of trap bowlders. The Farmington highlands had not become much covered Hire. i: Fic. 1. Diagram of the Farmington swamp, showing the bones in situ and the trench that led to their discovery. Drawn in absentia, from descrip- tions, sketches, and photographs, by Richard 8. Lull. Approximate scale 1/8 inch to 1 foot. A. Original excavation. B. Subsequent excavation in part. 1, femur; 2, pelvis; 8, femur; 4, humerus; 5, left scapula; 6, lower jaw; 7, skull; 8, right scapula; 9, tusk. with vegetation when the proboscidean under discussion died, for not a trace of organic matter or of oxidized till was found beneath the skeleton, and but very little vegetable matter is seen in the slightly modified or oxidized bowlderless clay surround- Fic. 2. FARMINGTON <— % MILE 3000 FEET 1000 Contour interval tis 5 feet Datum ts mean sea level L Topography by Albert H Bumstead The sharp ridge is made up of Triassic trap. Fic. 2, Contour map showing the vopography of the land where the Pope mastodon was found (+). Drawn by Albert H. Bumstead, through the kindness of Professor Hiram Bingham. 326 CO. Schuchert and R. S. Lull— ing the bones, nor was there a trace of the organic matter of the animal left (see fig. 3). - Further, there was no permanent lake with an abundance of vegetation formed at this time or later, for no freshwater shells (shell marl) were seen and the clays do not at all effervesce under hydrochloric acid. Car- bonate of lime is a good preservative of bones, but in this case there is none present. ‘The animal does not appear to have been mired where it was found; it probably died and lay decomposing in the marsh, where occasional carnivorous or scavenging animals came and dragged about some of the bones, though why one of the tusks should be about 23 feet away from the skull and on Iigher ground by 2 feet is hard to explain. On the other hand, the body could not have lain thus exposed many years, even in a cold climate, for most of the bones immediately above the till have well-preserved smooth and glossy surfaces. Nevertheless, nearly all of the skeleton of the feet is absent, while some of the other small bones, and especially those of the tail, were so rotten that they could not be lifted and preserved. The absence of these smaller bones may be due either to carnivores or to oxidation, or to both causes. Some of the parts, and especially the pelvis and top of the skull, lay so near the turf that the percolating waters with their humic acids and the penetrating roots of the plants had done considerable damage. It would seem, however, that the skeleton must have soon become buried, and this is suggested by the embedding clay, which is almost unaltered glacial till devoid of all bowlders. In other words, the clay of the general ground moraine on either side of the little valley, but more particularly that of the rounded hills to the north, was then being washed from between the bowlders into the lower land, rapidly covering the skeleton. These hills are even now covered in abundance with the much oxidized trap erratics, hardly any of which retain the glacial striations, but there is little blue clay here now. We therefore have the phe- nomenon of the unoxidized clay washed into the swamp or seasonal lake, while the remaining bowlders have ever since been exposed to the influences of the weather, the trap bowl- ders being the most altered. Above the till wherein the mastodon lay follows from 6 to 18 inches of glacial blue bowlderless clay that is but slightly modified by some plant remains, largely roots, that have penetrated into it from the swamp above. This is followed by 30 inches of a similar clay, slightly more modified by entombed plants and penetrating roots, but still with an abundance of briliant muscovite. At the top occurs about 18 inches of stringy turf that is full of water during most of the year, making the swampy ground. In other words, the Farmington 327 n Connecticut. eCanuUm 4% Mammut Amer Fic. 3. Avyo Lap[Moq oy} wo Av I [ev aegargy OTL ‘poloAOOUN SVAL TOJTOYS OY 104 je sou “4D we *y £q ydeasojoyg ISstp oy} Jo ydeasojoyg ‘e ‘DIT 328 C. Schuchert and R. S. Lull— mastodon was covered in the course of time by 5 feet of ground, 4 feet of which is modified glacial clay. The animal was therefore buried by the inwash | “of about 4 feet of glacial clay, since which time about 18 inches of turf has accumulated. As the skeleton was entombed shortly after the van- ishing of the Wisconsin ice sheet in the highlands about Farmington, one gets from the evidence given a hint of the recentness of these two sets of phenomena. The washing of the clay into the depression could have occurred in a “few hundred years, and the stringy turf apparently did not take much more time to form. Then, too, the skeleton shows no mineralization nor petrifaction and is but little discolored to a light brown by the waters of the swamp. Further, the several skeletons found about Newburgh, New York (chiefly the Warren skeleton of the American Museum of Natural History and the Otisville specimen at Yale) look like bones buried but a few hundred years. Both of these mastodons were buried in a shell marl,a preservative of bones, but still not a single shell species of which the marl is composed is extinct, according to the great conchologist Gould (whoidentified Limnea galbana, Planorbis parvus, Valvata tricarinata, Am- nicola galbana, and Cyclas galbana). The Warren mastodon lay in the marl of a small pond not over 40 feet in diameter, covered bya foot of red moss and 2 feet of peat bog. The entire skeleton lay in articulate position and was so well preserved that it was dug out by inexperienced men in two days. Further- more, between the ribs lay from 4 to 6 bushels of vegetable food, largely coniferous and much like spruce or hemlock. Did man and Mammut americanum live together in America? —Can it be that Mammut americanum vanished from Connecticut within a thousand or at most a few thousand years and yet was unknown to the North American Indians? If the prehistoric Indians knew and helped to exterminate these animals, and as they were the makers of neolithic implements, why do we not find ivory in their graves ? In this connection it should not be forgotten that John M. Clarke in 1887 dug up at Attica, Wyoming county, New York, bones of Mammut americanum associated with pottery and charcoal. Not much of the skeleton was present and the bones lay but 2°5 feet beneath the “natural surface.” Associated with them (ribs) were “four small fragments of charcoal” while in another part of the diggings beneath all (4 feet) of the vegetable muck and lying upon “compact laminated clay ” ‘was found a fragment of pottery, and from beneath and around it were taken about thirty fragments of thoroughly burned charcoal. These traces of ancient man were found fully 12 inches further down from the natural surface of the ground Mammut Americanum in Connecticut. 329 than the deepest of the bones taken from the other [there were two] sink-hole.” The pottery indicated a thick coarse vessel about 8 inches across, while the “thoroughly burned” chareoal varied in size “from two inches in diameter down.”* All in all, the evidence appears to show that the Wisconsin ice sheet vanished from Connecticut and New York not many thousands of yearsago. Further, the associated human evidence found with or beneath the Attica mastodon bones is a positive hint that should open our minds to the possibility that man was associated in America with Mammut americanum. There is still further paleontologic evidence suggesting the existence of man even earlier than the occurrence in New York. Professor Williston states that on a small tributary of the Smoky Hill river in Logan county, Kansas, Mr. Handel T. Martin found beneath but in contact with a right scapula of the extinct Bison occidentalis an arrowhead. The bones of the eight animals present and the human implement were not sur- face finds, but were secured by digging away several feet of the “upland marl,” a deposit that originally covered the fossils at least to a depth of 20 feet. In this marl also occurs Elephas primigenius, an animal well known to ancient man of western Europe.t Note on the Farmington specimen ; by Ricuarp 8. Lute. The Farmington mastodon has not yet been prepared for exhibition, so that two important elements, the skull and pelvis, being still in the plaster of Paris bandages which were put on at the time of exhumation, are unavailable for study. Certain broad generalizations can, however, be made which may be of interest. The skeleton is that of a fully adult individual, the evidence for which lies in the great size, the fully codssified epiphyses of the vertebre, and in the facts that the full quota of molars had come into use and that all of their ridges show signs of wear; though, as the more posterior ones have suffered but little in this regard, the animal must have been in the prime of life. I should judge the specimen to have been a male, though the one absolutely diagnostic character, that of the presence of one or two lower tusks as in the great Warren mastodon in the American Museum of Natural History, is lacking. The size again and the great development of the upper tusks are the only present evidences of sex available, though that of the ratio of pelvic aperture to pelvic breadth, larger in the female, may be learned when the element can be studied. * Forty-first Ann. Rept., N. Y. State Mus. Nat. Hist., 1888, 388-390. + Amer. Geol., xxx, 313-315, 1902. 330 0. Schuchert and R. S. Luli~Mammut Americanum. The limb bones are relatively slender in proportion to their length in comparison with those of the Warren mastodon or the Otisville specimen mounted at Yale, both of which are exceptionally powerful animals though differing markedly in stature and general appearance. In size, the new skeleton stands almost midway between them ; for instance, the ulna of the forearm measures 28 inches in length in the Otisville animal, 31 inches in that of Farmington, and 34 inches in the Warren specimen, and the other measurements which I have been able to compare run in much the same proportion. The Otisville specimen was not fully mature at the time of its death, as certain of the vertebral epiphyses are still but partially coalesced and the hinder ridges of the teeth show no appreciable signs of wear. As mounted, this specimen, a young male of abundant promise, stands 8 feet 2 inches at the shoulder, with a length of 13 feet 2 inches from the tip of the premaxillary bone to the curve of the tail. The assembled bones of the Connecticut specimen, however, give an estimated shoulder height at least 6 inches greater, though just how the other dimensions would compare one cannot say; but the contrast of bone proportions seems to indicate that those of the entire body would vary as well. The erection of the two skeletons side by side for comparison would thus be of the highest scientific importance. The tusks of the Otisville mastodon have not been found and little idea of the relative length of this element can be obtained. The size of the tusk socket would indicate a tusk but a half-inch less in circumference than that recently found at Farmington. The latter is 23 inches in greatest girth and measures 8 feet 10 inches on the curve and 6 feet 3 inches between perpendiculars. There is nothing to indicate that the Farmington mastodon does not belong to "the principal Pleistocene species, Mammut ( Mastodon) americanum, the known range of whose variations is amply sufficient to account for all of the points of contrast which have been mentioned. EL. W. Dean— Esters of Substituted Aliphatic Acids. 331 Arr. XX VIII.—On the Hydrolysis of Esters of Substituted Aliphatic Acids; by E. W. Dean. [Contributions from the Kent Chemical Laboratory of Yale Univ.—cclvi. ] 7. Hydrolysis in Acid and in Alkaline Solution of Ethyl Esters of Hydroxy Butyric Acids. Acip Hyprotysis. Preliminary Discussion. Tue effect of the hydroxyl group upon the velocity of acid hydrolysis has been shown* to be somewhat irregular in the eases of acids of the acetic and propionic series. Ethyl glycol- late is decomposed a little more rapidly than the acetate. Ethyl lactate, the alpha hydroxy propionate, has a greater reaction velocity at twenty-five degrees and a less at forty-five than has ethyl propionate. This is due to the lower temperature coeffi- cient of the hydroxy ester, but makes it immediately impossible to state that the presence of this particular substituted group facilitates the hydrolytic cleavage. Ethyl glycerate, which is the alpha, beta, di-hydroxy propionate, has relatively a very low rate of hydrolysis, showing that the substitution of two eroups may produce an effect differing in order of magnitude from that of one. In view of the above facts it was thought desirable to obtain more data bearing upon this subject, and consequently meas- urements have been made of the rates of decomposition of several hydroxy butyric esters. Those which were examined comprise the following list : Ethyl alpha hydroxy-butyrate, CH,CH,CH(OH)COOC,H, , Ethyl beta hydroxy-butyrate, CH. ,CH(OH)CH,COOC, He Ethyl alpha hydroxy-iso-butyrate, on *>0(OH)COOC,H.. 3» and Ethyl normal butyrate and ethyl iso-butyrate were of course hydrolyzed at the same time with these esters in order that the comparisons made should deal with the results of measurements made under identical conditions. Preparation of the Esters. Ethyl butyrate was the Kahlbaum preparation of commerce which was carefully purified by washing and by fractional dis- tillation. Boiling point, 118°5° to 1195°. The figure given in Beilstein is 119°9°. Ethyl iso-butyrate was prepared by the method of Pierre and * This Journal, xxxiv, 293; xxxv, 486. 332 EF. W. Dean—Esters of Substituted Aliphatic Acids. Puchot.* Iso-butyl alcohol was oxidized with chromic acid and the iso-butyl ester of iso-butyric acid thus formed. From this the ethyl ester was obtained by the successive processes of saponification by alkali, acidification, and treatment with ethyl alcohol. This last esterification was conducted accord- ing to the method of Wislicenus.t The purified product boiled at 110°, which is the figure given in the literature. Ethyl alpha hydroxy-butyrate was obtained by the esterifi- cation of acid purchased from Kahlbanm. After obtaining. unsatisfactory results from trials by most of the usual methods the method of Bogojawlensky and Narbutt was found to be well adapted to this case. The acid was boiled with absolute alcohol in the presence of finely powdered anhydrous copper sulphate. The resulting ester was purified by fractional distil- lation and a sample obtained boiling between 162° and 164° uncorrected. The temperature given in Beilstein is 165° corrected. Ethyl beta hydroxy-butyrate was purchased from Kahlbaum and was used without special purification. Its boiling point was 178° to 180° uncorrected; no value is given in Beilstein. As a check upon the reliability of this commercial product a small quantity was prepared by esterifying some of the free acid by the copper sulphate method. One or two portions of this, when hydrolyzed, gave constants approximately the same as those of the Kahlbaum preparation. The freedom of the latter from impurities of any other esters was also proven by the regularity of its velocity constants. The presence of another ester having a different rate of reaction would have caused a progressive > decrease in any series of these constants. Ethyl alpha hydroxy-iso-butyrate was also purchased from Kahlbaum and was proven sufficiently pure by analysis and by determination of its boiling point. The latter was 146° uncor- rected, that given in Beilstein is 150° corrected. Procedure. The unsubstituted esters of this list have rather low solubil- ities, so that it was found desirable to modify somewhat the procedure | which had been used for acetates and propionates. In cases where a solution of the ester six hundredths or more normal can be prepared in decinormal hydrochlorie acid, it is simplest to titrate twenty-five cubic centimeter portions directly * Anal. de Chim. et Phys., (4), xxviii, 366. + Ann. d. Chem. u. Pharm., clxiv, 181. t Ber. xxxviii, 3844. See also Clemmenson and Heitman, Am. Chem. Jour., xiii, 319. | This Journal, xxxiv, 293; xxxv, 486. E. W. Dean—Esters of Substituted Aliphatic Acids. 333 with approximately tenth normal baryta water. The initial titration will then be about twenty-five cubic centimeters, the final at least fifteen cubic centimeters greater. With a range of this magnitude the constant error of titration does not cause undue irregularity among the constants obtained. Saturated solutions of ethyl butyrate or ethyl iso-butyrate (in decinormal hydrochloric acid) are only two to three hundredths normal, and the range in the titrations, measured in tenth normal barium hydroxide solution, is from five to seven and a half cubic centi- meters. This is inconveniently small, and three modifications have been tried to increase it. First, fifteen cubic centimeter portions were titrated with twentieth normal alkali. The advantage here is slight but definite, the range increasing by one-fifth and the error remaining about the same. Next, a modification of the method of De Hemptinne was tried.* A twenty-five cubic centimeter portion of the reaction mixture was treated with twenty-five cubic centimeters of tenth normal baryta water and the titration completed with a twentieth normal solution of the same alkali. This doubles the range, but unfortunately largely increases the error of titration. The third and most satisfactory method involved the use of twentieth normal alkali and a hundred cubic centimeter burette. The latter was of the bore of the ordinary fifty cubic centi- meter burette but was made with a top bulb that brought its capacity up to a hundred cubic centimeters. When twenty- five cubic centimeter portions of the reacting mixture were titrated with this it was possible to have a total range of from ten to fifteen cubic centimeters with a very slight increase in the error. One drop of twentieth normal alkali was sufficient to produce a perfectly definite color change in the indicator, and there was little increase of error due to the draining and read- ing of this burette. The procedure may be briefly summarized as follows: Two hundred and fifty cubic centimeter flasks containing the solu- tions of the esters in decinormal hydrochloric acid were placed in an accurately regulated thermostat.t At measured intervals of time portions were removed by means of a twenty-five cubic centimeter pipette, run into cold water and their acidity deter- mined by titration with twentieth normal barium hydroxide solution. Phenolphthalein was employed as an indicator. Of the more insoluble esters, saturated solutions were pre- pared by shaking the acid vigorously with an excess of ester and then filtering. The paper was first moistened with the acid solution, which prevented the passage through it of any of the undissolved ester. The measurements of the final acidity were made upon portions ef the mixtures, heated several days * Zeitschr. f. phys. Chem., xiii, 561. + This Journal, xxx, 72 334. EL W. Dean—FEsters of Substituted Aliphatic Acids. in sealed tubes at about ninety degrees. The completion of the reaction was in this way assured. Calculations were made by the use of the well-known titra- tion formula* for reactions of the first order. In the tables are given values of the constants obtained and of the periods in minutes over which the courses of the various reactions were followed. Bracketed constants are not counted in the averages; their irregularity being undoubtedly caused by undue experimental error. TABLE I. Hydrolysis at 25°. N/10 HCl. Ethyl Ethyl Ethyl a- Ethyla-Oxy- Ethyl 8- Ester Butyrate Iso-butyrate Oxy-butyrate Iso-butyrate Oxy-butyrate Time 3675 3675 4320 3675 3675 53°3 40°5 37°2 19°9 10°2 51°8 39°4 40°1 21°3 11°3 49°2 40°2 40°4 20°4 11:0 50°3 39°6 40°7 21°3 10°7 10°K 53°6 39°7 40°8 22°0 10°7 55°d 39°4 40°7 21°8 10°7 54°8 40°0 39°5 21°9 10°8 Average 52°6 39°8 39°9 21-2 19°8 Average 51°7 40:0 ie 20°7 10°6 (duplicate) Hydrolysis at 35°. Time 2280 2940 1920 3000 5700 107° 92°7 86°3 50°9 28°6 108° 89:7 86°5 Bile. 27°8 109° 89°1 88.5 52°3 28°5 10°K IA@s 93°3 Size BLO 7 27°6 Mas 91°4 84°8 d1°4 27°4 ik|}ate 94-1 88'8 51°3 27-3 (117°) 92°3 (82°3) 51:2 27-4 Average 110° 91°8 87:0 515 27°8 Average 109. 90°9 ae 50°9 27°4 (duplicate) Hydrolysis at 45°. Time 840 840 720 1560 1560 248- 205° 203° 113° 61°7 245° 204° 201° 109° 67°4 243° 209° 203° rales 64°2 10°K 240° 206° 198: 114: 64:1 238° 205° OS 116° 64-0 234° 201° 186° 115: 63°5 230: (192.) 182° 113° 62°7 Average 239: 205° 195: 113° 63°9 Average 234: 201° S159 let 64°1 (duplicate) * This Journal, xxxiii, 27. EE. W. Dean— Esters of Substituted Aliphatic Acids. 335 Discussion of Results. Perhaps the most noticeable fact set forth by the above figures is the marked and varying retardation caused by the hydroxyl group in esters of this series. The irregularity in effect of this group was indicated in the previous experiments with acetates and propionates and is now definitely established. The comparison of velocities of corresponding hydroxy and unsubstituted esters shows that the hydroxyl group may pro- duce effects varying in different cases from a slight accelera- tion to a considerable retardation. Thus it seems to be defi- nitely proven that the effect of the hydroxyl group upon the velocity of ester hydrolysis is not in any sense an additive property. A second fact of interest is that substituted and unsubstituted iso-butyrates decompose more slowly than the corresponding normal esters. The difference here measured appears to be somewhat greater than that found by De Hemptinne* between the unsubstituted esters but is in the same direction. The beta ‘hydroxy ester hydrolyzes more slowly than the alpha, which is in accord with the rule indicated by all pre- vious experiments.t All the esters decompose more slowly than analogous ones of the propionic and acetic series. Temperature coeflicients seem to be rather irregular, perhaps on account of experimental error. There are at least no conclusions to be drawn from their variations. ALKALINE HybDROLYSIS. In connection with the experiments recorded in the first half of this paper it was thought desirable to make measurements upon the reaction velocities of esters of the above list in alka- line solution. Previous work* has indieated that this velocity is controlled by the strength of the acid from which the ester is derived and by another factor called steric influence. No mathematical relation for the latter has yet been worked out, but it seems to be closely related to that which controls the velocity of acid hydrolysis. The strength of the acid in the latter case has at most only a minor influence. Procedure. Part of the measurements were made by the use of the titra- tion method which has been carefully described in an earlier papert of this series. Experiments were also made by a modi- * Loc. cit. + This Journal, xxxiv, 69; xxxv, 486. } This Journal, xxxv, 605. 336 FE. W. Dean—Esters of Substituted Aliphatic Acids. fication of the conductivity method of Walker* which has been employed in the experiments of Findlay and Hieckmans+ on the velocities of saponification of esters of hydroxy and alkyloxy acids. The modification devised by the author simply adapts the excellent procedure recommended by Walker to the cases of the somewhat insoluble esters here examined. This proce- dure was in outline as follows: Fiftieth normal solutions of carbonate-free sodium hydroxide and of ester were first prepared. The conductivity apparatus was of the usnal form with a cell of the narrow Ostwald type. The constant temperature bath was a mixture of cracked ice and water which was vigorously and efficiently agitated with a spiral stirrer operated by an electric motor. The initial con- ductivity was obtained by measurements made on hundredth normal sodium hydroxide solution. The cell was then ecare- fully rinsed and dried. For the actual hydrolysis fifty eubie centimeters of the fiftieth normal alkali solution were placed in a clean and dry Erlenmeyer flask and suspended in the ice bath. The fiftieth normal ester solution was likewise cooled down to zero degrees in the same bath. At the proper time fifty cubic centimeters of the latter were removed from its container by means of a pipette, run into the alkali and the mixture vigorously shaken for a second or two. The time at which the pipette was half empty was recorded as beginning the reaction. As soon as was convenient a portion of the reaction mixture was transferred from the flask to the conductivity cell and bridge readings were taken at appropriate intervals of time. The flask containing the remainder of the reacting mixture was tightly corked and warmed in a water bath for an hour or so to bring the action to completion. After this it was re-cooled and measurements made upon por- tions of its contents to obtain the final conductivity. Calcula- tions were made in the same way as those of Walker except that his scheme for their simplification was not adopted. Dur- ing all the measurements a constant resistance of two hundred ohms was kept in the box, thus eliminating the possibility of slight errors due to this source. A trial experiment with ethyl glycollate showed that the results obtained checked almost exactly with those of the titra- tion method. Measurements were made with the esters of alpha hydroxy-iso-butyric acid and of beta hydroxy-butyric acid. The constants recorded for the other three esters are from the results of titration experiments. in Table II are values of the constants obtained and the time periods in minutes over which the reactions were followed. * Proc. Royal Soc., Series A, lxxviii, 155, (1908). + Jour. Chem. Soc., lxxxvii, 747. E. W. Dean—Esters of Substituted Aliphatic Acids. 337 TaBLe IT. Hydrolysis at 0° N/101 NaOH Ethyl Ethyla- Kthyla-, Ethyl p- Ethyl Iso- Oxy- Oxy-iso- Oxy- Ester Butyrate butyrate butyrate butyrate butyrate PING aes oe 2 Se 160 185 50 130 120 79 ("75 6°7 1°5 2:5 67 48 6°3 1°8 2°2 68 46 71 Neg 2-1 65 "46 54 1°6 19 62 45 6:4 1°6 159 46 5°8 1°6 18) “45 Eo 7 1°6 “49 Me 7/ INVCTAGC a. oes =o! 68 46 6:2 1°6 eI PANIC. AG Cle sea ies “Tal 47 6°5 ihovy 2°0 (duplicate) Table III gives a summary of results from Tables I and II and has in addition some values of affinity constants of acids taken from tables of Landolt, Bornstein & Roth (fourth edi- tion). Figures for ethyl acetate and for ethyl propionate are also tabulated for purposes of comparison. TaBLEe III. Summary. Temp. Alkaline Acid hydrolysis Coefficients hydrol- lS a 7) SSS SS sis Ester NOS KG aes Kenan 02 Ke K_ (10°)k(aff) 20° 30° 45° — -25°-85° 85°-45° (0°) (25°) Bthylsbatyrates 521) 10. 937) 12°09) 2-17 68) 1-49 Ethyl iso-buty- TOs seo a Seo cD) et 2038 esi) DIAS} °46 1°44 Ethyl] a-oxy- butyrate Stay 39:8) 187-0 195° QASIM 24 6°2 teas Ethyl oxy-iso- butyrate oe eee 21°0 51:2 Ws 2°43 2°19 16 10°6 Ethyl B-oxy- butyrate -__-- LOS ipo G4:0; a QebieVOrS atk aro Mi) cso Ethyl acetate _. 64°7 163 374° 2°51 2°29 1°2 1°8 Ethyl propionate 71°6 179 406 PHAR) ) POS ated 1°34 SUMMARY. I. When hydrolyzed in acid solution three hydroxy-butyrates were found to have considerably smaller velocities of reaction than the esters of the acids from which they were derived. Am. Jour. Scl.—FourtH SERIES, VoL. XX XVII, No. 220.—ApRIL, 1914. 24 338 FE. W. Dean—Esters of Substituted Aliphatic Acids. II. Ethyl iso-butyrate decomposes more slowly than the ester of the normal acid; a difference in the same direction but of greater magnitude occurs with the corresponding alpha hydroxy esters. III. Ethyl beta-hydroxy-butyrate decomposes more slowly than the alpha ester. IV. Esters of the butyric acid series appear in general to de- compose more slowly than propionates and acetates. V. The effect of the hydroxyl group upon the velocity of acid hydrolysis is not an additive property. VI. The velocities of saponification, measured in centinormal sodium hydroxide at zero degrees, vary in the following order with the numerical values representing approximate ratios : Ethyl alpha hydroxy-butyrate (13:5), ethyl beta-hydroxy- butyrate (4°5), ethyl alpha-hydroxy-iso-butyrate (3°4), ethyl buty- rate (1°5) and ethyl] iso-butyrate (1:0). The author wishes to acknowledge his indebtedness to Pro- fessors W. A. Drushel and R. G. Van Name for the assistance and advice they have so kindly given during these experiments. Foote and Bradley—Solid Solution in Minerals. 3389 Arr. X XIX.—On Solid Solution in Minerals. V. The Iso- morphism between Calcite and Dolomite; by H. W. Foorr and W. M. Bravery. Sotutions of liquids in liquids may, as is well known, con- veniently be divided into two classes depending on whether there is complete or incomplete miscibility in all proportions. In the same way, it is convenient to distinguish two classes of isomorphous mixtures or solid solutions. In one class, iso- morphism is complete and solid solutions in any proportions are possible, while in the other, each solid can take up but a limited amount of the other. The alums form solid solutions of the first class, and among minerals, anorthite and albite are similarly capable of mixing in all proportions, forming the plagioclase feldspars. A large number of salts form solid solu- tions of the second class, each salt taking up only a limited amount of the other, and among minerals, calcite and dolomite belong in this class. In the present investigation, we have endeavored to determine, at least approximately, to what extent solid solution may take place between these two minerals as they occur in nature. It will be well to consider first the possible effect of tem- perature on the mixing limits. This influence has very com- monly been disregarded in determining to what extent solid solution may take place. For instance, Retgers* has deter- mined the mixing limits of a large number of salts and has considered the results as representing fixed values without regard to temperature. Van’t Hoff,t however, pointed out that the mixing limits of solid solutions, like the composition of partially miscible liquids, should be a function of the tem- perature and this has been demonstrated experimentally in a number of cases. The determination of mixing limits might, therefore, have little significance unless the temperature were known. However, if the solubility of one solid in another is slight, temperature will have but little influence on the abso- lute amount dissolved and the mixing limit should be reason- ably constant. This is the case, as will be seen below, with calcite saturated with dolomite (or magnesium carbonate). When a solid dissolves a considerable quantity of another, temperature may affect the limit largely and in general a greater absolute change would be expected than where the solubility is slight. It is evident that for a given temperature each solid will take up the maximum amount of the other when both are * A series of eleven articles, Zeitschr. phys. Chem., 1889-1895. + Vorlesungen iiber Chemie, I, 49. 340 Foote and Bradley—Solid Solution in Minerals. deposited simultaneously. This is the case, for instance, when solid solutions crystallize as eutectics. If dolomite and ealcite could be obtained which had erystallized simultaneously from solution, this material would therefore be ideal for determining the mixing limits of both minerals. A very careful search through the extensive Brush collection has shown no specimen where this condition of affairs was realized, nor have we been able to obtain such a specimen from other sources. This was not unexpected, for it can be shown that at a given tempera- ture, both minerals could be deposited only when the ratio of lime to magnesia in the solution has one fixed value. On the other hand, dolomite or calcite alone could be formed from an infinite number of solutions, with varying proportions of lime and magnesia. Almost the only chance of obtaining such specimens appears to be from the concentration of a large quan- tity of solution containing salts of both metals, in which case, one mineral would first crystallize to be followed ultimately by the crystallization of both. There appears to be another way, however, in which crystals of one of the minerals may be obtained saturated with the other. If, for instance, a solution capable of depositing dolo- mite comes in thorough contact with calcite, the solution should become saturated with the latter and the dolomite resulting should contain the maximum amount of calcium carbonate. A specimen of dolomite, therefore, deposited on calcite, or of calcite on dolomite, may usually be expected to contain the maximum amount of the other salt in solid solution. On the other hand, the calcite or dolomite originally present would not, of necessity, be changed in composition to the limiting value. As will be seen below, one specimen of calcite depos- ited on a dolomite was apparently not saturated with the latter. In this case, the calcite crystal stood out from the mass of dolomite and the solution from which it formed was probably not saturated with the latter. In general, where the secondary mineral crystallizes in intimate contact with the primary one, the composition should approach closely to the limiting value. A number of specimens have been obtained showing the associations mentioned above, where either dolomite or ealcite has been deposited on the other, and in each case the secondary mineral has been analyzed. In some eases, the material required for analysis was so closely associated with the primary mineral that separation by means of heavy solution was neces- sary. In the case of one dolomite, No. III, the material was put through the heavy solution twice to remove particles of calcite carried down in the first treatment. Where the specific gravity of the samples is not given in the table, the material was in such good erystals that it could be separated in pure Foote and Bradley—Solid Solution in Minerals. 341 condition by careful picking. The analyses were made by the usual methods. All the minerals dissolved completely in hydrochloric acid. Iron and manganese were both precipitated by means of bromine and ammonia. Ordinarily, manganese and iron were not both present in appreciable amount and the precipitate, after ignition, was considered either as Fe,O, or Mn,O,. In the one case necessary, they were separated by the basic acetate method. Calcium was weighed as oxide after a double precipitation as oxalate. We were able to confirm the observation made by Gooch and Austin* and others that magnesia gives high results when precipitated from a cold solution as ammonium magnesium phosphate in the usual man- ner. On this account, the precipitate, after standing, was filtered and redissolved in hydrochloric or nitric acid. The solution was heated to boiling, and after adding a small amount of ammonium phosphate, was again made alkaline with ammonia. The precipitate was allowed to stand until cold before filtering. Carbon dioxide was not determined directly. Calcite. The specimens used for analysis were the following: 1. Small yellowish scalenohedrons of calcite deposited on well-crystallized dolomite. Locality unknown. 2. Fine erystals of calcite, slightly etched, a combination of prism with rhombohedron, deposited on a crystalline layer of dolomite. This, in turn, was deposited on large scalenohedrons of calcite. The dolomite of this specimen was also analyzed. (See dolomite No. III.) Ouray, Colorado. 3. Small scalenohedrons of calcite deposited on well-erystal- lized dolomite. Joplin, Missouri (?). 4. Small etched scalenohedrons on a crystalline deposit of dolomite. Guanajuato, Mexico. 5. A single crystal of calcite deposited on a mass of well- erystallized dolomite. Cave of the Winds, Niagara. ‘6. A large, water-clear erystal of calcite, associated with pink rhombohedral dolomite. Joplin, Mo. The analyses are given in Table I. Of the six analyses of calcite, all but one (No. 6) show a reasonably constant amount of magnesium carbonate. The average amount in the five analyses is 0°97 per cent and the greatest deviation from this value amounts to only 0:17 per cent. Coming as the specimens do, from a number of widely separated localities, and varying greatly in habit, it is certain that they were formed under varying ‘conditions. The con- clusion appears justified, therefore, that calcite is saturated by * This Journal, vii, 187, 1899. 342 Foote and Bradley—Solid Solution in Minerals. TABLE I, Analyses of Calcite Deposited on Dolomite. 1 2 Sp. gr. 2°718—2°722 Sp. gr. undet. Ouray, Colorado. a b Average a b Average Me@Ole ee <1°08 «0:98 = «1038S 0-79") 0-80 oreo CAC Oweese 99°06 99°08 99°07 SGa/al: 96°48 96°59 MniCOe ==. 0°12 0°10 Orll 2°92 2°94 2°93 100°26 100°16 100°21 100°42 100°22 100°32 3 4 Sp. gr. undet. Sp. gr. 2°741—2°769 Joplin, Mo. ? Guanajuato, Mexico a b Average a b Average Mo COn sere 1:03 1°06 1:04 1:07 1°12 1-09 CaCO = 99°26 99°21 99°23 93°05 93°24 93°15 nC OM aes 0°17 0°18 0°18 5°80 5°70 5°75 100°46 100°45 100°45 99°92 100°06 99°99 5 6 Sp. gr. undet. Sp. gr. undet. Cave of the Winds, Niagara Joplin, Mo. a b Average ha, —_—_— + Mo COs een rae 0°87 0°47 0°42 0°44 CAC ORs eres eae ee 99°38 99°65 99°65 99°65 MnC OFee sec seers 0°42 0°22 0°22 0°22 100°67 100°34 100°29 10031 approximately one per cent of magnesium carbonate at common -temperatures of crystallization. As we pointed out before, a nearly constant value is to be expected where the amount of material in solid solution is small, unless, indeed, the tempera- ture variation is great. We know of no reliable data at present to show whether this limit varies appreciably with extreme temperature conditions. An examination of some of the magnesian limestones would probably give information on this point. A number of cases have been reported in which cal- cites contained more magnesia than has been found by us, but there is no evidence that the material aualyzed was homogeneous. Thus, Eisenhuth* has analyzed two calcites containing respec- tively 3-01 and 1°52 per cent of magnesium carbonate. Both of his specimens contained insoluble matter, showing the material was not quite homogeneous, and it seems fully as * Zeitschr. Kryst., xxxv, 582, 1901. Foote and Bradley—Solid Solution in Minerals. 348 probable that there may have been a small amount of admixed dolomite as that the calcite itself contained this unusual amount of magnesium carbonate. The magnesium content of our No. 6 is lower than any of the others. In this case, the calcite was probably unsaturated. The crystal was a water- clear specimen projecting two or three centimeters from a mass of dolomite crystals which served as its base. The asso- ciation was not an intimate one and the solution depositing ealcite had evidently not become saturated with magnesium carbonate from the dolomite. ‘The analysis is not given to show the limiting value but to show that fairly intimate asso- ciation of the minerals is necessary in order that the limiting value may be reached. It is perhaps worth mentioning here that iron in appreciable quantities was not found in any calcite examined. No other samples of calcite were analyzed. Dolomite. The specimens of dolomite deposited on calcite were the following : I. A erystalline layer of dolomite deposited on massive eal- cite. Guanajuato, Mexico. II. Small rhombohedral crystals of dolomite deposited as a erust on scalenohedrons of calcite. Guanajuato, Mexico. III. An occurrence similar to that of No. II. On the dolomite, calcite was subsequently deposited (see calcite No. 2). Ouray, Colorado. The analyses are given in Table II. TaBLeE II. Analyses of Dolomite deposited on Calcite. I II Sp. gr. 2°865—2°914 Sp. gr. 2°891—2:907 Guanajuato, Mexico Guanajuato, Mexico a b Average a b Average me Se Sa NNW Gane oan cy OSL RATA ST MgCoO, EE ars 33°52 33°48 33°50 31°50 31°45 31°48 FeCO, See 6°14 5°95 6°05 8°43 8°49 8°46 CaCO, SE eee 60°68 60°61 60°64 60°67 60°76 60°71 100°34 100°04 100°19 100°60 100°70 100°65 III Sp. gr. 2°887—2-860 Ouray, Colorado a b Average Mo CO ee S057 30°85 30°71 KeCOr ras 3°45 3°40 3°43 MnCO, pat eases 3°41 3°07 3°24 63°03 63°04 63°03 100°46 100°36 100°41 344. Foote and Bradley—Solid Solution in Minerals. The analyses show that all specimens contained more or less ferrous carbonate, and one, manganese carbonate also, We consider that these components replace magnesium carbonate. The ratios are as follows : i II Til MeCOr en: omnes ‘397 373 364 RECO ih Saas 052 073 + ~—--080 Mn @ ONS 2 Se eee oy Sie saat 028 CaCO, ae -6064 (0 6omamese : CaCO The ratios - ’___ ealeulated from these results are: (Mg,Mn,Fe)CO, i I II IIL 1°349 1°361 1°492 These ratios show a somewhat surprising excess of calcium above the dolomite ratio, and the excess is variable. Since the material was homogeneous, but deposited directly on calcite, we see no reason why the ratios do not represent approximately limiting values for calcium carbonate in dolomite. From the fact that dolomite occurs so generally in the 1:1 ratio, Ret- gers® assumed that this ratio could not be much exceeded, but this appears not to be the case. The variable ratio is not nnexpected and should be due in large part to the influence of temperature at the time of formation. These were all the specimens analyzed, in which dolomite was clearly the secondary mineral. One specimen was obtained in which there was pri- mary formation of dolomite followed by the formation of a single large calcite crystal. (See calcite anal. No. 6.) About the base of the calcite, a small quantity of dolomite had subse- quently formed. Both deposits were similar in appearance and could not be separated. A sample, chipped off near the base of the calcite crystal and containing some of each deposit, gave the following results on analysis : Sp. gr. 2°834-2°868 Joplin, Mo. a b Average MegCoO, SERS lve 40°31 40°15 40°22 FeCO, ON Sit penta eae 1°91 1°93 1°92 CaCO, Side ayers ae 58-19 58°13 58°16 100°41 100°19 100°80 CaCO, (MeFe)CO, * Loc. cit., vi, 227, 1890. ‘ALhesato = ealculated from these results is 1:176. Foote and Bradley—Solid Solution in Minerals. 345 The proportion of calcium in this case is much larger than is commonly found in a dolomite, but it cannot be regarded as a limiting ratio, as some of the primary dolomite was present. Our results do not show whether the effect of increased tem- perature will be to increase or diminish the proportion of lime at the mixing limit. In this case,as in that of calcite, it is probable that the investigation of limestones containing mag- nesia would give some information. In conclusion, we wish to call attention to the fact that the quantitative isomorphous relations between other mineral car- bonates are quite unknown. To what extent, for instance, siderite or rhodochrosite can take up calcinm carbonate has not been determined. It seems not unlikely that some of these problems, at least, may be settled by artificial prepara- tions, and this method would have the very great advantage of working under known temperature conditions. Chemical and Mineralogical Laboratories of the Sheffield Scientific School of Yale University, New Haven, Conn., Feb., 1914. SCIENTIFIC INTELLIGENCE. I. CuHeEemistry anp Puysics. 1. Fused Magnesium Chloride as a Crystallizing Agent.—K. A. Hormann and K. HéscuEre have observed that magnesium chloride in the anhydrous, fused condition is an excellent solvent and means of crystallization for many inorganic substances. The salt melts at 708° C., a temperature easily obtained by means of a good Bunsen burner, to a very mobile liquid which dissolves many metallic oxides, and some of these crystallize very well upon cooling the solution. Other oxides, and particularly sul- phates, react with the fused magnesium chloride and yield vola- tile chlorides, as for example the chlorides of beryllium, zinc, iron and tin. In other cases compounds of the spinel class are formed by the combination of oxides with magnesium oxide. The latter is formed by the action of the moisture of the flame or the oxygen of the air upon the magnesium chloride. This action leads also to the formation of large brilliant octahedral crystals of magnesium oxide (periclase) which often become mixed with other products. This decomposition of magnesium chloride leads to the corrosion of platinum, gold, silver and copper vessels in which the fusions are made. The platinum is deposited again in the form of beautiful hexagonal crystals, the gold forms. iso- 346 Scientific Intelligence. metric crystals, the silver is deposited as crystalline chloride, while the copper gives red crystals of artificial cuprite, Cu,O. Covered platinum vessels are only slightly attacked by the fusions, and porcelain crucibles resist the action for many hours. Among the compounds prepared were crystallized magnesioferrite, MgFe,O, and products intermediate between this and magnetite, FeFe,O,, well crystallized magnesium orthoborate, Mg,B,O,, and magnesium uranate, Mg,U,O,. Cerium dioxide was crystallized by using the sulphate in the magnesium chloride fusion. The authors regard this as one of the most beautiful substances of inorganic chemistry, as it forms colorless cubic or octahedral crys- tals which are very hard, resembling diamond in brilliancy and luster, and having a refractive index above 1°9. By the use of a small proportion of praseodidymium sulphate mixed with the cerium sulphate, reddish yellow to deep red crystals were obtained. Zirconium dioxide also was obtained in the form of white tetrag- onal crystals. The rare earths, used as sulphates or oxides, gave well crystallized oxychlorides, for instance EKrOCl was the erbium compound. The authors have studied the absorption spectra of a number of these products.— Berichte, xlvii, 238. ibs Wis 2. Carbon Sulphide- Telluride and Carbon Sulphide- Selenide. —ALFRED Srock, P. Prarrorius, and E. WiLtFrotH have pre- pared the compounds CSTe and CSSe by passing an electric are under cooled carbon disulphide between a cathode of tellurium or selenium and an anode of graphite. The tellurium compound 1s very unstable and difficult to purify. It is decomposed by light and exists only at low temperatures. It was purified by distilling in a high vacuum at —35° C. and condensing at —80° C. At low temperatures it is a yellowish-red solid which melts at —54° C. toa brilliant red liquid, and this becomes blood-red at room temperature, then turns black and soon decomposes completely. It has a garlic-like odor. The selenium compound is much more stable than the other. At room temperature it is a liquid of intense yellow color. Its melting point is —85° C. and the boil- ing point 84° C. It has an irritating garlic-like odor. It is decomposed by light and by long standing at room temperature. At higher temperatures it decomposes more rapidly, forming car- bon disulphide, carbon and selenium. In the study of both the tellurium and selenium compounds careful search was made for the presence of the possible compounds CTe, and CSe,, but no evidence of their formation could be found.— Berichte, xlvii, 131, 144, H. L. W. 3. Double Bromides of Gold.—A. GutTsierR and J. Huser have prepared and described about eighty double salts of AuBr, with other bromides, most of which are bromides of organic bases derived from ammonia, but they include the potassium, rubidium and caesium salts, KAuBr,, RbAuBr, and CsAuBr,, the last two of which were described in this Journal by Wells and Wheeler in 1892. Examples of the organic base compounds analogous to the ammonium salts, NH,AuBr,, are the monomethyl ammonium Chemistry and Physics. B47 salt (CH,NH,)AuBr,, the tetra-ethyl ammonium salt (C,H,),N- AuBr,, etc. The authors have decided from their experience with a large number of salts that the double gold bromides of the organic bases are not well suited for the examination of such bases on account of their usual difficult crystallization and their frequent instability.—Zettschr. anorgan. Chem., \xxxvi, 3538. H.)L. W. 4. The Gravimetric Determination of Seleniwn.— Jurius Meyer has shown that the evaporation of nitric acid solutions of selenious acid leads to the loss of appreciable amounts of the sub- stance, and that the loss may be serious when the dry residue is heated for a long time on the water bath. Although several of the best authorities recommend the removal of nitric acid in the course of analysis by repeated evaporation with strong hydro- chloric acid, Meyer has found that this operation leads to very serious losses even in the presence of potassium or sodium chlor- ide. For instance, nearly one-half of 0°14 g. of selenium was lost by evaporating its nitric acid solution to dryness twice with con- centrated hydrochloric acid. The author recommends the em- ployment of hydrazine hydrate for the precipitation of selenium, and he states that the interfering effect of nitric acid may be overcome by the addition of ammonia and some hydrochloric acid.—Zeitschr. analyt. Chem., lili, 145. H. L. W. 5. Active Nitrogen.—E. TiepE and E. DomcKe maintain that Strutt’s phenomena attributed to active nitrogen, which are pro- duced by passing electric sparks through nitrogen at low pres- sures, do not occur when pure nitrogen is employed, and that they are due to the presence of oxygen in the gas used. When they used pure nitrogen prepared by heating barium or sodium azide no luminescence occurred. By the use of hot metallic cop- per they were able to purify commercial nitrogen so that the after-glow did not occur, but they showed that the addition of oxygen to the gas brought about its appearance. The curious fact was brought out that metallic copper will remove the oxygen from its mixture with nitrogen at temperatures between 325° C. and 570° C., but that when the temperature is between 570° C. and 700° C. some oxygen remains in the gas, evidently on account of the dissociation of copper oxide at the low pressures (2 to 12™™) employed.— Berichte, xlvii, 420. H. L. W. 6. Rays of Positive Electricity and their Application to Chem- ical Analyses ; by Sir J. J. Toomson. Pp. vii, 132, 50 figures and 5 plates. London, 1913 (Longmans, Green and Co.).—This book will doubtless be received with enthusiasm by most physi- cists because it gives a consecutive account of the brilliant series of researches on positive rays carried out during the past seven years in the Cavendish Laboratory. Some idea of the scope of the work may be formed from the following list of headings of the sub-divisions of the text. (The material is not arranged in numbered chapters.) ‘Rays of Positive Electricity. Double Cathodes. Rectilinear Propagation of the Positive Rays. Elec- 348 Scientific Intelligence. trostatic Deflection of the Particle. Effect at Very Low Pres- sures. Discussion of the Photographs. Negatively Charged Particles. Atoms Carrying Two or More Positive Charges. Methods for Measuring the Number of the Positively Electrified Particles. Retrograde and Anode Rays. Anode Rays. Dop- pler Effect Shown by the Positive Rays. Spectra Produced by Bombardment with Positive Rays. Disintegration of Metals under the Action of Positive Rays. On the Use of the Positive Rays for Chemical Analysis. On the Nature of X, the Substance giving the ‘3’ Line. Evolution of Helium and Neon.” In his preface the author says;—“I have described at some length the application of Positive Rays to chemical analysis; one of the main reasons for writing this book was the hope that it might induce others, and especially chemists, to try this method of analysis.” It will be very surprising if this hope is not vain, at least in so far as it relates to chemists, because practically no data are recorded which would enable a beginner in the subject either to order the necessary apparatus or to apply the general principles laid down, without spending a very large amount of time in letter writing or in performing preliminary experiments. The proof seems to have been read hurriedly since a number of typographical errors have been overlooked and since several of the figures lack the reference letters used in the text. The book will be chiefly use- ful to those who are already familiar with the original articles published in the Philosophical Magazine and elsewhere, and who desire to review the whole subject without undue labor. H. 8. U- 7. Photo-Electricity ; by H. Stantey ALLEN. Pp. x, 221; 35 figures. London, 1913 (Longmans, Green and Co.).—“The present book is based on a course of advanced lectures delivered by the author at King’s College, London, during the Lent Term of 1910; but the greater part has been entirely rewritten so as to incorpo- rate the results of the large amount of research carried out in the three years since these lectures were given.” The introductory chapter, which gives an outline of the whole field, is followed by chapters on the emission of electrons in a vacuum, on the veloc- ity of the electrons, and on the photo-electric current in gases at various pressures respectively. After discussing photo-electric substances in the various physical states the author turns his atten- tion to the influence on the photo-electric discharge produced by temperature, and by the character, intensity and plane of polariza- tion of the exciting light. The remaining chapters deal with the theories of photo-electric action, photo-electric fatigue, fluores- cence and phosphorescence, and photo-chemical actions and pho- tography. The volume ends with both author and subject indices. The text is a valuable contribution to this branch of physics because in it are brought together, in English, for the first time, the experimental results and hypotheses which are widely scat- tered in the various scientific journals. Furthermore, the appear- ance of this book is especially timely since the most recent work Chemistry and Physics. 349 of Hallwachs, Fredenhagen and Kiistner bids fair to simplify and perhaps revolutionize certain chapters of the subject, so that the volume will greatly facilitate the comparison and adjustment of the earlier and latest ideas and results. HaseUe 8. An Introduction to the Mathematical Theory of Attrac- tion; by Francis A. Tarteron. Vol. II, pp. xi, 207. London, 1913 (Longmans, Green & Co.).—In the preface to the first volume, which was published about fourteen years ago, (see vol. Vill, page 88, 1899) the author says: “... I hope, at some future time, to make this book more complete by the addition of chap- ters dealing with Spherical Harmonics, Conjugate Functions, and the Theory of Magnetism for bodies having finite dimensions.” This hope is fultilled by the present volume, which begins with chapter eight and ends with chapter twelve. The original plan has been departed from in two respects, namely, the subject of conjugate functions has been omitted and a chapter on Maxwell’s theory of light has been added. The more recent developments of the electromagnetic theory of light are not touched upon. The treatment lays more stress on the mathematical than on the physical aspects of the topics discussed. The analysis would have been simplified and made more elegant and up-to-date if the author had used vector methods instead of sealar notation. The index is immediately preceded by a “Note on Thomson and Dirichlet’s Theorem.” H. S. U. 9. The Chemistry of the Radio-elements. Part II. The Radio-elements and the Periodic Law, by FREpEriIcK Soppy. Pp. v, 46. London, 1914 (Longmans, Green and Co.).—This little book deals with the following topics: The periodic table. Chemical and electro-chemical advances. The connection between the sequence of changes and the chemical properties of the pro- ducts. The branching of the disintegration series. Nature of the end products. Atomic weight of lead. The origin of actinium. The spectra of isotopes. Neon and metaneon. Nature and properties of isotopes. The structure of atoms. Nature of the argon gases. The definitions of the new words “isotope”’ and “isotopic” may merit quotation. ‘...a group of two or more elements occupying the same place in the Periodic Table, and being in consequence chemically non-separable and identical, will be referred to as a group of isotopes, and, within the group, the separate members will be referred to as isotopic.” “Thus ionium, thorium, and radio-thorium are isotopes, and mesothorium | is isotopic with radium.” Hs) Sa Us 10. Modern Seismology ; by G. W. Waker. Pp. xii, 88; 13 figures and 13 plates. London, 1913 (Longmans, Green and Co.). —The first five chapters of this book relate to seismometry and the remaining five to seismogeophysics. More specifically, the first 36 pages treat of the general dynamical theory of seismo- graphs and the installation, standardization, sensitiveness, damp- ing, mechanical and electromagnetic registration of the five chief types of seismographs in actual use. The rest of the text is 350 Scientific Intelligence. devoted to the theory of a solid isotropic earth, to the interpre- tation of seismograms, to the determination of epicenter and focus, and to statistical problems. The presentation of the subject is clear and logical, the illustrations are well selected and neatly reproduced, and the entire volume should be very useful, both as an introduction to this important field of investigation and as a practical guide to the use of seismographs. A pocket on the inside of the back cover of the book contains three specimen plates of seismograms recorded by Galitzin instruments at Eskdalemuir at the times of the Dardanelles and Zante earth- quakes. H. S. U. 11. Alternating Currents and Alternating Current Machinery; by Dueap C. Jackson and Joun Price Jackson. Pp. ix, 968 ; 526 figures. New York, 1913 (The Macmillan Co.).—Since the appearance of the first edition in 1896 the subject has grown rapidly so that the present volume has been rewritten and greatly extended as compared with the earlier edition. (See this Journal, vol. il, page 455, 1896.) Some of the improvements of the text may be briefly stated as follows: More attention is paid to the transient state in electric circuits than formerly. A considerable amount of related matter has been introduced in respect to vec- tors, complex quantities and Fourier’s series. The treatment of power and power factor has been given great attention, and an entire chapter is now devoted to the hysteresis and eddy current losses which are developed in the iron cores of electrical machinery. The discussion of synchronous machines and of asynchronous motors and generators has been amplified and made more complete. Finally, the treatment of the self-inductance and mutual-inductance of line circuits and skin effect in conduc- tors has been extended and supplemented by the consideration of electrostatic capacity of lines and the influences of distributed resistance, inductance, and capacity. The manner of presentation is intended to serve a twofold pur- pose, namely, to fulfil the requirements of classes in engineering schools and to serve as a reference book for electrical engineers. Scattered throughout the volume may be found 208 problems for solution by the student, and also a fairly large number of foot- notes referring to other standard texts and original journal articles. The book under consideration is apparently the best and most up-to-date on the subject in the English language. 36 feb Wie 12. The mutual Repulsion of rigid Parallel Plates separated by a Film of Air ; by ©. Barus. (Communicated.)—By the ap- plication of displacement interferometry to the horizontal pen- dulum,* I find that two parallel rigid plates whose distance apart is of the order of 1™™ and less repel each other in air, with a force far in excess of their gravitational attraction. This force increases rapidly (certainly as fast as the inverse square) as the distance of the plates decreases and vice versa, but can be recognized beyond a millimeter of distance. For brass plates 20°" in diameter and * The full method will be shown later in this Journal. Geology and Mineralogy. 351 1™™ apart, the repulsion in question is of the order of °5 dyne and therefore equivalent to a pressure of °0015 dynes/em or roughly 10~° atmospheres. It is in excess of any electric repulsion due to the absolute voltaic potential of the discs. The suspended plate reaches its position of equilibrium gradually, the motion progress- ing at a retarded rate through infinite time, in a way character- istic of the viscosity of the film of air between the plates. I have estimated the intensity of the force both from the repulsions of a vertical plate suspended from the horizontal pendulum on opposite sides of a fixed parallel identical plate ; also by charging pairs of plates to a given difference of potential for a given distance apart. So far as can be seen, the repulsion is caused by the condensation of air on the surface of the plates by molecular and not by gravitational force (which is too small). Hence, the method employed should enable the observer to find the density of the concentration in terms of the distance from the plate and the law of attraction of the plate in terms of distance, within the small distances in question. In other words, a method for the direct investigation of molecular force is here apparently given. Brown University, Providence, R. I., Feb. 24. Il. Gronocgy anp MINERALOGY. 1. Third Annual Report of the Director of the Bureau of Mines, Joserpu A. Houmess, for the fiscal year ended June 30, 1913. Pp. vii, 118. Washington, March, 1914.—The Bureau of Mines was established by act of Congress in 1910, and began its work on July 1st of that year, taking over the investigations lying in its field which had been carried on previously by the U.S. Geological Survey. 0212-2) 2 2 ee XXIX.—Solid Solution in Minerals. V. The Isomorphism between Calcite and Dolomite; by H. W. Foorr and Wal: Brapley. sc eS eee 339 SCIENTIFIC INTELLIGENCE. Chemistry and Physics—Fused Magnesium Chloride as a Crystallizing Agent, K. A. Hormann and K. Héscuetz, 345.—Carbon Sulphide-Teliuride and Carbon Sulphide-Selenide, A. Stock, P. PRamtTorius, and EK, WILLFRO'TH: Double Bromides of Gold, A. Gursrer and J. Huser, 346.—Gravimetric Determination of Selenium, J. MzyreR: Active Nitrogen, E. TimpEr and H. Domcxke: Rays of Positive Electricity and their Application to Chemical Analyses, J. J. THomson, 547.—Photo-Hlectricity, H. S. ALLEN, 548,— Introduction to the Mathematical Theory of Attraction, F. A. TARLETON: Chemistry of the Radio-elements, Part II, F. Soppy: Modern Seismolog ey, 349.— Alternating Currents and "Alternating Current Machinery: Mutual Repulsion of rigid Parallel Plates separated by a Film of Air, C. Barus, 350. Geology and Miner alogy—Third Annual Report of the Director of the Bureau of Mines, 351.—Canada Department of Mines, 352.—Report of Topo- graphic and Geologic Survey Commission of Pennsylvania, 1910-1912, 303.—Graphite Deposits of Pennsylvania, B. L. MinteR: Geological Sur- vey of New Jersey: Geologic Atlas of the United States, Niagara Folio, New York, EH. M. Kinpie and F. B. Taytor: Union of South Africa, Mines Department, Annual Reports for 1912, 354.—Inter-State Conference on Artesian Water: The Ocean: General Account of the Seience of the Sea, J. Murray.—Descriptions of Land, R. W. Cautuey: Interpretation of Anomalies of Gravity, G. K. GinpErt: Mud Lumps at the Mouths of the Mississippi, 356.—La Face de la Terre, EK. Suzss: Water and Volcanic Activity: Zeitschrift fir Vulkanologie, 357.—Igneous Rocks and their Origin, R. A. Dany, 358.—Bergalite, 359.—Notices of recently described Minerals, 360. Miscellaneous Scientific Intelligence—Carnegie Institution of Washington, 361.—Annual Report of the Superintendent of the United States Coast and Geodetic Survey: Newcomb-Englemanns Populire Astronomie, 363.— Milton’s Astronomy, The Astronomy of Paradise Lost, T. N. OrosarRp: Trigonometry, A. M. Kenyon and L. Incoup, 364.—Who’s Who in Science, International, 1914: Franklin Institute award of the Hlliott Cres- son Medal: The Cambridge Manuals of Science and Literature: Tables and other Data for Engineers and Business Men, C. E. Ferris, 360. Obituary—J. Murray: E. S. Honpen: W. W. Batiey: G. Mprcaii: A. GUnTHER: H. B. Woopwarp: A. R. CLARKE, 566. — _ Smithsonian Institution. Cu, a Wal; Lax, . VOL. 2 — : MAY, 1914. a Established by BENJAMIN SILLIMAN in 1818. THE AMERICAN JOURNAL OF SCIENCE. Epitorn: EDWARD S. DANA. ASSOCIATE EDITORS Proressors GEORGE L. GOODALE, JOHN TROWBRIDGE, W. G. FARLOW anp WM. M. DAVIS, or Camprivce, Proressors ADDISON E. VERRILL, HORACE L. WELLS, LOUIS V. PIRSSON, HERBERT E. GREGORY anD HORACE S. UHLER, or New Haven, Proressor HENRY S. WILLIAMS, or Irwaca, Proressor JOSEPH S. AMES, or Baurtimore, Me. J. S. DILLER, or Wasuineton. FOURTH SERIES VOL. YW HOLE NUMBER, CLXX) arte. soma ins 7, No. 221—MAY, 1914. 4 MAY 2 1914 | Metional ale lh WITH PLATE IX, Muse i NEW HAVEN, CONNECTICUT. LOL 4. THE TUTTLE, MOREHOUSE & TAYLOR CO., PRINTERS, 123 TEMPLE STREET. Published ae 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). HODGKINSONITE, A NEW MINERAL. We have been fortunate enough to secure the best specimens of this very rare mineral. It is from the celebrated Franklin Furnace mines and is arare compound, the formula of ‘which is Mn(ZnOH).SiO,. It crystallizes under the monoclinic system and is pink in color associated with barite. Ina few specimens itis associated with the rare minerals Pyrochroite and Gageite. The whole makes a very pretty specimen. Prof. Charles Palache has ana- lyzed and will soon publish a description of-it. The quantity found is scarcely enough to supply the scientific institutions who will want a speci- men. Prices range from $1.00 to $25.00. A NEW OCCURRENCE—Fluorescent Willemite with Rhodocrosite.. The Willemite occurs in transparent crystals running in size from 2 milli- meters in diameter to almost a hair in thickness. They occur in cavities on Rhodocrosite, making a very beautiful specimen. Under a current they show a strong fluorescence. . Prices range from $1.00 to $5.00. SYNTHETIC GEMS. It is remarkable the interest that has been taken by scientists in these wonderful scientific discoveries. The Corundums are now produced in Pigeon blood, Blue, Yellow, Pink and White. Also the new Indestructible Pearls in strings with gold clasps. These are identicalin hardness and rival in color and lustre the real gems. They can be dropped and stepped on without injury and are not affected by acids. My collection of the above is unrivalled, and prices of the same are remarkably low. : : OTHER INTERESTING DISCOVERIES AND NEW FINDS Will be found in our new Catalogues. These consist of a Mineral Catalogue of 28 pages; a Catalogue of California Minerals with fine Colored Plates; a Gem Catalogue of 12 pages, with illustrations, and other pamphlets and lists. These will be sent free of charge on application. Do not delay in sending for these catalogues, which will enable you to secure minerals, gems, etc., at prices about one-half what they can be secured for elsewhere. REMOVAL NOTICE. Finding that at our recent location we were placed at a disadvantage in receiving and shipping our numerous consignments, we decided to return to our old location, where we are surrounded by the greatest gem and financial district in the world, and are near the vast shipping interests, Custom House, ete., etc. Our old offices have been entirely renovated in a beautiful manner, new cases have been fitted and all who take the trouble to call on us will find themselves well repaid by the beautiful display of minerals, gems, curios, etc., etc. ALBERT H. PETEREIT 81-83 Fulton St. New York City. mt" Pu va 5 Am. Jour Sci., Vol. XXXVII, May, 1914 Plate IX, ara ut Ponta | om _Tres Irmaos ms Santa> Aare Metis Caigara™ So Carnaubinha, COAST OF NORTHEASTERN BRAZIL rhe Br FROM THE HYDROGRAPHIC CHARTS WITH A PORTION IN DETAIL FROM A TRAVERSE BY G.A WARING o 2 tna 22 IE ge Maracajahd SHOWING REEFS io or "es ae —Calcareous sandstone reefs suas | ~ON- Cemented sandstone reefs aut Coral reefs yt Barra do*- Maxaranguape EQUATOR A 100 MILES fo} bi 10 MILES ae ae aeons ii iin Reser He i \ { t e ‘ THE AMERICAN JOURNAL OF SCIENCE [FOURTH SERIES.] 0% Arr. XXX.—Reef Formations of the Northeast Coast of Brazil; by Geratp A. Waring. With Plate LX.* Location and character of the region. Calcareous sandstone. Reefs. a General features. Aaqgoman INStitaa~ 5 yA “aN Dred "tle Mode of formation. fe "Dy Reefs north of Rio Grande do Norte. f? 2 Other occurrences of calcareous sandstone. i { 3 AF i 9 Tron-cemented sandstone. \ MAY 2 4 914 Cliffs. XQ W ae ‘ 7 at « AS ‘ Reefs. ~~ aig | Mi (Sey Coral reefs. lal Wiese Summary. LocATION AND CHARACTER OF THE REGION. Lyine wholly within the tropic zone, northeastern Brazil forms the shoulder of the South American continent. To the northwest is the Amazon region, noted for its abundant rains and tropic vegetation. To the south also the country is well supplied with water; but the region between about 3 and 15 degrees S. lat. isarid and is subject to periods of great drought. Few rivers drain this area, and except for a narrow belt along the coast, water is scarce during the greater part of the year. All of this coastal extent is low. for many miles the shore is bordered by sand dunes, which extend back afew miles toa gently rising plain, covered with a stunted forest of tall brush and cactus, and interrupted here and there by small isolated peaks and mountains. There are few settlements other than fishing villages along this barren stretch of coast. Nearly all the streams that enter the ocean discharge water only during * The writer is indebted to Dr. J. C. Branner for encouragement in the preparation of this article. Am. Jour. Sct.—FourtH Series, Vou. XX XVII, No, 221.—May, 1914. 26 368 G. A. Waring—Reef Formations of the periods of storm, and are hence of uncertain flow; for as is true in other arid regions, the rainfall of northeastern Brazil is very irregular in amount and distribution. CALCAREOUS SANDSTONE. Reers. General features. At certain places along this coast there are sandstone reefs whose character and appearance is considered by scientists and others who have seen them to be almost unique. The only other known occurrences of a similar formation appear to be at a few places along the eastern elite of the Mediterranean and on the borders of the Red § Sea, and little information is available concerning the reefs of these localities. The best known reef on the Brazilian coast is that which protects and in fact forms the harbor of Pernambuco, in about 8 degrees south latitude. This reef has been described or mentioned by numerous writers. In 1836 Charles Darwin in writing of it said that “at first sight it is unser to credit that it is the work of nature and not of art.” Dr. J. ©. Branner, eiedilank of Stanford University and head of its geological department, who has worked on the geology of Brazil for many years, has published an exhaustive study of the reefs.+ The principal features of the reef have also been summarized by him in a later paper,t and may be stated as follows: The reefs are long, narrow, and nearly straight, and are situ- ated in most instances near the mouths of efreams and nearly parallel to the present beach. At some places where reefs have been formed, the reef is not connected with the land; in other places one end is exposed on the beach or is buried beneath the sands, while in still others the reef lies as a whole along the present beach. The axis of each reef is practically horizontal and lies at about high-tide level. Its top, however, slopes gently seaward and in section the material also exhibits bedding planes that dip at low angles toward the sea. While the better -developed reefs are as a rule nearly continuous, in many places they are cracked, broken and undermined by the surf, and many of the smaller reefs have been greatly broken foo) up by this means. The lithified material is only ten or fifteen * Journal of Researches into the Geology and Natural History of the vari- ous countries visited by H. M.S. Beagle . . . from 1832 to 1836. London, 1839, p. 593. + The Stone Reefs of Brazil, their geologic and geographic relations, with a chapter on the coral reefs, Bull. Mus. Comp. Zoology, vol. xliv, Geol. Series vii, Cambridge, 1904. t{ Stone Reefs on the Northeast Coast of Brazil, Bull. Geol. Soc. Am., vol. Xvi, pp. 1-12, Feb., 1905. Northeast Coast of Brazil. 369 feet thick at most, and is underlain by the usual coastal sands and clays. The outer edge of a reef is in many places nearly vertical, due to the breaking off of masses which now lie beside it in inclined positions and partly protect the exposed face from the force of the surf. The inner face is commonly broken much less, but its edge is irregularly etched. In many places the surface of the reef is also deeply etched into fantastic points and pinnacles, and contains shallow pools; but ordinarily it is fairly smooth and is partly covered by organic erowths, which are more abundant near the seaward edge. The material is a sandstone that contains quartz grains, with oceasional beds of pebbles, a little mica, and shells and other organic remains in varying amounts, all being firmly cemented together by lime carbonate. The material is mainly so hard that it rings under the hammer and fractures across the quartz grains and pebbles ; but in some places it is relatively soft. The shells appear to be those of kinds at present living in the adjacent water, and the other constituents of the rock are those of the present beaches. Mode of formation. Difference in hardness seems to be the only important feature that distinguishes the material of the reefs from that of the beaches. The conclusion, therefore, is that the reefs are beaches or sand spits lithified in place. This origin of the reefs and their mode of formation was worked out by Dr. Branner after a detailed study of the various reefs south of the port of Rio Grande do Norte. The various phases of the problem are treated fully in his publication on the subject,* aud his conclusions can only be summarized here. The main features brought out by his study are as follows: The reefs are in reality hardened beaches or spits, that have been encroached upon from both sides by the sea. Their straightness implies that the original beaches or spits were straight, and seems to be well accounted for by the usual process of straightening of an irregular shore line by the cut- ting down of headlands and the filling of embayments. The natural features of the coast, notably the lack of good harbors, the presence of low headlands which protect shallow bays, and of stream mouths partially or wholly closed by sand bars, are some of the factors which point to this conclusion. The reefs are therefore believed to be portions of mature, nearly straight, beach lines. The lime carbonate that har dened these beaches appears most probably to have been dissolved from calcareous beach sands by streams entering behind them, and to have been redeposited along the zone of contact with the salt water as the fresh water percolated seaward through these * Bull. Mus. Comp. Zool., Geology, vii, 1904. 22 370 G. A. Waring—Reef Formations of the sands. The solvent action of the fresh water for lime carbo- nate was increased by organic acids derived from decaying vegetation, and the redeposition of the lime was mainly due to the inability of the increasingly salty water to hold it longer in solution. The conditions necessary to such an hypothesis are all present along the coast where the reefs exist. The annual rainfall is so irregular in amount and distribution that the streams carry little or no water during a large part of the year and their outlets become partially or wholly closed by drifting sand. Vegetation, especially the mangrove and the water hyacinth, grow luxuriantly in the tidal portions behind many of these closed river mouths, and the decay of dead leaves and other fragments of vegetation doubtless charges the water with organic acids. This water gradually escapes to the sea by percolation through the sands that block the way, and at times by flowing long distances around the ends of sand bars. The sand contains many fragments of shells, coral and other calcareous animal remains, so that it may “easily yield the requisite lime carbonate for the cementing process. Studies made by the Challenger expedition have shown that the ocean water is unusually dense in an area that borders the Brazilian coast approximately along the portion where the reefs are found.* This greater density of the ocean water tends to cause more rapid deposition of the lime, due to the inability of the dense water to readily hold it in solution, and is believed to be another important factor which caused the formation of the reefs. Reefs north of Rio Grande do Norte. The structure and the mode of formation of the reefs have been thoroughly discussed by Dr. Branner, and the present paper is not written in an attempt to add anything to his con- clusions in those respects. Dr. Branner’s studies, however, were confined to the reefs at and south of Rio Grande do Norte, and as in February, 1913, the writer traversed a portion of the coast north of that port, the following notes concerning the region are presented. These notes are intended mainly to record the calcareous sandstone reefs noted, but observations on the other kinds of reefs seen are also given. The harbor at Rio Grande do Norte is formed by an especi- ally well-developed reef, descriptions of which have been pub- lished.t Since no reefs equalling it in development were seen farther north, the illustration (fig. 1) is here presented, for * Challenger Reports, Physics and Chemistry, vol. i. +The stone reef at the mouth of the Rio Grande do Norte, by Dr. J. C. Branner and ©. E. Gilman, American Geologist, Dec., 1899, pp. 342-344. The stone reefs of Brazil, by J. C. Branner, Bull. Mus. Comp. Zool., Geol. Series vii, pp. 35-39, 1904. Northeast Coast of Brazil. 371 together with the representation of this reef on Plate LX, it well brings out two characteristics of the better- developed . reefs, 1.€., “their relations to present beach lines, and their posi- tions relative to stream mouths. North of Rio Grande do Norte the first typical caleareous sandstone was found at the mouth of Rio Ceara-Mirim, where a double reef of the material is exposed at low tide (fig. 2). The two reefs are parallel, about 100 yards apart and curve Fic. 1 Fic. 1. Reef at mouth of Rio Grande do Norte. gently with the present beach. The inner reef is the shorter, and some parts of it project only a few inches above the sand of the flat beach that is laid bare by the receding tide. Its rock is brighter in color and softer than the usual pale yellow to gray material, but otherwise it has the appearance of the typical reef rock. It is the most probable example noted of a recently formed deposit. A tidal lagoon, fringed with man- erove, is formed behind a bar which separates it from the beach. The outer reef consists of the usual hard, gray material. About 450 yards beyond the place where its north- ern end disappears beneath the water, another stretch of the same material, probably its continuation, appears close to the shore and extends around a gently curving beach, gradually receding and disappearing beyond the surf. Proceeding northward, the reef rock next appears as a few scattered blocks in the surf about one mile south of a low point near the village of Pitangy. About 500 yards beyond these blocks, the material is exposed as a continuous reef, with bed- 372 G. A. Wuring—Reef Formations of the ding planes dipping 3 degrees seaward, and lies partly as a barrier, partly along the beach 1, suggesting that a little bay has been washed out behind it as shown i in fie. 3a. Its northern end is buried in the beach. Near the village of Jacuma, a few miles northward, a small reet of the rock extends from the beach into the surf. Ata low point beyond, a similar reef, which first appears in the water, extends in a straight line in such a direction that it Fic. 2. = Rar CT gate en es ed a — = Nate Fic. 2. Bird’s-eye view of reef at mouth cf Rio Ceard-Mirim. passes for some distance along the beach and then disappears in the water beyond the point (fig. 36). No signs of a calcareous sandstone reef were found near the mouth of Rio Maxaranguape, 42 miles farther north, though this stream has a small perennial flow and contains much man- grove in its lower portion. The South Atlantic equatorial current flows westward from Africa and divides at Cabo Sao Roque, part going northward and part southward along the Brazilian coast. The southward-flowing current may have been of influence in preventing the formation of a calcareous sandstone reef at the mouth of the Maxaranguape, which is only one mile south of the cape. The position of the Maxaran- guape, near a projecting part of the coast, also appears not to be favorable to the formation of a calcareous reef, for nearly all of the prominent reefs described by Dr. Branner extend along inward-curving beaches, northward from river mouths.* * This characteristic is well brought out by the several detailed maps of reefs accompanying Dr. Branner’s report (Bull. Mus. Comp. Zool., vii, 1904). The northward discharge of river mouths, produced by drifting sand. is mentioned by Olaf Pitt Jenkins, Geology of the Region about Natal, Rio Grande do Norte, Brazil, Proce. Am. Phil. Soc., vol. lii, No. 211, p. 4, Sept.- Oct., 1913. Northeast Coast of Brazil. 373 The first observed occurrence of the typical reef rock north of Jacuma is near the mouth of Rio do Fogo, 24 miles away in a direct line. About one-half mile north of the mouth of this small stream a reef of the usual material appears off-shore and trends N. 20° W. (mag.) while the beach trends N. 6° W. Hence after a short distance the reef attains the beach, at a Fic. 3. SS \Riodo Fogo = Fie. 5. Positions of calcareous sandstone reefs relative to the present beach: a, near Pitangy;: 6, near Jacumdé; c, near Rio do Fogo; d, north- west of Ponta Reducto. place where the beach turns a little more to the west (fig. 3e). At this place the reef also turns, but at an angle instead of in the nearly universal gentle curve, and bearing } ~N. 16° Wheexe tends for about six-tenths of a mile in a straight line nearly parallel with the beach, finally disappearing in the surf.