FCM RL, 0 Wy A ate Hd om om de YF ae Ne Ee RT Pee CAH ORR CR a RL LS LE Uocereah ete Ibe Ae Se ne OH fel ote ee hy a Aye = Ante le he eine tee. ar & Se Nadie Le hewbeite te Nt Me mm Be emtet A te A hte aft te heme Se de mene PA neta aeDeny B20 vy etn shobe /yite Noten! wtengher yo ba ate et AH BG indy Cowl OT, - WAVE OR, To peor eset oo ' : Ft a a Saw nee wg oy : ace? : CPAP CG Pee se ee CRY el Se ol eee ee to eae fa eae os * 4 re | Oe tl “gil 3 Nie Ml ' i S oe Dae ‘9 ; Fae Ss WA BALLS MEN ely at Wa Seana ie ; : : i, ‘ | : 4 = i Py SE ? oem, oh i es As tee Say ag tY Be fs eS y Sat ve uk ‘ie hy Se (aes soo THE AMERICAN JOURNAL OF SCIENCE. Epviron;: EDWARD S. DANA. ASSOCIATE EDITORS Proressors GEORGE L. GOODALE, JOHN TROWBRIDGE, W. G. FARLOW ann WM. M. DAVIS, or Camprincez, Proressors ADDISON EH. VERRILL, HORACE L. WELLS, LOUIS V. PIRSSON, HERBERT EF. GREGORY anp HORACE S. UHLER, or New Haven, Proressor HENRY S. WILLIAMS, or Irwaca, Proressor JOSEPH S. AMES, or Battrvors, Mr. J. S. DILLER, or Wasuineton. FOURTH SERIES VOL. XL—[ WHOLE NUMBER, CXC.] WITH THREE PLATES. NEW HAVEN, CONNECTICUT. Jy uae %33856 ies ty \ f int f _ ie % ' ; i nee ea t g 5 8 Z y ‘ Ta ” Py 4 f pe THE TUTTLE, MOREHOUSE NEW HAVEN AYLOR COMPANY, CONTENTS TO VOLUME XL. Number 235. 2 age Art. I.—F actors in Movements of the Strand Line and their Results in the Pleistocene and Post-Pleistocene ; by PREPARE Ry ags sey eee Se elo ope hae oye Saad a 1 Il.—Heat of Formation and Polymerization of some Oxides and Determination of the Heat of Combination of Water by Fusion with Sodium Peroxide ; by W. G. Mixter __ 23 III.—A Study of the Relations existing between the Chemi- eal, Optical and other Physical Properties of the Mem- bers of the Garnet Group ; by W. E. Forp...-.-.----- 33 IV.—The Lower Ordovician (Tetragraptus Zone) at St. John, New Brunswick, and the New Gerus Protisto- Brapiis byewe It. MCMEARN S 2.2200. /2 2222222 eee 49 V.—A Study of the Recent Crinoids which are Congeneric wath Mossil Species ; by A. H. Crarm _--.-..----.---- 60 VI.—Relation between the Maximum and the Average Bathy- metric Range, etc., in the Subfamilies and Higher Groups of Recent Crinoids ; by A. H. Crark _--.------------ 67 VII.—Separation of Potassium and Sodium by the Use of -Aniline Perchlorate and the Subsequent Estimation of Phossacnuimes by pO. Wis Eiht ts 6 rai a SS ade SSPE ee 75 SCIENTIFIC INTELLIGENCE. Chemistry and Physics—Prout’s Hypothesis, W. D. Harxins, 78.—Action of Chloroform upon Metallic Sulphates : Explosives, A. MarsHALL : Chemical Technology and Analysis of Oils, Fats, and Waxes, J. Lewkowi1tsca, 79.— Annual Reports on the Progress of Chemistry for 1914 : X-Ray Band Spec- tra, H. WaGner, 80.—Elements of Optics, G. W. ParKker: Dielectric Phe- nomena in High Voltage Engineering, F. W. PrEx, Jr., 82.—Radium Uranium Ratio in Carnotites, 8. C. Linp and C. F. Wuittemorse, 83. Geology and Mineralogy—Climate and Evolution, W. D. MartrHew, 83.— Publications of the U.S. Geological Survey, G. O. SmirH, 85.—U. S. Bureau of Mines: Canada Department of Mines, 87.—Lavas of Hawaii, W. Cross, 88.—Brief Notices of some Recently Described Minerals, 89.— Amateur’s Introduction to Crystallography, W. P. Brae: Die 32 kris- tallographischen Symmetrieklassen und ihre einfachen Formen, H. A. Woturine: Annual Tables of Constants and Numerical Data, Chemical, Physical and Technological, 91. Botany—Transpiration and the Ascent of Sap in Plant, H. H. Dixon, 91.— Manual of Weeds, ApA E. Groreia: Plant-Breeding (Bailey), A. W. Git- BERT, 92.—Fundamentals of Plant-Breeding, J. M. Counter: Principles of Fruit-growing, with Applications to Practice, L. H. Barty, 93. Miscellaneous Scientific Intelligence—The Carnegie Foundation for the Ad- vancement of Teaching; Ninth Annual Report, H. S. PrircuEett, 93.— Publications of the Carnegie Institution of Washington: Crocker Land Expedition, 94.—Spencer Fullerton Baird ; A Biography, etc., W. H. Datu, 95.—Publications of the British Museum of Natural History: Rumford Medal of the American Academy of Arts and Sciences, 96. Obituary—A. H. CourncH: H. Mtuuer: A. S. Stein, 96. 1V CONTENTS. Numer 236: Page Art. VIII.—The Igneous Origin of the ‘“ Glacial Deposits” on the Navajo Reservation, Arizona and Utah ; by H. E. GREGORY 2222-222 08 -as ea ee 97 IX.—The Energy of a Moving Electron ; by L. Pace. -_-- 116 X.—A New Nebraska Mammoth, Elephas hayi; by E. U. BARBOUR 2-2. /o2--- 42-3 2-2 ee ee eer XI.—A New Gavial from the Late Tertiary of Florida; by EK... H. SELUARDS: 22.23.2502 42 22 tee oho: XII.—Chlamytherium septentrionalis, an Edentate from the Pleistocene of Florida; by E. H. Setiarps .---------- 139 XIII.—Bournonite Crystals of Unusual Size from Park City, Utah ; by F. R. Van Horn and W. F. Hunt-_-------- 145 XIV.—The Age of the Castile Gypsum and the Rustler Springs Pormation ; by J. A; WnpENn= 52252222 e= eee 151 XV.—On the Determination of Lead as Sulphite ; by G. 8. JAMIESON .2 02. 2-2 220-25 530 42 aoe ee 157 XVI.—The Crystallization of Haplobasaltic, Haplodioritic and Related Magmas; by N. L. Bowmwn _------------- 161 XVII.—The Migrations and Geographic Distribution of the Hossil Amphibia ; by R. 1.’ Mooprm_ 222 22222 186 XVIII.—The Microscopical Characters of Volcanic Tuffs— a Study for Students ; by L. V. Prrssons2=) === 191 XIX.—Northfieldite, Pegmatite, and Pegmatite Schist ; by B. K. Emerson 220220). 222. 22 eee SCIENTIFIC INTELLIGENCE. Geology and Mineralogy—Water Reptiles, Past and Present, S. W. WILLI- ston: Illinois Coal Mining Investigations, 217.—Topographie and Geologic Survey of Pennsylvania, R. R. Hick: West Virginia Geological Survey, I. C. Wuire: Wisconsin Geological and Natural History Survey, E. A. BirGE, 218.—Geological Survey of New Jersey, H. B. KtUmmez: Native Silver in Glacial Material at Columbia, Mo., W. A. Tarr: Mineral Re- sources of New Mexico, F. A. Jones, 219.—Geological Investigations in the Broken Hill Area, D. Mawson: The Turquoise; A Study of its His- tory, Mineralogy, Geology, Ethnology, Archzology, Mythology, Folklore and Technology, J. E. PoGur: Chiastolites from Bimbowrie, South Australia, D. Mawson, 220. Miscellaneous Scientific Intelligence—Ancient Hunters and Their Modern Representatives, W. J. Souzas, 220.—Insects and Man, C. A. EALAND, ' 221: Thirteenth Report on the Sarawak Museum, 1914, J. C. Mourton, 222. Obituary—Dr. JosrpH A. Houmas, 222. CONTENTS. ; v | Number 237. Page Arr. XX.—A Shaler Memorial Study of Coral Reefs; by Riese DAIS) Assets ees 2 ee Soaeisha ae ee 223 XXI.—Notes on Black Shale in the Making; by W. H. 4) WIDISTEIO LMS, 12 eee 1 ey A os ee Pee se RE 272 XXII.—Anodic Potentials of Silver: I. The Determination of the Reaction Potentials of Silver and their Signifi- PANCCHMD YO gill WED Sosa eso toes eta Sel Le SStS 281 XXIII.—Use of Compensators, Bounded by Curved Surfaces, in Displacement Interferometry ; by C. Barus -------- 299 XXIV.—Radioactivity of Spring Water; by R. R. Ramsey 309 XXV.—Stream Piracy of the Provo and Weber Rivers, Rian ppyeGe Hi. ANDERSON,. 2252222 2l222s2.besete. 34 SCIENTIFIC INTELLIGENCE. Geology—Geological Survey of West Australia, 316.—Physiographic Geology of West Australia, J. T. Jutson, 317. Miscellaneous Scientific Intelligence—The Social Problem; a Constructive Analysis, C. A. ELLwoop, 317.—Societal Evolution; a study of the Evo- lutionary Basis of the Science of Society, A. G. KrLier: British Associa- tion for the Advancement of Science: American Association for the Advancement of Science, 318. vi CONTENTS. Number 238. Page Art. XXVI.—The Mammals and Horned Dinosaurs of the Lance Formation of Niobrara County, Wyoming ; by IOS, vn... 2 aa eae oa ee 319 XXVIIL—A Note on the Qualitative Detection and Sepa- ration of Platinum, Arsenic, Gold, Selenium, Tellurium and Molybdenum; by P. E. Brownine.---.-----.---- 349 XXVITI.—On Aventurine Feldspar; by O. ANDERSEN. (With Plates I-II]).--- 2255 22. 2 3 See XXIX.—Anodie Potentials of Silver: IJ. Their Role in the Electrolytic Estimation of the Halogens; by J. H. REEDY 2225 2.223204 3 ee 400 XXX.—Nephelite Syenites of Haliburton County, Ontario ; by W.. Go Foy .25 O20 {V6e eee 413 XX XI.—Post-Glacial History of Boston; by H. W. Samer 437 SCIENTIFIC INTELLIGENCE. Chemistry and Physics—Improvement of High Boiling Petroleum Oils, A. M. McAresz, 443.—Arsenious Oxide as an Alkalimetric Standard, A. W. C. Menzies and F. N. McCartray: A New Method for the Qualitative Sepa- ration and Detection of Arsenic, Antimony and Tin, F. L. Hany, 444.— An Alleged Allotropic Form of Lead, H. HeLurr: Reflection of Gas Molecules, R. W. Woop, 445.—The Stark Effect for Solids, C. E. Mrn- DENHALL and R. W. Woop, 447. Obituary—F. W. Putnam: J. H. Van Amrinen: K, E,. Gutue: P. WaRwicnH: J. VON Payer, 448, CONTENTS. vil Number 239. Art. XXXIJI.—Experimental Studies and Observations on icensiructure;\ by O: Di von WNGELN -2 2-2-2222. -2-- 449 XXXIIL—A Mounted Specimen of Dimetrodon incisivus Cope, in the University of Michigan; by E. C. Case __ 474 XXXIV.—A Fossil Ruminant from Rock Creek, Texas, Preptoceras mayfieldi sp. nov.; by E. L. Troxett ----- 479 XXXV.—The Separation and Estimation of Aluminium and Beryllium by the Use of Acetyl Chloride in Acetone ; iho TEL, TDs dy IGS pe ge ae, eine oe ee a ee 482 XXXVI.—On the Interferences of Crossed Spectra and on Trains of Beating Light Waves; by C. Barus-------- 486 XXXVII—The Brandywine Formation of the Middle Atlantic Coastal Plain; by W. B. Cuark -.---------- 499 XXXVIII.—On Two Burners for the Demonstration and Study of Flame Spectra; by P. E. Brownie ---. ---- 507 XXXIX.—On the Preparation of Glycocoll and Diethyl Car- bonate ; by W. A. Drusuet and D. R. Knapp -___---- 509 XL.—On the Preparation and Properties of Hydracrylic Esters ; by W. A. Drusuet and W. H. T. Hotpen._-- 511 SCIENTIFIC INTELLIGENCE. Chemistry and Physics—Volumetric Estimation of Lead, F. D. Mitzzs: Search for an Alkali-Metal of Higher Atomic Weight than Czsium, G. P. Baxter, 514.—Experimental Organic Chemistry, J F. Norris: Labora- tory Experiments in Organic Chemistry, E. P. Cook: Elements of Physical Chemistry, H. C. Jonzes: Alcoholometric Tables, E. THorps, 515.—Brief Course in Metallurgical Analysis, H. Zrmeru: Characteristics of Long Direct-Current Arcs, W. Grorrian, 516.—Prinzipien der Atomdynamik, J. StarRK: Ten Years’ Work of a Mountain Observatory, G. E, Hats, 517. Electrical Nature of Matter and Radioactivity, H. C. Jonrs: Book of Wire- less, A. F. Couiins, 518.—Plane Geometry, C. I. Patmrr and D. P. TaYtLor, 519. Geology and Mineralogy—Publications of the United States Geological Sur- vey, G. O. SurrH, 519.—Relation of the Cretaceous Formations to the Rocky Mountains in Colorado and New Mexico, W. T. Lez: Conceptions regarding the American Devonic, J. M. CLARKE: Fauna of the San Pablo Group of Middle California, B. L. CLarK: Cretaceous Sea in Alberta, D. B. Downe, 521.—Wabana Iron Ore of Newfoundland, A. O. Hayrs: Yukon- Alaska International Boundary, between Porcupine and Yukon Rivers, D. D. Carrnes: Ordovician Rocks of Lake Timiskaming, M. Y. Witiiams: Structural Relations of Pre-Cambrian and Paleozoic Rocks North of the Ottawa and St. Lawrence Valleys, E. M. Kinpz and L. D. Burtine, 522.— Geology of Franklin County, A. M. Minter: Revision of the Tertiary Mol- lusca of New Zealand, H. Surrr: Third Appendix to the Sixth Edition of Dana’s System of Mineralogy, 523. Miscellaneous Scientific Intelligence—Publications of the Carnegie Institu- tion of Washington, 523. Obituary—J. H. Fapre: T. Bovert: H. G. J. Mosetey: D. T. GwyNnNE- VauGHan: W. Watson: A. J. DuBois, 524. Vill CONTENTS. Number 240. Page Art. XLI.—A Metallographic Description of Some Ancient Peruvian Bronzes from Machu Picchu; by C. H. MATHEWSON 1... .-.-25-2 0022 22 525 XLII.—A New Cephalopod from the Silurian of Pennsyl- vania; by Rura RampEer Moon 2222 2222 322 eee XLITI.— Activity of Mauna Loa, Hawaii, December—January, 1914-15; by T. A. JacGar,) JR, 222-22 2. ee XLIV.—Decomposition of Mineral Sulphides and Sulpho- Salts by Thionyl Chloride; by H. B. Nort and C. B. CONOVER .. 2.02 -- 232 2 ee ee 640 XLV.—On Simple and Mixed Alkyl Phosphates; by W. A. DRUSHEL . 2.02525 Figs eee 643 XLVI.—Two New Fresh-water Gastropods from the Meso- zoiec of Arizona; by W. LURoBINSONj@ c=: 2==2 === 649 XLVII.—The Ordovician Cynthiana Formation ; by A. M. MILLER’ ...- liu. See ee eee 651 SCIENTIFIC INTELLIGENCE. Chemistry and Physics—Annual Report of the International Committee on Atomic Weights: Analysis of the Non-Ferrous Metals, F, Ipporson and L. Aircnison: A Course in Quantitative Chemical Analysis, N. Kyieur, 658.— Household Chemistry for the Use of Students in Household Arts ; H. T. Vurri: Chemical German, F. C. Partiips: A Compend of Medical Chemistry, H. Lerrman, 609.—Sounds Resulting from Firing Modern Can- non and Rifles, M. Acnus, 660.—Fluorescence and Resonance of Sodium Vapor, R. J. Strutt: Elementary Lessons in Hlectricity and Magnetism, S. P. THompson, 661.—A Treatise on Light, R. A. Houston, 662. Geology—Strength of the Harth’s Crust, J. BARRELL: A Text-Book of Geol- ogy, L. V. Prrsson and C. ScHucuErtT, 663.—Grundlagen der physicalisch- chemischen Petrographie, H. KE. BornKE, 664.—Papers from the Geological Department of Glasgow University, 666. Miscellaneous Scientific Intelligence—National Academy of Sciences, 666.— Memoirs of the National Academy of Sciences: Craniometry of the Southern New England Indians, Vera M. Knicut, 667.—Publications of the British Museum of Natural History, 668.—Contributions from the Princeton Observatory: Publications of the Cincinnati Observatory, J. G. Porter: Nature and Science on the Pacific Coast, 669.—Mining World Index of Current Literature, G. E. Stsnmy: Les Prix Nobel en 1918: Lee- ward Islands of the Hawaiian Group, C. EiscHnerR, 670. Obituary—R. Meutpoua: BE. A. Mincuin: T. AtBrecat: W. Tassin, 670. InpEx to Volume XL, 671. VOL. XL. | JULY, 1915. Established by BENJAMIN SILLIMAN in 1818. THE AMERICAN JOURNAL OF SCIENCE. Epiror: EDWARD S. DANA. ASSOCIATE EDITORS Proressors GEORGE L. GOODALE, JOHN TROWBRIDGE, W. G. FARLOW ann WM. M. DAVIS, or Camsrivaz, 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 Irnaca, Proressor JOSEPH S. AMES, or Battrore, Mr. J. S. DILLER, or Wasuineron. ee mo HSE FOURTH spat le O- ba VOL. XL—[WHOLE NUMBRR, OLX}, “lenal Must Not Pas TULN 1A9ioe = NEW HAVEN, CONNECTICUT. £9 -1-5:. THE TUTTLE; MOREHOUSE & TAYLOR C©O., PRINTERS, 128 TEMPLE STREET. a a ener __ 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, a letters, or bank checks (preferably on New York banks). IMPORTANT TO COLLECTORS I take pleasure in announcing that the large collection of minerals recently received has been thoroughly gone over and properly labelled and is now ready for sale. This collection consists of over 1,000 specimens of excellent quality, some of them from old finds, and almost all- very well crystallized. Let me know what you desire and I shall be pleased to send you a selection on approval. Indian Relies In addition to the minerals, a large assortment of Indian relics were also included :— arrow heads, tomahawks, spears, celts, ceremonials, pipes, serapers, pestles, implements, obsidian knives, etc., ete. Also an assort- ment of Mexican relics of all descriptions, \ Are You Interested in Gems ? If so, you will find my stock now richer than ever before in beautiful examples suitable for both jewelry and specimens. Among the synthetic gems in my stock, all of which are of the finest quality, I have the follow- ing: Rubies, pink, white, blue and yellow Sapphires, and the latest dis- covery, the alexandrite. I have an unusually large stock of common and rare Semi-Precious and Precious Stones, both cut and in the rough. Iam able to supply any gem desired, in best quality and all sizes. New Minerals HODGKINSONITE :—I have been fortunate enough to secure the best specimens of this very rare mineral. It is from the celebrated Franklin Furnace Mines anf is a rare compound, the formula of which is Mn(Zn OH),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. — BETAFITE:—A member of a group of cubic minerals, niobo- tantalotitanites of uranium, etc., including also blomstrandite (of G. Lind- strom 1874) and samiresite (q. v.); they are closely allied to pyrochlore and hatchettolite, but differ from the former in containing titanium. Betafite is a hydrated niobate and titanate of uranium and occurs in pegmatite near Betafe, Madagascar. Named after the locality. Any of the above which may be desired for selection I shall be glad to send on approval to patrons and customers. Information jand prices of © individual specimens cheerfully furnished upon request. ALBERT H. PETEREIT — 81-83 Fulton St., New York City ~~ THE AMERICAN JOURNAL OF SCIENCE [FOURTH SERIES.] Arr. I.—Fuactors in Movements of the Strand Line and their Results in the Pleistocene and ‘Post-Pleistocene ;* by JosepH Barrett, New Haven, Connecticut, including a letter on Botanical Evidences by M. L. Fernald, Cambridge, Massachusetts. TABLE OF CONTENTS. TSR ROCIUCHION (Set oe See Ae SOS oS ae Soe ee ae ee eee ae eee ere 1 The interpretation of composite rhythms ---._------------------------- 3 Indications of osciliations given by subaqneous proiiles ---...-..----.-. 4 Pliocene and Pleistocene marine terraces___-.. ----------------------- 9 Post-glacial emergent cycle marginal to the glaciated areas ___.___.----- 13 Botanical evidences by M. L. Fernald.-.-..._------------_-------- 7 Possible effects of radial shrinkage. ____-.-___-----__------------------ 20 Conclusions Pa ae Ho te ae YIN eI SE aS a eiskj be Skee eaeed dk kok 21 Introduction. Tue problems of the nature of the Pleistocene and post- Pleistocene movements of the strand lines and their causes are important from a number of standpoints. They constitute one of those common fields in which stratigraphy and physical geology meet. The evidence shows a complex series of move- ments, indicating a complexity of causes. Hach locality may ive indications of the existence in the recent geologic past of both higher and lower stands of the sea, recorded by such unlike features as elevated seaplains and ‘cliffs, on the one hand, by drowned river valleys on the other. Minor rhythms of movement are superposed upon larger rhythms. What is the proper sequence, what the relative duration, and what the correlation of successive stages of emergence and submergence between different localities? As the last phase in this series of oscillations, what is the direction and amount in different regions of the last movement of the strand line ? * Written at the suggestion of Dr. T. Wayland Vaughan and presented with lantern slides at the meeting of the Geological Society of Washington, on March 24, 1915. Am. Journ. Sct.—Fourts Serizs, Vou. XL, No. 235.—Juzy, 1915. 1 2 Barrelli— Movements of the Strand Line But after the complexities of the record have been correctly deciphered there come the questions of interpretation. The making and unmaking of the great ice sheets must, by deduc- tion, have been the cause of very appreciable fluctuations in the volume of ocean water. To what extent does the record of movement coincide with this cause, or to what extent does the evidence show inductively that there are other equally important or more important causes? The weight of the ice sheets depressed the crust beneath them, but what was its effect on the zone of crust immediately beyond? Among other questions, if it is found that the Pleistocene and post- Pleistocene movements are to a considerable degree unrelated to glaciation, to what degree are they due on the one hand to isostatic movements, including possibly intermittent changes in the volume of ocean basins, movements merely toward crustal equilibrium ;. te what degree are they due to deep- seated radial shrinkage of the earth, generating rhythms of ‘tangential compression in the crust? These factors, together with the effects of changes in crustal density, the effects of erosion, of sedimentation, and possible changes in the volume of ocean water have operated through all past time, but their effects can be most minutely studied and correctly evaluated for the recent geologic past. When, in the future, comprehensive answers can be given to these questions, much will have been learned in regard to the nature of crustal movements, and, as the present is the key to the past, better detailed interpretations will be possible for the stratigraphic record of previous geologic periods. Within recent years new points of view have been devel- oped by a number of geologists in regard to present and recent movements of the sea level. Among them, D. W. Johnson has shown the necessity for a re- examination of the supposed evidences of present continuing submergence of the Atlantic shore. Daly has suggested that the rise of sea level with deglaciation may account for the appearance of subsidence of coral islands, without requiring any real movement of the erust. Vaughan has come to the view that the development of recent off-shore reef corals is closely related to a recent rise of sea level and that, in general, reef corals flourish broadly only during and immediately following submergent phases of movement. W. M. Davis has developed especially, through a critical re-examination, the erulende given by the topography of the Pacitic Islands. The writer has not entered upon this subject as a special problem for research, but in the pursuit of other investigations has several times come into contact with it. But coming unex- pectedly upon a subject from an angle is apt to give new sug- in the Pleistocene and Post-Pleistocene. 3 gestions and somewhat novel points of view. It is from this standpoint that the following consideration of the subject may be of value. These points of view must be followed further and tested, however, before their results can be correctly and finally evaluated. In that brief treatment which is necessary in order to compress a large subject into the space of a short article detailed demonstrations must be avoided. The article is planned to be suggestive rather than conclusive. The pur- pose is to outline the controlling factors, putting the emphasis upon those aspects which have presented themselves to the writer as somewhat novel. It is not the plan here to demon- strate fully any single thesis, nor to treat in proper proportion all of the composite factors. The Interpretation of Composite Khythms. The movements of the strand show a rhythmic nature. Smaller are superposed as undulations upon the larger oscilla- tions, as, on a smaller and far briefer scale, are superposed the rhythms of waves due to wind upon those produced by tidal forces. In the interpretation of these rhythms, as Gilbert has so clearly shown, attention has been focused upon the marks of previous inroads of the sea, not upon the limits of its retreats, for these are now concealed.* To extend the analogy used by Gilbert,—if at any time we stand upon the shore and look at the lines upon the sand which mark the arrest and turning of previous waves, we see that the highest is the oldest and the successive marks form a descending series in time and inten- sity. We might generalize from this that the motion of the water was subsiding and the sea would therefore soon be at rest ; but no sooner might this conclusion be announced than greater waves roll in, each obliterating the marks of the lesser waves preceding it, effacing also a part of the previous descend- ing series and beginning the record of a new descending series. If we visited the shore at those rare times when a combination of spring tide and heavy on-shore storm raised the water level highest, we would at the moment of culmination see no series but only the single line of the highest wave. A visit at another time would show highest and faintest, as lines of shells and wrack, the marks made long previous duri ing combined highest tide and storm. Below this would be the record, weeks or months younger, developed during the more moderate spring tides. Below this and still younger would be seen the marks of the waves made during the last high tide. Below this would be the series made by the combination of larger and smaller sets of wind-caused waves, the sequence recorded by *G. K. Gilbert, Continental Problems, Bull. Geol. Soc. Am., vol. iv, pp. 187-189, 1893. 4 Barrell—Movements of the Strand Line each descending series, destroyed and covered up while we watched by the following ascending series. The tide and storra might meanwhile be rising higher, and yet the writing of the past would appear as of a subsiding nature. The lesson to be drawn is that the record of a descending series measures only the descending sequence from the last maxima of the greater rhythms. It does not in itself tell the present trend of the oscillations. Since the end of the Oligocene the strand has, on the whole, been retreating, the continents rising higher, the climates grow- ing colder. There seems, in the Pleistocene, to have been a cul- mination of crustal and climatic oscillations, closed by a descending series. Men have taken hope that the ice age is past and have looked upon the Quaternary revolution as clos- ing. But the study of the rhythm of the waves robs us of that assurance. At several times in the Pleistocene that view, as based upon apparent subsidences of crustal and climatic move- ments, would have been far better justified by the evidence than it is at the present moment. The high latitudes, unlike their state through the most of geologic time, are still mantled with glaciers. The shore lines are now in a stage of earliest youth. We live in fact within the Age of Ice, within an age of crustal unrest and revolution; the geologic morrow may bring forth greater and more compelling changes than the geologic yesterday. With this understanding of the nature of composite rhythms, always seen in retrospect as ending with a descending series, we should study the record of the ocean waves made on “ Time’s great continental strand.” Indications of Oscillations given by Subaqueous Profiles. The first line of investigation which led the writer to an in- tersection with the problem of recent strand-line movements was in a study of present shore action as a basis for the development of ‘“ Some Distinctions Between Marine and Ter- restrial Conglomerates.”* The pressure of other work and the desirability of making quantitative measurements has _pre- vented the publication of more than an abstract of this article, which Jed up to the conclusion that “the truly terrestrial forces produce vastly more gravel, spread it far more widely, and provide more opportunities for deposition than do the forces of the littorai zone.” This conclusion was reached by a re-examination of the data used by Geikie, de Lapparent, and Penck for determining the relative rates of erosion. The result was to diminish still further the ratio of marine to ter- restrial erosion as given by those authors. It was estimated, by multiplying the length of shore line by the thickness * Abstract, Bull. Geol. Soc. Am., vol. xx, p. 620, 1908. in the Pleistocene and Post-Pleistocene. 5 eroded, that marine denudation for the whole earth at present amounts to probably between 02 and -10 cubic miles per year, whereas the rivers bring to the sea annually probably between 1°50 and 3:00 cubic miles of rock material. These wide limits must be used, since for the earth as a whole the data are as yet not on a quantitative basis. But to have any application to the geologic past the variation in rates of erosion of both marine and subaerial agencies, and the high or low values of the present rates, must be considered. Now an examination of present shores shows that they are characteristically young and the sea work is mostly erosion at certain parts of the shore, not upon the bottom. Beyond some miles off shore, deposition is in general taking place. In the oscillations of sea level the last movement therefore has, on the whole, been one of sub- mergence and this phase is favorable for a large proportion of gravel in the marine deposits now in evidence near the shores. Turning to the application to the present subject, it appears, by reversing the argument, that recent movements of the strand line may be elucidated by a method of study which analyzes the place and character of marine degradation and aggradation ; a study pursued in a somewhat similar fashion to that study of fluviatile degradation and aggradation which has thrown such great light on the succession of crustal movements. Both rivers and sea work with respect to a base level. Their first effort is to bring a profile to a graded slope, eroding in some places, depositing in others. For the one as for the other, the nature of the work shows the direction, amount, and relative duration of recent changes of level. Valuable studies on the principles controlling the character of the shore line have been made by Davis and Gulliver, but what is here emphasized is the study of the water bottom, both near shore and offshore, its sedimentary character andits form. The basis for study consists especially of the hydrographic charts issued for the use of mariners. The writer has used particularly those published by Great Britain and the United States. The form of the bottom may be advantageously studied, not only from the plan, but by means of the projected profile. For this a section plane is taken at right angles to the submarine con- tours, and, to smooth out the minor effects of currents and waves and to make up for the scantiness and possible errors of the data, all soundings within a certain width are projected upon it and located as dots. A smoothed curve is then drawn through the soundings. These profiles are illustrated in figures 1 and 2. In such studies distinction must be drawn between the sub- aqueous profiles of aggradation and degradation. Although in the course of unlimited time the two would tend to approach 6 Barrel— Movements of the Strand Line each other, yet the recurrent movements of the crust tend to eliminate this mark of old age, by creating new base levels of planation. The profiles for various grades of sediment will also be of unlike depth. Sand derived abundantly from active shore erosion may build out a subaqueous terrace at the same time that finer sediment, partly of river origin, may not be permitted to settle on the older and deeper platform beyond. After a profile is graded, then the advance of the sea inland or the retreat of the sea, provided the level is constant, requires erosion or deposition on all parts. This graded profile is Fie. 1, (2) 50 (00 150 Nautical miles 200 n i 20 renee SL ~ oa = 40 a : Ss Schr ke = rN = +——* 60 Fic. 1. Projected profile of the Argentine coast, South America. Section 50 miles wide, 25 miles on each side of a line, bearing S. 24° HB. through lat. S. 38° 45’, long. 59° 20’. Vertical scale 890 times the horizontal scale.* concave upward and may be ealled the profile of equilibrium. The depth of the profile varies greatly according to the power of the waves. Off the flat shores of quiet seas it ranges, at a distance of about ten miles, between ten and twenty fathoms. Vaughan, from his experience in the West Indies, reaches similar conclusions and considers that about ten fathoms is there the depth to which submarine erosion is vigorous. Off abrupt coasts exposed to on-shore heavy winds the profiles make it clear that it reaches to thirty or forty fathoms at the same distance. As observed off many continental shores these slopes at the present time may continue almost flat for twenty or thirty *Mr. Willis has noted in the discussion of this paper that the land surface of Argentina has been greatly affected by warping in late geologic time and that the possibility of warping should also be entertained in considering the significance of the subaqueous terraces. This is a valuable suggestion and no doubt has controlled the general form of the coast line. The development of these upper terraces at uniform levels over broad regions, however, at approximately 30 and 50 fathoms appears to indicate that their surfaces have been controlled by wave action working at these two levels at two different times later than the warping. in the Pleistocene and Post-Pleistocene. 7 miles from shore. Then comes a convexity and a descent, in places to great depth, in other places, as illustrated by the shelf seas of Argentina, to broad flats about twenty fathoms deeper. Daly has been independently following the very same line of investigation, and he states in a personal communication that he has reached the conclusion that effective wave action stops much above the one hundred fathom line, notwithstanding that this has become conventionally accepted as the limit. With this conclusion the writer is in accord. Vaughan also has called attention to precisely the feature of Fie. 2. 0 MILES 10 20 ie West coast Madagascar Lat S.18°53' to 19°03’ O FATHOMS Southeast coast Madagascar BetweenC.Ranovalona and Galleon Bay Lat. 5. 25°05" Fic. 2. Subaqueous projected profiles, coasts of Madagascar. Most of the coast shows graded, subaqueous profiles in adjustment with the present level of the sea, the wave action being relatively weak on the western side, strongest on the southeast. The figure shows how the different exposures of a coast must be considered together in order to eliminate the varying effect of the waves according to their intensity. This permits a conclusion regarding the relation of the profile to present and past water levels. Vertical scale 100 times horizontal. the ocean profiles here described, noting the steep descent from about the 30-fathom to the 50-fathom curve, with gentler slopes above and below, as seen especially on the North American and Australian platforms.* E. C. Andrews, a *Sketch of the Geologic History of the Florida Coral Reef Tract. and Comparison with Other Coral Reef Areas, Jour. Wash. Acad. Sci., p. 33, 1914. The Platforms of Barrier Coral Reefs. Abstract, Bull. Am. Geog. Soc., vol. xlvi, p. 427, 1914. 8 Barrell— Movements of the Strand Line pioneer in thought in many fields of Australasian geology, showed in 1902 that the Great Barrier Reef of Australia stands on a platform which marks a former lower sea level. As Vaughan has stated, “‘ The problem of the depth of barrier platforms is a world- wide one, for it is only an aspect of the general problem of the history of continental shelves.”* The study of the form of the subaqueous profiles shows, then, that they are not generally in equilibrium with wave action throughout their length. The character of the departure indicates a recent submergence. Where the shore is abrupt, as off the coast of Maine, the profile is overdeepened and the power of the waves is concentrated upon the exposed shore. Where the coast lands are very flat, as on the coastal plain, the submergence has inundated a strip which was previously land. The waves spend most of their energy on the flat bottom, miles off shore ; they drag and throw up bars, partially recover their form and advance farther inshore to where barrier beaches are thrown up and cut off farther advance. The overflat bottom profile thus becomes undulatory by theshifting of material which is finally swept in part to the beach, in part to deeper water, and a normal profile becomes established. Thus it is seen that according to the form of the land submerged the profile may be overdeepened or overshallowed and the wave work becomes opposite in trend. But the cause,—submergence, may be equally clear in each. Vaughan has developed additional criteria from the study of the shore and offshore physiography of the West Indies. He has pointed out that solution wells in the limestone show former emergence. These were made by fresh waters, and in the Bahamas are now submerged to a depth reaching 33 fathoms. There was also a still-stand of the land with the shore about ten fathoms below present level, as shown by the levels of the submerged terraces on the outside of the barrier reefs. at from 8 to about 13 fathoms, with solution wells inside the barrier showing a depth from 6 to 7-5 fathoms. The filled channels of Havana harbor further substantiate the conclusion. ’ Here there is a concurrence of evidence. From many regions, however, the other lines of evidence will be absent. There the conclusion must rest upon the testimony of the sea floor - alone. When understood it is as convincing, however, as are any of the more familiar lines of evidence. As an implication from the detailed evidence of the sub- aqueous profiles, it appears that the outer flats of the con- tinental shelves still preserve to a greater or less degree the form determined by a sea level about twenty fathoms below * Sketch of the Geologic History of the Florida Coral Reef Tract, p. 33, 1914. in the Pleistocene and Post-Pleistocene. 9 that existing at present. Still farther out there are probably terraces cut still lower. But such older steps must be com- monly mantled over by the terraces of later construction, their forms would tend to become smoothed out and they are not developed by sufficient soundings to permit as yet of a detailed study. In regions where rivers do not bring much sediment from the land, however, such fossil terraces may be best pre- served. For the study of the negative phases of oceanic movements,—that side of the rhythms of which, as previously pointed out, so little is known,—they should be of high importance and be best developed where the evidence given by drowned river valleys is wanting. Pliocene and Pleistocene Marine Terraces. The second line of investigation which has led the writer toward the problem of recent movements was in connection with areal geologic work in southern New England. An examination of the topography indicated that there were in this region a series of descending baselevels, more than could be fitted to the so-called Cretaceous, early Tertiary, and late Tertiary baselevels. The necessity of determining the number, the slope, and the age of these, as a key to the post-Jurassic history of the region led to a development of methods the results of which have been published in abstract.* Each baselevel would be recorded in the interior by surfaces of subaerial denudation, on the seaward side by surfaces of marine plana- tion. The latter were cut as benches across the harder rocks and are consequently better preserved than the former. They were most strongly developed at the maximum elevation of the sea during each oscillation. A method of projected profiles restores these ancient levels and hides the later dissection earried on by subaerial agencies. The length of the several profiles of marine planation is related to the duration of the baselevels; their difference in elevation gives, on the other hand, the amount of elevation of the land as the result of each cycle of motion. It is thought that the Goshen level, attain- ing an elevation of 1380 feet in northwestern Connecticut, dates from probably the earlier Pliocene. The terraces below, facing the sea, descend by steps which are a little over 200 feet apart in elevation and many milesbroad. The series resembles a flight of stairs, but one in which the rises and treads are both gently sloping and the whole so dismantled by subaerial erosion that what the eye sees in viewing the landscape is almost wholly the dissection due to later cycles of destruction. The * Piedmont Terraces of the Northern Appalachians and their Mode of Origin ; Post-Jurassic History of the Northern Appalachians, Bull. Geol. Soc, Am., vol. xxiv, pp. 688-696, 1913. 10 Barrell—-Movements of the Strand Line shore lines of the terraces give the maximum elevations, the five Pliocene terraces reaching approximately to 1380, 1140, 920, 730 and 520 feet respectively. The Pleistocene terraces below were, in comparison, very imperfectly developed, but planation can be traced at stages whose mean elevations on their inner margins are 360, 220, and 100 feet. That at 220 feet is best developed, but none show the clear character of the Pleistocene terraces farther south on the coastal plain, since the Connecticut terraces were developed on formations of resistant rock. Furthermore, all are older than the last glacia- tion and have been obscured by that event. Their narrow breadth and originally imperfect development in comparison with the older and higher terraces appears to be a measure of an acceleration of the diastrophic rhythm in Pleistocene time. This crustal unrest is continued to the present; the present is diastrophically as well as climatically a part of the Pleistocene. The broader Tertiary rhvthm shown by the wide terraces facing the sea must be complicated in the Pleistocene by the special effects of the weight of the ice sheets, the changes in volume of ocean water related to glaciation, and the percep- tion of the briefer minor diastrophic movements whose records are lost to sight among the larger movements of the past. These complications do not hide the facts, however, which lead to the conclusion that an abnormal crustal unrest beginning in the Pliocene has marked especially the entire Pleistocene period. The record which has been described shows the rhythmic oscillation during the upwarping and emergence of the Atlantic shore of the United States between latitudes 38 and 43 degrees. The record is different on the Pacific shores. There it is doubtless more largely related to the orogenic forces existent in latest times in the Pacific monntain system. On the eastern shore the deposits older than the Pliocene show a seaward tilt, greater for each older formation. This is a mark of progressive crustal warping. For the Pliocene and Pleisto- cene levels this warping, however, diminishes progressively. These later changes are more largely broad and parallel move- ments of level. They represent apparently more largely movements of the sea itself. To what extent, however, the net result is due to local movements of the crust cannot be determined until the series of strand lines has been studied on many shores and their resemblances and differences compared. These high-level terraces represent the height of successive submergences. Let attention be turned next to the other side of the rhythm, that of the record of emergences, as shown by river valleys which now are drowned. Reaching far back into the Pliocene each shore line is found to be marked by large embayments developed upon regions of softer rocks. The in the Pleistocene and Post-Pleistocene. 11 shore lines represent therefore the rhythmic rise of the sea upon a land surface previously dissected by subaerial erosion. The baseleveling on the softer rocks near the larger streams each time during the emergent phase had reached to an advanced stage, but upon the broad formations of harder rocks it had not progressed beyond the stage of youth. The sub- mergent phase, however, lasted long enough for the sea to produce notable planation against the outstanding harder masses. Subaerial baseleveling at these stages seems also to have made notable progress, marking times of diastrophic rest. The submergence was in each cycle great enough to permit concentration of marine erosion against the headlands, the waves not spending their force upon the bottom. The motion of the sea-level was therefore cyclic, but not smoothly undula- tory. If time be laid off as a horizontal axis, the curve of sca- level was nearer to the form of the common cyloid than to the sinusoid. The variety of the cyloidal curve suggested ap- proaches that described by a point on the circumference of a rolling wheel. This curve is characterized by a nearly flat top and a cuspate bottom. Conceive the point to be somewhat within the wheel and the pointed bottom of the curve would be somewhat broadened out. Conceive the wheel further- more to be rolling down an irregular grade. Such a more complicated curve as would now be generated by the motion of the point may be nearer to the actual record of the sea-level movements against the Atlantic shore. It does not quite rep- resent the discontinuous nature of the movements, but it does make graphic the conclusion that the emergent phases of the oscillations are relatively rapid and brief, the submergent phase prolonged, and marked by a slowly rising level of the sea. The retreats of the sea are thought to have been more rapid than the advances. In Maryland broad and gently sloping terraces of Pleisto- cene age face the sea, bounded on the land side by low escarp- ments. They mark submergent phases of rhythmic crustal movements and show planation at levels of 220, 100 and 40 feet respectively, in descending order of age. The gentle sea- ward slopes are more or less in adjustment with the slope ot the subaqueous profile of wave-worked bottoms. Within the valleys the deposits are regarded as fluviatile. Facing the sea they are doubtless marine, though the porous nature of the thin mantle of deposits has not permitted the preservation of marine fossils. The terrestrial fossils fix the age as Pleisto- cene. The oldest and highest deposits are known as the Sun- derland formation, and are apparently of early Pleistocene age. The youngest, known as the Talbot, are judged to be late Pleistocene. The materials of all three, Sunderland, Wicomico, 12 Barrell— Movements of the Strand Line and Talbot, carry a large proportion of ice-borne bowlders, too large to be transported by the moderate development of river- ice now carried by the streams. They indicate thicker ice and a much colder period than at present.* These stages corre- spond in general to what in New Jersey, Salisbury has named the Bridgeton, Pensanken, and Cape May formations. The oldest glacial deposits of New Jersey, represented by the patches of extra morainic drift, over the Triassic area are preserved on flat hill tops which range from 200 to 220 feet or more, in elevation. Some attenuation of the drift, ascribed to creepage, is observed below these levels, but it is inferred that the drift was laid down when the Triassic Plain was a gen- tly undulating surface, corresponding with that whose dissected remnants are now 225 to 275 feet above the sea. The larger streams now flow at levels more than 100 feet below.t The combination of evidence from Maryland and New Jersey sug- gests strongly, though it lacks perhaps actual demonstration, that certain early Pleistocene stages of cold climate and gla- ciation occurred when the ocean level stood at the higher parts of the phases of cyclic oscillation. The statement fails of com- plete demonstration because the glacial deposits of New Jer- sey may be somewhat different in age from the Sunderland deposits of Maryland. The development of the plains at this level by subaerial erosion during crustal rest had been accom- plished, at least in large part, when the ice sheets ended consid- erably north of their present limits, since the writer has iden- tified baselevels of erosion at these levels in Connecticut. But the glaciation tock place either before these baseleveled sur- faces of soft Triassic rock had been uplifted at all or had been uplifted long enough to become dissected. The important point, however, is that the development of the cold climates, culminating in glaciation, does not appear to have required a low level of the sea, or high elevation of the land of this region. Such abstraction of ocean water as took place at these times was, therefore. more or less compensated for by occurring at the long submergent stages of the diastrophic cycles. Nothing is said here as to the attitude of the land in the centers of glacial accumulation. It is true that the central regions of glaciation are now in a depressed condition, as shown by the drowned topography. This may be the result of a lack of complete isostatic recovery after the removal of the ice load. This permanent depression is a factor, however, which would not have operated in Maryland. In conelu- sion, therefore, the evidence of the region marginal to gla- *G. B. Shattuck, The Pliocene and Pleistocene Deposits of Maryland ; Maryland Geological Survey, p. 85,.1906. +R. D. Salisbury, The Glacial Geology of New Jersey, Final Report of the State Geologist, vol. v, pp. 755-760, 1902. in the Pleistocene and Post-Pleistocene. 1B ciation leans toward the view that the amount of water abstracted for the formation of the ice sheets was not a major factor in the control of Pleistocene sea levels. From the estimates which have been given by Woodward, Penck, and Daly it would seem neccessarily to have been a true and important factor, but the evidence of the Pleistocene oscilla- tions of the Atlantic shore suggests that the diastrophic rhythm continuing with accelerated movement from the Pliocene con- stituted a more controlling factor. Post-glacial Emergent Cycle Marginal to the Glaciated Areas. The third line of investigation which led to an intersection with the problem of Pleistocene and post-Pleistocene crust movements is that connected with the Strength of the Crust, a subject now in progress of publication in the Journal of Geology. If the hypothesis set forth there is valid,—that a thick and strong lithosphere rests upon a thick zone of com- parative weakness, an asthenosphere ; then the weight of a continental ice sheet should tend to depress the crust into this weak zone. ‘The crust up to a certain limit would yield as an elastic plate, the asthenosphere would reach its elastic limit at a far earlier stage and from that point yield by flowage, a flowage, it is thought, which is akin to the recrystallization which explains glacial flow. But as this subcrustal zone is not a fluid, it cannot transmit the excess pressures to unlimited distances. Broad and low pressure ridges would tend to be raised therefore beyond the limits of the ice sheets. A rail- road embankment sinking into a marsh and elevating slightly the adjacent portions of the marsh offers an instructive analogy. Upon the removal of all of the embankment, including that which had sunk below the marsh level, a re-elevation of the central tract toward the original level would occur. The lateral pressure ridges might rise at first with the central part and then subside. This double motion would be favored if the deeper levels of the marsh were the more nearly fluid and if the removal of the load was very rapid, exceeding the rate of readjustment for the upper levels. Does the expectation raised by this hypothesis correspond with the known depression of the crust under the Pleistocene continental ice sheets and the recovery from that depression which followed the retreat and disappearance of the ice? The question serves in a measure as a test of the validity of the initial hypothesis regarding the distribution of strength in the outerearth. As giving such a test the problem was investigated, though, as it is somewhat apart from the principal subject, it i$ not intended to publish the results as a part of the series on the Strength of the Crust, but rather later as a separate article. 14 Barreltl— Movements of the Strand Line The results of that investigation in so far as they bear upon the post-glacial crustal movements may here be touched upon, but within a short article the detailed arguments cannot be given. These results cannot therefore logically be used here as proofs regarding the nature of recent movements, but may serve as suggestions: the deductive conclusions from a hypoth- esis must be subordinated to the direct inferences from established facts, but the deductions may serve, nevertheless, to emphasize, explain, and coordinate such inferences. This theory of the distribution of crustal strength was applied to the data regarding water levels during the retreat of the ice, and their correlations as given by Woodworth for the Champlain and Hudson valleys.* The accuracy and clearness of presentation of this difficult snbject by Woodworth when combined with this theory permitted the determination of the sea level in latitude 40° north for each stage of the ice retreat. This is at a distance where the weight of the ice should have had no direct influence and “represents the movements in a belt beyond the terminal moraines. As previously noted, however, the results of such an investigation are to be regarded as suggestive rather than conclusive. When the ice had retreated about to lat. 42° 45’, some 150 miles back from the terminal moraine, the sea level near the ice front stood 480 feet above the present level. The projec- tion of the curve of flexure points to a sea level at latitude 40° 40 feet lower than at present. When the ice had retreated to lat. 48° 80’ the sea level there stood at +560 feet; at latitude 40° the appropriate curve suggests a sea level at —90 feet. When the Champlain marine water body existed, its slope and the appropriate curvature point to a water level of —220 feet south of the terminal moraine. These figures have of course no exact value, but if they are in the direction of the truth they mean that south of the hmit of glaciation the coastal plain of the Atlantic shore was higher than at present and rising, the sea sinking, during the retreat of the ice, and that in later post-glacial time a marked submergence of this region has taken place to the present level. This emergence and submergence are of an opposite character to the effect of the addition of water to the ocean by deglaciation and are distinct from the marked isostatic rise of the glaciated region toward the present level, a rise which accompanied the removal of the load of ice. Are there independent lines of evidence pointing in this direction, evidence strong enough to make a real case and not depending upon the grist of a mathematical mill, grinding out * Ancient Water Levels of the Champlain and Hudson Valleys, Bull. 84, N. Y.,State Education Department, 1905. in the Pleistocene and Post- Pleistocene. 15 the results of certain hypotheses? Such independent evidence will now be considered. Woodworth has called attention to evidence that at some time since the retreat of the ice the land has stood higher than at present. South of Albany many of the tributaries appear to be now filling up excavations below sea level which at some previous time they had ent into post-glacial deposits.* In the gorge of the lower Hudson River silt extends to variable depths, 280 to 320 feet at New York City. At the Storm King Crossing of the Hudson River, borings for the new aqueduct encountered the highest stratum of bowlder beds at a little over 200 feet in depth. These appear to be deposits not excavated by river action since the ice retreated. The evidence shows, therefore, that a stage of emergence has occurred since the retreat of the ice but that this emergence above present level was not over 200 feet asa maximum. It may have been less. Evidence that a minor cycle of emergence and submergence has occurred since the retreat of the ice is also found along the shores of southern New England and the south. The subnier- gence to present level appears not to date back more than a few thousand years at the most, as is shown by the extreme youth of the shore line. Headlands have barely begun to be cut into, embayments have only begun to be filled. Sediment carried in by the tides forms salt marshes which are still flooded at spring tides. Deltas at drowned river mouths have barely begun to form. It happens, however, that near the margins of glaciation the latest submergence has brought the present sea level only slightly above that which existed at the beginning of the retreat of the ice. The physical evidence for a rapid cycle of post-glacial emergence of considerable amount, involving a tract beyond the limits of glaciation, is therefore drowned and the amount and importance of the emergence is perhaps best given by the distribution of plants along the Atlantic shore from New Jersey to Newfoundland. Britton, Hollick, and others have discussed the significance of plant distribution, especially of southern types of the Coastal Plain, but the most important contribution on this subject has come from Fernald in his analysis of the flora of Newfoundland.t He points out that 118 species of plants belonging to the Pine Barren and Coastal Plain floras of New Jersey and the south are known from remote outlying stations along the coastal strip of New England and the Maritime Provinces. Most striking, how- ever, is the evidence from Newfoundland. He shows that 60 '*J. B. Woodworth, loc. cit., pp. 229-234, 1905. +M. L. Fernald, A Botanical Expedition to Newfoundland and Southern Labrador, Rhodora, vol. xiii, pp. 109-162, 1911. 16 Barrell—Movements of the Strand Line per cent of the plants are boreal types. Only a few plants are present out of the hundreds which constitute the flora typical of Canada in the same latitude. On the other hand 35 per cent of the plants of Newfoundland are southwestern types and 7-7 per cent of these are plants characteristic of the coastal plain of New Jersey and the south. Fernald shows, by means of the evidence of the flora, the ineffectiveness of winds and currents to transport across the Gulf of Saint Lawrence to Newfoundland the typical Canadian plants, though the boreal types may have entered across the comparatively narrow straits which separate the island from Labrador. Fernald argues that it would be still more difficult for winds and currents to act as agents for the migration of the Coastal Plain flora. Not only are the distances vastly greater, but the winds and currents move actually in the wrong direction to carry plants from the southwest. Migrating birds can hardly be invoked. The nearest land to the southwest of Newfoundland is Cape Breton Island at a dis- tance of 70 miles and many of the species of plants which are concerned are rare and local at all points northeast of New Jersey, a distance of over 850 miles. Studies of migrating birds, especially in connection with the flora of the Farée Islands, have shown that they reach the end of their flights with intestines empty. A remaining hypothesis is, that after the last retreat of the ice, the Coastal Plain stood higher and offered northeastward a line of passage broken only at a few places by channels of deep water. A climate warmer than the present would favor this migration of the flora of the southern Coastal Plain. An elevation of at least 100 or better 150 to 200 feet, enduring for some thousands of years, would apparently be needed to give the necessary conditions for plant migration. An elevation of 200 feet would still involve, however, the crossing of 70 miles of water in Cabot Strait. An elevation of the Coastal Plain seems the readiest way of explaining the existence of such isolated stations of southern plants as that of Magnolia virginica at Cape Ann, Massachusetts, but migration from the southwest is difficult to apply to Newfoundland. For this it may be necessary to invoke another aspect of the same hypothesis of elevation. The Great Banks extend 200 miles southeast of Cape Race, to latitude 48 degrees. The Gulf Stream flows past not far distant on the south and must serve to ameliorate the climate. Nearly to the margin of the Banks the depths range no deeper than trom 80 to 40 fathoms. If, during and for a time follow- ing glaciation, this region was above sea level and unglaciated, it may have served as a haven of refuge for plants of the Coastal Plain, requiring a temperate climate. in the Pleistocene and Post- Pleistocene. ive In his first publication Professor Fernald had given much weight to the suggestions derived from Penck and Daly,—that the ice of the continental glaciers abstracted so much water from the ocean that an emergence would be expected of part of the continental shelves. But against this cause as an agent operative in this instance it is to be noted that in direct propor- tion as the climate became warmer and the ice withdrew such an emergence would have disappeared. The character of the flora suevests, however, that the migration to the present isolated localities must have taken place during a period of climate even warmer than the present and at a time after the ice sheets had given up their water. This aspect tends to rule out the availability of the hypothesis that emergence was con- trolled only by the level of the ocean water as controlled in turn by glaciation. In response to questions by the writer regarding the indications of the plants as to the climatic con- ditions which would favor their migration and successful com- petition with other floras, Professor Fernald, under date of January 15, 1915, wrote a letter, which because of its great value on the problem of recent crustal movements is quoted here nearly entire. Botanical Hvidences.—A. Preliminary Statement of Results of Studies on the Northeastward Distribution of the Coastal Plain Flora, by M. L. Fernald : The question you raise, of the probable period and climatic conditions of the migration of southern coastal plain plants to Newfoundland, is one on which there is now accumulating a vast amount of evidence. Since I published a preliminary statement of the case I have had parts of two seasons in Newfoundland where I secured much more similar material, a season divided between Prince Edward Island and the Magdalen Islands, which have a flora much more southern than boreal, with many New Jersey Pinebarren species, a summer on Cape Cod and Black Island, where there are still larger proportions of such cases. One of my research students, Mr. Harold St. John, has spent a season on Sable Island, getting evidence in the same line, and another season on the Magdalen Islands, and he is now working on the results as his doctor’s thesis. Another graduate student, Mr. Sidney F. Blake, has spent a season on Northumberland Strait and the sands of northeastern New Brunswick, where he secured further material upon which his thesis will be based: Briefly, we have in southern Newfoundland (south of the North Peninsula) a large number of species of southern origin, some even of tropical affinity. Further, on the shores of Dawson’s warm Acadian Bay.(see Dawson, Canadian Nat., ser. 2, vil, p. 277, 1875), including the region from ‘Cape Breton to the south side of Baie des Chaleurs, i.e., eastern Nova Scotia, eastern New Am. Jour. Sci.—Fourtu Series, Vou. XL, No. 235.—Juty, 1915. 2 18 Barrell— Movements of the Strand Line Brunswick, Prince Edward Island, and the Magdalen Islands, we find many such species, some identical with, others different from those reaching Newfoundland, and on Sable Island there are some similar cases, some of the species from the south getting to Sable Island but not to the other areas. Now it is inconceivable that these southern types should have extended to these remote out- lying areas under conditions much colder than at present, and judging from their present very restricted occurrence in these northern regions it looks as if they must have reached them dur- ing a period somewhat warmer than at present. They either migrated northward on the continental shelf prior to the Wis- consin glaciation and persisted outside the subsequently glaciated area, finally taking possession of their present isolated habitats on the receding of the ice or, as seems to me perhaps equally prob- able, the continental shelf, including the present southwestern half of the Gulf of St. Lawrence, must have been considerably elevated after the Wisconsin, long enough for the migration— including some mammals, and even fresh water and land snails. The question is complicated by a mass of evidence not yet published upon: briefly that in the regions above enumerated, there is also a remarkable proportion of the flora, quite unknown elsewhere in North America, which is strictly identical with the flora of the Atlantic area of Europe, even including the Azores. We have startling cases (several scores of them) of species strictly indigenous in Newfoundland and known otherwise only in the region from the North Sea to Portugal, or to western France, and in some cases also on the Azores or even Madeira. Similar and very striking cases have come to light on Sable Island, the Magdalens, Prince Edward Island and eastern New Bruns- wick, indicating pretty clearly a large flora (and snail fauna, too) which must have crossed on the mid-Tertiary Jand and been stranded at the present margins of the Atlantic. Now these cases imply some tract along the coast, especially of Acadia, and Newfoundland, which held this flora and fauna continuously through the Pleistocene. : 5 : : ; é . ; Another point: in a few cases plants clearly indigenous on these isolated areas, Newfoundland, Sable Island, Prince Edward Island, etc., are species identical with, or representative of, the Australasian flora, and seem to be harking back to the Eocene. So you see, that it is not very easy to give an offhand brief answer to your inquiry about the botanical evidence. My students, Messrs. St. John and Blake, and I are wading aboutin a flood of such evidence, trying to work out the exact identities and affinities and come to safe anchorage, but it is a vast problem and every species has to be studied intensively, often requiring a com- plete monographic study of a world-wide group before we are sure of our ground. It will be a long time before we are able to publish our detailed results. If you wish to use the above as a preliminary statement you may do so. an the Pleistocene and Post-Pleistocene. - 19 Evidence somewhat similar to that which has been so strik- ingly presented by Fernald is found on the shores of north- western Europe, of land emergence above the present level involving regions beyond the limits of glaciation and occurring in a warm stage following the retreat of the ice. This is presented in comprehensive form by W. B. Wright of the Geological Survey of Ireland in his recent book on “The Quaternary Ice Age.” Before the glacially depressed region had recovered in Scandinavia all of its present elevation, there was a rapid uplift to the south which cut off the connection of the Yoldia (enlarged Baltic) Sea with the Atlantic and pro- duced the Ancylus lake. The climate was notably warmer than at present, as recorded by the advance of the hazel (Corylus avellana) and other plants to regions well beyond their present northern limits. The same climatic optimum has been proved in Finland, for at a certain floral horizon (the upper portions of the Birch—Fir zone) certain warmth-requiring plants made their appearance in localities far north of their present northerly limits.* In southern England forest beds which are correlated with this stage are found submer ged now to a depth of 55 to 60 feet, indicating a post-glacial “uplift of land to a height well above the pr esent sea level. Wright considers the breadth of this movement as indicative of a general lowering of sea level and seeks its cause in a recrudescence of some one of the ice sheets, possibly that of the Antarctic continent.t But the general warmth of climate ex- isting at that time does not favor such a view. So far as the evidence goes two other hypotheses are preferable. One of these hypotheses is that the post-glacial emergence and following submergence represent a diastrophic emergent eycle entirely unrelated to glaciation, one of the movements in the complex rhythms which have been traced in southern New England through Tertiary and Quaternary times. If so, it should be in part a sea-level movement, and traceable widely over the earth. The other hypothesis involves less widespread effects. It is that which was discussed in the pages leading up to the botan- ical evidence. This hypothesis is, that the weight of the ice sheets caused crustal depression directly below the load, but moderate elevation in a wide zone beyond the load. Upon the removal of the ice it appears the first isostatic upwarping carried up higher this marginal upwarped zone with it. Being already an upswollen tract the broader regional movement carried it up to a level where it became unstable and a slow settling back occurred as an after effect, coincident with the *W. B. Wright, loc. cit., p. 437, 1914. + Loe, cit., p. 417, 1914. 20 Barrel— Movements of the Strand Line last stages of upwarping over the centers of glacial load. The actual evidence at hand does not decide between these hypo- theses. The association with the close of glaciation appears to favor a genetic connection with deglaciation, but on the other hand it remains to be demonstrated why the extra-marginal zone should rise together with the region directly glaciated, or that the cycle was restricted to such an extra-marginal zone. Possible effects of radial shrinkage. The fourth line of intersection with the problem of recent movements was the result of a query raised by Professor Schuchert in 1909 as to the quantitative effect of earth shrink- age in producing an increased speed of diurnal rotation, and a greater oblateness of the earth’s figure; tending to raise the level of ocean waters toward the equator, and to depress the ocean level near the poles. Calculations made by the writer for Schuchert in regard to the possible magnitude of this effect suggested that it might be of very real influence.* A radial shrinkage of a mile, it was estimated, would increase the differ- ence between equatorial and polar radius by about 90 to 100 feet. If this difference were taken up wholly by adjustment of ocean level and not by internal earth adjustment, a supposi- tion which can be true to only a limited degree, then a shrink- age of a mile in radius would raise the water level at the equator about 35 feet, while at the poles it would sink about 60 feet. In latitude 85 degrees the water level would suffer no change. The rigidity of the earth in an east-west direction is also found to be different from that in a north-south direction and may have some influence on mode of crustal yielding. But to be of influence in a result, a factor must not only be real; it must, furthermore, be of quantitative importance. Such oreat folding and thrusting movements as have occurred in the later Ter tiary suggest that earth shrinkage is an important factor. Schuchert has called attention to the greater persist- ence of equatorial epeiric seas and the readvance of waters more frequently from lower toward higher latitudes. The southward bends of certain rivers such as the Delaware and Susquehanna along the strike of soft formations has received various explanations. In so far as land emergence is produced by sea-bottom subsidence or is coincident with move- ments of crustal shortening, the retreating sea should also slightly change its slope. It is possible that such a slight change in slope of sea level at the time the lower river courses were established on a newly emerged coastal plain may be at * Charles Schuchert, Paleogeography of North America, Bull. Geol. Soc. Am., vol. xx, pp. 505-508, 1910. an the Pleistocene and Post-Pleistocene. 21 least a partial cause of the southward deflections of rivers which were established along the Atlantic margin in late Ter- tiary time. Such a possibility should at least be added to the eroup of multiple hypotheses which have been brought to bear upon the problem. In the emergence of a sea bottom, the more important control of the slopes is due to the slope of the previous surface of sedimentation and the later local warpings of the crust, but a component tending to make for deflections toward the equator may be due to more broadly acting forces, and at certain times and places it may rise to the role.of an important factor. Conclusions. To review in conclusion the factors which appear to enter to a greater or less degree into Pleistocene and recent crust movements,—First, we must recognize the climatic factors connected with Pleistocene glaciation. These consist in the abstraction of the vast quantities of ocean water which accom- panied the development of the ice sheets; the direct depression under the load of ice; a possible compensatory elevation in a somewhat broad zone beyond; a deferred, intermittent, and possibly oscillatory readjustment upon the removal of the bur- den of ice. But these factors related to glaciation must not be used alone. There are other factors unrelated to climatic causes which may be of equal importance. The second group of causes includes the movements which tend to maintain the equilibrium of large sections of the crust, affect the whole ocean level and locally warp the lands, but do not involve earth shrinkage. It is thought that they find their cause in a lack of equilibrium of pressures upon that zone of the earth’s body which lies just below the level of isostatic compensation. They are then the isostatic factors. The general rise of lands which has marked the later geologic times may most probably be placed in this category, due either to changes of external load, or to changes in crustal density. This would include a possible enlargement or deepening of portions of the ocean basins. As continental rejuvenation is the chief effect, these adjustments toward the maintenance of isostasy may be called epeirogenic factors. The third group of causes are thought to be found in great compressive movements in the lithosphere, which seem to be due in turn to a shrinkage of the centrosphere. Locally they may work against isostasy, more broadly they may start iso- static movements. As their ultimate expression is in moun- tain building along lines of weakness they may be classed as orogenic factors. 22 Barrell— Movements of the Strand Line. Fourth are the planation factors, working in opposition to isostasy. They are the forces which tend to erode the lands, fill the seas with sediment, and raise their surface level. Added to this is whatever slow and secular gain the ocean makes in the volume of its water. The combinations of isostatic and compressive forces have together operated through geologic time. The causes are com- plex and the result is seen in the composite diastrophic rhythms which are expressed in all ages but which can be studied best and more definitely in the record of the latest times. To sift apart the factors, comprehensive investigations must be prosecuted over different parts of the earth. The sequence and amount of oscillations in the tropics must be linked to those of higher latitudes. Changes of sea level must be sepa- rated from local changes in the level of the crust. Multiple working hypotheses must be tested by the application of new and significant facts. When, as Gilbert says in the introduction to his monograph on Lake Bonneville, ‘‘the work of the geologist is finished and his final comprehensive report written, the longest and most important chapter will be upon the latest and shortest of the geologic periods”. W. G. Mixter— Polymerization of Oxides. 23 Arr. I].—The Heat of Formation and Polymerization of some Oxides and Determination of the Heat of Combination of Water by Fusion with Sodium Peroxide; by W. G. MrxtrEr. [Contribution from the Sheffield Chemical Laboratory of Yale University. | Tue heat of formation of an oxide may be determined directly by combustion, or it may be derived from the heat of the reac- tion of a metal and its oxide with water, or acids, or sodium peroxide. The thermal result is usually for an oxide formed at a high temperature and above that at which a given hydrox- ide glows. In the article* on chromium and aluminum sesqui- oxides it was shown that the heat of polymerization of the former is not small. The present paper contains results ob- tained with other oxides and also determinations of the heat effect of combined water in fusions with sodium peroxide. The heat of combination of some oxides with water is found by direct treatment with water, or the same result may be obtained bv dissolving the oxide and the corresponding hydrox- ide in an acid. When an hydroxide or oxide becomes anhy- drous at low temperatures the method of fusion with sodium peroxide applies as shown later. ‘The heat effect of the com- bined water is the difference between the heat of fusions of mixtures of sodium peroxide and the hydrous and anhydrous compound, if the latter is not a polymer. This value subtracted from the heat of the reaction of water and sodium oxide is the heat of combination of the water with the anhydrous com- pound. As is well known, some oxides retain water at temper- atures below that at which they glow and polymerize. The heat of combination of such water can only be estimated approximately. The heat of the reaction between water and sodium peroxide is derived from the heats of formation of the peroxide, water and sodium hydroxide, thus (2Na + 20) + (2H + O) = 2(Na + O + H) + O = 15'5 Cal. 119°8 68°4 (Thomsen) 203°7 (Thomsen) and 19-4 Cal.t are required to separate the oxygen. Hence * This Journal, xxxix, 295, 1915. + De Forerand found in 1898 (C. R., exvii, 514), 2Na + 20 =119°8 and Na.O + O =19°4. Recently (C.R., clviii, 991, 1914) he obtained 119°7 and 19°03 Cal. respectively. I have used the value-19°4 Cal. in previous work and the results have accorded well with those obtained by combustion in oxygen or other methods. This value is used in calculating the heat effect of oxygen liberated in fusions with sodium peroxide. For each cubic centi- meter of oxygen at 0° and 700™™ pressure 1-734 small calories are added to the observed heat of an experiment. 24. W. G. Mixter—Polymerization of Oxides. 15°5 + 19-4 = 34-9 for the heat of combination of sodium oxide with water. Beketoff, however, found that Na,O + aq = 55°5 and Rengade, 56:5. The mean 56 — 19-9 (the heat of solution of 2NaOH), = 36:1 for the heat effect of Na,O + H,O. It is impossible to say which is the better one of the two values, but 36-1 accords best with the results obtained with gypsum and brucite and is the one used in this paper. When a combustible substance, such as sulphur or carbon, is mixed with the peroxide and an hydroxide, and the mixture ignited, sodium oxide is formed thus: S + 3Na,O0, = NaSO, + 2Na,0 and the sodium oxide unites with the combined water to form hydroxide. If the mixture contains brucite the reaction then is S + 3Na,O, + MgO.H,O = Na,SO, + MgO + 2NaOH + Na,O and the heat effect in excess of that due to the sulphur is (Na,O + H,O) — (MgO + H,0O) The method of fusion with sodium peroxide on hydrates was first tried with hydrous calcium sulphate because Thomsen found the heat of combination of water with anhydrous calcium sulphate. For the work CaSO,.2H,O was precipitated from a cold solution of calcium chloride by sulphuric acid. After standing a day it was washed and dried over oil of vitriol in a vacuum. It lost on heating 20°9 per cent; caleulated for 2H,0, 20°9 per cent. The anhydrous sulphate used in experi- ments 1 and 2 was obtained by heating the hydrous salt at 160° until the weight was constant, while that used in experiment 3 was heated to 200°. In experiments 4 and 6 the CaSO,.2H,O made was used, and in 5 a good specimen of selenite. It was impossible to mix thoroughly by shaking together in the bomb, sulphur, the precipitated CaSO,.2H,O and sodium perox- ide. The fine sulphur apparently was electrified, for some of it adhered to the bomb and did not burn completely when the fusion occurred. In some experiments a little free sulphur remained, and in all cases where the mixtures were made in the bomb the results were Jow and considerable silver sulphide was formed. The mixture for experiments 3 and 5 were made in a mortar. In 3 and 6 lampblack was used in place of sulphur and the mixing was made by shaking the ingredients in the bomb closed by a glass plate. The experimental data are given in Table I. The mean of 1, 2 and 3 for CaSO, is 323 cal., which gives 43°7 Cal. for 1 gram molecule. It was found that calcium W. G. Mixter— Polymerization of Oxides. 25 TaBLeE I 1 2 3 4 5 6 CHISOS (c)) as ses Selene 4:026 4°865 5:166 ° @asOg2H5 (b) 2222.2. ---- 3085 3°000 5°283 g S 122 2303 eRe eee 1:000 1:000 1-000 1-000 Wampblack (C)_---+---_-_- 1290 1274 igs) 2-2 cee as ae 13°5 14: 24: 9 18° 22° Water equiv. of system---- 4122 4056 4112 4138 4125 4166 Temperature interval __-_--- 1612 1-709 3°745 WUT a 1:°764 +:076° Oxyeen! set free-.- +. ..--- - 0 0 0 0 0 0 entreteeha tS... j2222 3222 6645 6932 -5400 7309 7277 16981 cal iG BRO ae ee. ae —5300 —5300 —5300 —5300 a U6 CEN OE cae —13722 © 13551 ff 3G 0 EG TAGY Lee bepaeegs —40 —42 —40 —40 —40 —40 oe pameesran(C) ess a5 1305 1592 1638 is Aa 2991 (6) eet eee 2013 1937 3990 Sere itec ofa <5 324 327 317 a cc Meme tecnica 2c be ee, 653 646 648 oxide gives no heat effect in fusion with sodium peroxide and sulphur, hence the heat effect of CaSO, is from the reaction of this compound with sodium oxide as shown in equation 1 be- low. The mean of 4, 5 and 6 for CaSO,.2H,O is 649 cal., o1 111-6 Cal. for 1 gram molecule. The calculated heat effect of 1 gram of sulphur reacting with sodium peroxide is 5300 cal. and this value is used because Thomsen’s values are adopted in the equations 1, 2, 3 and 4, given later. The lampblack used was of the lot used in pre- vious work.t (1) eo (2Na +S +40) + (Ca+ 0) 100°3 328°6 145, + 40°9 Cal. The experimental result is 43-7 Cal. Including sulphur and sodium peroxide in the calculation, we have S+3(2Na +20) + (Ca+8 +440) =2(2Na +8 + 40) + 359°4 332°4 657°2 (Ca + O) + (2Na + O) + 210°7 Cal. 145 100°3 The calculated heat of the reaction between sulphur and sodium peroxide is 169°8 Cal., and adding this to 43-7 gives 213°5 Cal. for the observed heat effect. For the hydrous sulphate we have + Loe. cit. 26 W. G. Mixter—Polymerization of Oxides. S + 3(2Na + 20) + (Ca+S + ia: + (2Na + O) = 359°4 473° 100°3 2(2Na + 8 + 40) + (Ca + O) + 1(Na + O + H) + 276-1 Cal. 657°2 145 407°5 The experimental result is 169°8 + 111-6 = 281-4 Cal. The results with CaSO, and CaSO,.2H,O are somewhat higher than the values derived from Thomsen’s data. Whether the difference is due to errors or to CaO set free from the sulphate combining with sodium oxide, or to other causes, can- not be determined. It has been stated that CaO made by heating the carbonate gives no heat effect in the fusion. CaO separated from the sulphate may be in a different molecular state from the intensely heated oxide. The difference found between the heat effect of CaSO.2H,O and CaSO, is 111°6 — 43-7 = 67:9. The number taken for the heat of the reaction of H,O + Na,O is 36:1 (p. 24). For free 2H,O it is 72°2. Hence for the heat of conisnaien of 2H,O with CaSO, we have 72:2 — 67-4 =4:8 Cal. Thomsen’s value is 4°7. The work with gypsum has been given in detail as it illustrates well the method of fusion with sodium peroxide for determining the heat of combination of water. Brucite, MgO.H,O.—The mineral used was a good erystal- line specimen. An analysis gave the theoretical quantity of water, 0°3 per cent FeO, and traces of Mn and Ca. Two eal- orimetric tests were made of the MgO from heating the brucite with the result of 0, and 13 cal. per gram. One of pure MgO gave 13 cal. Three determinations of the heat effect of brucite in mixtures of sulphur and sodium peroxide gave 476, 438 and 456 cal. With lampblack in place of sulphur the result was 461 cal. The mean 458 xX 58:3 = 26-7 Cal. This number subtracted from 36:1 gives 9°4 Cal. for the heat effect of MgO + H,0. From Richards’ result for the heat of solution of magnesium in hydrochloric acid, Thomsen’s values for the heat of solution of the chloride, etc., and Van Wartenberg’s Mg + O = 143°6 is derived 7-0 Cal. This value is for preci- pitated Mg(OH),, which is commonly regarded as amorphous. Lerric Hydroxide.—Ferric hydroxide after drying in vacuo was found to have the composition Fe,O, + 1°96H,O. When mixed with sodium peroxide the mixture became warm, hence the preparation was not available for the calorimetrie work. Next ferric hydroxide was precipitated from a boiling solution of the chloride by ammonia and the whole was kept hot ona steam bath for several hours. The precipitate was washed with hot water and dried in the open air. After pulverizing it was left nine days over solid potassium hydroxide. This preparation, designated as A’, lost on ignition 11°3 per cent and W. G. Mixter—Polymerization of Oxides. 27 after remaining a year and a half in a rubber-stoppered bottle the loss was 11-6 per cent. It is marked A*. The substance was free from chlorine. Preparation B was made from Kahl- baum’s ferric hydroxide, having the composition approximating to Fe(OH),, by heating for three days at 200-230. It contained a little chlorine and lost on ignition 1:9 per cent. Next an attempt was made to dehydrate ferric hydroxide at 100° by the well-known method of boiling. For this purpose an excess of ammonia was added to a boiling solution of ferric chloride in a flask connected with a reflux condenser and the boiling over a Bunsen burner was continued eight days. The solution, of course, became slightly acid. A portion of the pre- cipitate was washed by decantation until it became a slime and ceased to settle. ‘Then some ammonium nitrate was added to coagulate it and the washing was continued. Finally the pre- cipitate was collected on a filter and dried at 100°. It was free from ammonia, but contained a trace of chlorine. It lost on ignition 2°7 per cent. The preparation is designated as C. The dehydration just described may be due to contact with the glass, which was above the temperature of the boiling solution. One experiment made supports this view. Ferric hydroxide was precipitated as before and the flask containing it was surrounded by steam for eight days. The solution became acid as before. The precipitate was washed, dried at 100°, and then digested with ammonia to break up oxychloride present and washed again. Finally it was dried, pulverized and heated in a weighing bulb for 24 hours at 100°, losing 0-1 per cent the last 12 hours. This preparation lost on igni- tion 9°27 and 9°32 per cent. Preparation D was made from Kahlbaum’s ferric hydroxide, which was in the form of fine grains. It was washed with hot water, digested with ammonia and washed again with hot water. Next the substance was dried at 100°, sifted, and only the fine powder used. A portion of the powder, after heating at 100° to expel hygroscopic moisture, lost on ignition 6-2 per cent. The remainder of the powder, 58 g., was heated four days at about 160°, losing the last day 0°08 g. This prepara- tion lost on ignition 2°65 and 2°61 per cent. It contained no chlorine. In a former paper* it was shown that 1 g. of ferric oxide which has been subjected to a dull red heat reacts with sodium oxide to form sodium ferrite with a heat effect of 363 cal. Hence, after finishing the work on the various samples contain- ing water, it seemed desirable to learn whether ferric oxide which had been exposed to a higher temperature than a dull red would give a different result than the one given above. *This Journal, xxxvi, 55. 28 W. G. Mixter—Polymerization of Oxides. For this purpose the preparation designated as E was made as follows: a portion of D (which was a soft red powder con- taining 2°6 per cent of water) was heated very gradually to approximately 1000°* and then kept at this temperature two hours. The oxide resulting was in the form of a dark gray friable mass. It was easily rubbed to a powder and was passed through a mesh less than 1/200 of an inch. The powdered oxide, after heating twice as described, showed no change in appearance and each time formed a coherent mass. It was found not to change weight during the last heating and to be free from ferrous oxide. The dark gray powder turned dark red when finely ground. It was not too coarse for the calorimetric experiments. The mean of the results for E is 861 cal. This does not indicate polymerization of ferric oxide between a dull red heat and 1000°. The results obtained with ferric oxide holding various amounts of water are given in considerable detail in Table II, as they show that the experimental errors are probably small. The values obtained for D and A’ are the best because of the purity of the preparations used, and only these values will be discussed. The mean for D is 425 and for A’ is 529. The 2°6 per cent of water is undoubtedly more firmly held by the ferric oxide than the 11°6 per cent. If it is assumed that the heat of combination of 1 g. of water in both A and D is the same, the error will not be great, for the proportion of water in D is small. Furthermore, it is assumed that the ferric oxide is in the same molecular state in both Dand A®*. Then if x equals the heat effect of 1 g. of ferric oxide in the fusions and y that of 1 g. of water combined with the ferric oxide, we have O-974e + 0°026y = 425 0°884% + 0°116y = 529 x = 398 Cal. for 1 g. of ferric oxide. The heat of fusion of 18 g. of wateris 1:44 Cal. and 36:1 — 1-440 + 18 = 1-9 Cal. for the heat effect of 1 gr. of solid water reacting in the sodium peroxide fusions. Assuming that in D the solid water is not combined with heat effect, we have 425 — (0°026 X 1°19 Cal. +) 0°974 = 385 eal. for 1 g. of ferric oxide. Hence, on any assumption the ferric oxide in D which had been heated only to 160° polymerizes at high temperatures. The difference between 398 and 363 is 35 cal. and for a gram molecule of ferric oxide, 160 g., it is 56 cal. * Silver, m. p. 960°, melted in the electric furnaces used, but gold, m. p. 1060°, did not melt. : of Ourdes. zon O t . YMEPIaa W. G. Mixter— Pol LOE ccs 669 66g 8661 CP8T 008§ L&8c 00T 0 619+ 8Pgt to Vo— 0§— 0g — *66SEE— xL9CTL— eePll— O8ell— “189 6C6CT CPhPSt OLIPL 678€1 o009-S GEG-S 8PS:S O&&-E C6LP 6917 L8TV OSTP -96 “G6 GG +6 686-1 60T-E 220-T 190-1 STP-G 008-¢ “3 988.9 OF-S U—-+-.—__~Y Ve —-+--— SY o00OT FH0GB 9-11 O° 48 ®OrOA P-88 °O°98.0 Gk aV [RO JEPTT avs youtqdurey Jo yo, Mou oT » 667 CGP cly COV VSP SOP 60S 867 VS9o = 89FG—NG—C‘<«OGTGGsCiCDTGYSC(‘<é‘éiOGSC“> ~ XN S i 2 — ‘ x x x = ooo 0 0o90 0 000 oc; DF CoOaoaoeceoaese 8 & onounownownowndwno0owndcendownddcdd Cal GS CONGUE G) Be) SO) ST) 08) Cae SE COSC) GTS) Cay het 1400, 1500! 1600, 1700 1800 1900 2000: 2500 3000) ) and the Average Depth ) of the Families of Recent Crinoids. Comparison of the Average Range ( Fie. 3. of Habitat ( D. U. Hili—Separation of Potassium and Sodium. 75 Arr. VII.—The Separation of Potassiwm and Sodium by the Use of Aniline Perchlorate, and the Subsequent Esti- mation of the Sodium ; by D. U. Hit. [Contributions from the Kent Chemical Laboratory of Yale Univ.—celxviii. ] Tue chief purpose of the experiments to be described was to determine the availability, as a method for estimating sodium quantitatively, of the precipitation of sodium chloride from solution in alcohol by means of gaseous hydrogen chloride, after the removal of potassium as perchlorate. Kreider and Breckenridge* have determined the delicacy of the qualitative test for sodium made in this way, and have found that “even in 40°°* 0:0002 grm. of sodium oxide can be seen distinctly ; from which fact,’ they conclude, “it is evident that this method can be applied to the quantitative determination of sodium.” The possibility of substituting aniline perchlorate for per- chloric acid as precipitant of the potassium has, also, been tested in the present investigation. Aniline perchlorate forms erystals of definite composition, without water of crystalliza- tion, so that the amount to be used can be readily determined by weighing. A greater advantage which it was thought the use of the aniline salt would have over the acid, viz., that it could be more easily prepared, proved to be illusory. Aniline expels the ammonia from ammonium perchlorate when the two substances are boiled together, but the reaction is slow and during the boiling the aniline develops a dark color. No con- venient way of getting rid of this color could be found. Other methods for preparing aniline perchlorate, starting with the ammonium salt, suggest themselves, as, for instance, the con- version of ammonium perchlorate into barium percholate by boiling with barium hydroxide, and then into aniline perchlo- rate by dissolving in dilute alcohol, adding aniline, precipitat- ing the barium as chloride by hydrochloric gas and ether, and evaporating off the excess of hydrochloric acid; or by mixing exactly equivalent amounts of barium perchlorate and aniline sulphate.t But all of these methods are more laborious than the preparation of perchloric acid by Willard’st method, i. e., the oxidation of ammonium perchlorate to perchloric acid by dilute aqua regia. Indeed, the easiest way to prepare the ani- line perchlorate is to make the acid first by this very simple method, and then to add aniline to a water solution of the acid until some of the oil remains after shaking, and boil vigor- * This Journal (4), ii, 263-8, 1896. + Spallino, Ann, Chim. applicata, i, 435; Chem. Abs. viii, 2701, 1914. ¢ Jour. Am. Chem. Soc., xxxiv, 1480-5, 1912. 76 D. U. Hill—Separation of Potassium and Sodium. ously to expel the excess aniline with the steam before it has time to darken. Colorless crystals are obtained in this way. Experiments were made first to determine the completeness of the precipitation of sodium chloride from 97 per cent alcohol by gaseous hydrogen chloride. The method was as follows: 0°1000 grm. of purified sodium chloride was dissolved in 1°5°™* of water, 48°5°™* of absolute alcohol was added (this amount of 97 per cent alcohol will not hold much more than 0-05 grm. of sodium chloride in solution), and gaseous hydrogen chloride (conveniently evolved by the action of concentrated sulphuric acid upon massive ammonium chloride in a Kipp generator) was passed into the cooled solution through an inverted funnel. When the solution appeared to be saturated with the gas, the precipitate was collected on asbestos in a perforated crucible, dried at about 110°, and weighed. Table I shows the results of these experiments: TABLE I. NaCl taken NaCl found Error on NaCl Error on Na,O grm. grm. grm. grm. (1) 0°1000 0°0994 —0°0006 —0°0003 (2) 0:1000 0°0989 —0°'0011 —0'0006 (3) 0:1000 0:0994 — 0:0006 —0-0003 (4) 0°1000 0:0992 —0:°0008 —0:0004 A series of experiments was then made in which both potassium and sodium were estimated. Equal amounts of recrystallized potassium chloride and sodium chloride were weighed out and dissolved in 1:5°™* of water. The amounts used are shown in Table II. An excess of aniline perchlorate (about 5 grm.) dissolved in 48°5°™* of absolute alcohol was then added, and the precipitate of potassium perchlorate was filtered off on a perforated crucible with the aid of suction, and washed with about 20°° of 97 per cent alcohol. The pre- cipitate was dried at 110° and weighed. The filtrate was satn- rated with hydrogen chloride from the Kipp generator as in the previous experiments, the precipitate of sodium chloride collected on a perforated crucible, washed with a saturated solution of hydrogen chloride in 97 per cent alcohol, dried and weighed. The results of this series of experiments are shown in Table II. The errors on potassium are greater than those obtained by Kreider,* who took pains to convert the potassium chloride completely into perchlorate by evaporating it twice with per- chloric acid to a syrup-like consistency, and in washing used 97 per cent alcohol containing perchloric acid, finishing with a very little pure alcohol. In the case of sodium the experi- * This Journal (8), xlix, 4438-8, 1895. D. U. Hili—Separation of Potassium and Sodium. V7 TABLE II. KCltaken KCI1O, Error on NaCl taken NaCl Error on grm. found K.O erm. found Na,O grm. grm. grm, grm. (1) 070500 0'0932 +0°0001 0°0500 00496 —0°0002 (2) 0°0400 0:0729 —0°0005 0°0400 0:°0392, —0'0004 (3) 0:0400 0:0727 —0:°0006 0:0400 0°9393 —0°0004 (4) 0°0400 0:0728 —0-0006 0:0400 0°0394 —0°0003 5) 0°0300 070558 0:0000 0°0300 0°0293 —0-°0004 (6) 0°0300 0°0556 —0°0001 0:0300 0:0295 —0°0003 (7) 0 0300 0°0543 —0°0005 0°0300 0:0295 —0:0003 (8) 0:0300 0:0540 —0°0006 0°0300 0:0296 —0°0002 mental results indicate the presence of a small constant nega- tive error due apparently to some solubility of sodium chloride in the saturated solution of hydrogen chloride in alcohol, as well as to losses in manipulation. Unfortunately the error per cent cannot be lessened by working with larger amounts of the salts on account of the limited solubility of the chlorides in alcohol. During the past two years the methods described above for precipitating potassium by aniline perchlorate, and then sodium chloride by hydrochloric acid gas, have been successfully used as a means of detecting the presence of these elements by the class in qualitative analysis in this laboratory. In order to lessen the amount of gaseous hydrogen chloride used, the sodium, after the removal of the potassium, is precipitated by means of sulphuric acid instead of hydrogen chloride, where possible. A few drops of a dilute solution of sulphuric acid in alcohol will generally precipitate even very small amounts of sodium as sodium sulphate, but if too much acid is added the acid sulphate is formed and dissolved in the alcohol. Hence, if no precipitate is obtained with sulphuric acid, gaseous hydro- gen chloride is passed into the solution to saturation, precipita- ing the sodium as chloride practically completely. 78 Scientific Intelligence. SCIEN TLEIC INTEL EWE HN Cie Te Curmistry AND Puysics. 1. Prout’s Hypothesis.—This old idea that the atomic weights of the elements heavier than hydrogen are exact multiples of the latter has been elaborately discussed by Wiriiam D. Harxins and Ernest D, Wirson in two articles entitled, “The changes of mass and weight involved in the formation of complex atoms,” and ‘The structure of complex atoms. The hydrogen-helium system.” Taking the first 26 atomic weights above hydrogen in the order of magnitude, it is found on the basis of hydrogen as unity that nearly all of them show negative variations from whole numbers. Thus, the atomic weight of helium is 3°97, lithium is 6°89, and boron is 10°91, which give the variations —0-03, —0°1] and —0-:09 respectively. When these variations are calculated as percentages, the striking fact is observed that with the exception of Be, Mg, Si and Cl, which show positive variations, the average negative variation of 21 atomic weights is 0°77 per cent, while the six elements from boron to sodium show values of 0°77, 0°77, 0:70, 0°77, 0°77 and 0°77 per cent. The values of P, V, and Cr are 0°71, 0°77 and 0°77 per cent, while the others vary between 0°40 per cent for Al and 1:06 per cent for Fe. The negative devia- tion is, therefore, not a periodic bnt a constant one. The authors call this negative variation a packing effect, meaning that when several hydrogens are united to form a heavier atom there is a loss in weight amounting to about 0°77 per cent of the whole. When oxygen as 16 is taken as a basis for the atomic weights of the 21 elements under consideration their variations from whole numbers are very slight, as is well known. The “ packing effect ” occurs in the oxygen as well as in the others, so that in this case it is eliminated, and Prout’s hypothesis in this modified form applies to these atomic weights. The deviation from whole num- bers averages only 0-05 unit for the 21 elements, and the proba- bility that such values should occur by accident is so slight as to be unworthy of consideration. There are many points discussed by the authors in their inter- esting articles that will not be mentioned here, as it is intended to present only such features of the discussion as have a very direct bearing on Prout’s hypothesis. No explanation is advanced in regard to the atomic weights of the elements magnesium, silicon and chlorine, which are exceptional among the lighter atomic weights in not approaching close to whole numbers, and in show- ing positive variations when hydrogen is used as unity. Further than this, the authors have found that the remaining 42 elements with heavier atomic weights, from nickel upward, show no tend- ency to approximate to whole numbers, since the average of their deviations is very close to 0°25 unit, which would be expected Chemistry and Physics. (@) from variations governed by chance. No satisfactory explanation of this last circumstance is given, and it appears that Prout’s hypothesis modified by the “‘ packing effect ” can be applied only to about one-third of the elements whose atomic weights are pretty accurately known, and that these are among the lighter atomic weights.—Jour. Amer. Chem. Soc., xxxvil, 1367, 1383. H. L. W. 2. The Action of Chloroform upon Metallic Sulphates.—A. Conpbvucné has obtained interesting results by studying the action of chloroform vapor, carried along by means of a current of car- bon dioxide, upon heated metallic sulphates. The reaction sup- plies a method for the production of anhydrous chlorides. The change is indicated by the formation of white vapors, and it takes place at a rather low temperature: 250° C for CuSO,, 300° for NiSO, and FeSO,, 350° for MuSO, and PbSO,, 400° for Al,(SO,),, 450° for MgSO,, 500° for BaSO,, CaSO, and Na,SO,. Practically, however, a higher temperature is required to make the reaction complete, and on the other hand, the vapor of chloro- form decomposes between 400° and 500° with the production of carbonaceous deposits, so that the chlorides are not cbtained in a pure condition when the temperature used is too high. The reducing action of the chloroform may have an influence upon the chloride produced in case a metal forms two chlorides. For example, CuSO, yields pure CuCl, at 280-300°, while at about 400° CuCl is produced. In the cases of FeSO, and Fe,(SO,), mixtures of FeCl, and FeCl, are always obtained, the proportions of which depend upon the temperature employed.— Comptes Rendus, clviii, 1180. 1 ibe Wile 3. Explosives; by ArtHur Marsuaryi. Large 8vo, pp. 624. Philadelphia, 1915 (P. Blakiston’s Son & Co. Price, $7.50 net).— The author of this important work on the manufacture, proper- ties, tests, and history of explosives, is the Chemical Inspector of the British Indian Ordnance Department. No comprehensive work on this subject has appeared in English within a period of 20 years, during which time very great changes and developments have taken place in the inaustry, so that the present book will undoubtedly supply a real want. The book treats the subject very fully and ably, it contains 137 illustrations, and gives a very interesting and useful account of this great branch of chemical industry. H. L. W. 4. Chemical Technology and Analysis of Oils, Fats, and Waxes ; by J. Lewxowirscu, edited by Groner H. Warsur- ron. Vol. III, Svo, pp. 483. London, 1915 (Macmillan and Co., Limited. Price, $6.50 net).—The present, third volume, com- pletes the fifth edition of this important work, which has been entirely re-written and enlarged. The previous volumes have been favorably noticed already in this Journal, hence it is only necessary to point out that the present volume deals chiefly with manufacturing operations, and that among the important. and interesting topics treated are edible oils and fats and their sub- 80 Scientific Intelligence. stitutes, lubricating oils and greases, hydrogenated fats, varnishes, the candle industry, soap manufacture, and glycerine manufac- ture. H. L. W. 5. Annual Reports on the Progress of Chemistry for 1914. 8vo, pp. 303. London, 1915 (Gurney & Jackson, London ; D. Van Nostrand Company, New York).—This is the eleventh vol- ume of these reports which are issued by the Chemical Society. It contains eleven articles dealing with the various branches of chemistry, among which radio-activity is included. Each of the articles has been prepared by a specialist in the subject discussed, and the book is a very useful one in furnishing concise and inter- esting accounts of the important results of investigations pub- lished during the year under consideration. H. L. W. 6. X-Ray Band Spectru.—As is well known, the negatives of X-ray spectra taken by de Broglie show very clearly two bands in the region of very short wave-lengths. These bands have sharp limits on the less refrangible side, they have never been resolved into lines, and they maintain their spectral positions unaltered when the material of the anticathode is changed. On account of these properties it has been suggested by the Braggs and Siegbahn that the two bands owe their origin to the silver and bromine in the photographic films. In testing this hypothe- sis EK. WaGner has recently brought to light some interesting and important facts. The spectrometer used was of the rotating crystal type and was especially designed to minimize the mechanical vibrations arising from the driving mechanism. The most novel feature of the system consisted in coupling segments of the axles with short pieces of rubber tubing. ‘The angular speed was very uniform and amounted to about 10 degrees of arc per minute. Selected rock-salt crystals were used as space gratings, the natural cleav- age planes containing the axis of rotation of the spectrometer table and being equidistant from the emergence slit of the lead collimator tube and the center of the photographic plate. The “hard” and “soft” X-ray tubes were provided with tungsten and platinum or palladium anticathodes respectively. The nega- tives obtained with exposures of seven or eight hours are very clear and sharp. One horizontal strip of the 7-hr. negative obtained with a soft tungsten bulb shows three bands, the central image, and a num- — ber of metallic lines. The adjacent strip was produced simultane- ously by the radiations which emerged from a sheet of aluminium 1°4™™ thick which was placed 1°™ in front of the plate. The band of intermediate wave-length, which was very intense on the strip first mentioned, is not recorded on the second strip. The persist- ence and location of the least refrangible band show that it is the second order image corresponding to the band of shortest wave- length. The extreme bands were supposed to be due to silver and the intermediate band was ascribed to bromine. If this hypothesis is correct we should expect to find that a suitable metal Chemistry and Physies. 81 would give rise to a similar band, if the metal (in the form of foil) were placed close to the sensitive film instead of being located in the gelatine film itself. Layers of pure tinfoil were pressed flat against the back of the photographic plate and exposures were made with the sensitive film turned away from the incident radia- tion. Under these circumstances the soft, secondary radiation starting in the tinfoil would be absorbed by the glass and thus prevented from confusing the negative. As anticipated, a tin band of slightly shorter wave-length than the supposedly silver band is a prominent feature of the photograph. The new band ends abruptly on its longer wave-length side and is similar in all respects to the three bands of the first negative. In order to decide whether the tin band was due to secondary Roéntgen rays or to switt electrons, the last experiment was repeated under slightly different conditions. A sheet of aluminium foil, of such a thickness (0°1™™) as to absorb all corpuscular radiation but to readily transmit X-rays, was interposed between the tinfoil and the sensitive film. For sake of comparison, a longitudinal slot was cut in the aluminium screen. The tin band was of the same intensity behind the aluminium as on the unscreened strip of the negative. ‘Therefore the tin band was due to the strong second- ary or fluorescent Réntgen radiation, excited in the tin by the dis- persed primary rays. The fact that the secondary rays give a continuous band instead of one or more homogeneous lines shows that the band corresponds to the entire spectral interval within which the fluorescent radiation can be excited. It is worthy of note that the intensity of the secondary radiation near the edge of the tin band was of the same order of magnitude as the inten- sity of the primary beam. Accordingly the primary rays must experience very strong absorption in the tin. When an experi- ment was performed with sheets of tin placed a few centimeters infront of the photographic plate, it was found that the absorp- tion band had exactly the same edge and complementary inten- sity distribution as the fluorescent emission band. Having obtained this information about tin, it was easy for the investigator to show that the supposedly silver bands were really due to the silver in the photographic emulsion. A narrow strip of pure silver foil (0°013™™ thick) was placed at a distance of 1°" before the plate and an exposure taken. The characteristic edge of the silver band dis- appeared almost completely and, for each wave-length, the selec- tive absorption in the silver foil had produced precisely the complementary photographic action as the selective intensification of the silver inside the gelatine. Although Wagner has not yet demonstrated experimentally that the “bromine band” owes its origin to the presence of bromine in the sensitive film, neverthe- less the preceding work leaves but little doubt as to the correct- ness of the hypothesis. It may be concluded, therefore, that the photographic action of Réntgen rays of very short wave-length is due to the highly selective fluorescence of the silver and bro- mine atoms. On the other hand, the ability of the plate to record Am. Jour. Sct.—Fourts Series, Vou. XL, No. 235.—Juty, 1915. 6 82 Scientific Intelligence. radiations of longer wave-length than the bands (such as X-ray lines) depends upon the silver bromide molecules, just as for ordinary light. In an earlier investigation Wagner found a close connection between the wave-length A, of the sharp edges of the bromine, silver, and tin bands and the wave-length da of the corresponding fluorescence lines of Moseley’s A-series. Moreover, conformable to the law of Stokes, the radiation of longest wave-length able to produce fluorescence was always more refrangible than the excited Kline. Finally, the ratio A4/Aa (“der Stokessche Sprung”) was approximately constant. In addition to the three substances mentioned above the investigation has been extended by Wagner to the following metals, namely : cadmium, copper, iron, nickel and palladium. All of the relations between the wave-lengths were found to hold, and a number of new facts of theoretical sig- nificance were brought to light. For further details, however, the original paper must be consulted.—Ann. d. Physik, vol. xlvi, p. 868, March, 1915. H. S. U. 7. Elements of Optics; by Grorcre W. Parker. Pp. 122, with 64 figures. London, 1915 (Longmans, Green, and Co.).— That the text is very elementary in character may be inferred from the following quotation, namely: “This little book is intended for those students whose knowledge of Mathematics is limited to an acquaintance with Elementary Geometry, the solu- tion of Simple Algebraic Equations, and a few fundamental propo- sitions in Trigonometry.” The author’s style is clear, and the material is so chosen as to be interesting as well as instructive. The method of rays is used throughout and the illustrative figures are printed as white lines on black background. A comparatively large number (106) of “exercises” for solution by the student are incorporated in the text and the answers are given at the end of the volume. ‘The book is undoubtedly good, as far as it goes, but the impression of incompleteness is given by the unusually small number of topics discussed. Nothing is said, for example, about photometry, astigmatism, etc. H./S..U 8. Dielectric Phenomena in High Voltage Engineering ; by F. W. Peer, Jr. Pp. xv, 265, with 190 figures. New York, 1915 (McGraw-Hill Book Co.).—It is the object of the author to give in this book the properties of gaseous, liquid and solid insulations, and methods of utilizing these properties to the best advantage in the problems of high-voltage engineering.” ‘Much original work is given, as well as reference to other investigations.” ‘The author’s extensive research was made possible by facilities afforded by the Consulting Engineering Department of the Gen- eral Electric Company . . .” A general idea of the contents of the volume may be obtained from the titles of the chapters, namely : “The Dielectric Field and Dielectric Circuit (Mathematical Consideration) ; Visual Corona ; Spark-over; Corona Loss; Corona and Spark-over in Oil and Liquid Insulations; Solid Insulation ; The Electron Theory ; Geology and Mineralogy. 83 Practical Corona Calculation for Transmission Lines” ; and “ Prac- tical Considerations in the Design of Apparatus where Solid, Liquid and Gaseous Insulations Enter in Combination.” The appendix (pages 238 to 256) comprises a large number of tables of numerical data pertaining to corona losses. The graphs and diagrams are clear-cut and the reproductions of photographs are excellent. In general, special attention seems to have been given to making the text as accurate, useful and practical as possible. Chapter VIII, on the electron theory, alone marks an exception. For example, on page 193 may be found the following slips : “|. . a wire-carrying current”; “EKach ion in a gas acts’ as nuclei . . .”; “Ion is a general term used for. . . electrons. . . .” H. S. U. 9. The Radium- Uranium ratio in Carnotites; by 8. C. Linn and C. F. Wuitrrmorre. Bureau of Mines, Tech. Paper 88 (Mineral Tech. 6).—The authors have carried through an investi- gation of the carnotite of Colorado and Utah as to the radium- uranium ratio. The results obtained can best be given by quoting at length the summary with which the paper closes: 1. Samples of carnotite representing large quantities of ore (a few hundred pounds to several tons) show a radium-uranium ratio identical with that of pitchblende (3°33 x 10~’); this ratio is also in accord with the value calculated from radiation data. 2. Samples from small quantities of ore (hand specimens up to a few pounds) tend to exhibit abnormal ratios. In one instance the ratio was as low as 2°48 10~’, and in another as high as 4°6X107". 3. The most plausible explanation for these abnormal ratios seems to be that of transposition of radium within the ore bed, producing local differences which are equalized in large samples. 4. The “emanating power” of carnotite is high, and varies from 16 to 50 per cent. 5. In order to obtain concordant results by the Boltwood emanation method it was found desirable to determine the ema- nation liberated by solution in the same sample from which the emanating power had just been determined, thus making the two determinations strictly “complementary.” 6. Radium may be easily determined in one operation by the emanation method, either by solution or by ignition from tubes in which it has been sealed for one month to reach equilibrium. 7. In contrast with the success of the solution and the ignition methods for de-emanating carnotite, the method of fusion with sodium and potassium carbonates and the fusion-and-solution method both gave low results and were abandoned. Il. Gsonogy anp Mrneraoey. 1. Climate and Evolution; by W. D. Matrurw. Ann. New York Acad. Sci., vol. xxiv, 1915, pp. 171-318, figs. 1-33.—The title of this important work does not convey the intent of the author and should have been “The geographic dispersal of 84 Scientific Intelligence. animals as affected by climate and evolution” or “The theory of land bridges as negated by climate and evolution.” As is well known, the author is an ardent believer in the permanency of continents and ocean basins as they now exist, though the study in hand aims to prove the hypothesis for Cenozoic time only. However he states that “If the distribution of animals be inter- preted along the lines here advocated, there is no occasion for a Gondwana Land even in the Paleozoic” (191). The reviewer thinks his conclusion sound when restricted to the Cenozoic, but to say there was no Gondwana in early Mesozoic time, and especially none in Permian time, is to drag into this painstaking and most excellent study an unnecessary and unproved conclu- sion. The work is replete with facts and new ideas regarding the dispersal of animals (mainly mammals), interpreted on the basis of periodic changes of climate from moist, uniform, and warm to arid, to zonal, and glacial ones. ‘The writer seeks to prove that the present distribution of life in the various continents can be best explained by radial dispersal from Holarctic centers with variable climates (Europe, Asia, North America) into the periph- eral lands (South America, Africa, Australasia). His main prin- ciple of dispersal is that in the evolution of a race “it should be at first most progressive at its point of original dispersal, and it will continue this progress at that point in response to whatever stimulus originally caused it and spread out in successive waves of migration, each wave a stage higher than the previous one. At any one time, therefore, the most advanced stages should be nearest the center of dispersal, the most conservative stages far- thest from it. It is not in Australia that we should look for the ancestry of man, but in Asia” (180). Finally our knowledge of fossil land animals is almost wholly of those of the lowlands, with but rare glimpses of an upland form (274). As the oceanic islands have life derived from the adjacent continents, the author explains the arrival of this life over sea as due to natural rafts. He argues that for every raft seen a hundred have probably drifted out unseen, and if we concede that 1000 have probably occurred in three centuries, then 10,000,000 (by an error he states 30,000,000) would have occurred in the Cenozoic. He further estimates that only 1,000,000 will have living animals upon them, of these only 10,000 will reach land, and in only 100 of these cases will the species establish a foothold. This is quite sufficient to cover the dozen or two of Mammalia on the larger oceanic islands ” (206-207). Undoubtedly there is some truth in these figures, especially for very small animals, but such rafts can hardly have been a marked factor in the dispersal of land animals. Doctor Matthew does not believe in the fracturing of continents as evidenced by the separation of Madagascar from Africa, nor does he hold to the idea that where mountains are now seen to terminate abruptly facing the ocean (as in the Maritime Provinces Geology and Mineralogy. 85 of Canada, in Great Britain, Belgium and France, such coasts being known as Rias coasts) they have sunk into the depths. Regarding the continental shelf being “so marked, obvious and universal a feature of the earth’s surface that it affords the strongest kind of evidence of the antiquity of the ocean basins and the limits beyond which the continents have not extended” (308-309), the reveiwer holds that the present continental shelf is of modern construction, certainly of late Cenozoic making, and simply represents the land wash within the zone of wave and tidal action. With every shrinkage of the earth and subsidence of the oceanic areas the margins of the continents are locally or regionally warped downward and new continental shelves are developed upon the sunken areas. The possible increase in the amount of water during geologic time is left out of consideration (309) and nowhere is there a word as to why most of Africa and eastern South America have broken-down coasts instead of uplifted and folded margins. The reviewer heartily recommends the work to paleontologists and zoogeographers, as the author is believed to be sound in his general premises regarding the distribution of land animals during Cenozoic time. Some years ago the reviewer under- took a similar study, starting out with the theory that Africa and South America were still united in early Tertiary, but grad- ually came to the conclusion that these lands had been broken through by the Atlantic in Lower Cretaceous (Upper Comanchian) times. @: 8: 2. Publications of the United States Geological Survey, GrorcE Oris Suitu, Director.—Recent publications of the U.S. Geological Survey are noted in the following list (continued from vol. xxxix, pp. 316-318) : Topoerapuic ATLAS—Sixty-seven sheets. PRoFEsSIONAL Paprrs.—No. 88. Lavas of Hawaii and their relations; by WuirmMANn Oross. Pp. 97; 4 pls. See p. 88. No. 90. Shorter Contributions to General Geology. I. The Stratigraphy of the Montana Group with special reference to the position and age of the Judith River Formation; by C. F. BowrEn. Pp. 95-153; 1 pl. J. The Cretaceous-Eocene contact in the Atlantic and Gulf Coastal Plain; by Ltoyp W. SrErHENsoN. Pp. 155-182; 9 pls., 8 figs. K. The History of a portion of Yampa River, Colorado, and its possible bearing on that of Green River; by E. T. Hancock. Pp. 183-189; 2 pls. L. The Inor- ganic constituents of Echinoderms; by F. W. Crarke and W. C. WHEELER. Pp. 190-199. No. 95-A. The composition of muds from Columbus Marsh, Nevada; by W. B. Hicks. Pp. 11, 1 fig. Bu uetins.—Nos. 559, 560, 563, 567. Results of Spirit Level- ing. R. B. Marsuart, Chief Geographer. No. 559. Michigan, 1911 and 1913. Pp. 79; 1 pl. No. 560. Minnesota, 1897-1914. Pp. 190; 1 pl. No. 563. Maryland, 1896 to 1911, inclusive. Pp. 80; 1 pl. No. 567. Idaho, 1896-1914. Pp. 130; 1 pl. 86 Scientific Intelligence. No. 582. Mineral Deposits of the Santa Rita and Patagonia Mountains, Arizona; by Frank C. ScurapeEr, with contribu- tions by James M. Hitn. Pp. 373; 25 pls., 46 figs. No. 589. The calcite marble and dolomite of Eastern Ver- mont; by T. Netson Datz. Pp. 66; 2 pls., 11 figs. No. 594. Some mining districts in Northeastern California and Northwestern Nevada; by James M. Hitt. Pp. 200; 19 pls., 4 figs. No. 596. Geology and coal resources of North Park, Colorado; by A. L. Brrxry. Pp. 121; 12 pls., 1 fig. No. 580. Part I-L. Salines in the Owens, Searles, and Pana- mint Basins, Southeastern California; by Horr 8S. Gate. Pp. 251-323; 3 pls., 31 figs. Part I-P. Publications by Survey authors on metal and non-metals except Fuels. Compiled by IsaBeL P. Evans. Pp. 413-445. No. 581. 1918. Part II-E. The Coalville Coal Field, Utah; by Carrot, H. Wrecrmann. Pp. 161-187; 6 pls. No. 620-A. A gold-platinum-palladium lode in Southern Nevada; by ApotpH Kworr. Pp. 18; 1 pl., 1 fig. Water Suppty Papers.—Nos. 312, 331, 353, 354. Surface Water Supply of the United States; prepared in codperation with the respective States. No. 312. 1911. Part XII. North Pacific coast drainage Basins; by F. F. Hensuaw and others. Pp. 706; 4 pls. No. 331, 1912. Part XI. Pacific Coast Basins in California; by H. D. McGuasuan and G. C. Srrvens. Pp. 442; 2 pls. Nos. 3538, 354, 19138. Part II]. Ohio River Basin; by A. H. Horton, and others. Pp. 264; 5 pls. No. 354. Part IV. St. Lawrence River Basin; by W. G. Hoyt, and others. Pp. 136; 2 pls. No. 340. Stream-Gaging Stations, etc., 1885-1913 (compiled by B. D. Woop). F. Part VI. Missouri River Basin. Pp. viii, 63-81. G. Part VII. Lower Mississippi River Basin. Pp. viii, 83-93. Part VIII. Western Gulf of Mexico drainage Basins. Pp. viii, 95-104. Part IX. Colorado River Basin. Pp. viii, 105-116. Part X. The Great Basin. Pp. viti, 117-129. No. 345-H. Ground-Water Resources of the Niles Cone and adjacent areas, California; by W. O. Clark. Pp. iv, 127-168; 9 pls., 16 figs. I. Gazetteer of surface waters of lowa; by W. G. Hoyt and H. J. Ryan. Pp. 169-225. No. 375-A. Ground Water for irrigation in the Sacramento Valley, California; by Kirk Bryan. Pp. 49; 2 pls., 6 figs. Nos. 349, 350, 367, 368. Profile Surveys prepared under the direction of R. B. Marsuatt, Chief Geographer. No. 349. Willamette River Basin, Oregon. Pp. 8; 3 pls. No. 350. Bear River Basin. Idaho. Pp. 7; 1 pl. No. 367. Missouri River from Great Falls to Three Forks, Montana. Pp. 8; 1 pl. No. 368. Wenatchee River Basin, Washington. Pp. 7; 1 pl. No. 338. Springs of California; by Gmratp A. WARING Pp. 410; 13 pls., 4 figs. No. 341. Underground Waters of the coastal plain of Georgia; by L. W. SreruHenson and J. O. Veatcu. And a discussion of Geology and Mineralogy. 87 the quality of the Waters; by R. B. Dotz. Pp. 539; 21 pls, 4 figs. No. 343. Geology and Water Resources of Tularosa Basin, New Mexico; by O. E. Meinzer and R. F. Hare. Pp. 317; 19 pls., 51 figs. No. 365. Ground Water in Southeastern Nevada; by EverEtrr CARPENTER. Pp. 86, 5 pls., 3 figs. 3. The United States Bureau of Mines, Josrru A. Homes, Director.—The following Bulletins have been issued since the last summary (vol. xxxix, p. 224): No. 80. A primer on explosives for metal miners and quarry- men; by Cuartes E. Munroe and CLarence Hawt. Pp. 125; 15 pls., 17 figs. No. 81. The smelting of copper ores in the electric furnace; by D. A. Lyon and R. M. Kzrnzy. Pp. 77; 6 figs. No. 84. Metallurgical smoke; by Cuartes H. Futron. Pp. 94; 6 pls., 15 figs. No. 87. Houses for mining towns; by JosepH H. Wuire. Pp. vi, 58; 8 pls., 9 figs. No. 88. Petroleum Technology 20. The condensation of gasolene from natural gas; by G. A. Burret, F. M. Serer, and G. G OpErFELL. Pp. 106; pls. vi, 18 figs. No. 90. Law Serial 3. Abstracts of current decisions on mines and mining, December, 1913, to September, 1914; by J. W.THomeson. Pp. xvii, 175. Numerous Technical Papers have also been published. See p. 83. 4, Canada, Department of Mines.—The following are some of the more important of recent publications (see vol. xxxix, 481). (1) Geological Survey Branch; R. W. Brock, Director. Memoriat.—No. 56. Geology of Franklin Mining Camp, British Columbia ; by Cuarites W. Dryspate. Pp. vii, 246 ; 23 pls., 16 figs. Memoir 57. Corundum, its occurrence, distribution, exploita- tion, and uses; by Atrrep EH. Bartow. Pp. vii, 377; 2 maps, 20 pls. This is a highly valuable discussion of a subject of prime interest alike from the theoretical and the technical standpoints; a notice appears later. No. 59. Coal Fields and Coal Resources of Canada; by D. B. Dowrine. Pp. 174 ; 7 maps, 9 figs. No. 61. Moore Mountain District, Southern Alberta (second edition) ; by D. D. Carrnzs. Pp. 62; 2 maps, 1 fig. No. 65. Clay and Shell Deposits of the Western Provinces, Part IV; by H. Riss. Pp. 83; 8pls., 18 figs. No. 66. The same, Part V ; by J. Kezriz. Pp. 74; 8 pls. Also the Mining Museum Bu xetins : No. 11. Physiography of the Beaverdell map-area, etc.; by L. REINECKE. No. 12. On Hoceratops canadensis, gen. nov., with remarks on other genera of Cretaceous horned Dinosaurs ; by L. M. Lambe. No. 14. Glacial Drift in the Magdalen Islands; by J. W. GoLpTHWAIT. 88 Scientific Intelligence. (2) Mines Branch; Evucenr Haaner, Director.—Summary Report for the calendar year ending December 31, 1913. Pp. x, 214; 51 pls., 24 figs., 1 map. Annual Report on the Mineral Production of Canada during the calendar year 1913. Joun MclLxutsu, Chief of the Division of Mineral Resources and Statistics. Pp. 363. Bulletin No. 9. Investigation of the Peat Bogs and Peat Industry of Canada 1911-12; by A. V. Anrep. Pp. vii, 47; 29 pls. 6 figs, 11 maps. No. 10. Notes on Clay Deposits near McMurray, ‘Alberta ; ; by Sypnry C. Exzs. Pp. 15. Report on the Non-metallic minerals used- in the Canadian manufacturing industries ; by HowreLtts FrecuETTE. Pp. viii, 199. Peat, Lignite, and Coal: their value as fuels, etc; by B. F. Haanet. Pp. xv, 261, 19 pls., 39 figs., 20 tables. Petroleum and Natural Gas Resources of Canada. In two vol- umes. Vol. I; by FrepERIcK G. Ciapp and others. Preliminary Report on the bituminous sands of Northern Alberta ; by 8. C. Exits. Pp. iv, 92, 55 pls., 5 figs., 1 map. 5. Lavas of Hawaii and their Relations; by WuitmMan Cross. Prof. Paper 88, U. 8. Geol. Survey, 1915, pp. 97, 4 pls.— This paper is not only of local interest in furthering our knowl- edge of the petrology of the Hawaiian group, but also an import- ant contribution to the general literature of the subject. It represents the results of a considerable amount cf work in the field by the author in studying the occurrences of the rocks com- posing the volcanoes, and in the collection of material, in the petrographical investigation of this material as well as that col- lected by others, and of a large number of chemical analyses. The precise data thus assembled are of great and permanent value as a contribution toward a better understanding of the petrogenesis of Hawaiian lavas. The different islands are taken up separately and the rocks occurring upon them described. It would be beyond the limits of this notice to give these details, but it may be said of the rocks as a whole that while the author finds that basalts of the ealeic series are the dominant types, yet occurrences of rocks of clearly _ alkalic character are well represented, with some of intermediate nature. In regard to the origin of these rocks, Cross assumes that they have been formed by differentiation from a general parent magma of the composition of a normal calcic basalt, and that during the active period of each volcano differentiation was seldom if ever able to produce partial magmas of extremely salic or femic character. This might be due to short periods of quiet in the magma chamber insufficient to permit of much differentia- tion. He is not inclined to view the sinking of crystals as a dominant factor in the process, though this may have played some part. With decreasing activity and, perhaps, some contrac- tion and limitation of magma chambers there was more differen- tiation with correspondingly more salic and femic lavas. In the final period of dying activity, when parasitic eruptions followed, Geology and Mineralogy. 89 occurred the most extensive differentiation affording the comple- mentary alkalic felsic and mafic types. Every petrographer will find these discussions of the petrogenesis and classification of the rocks of great interest, based as they are upon such accurate data and presented so clearly. ie Va) Ps 6. Brief Notices of some recently described Minerals.— BarTHITE is a zinc-copper arsenate, described by W. Henglein and W. Meigen, from Guchab in the Otavi mountains, German South West Africa. It occurs as a crystalline incrustation on quartz crystals which form druses in a reddish dolomite. Hard- ness 3; specific gravity 4:19; color grass-green ; luster greasy. An analysis gave : As205 P20; ZnO CuO H,0 insol. 64:0 1:0 28°3 8:5 372 iil == alles for this the formula 3ZnO.Cu0.3As,0,.2H,0 is calculated.— Centralbl. Min., 353, 1914. UssInGITE is a new silicate, described by O. B. Boggild, from the remarkable locality at Klangerdluarsuk, Greenland. It is referred to the triclinic system on the basis of cleavage frag- ments. Hardness 6 to 7; specific gravity 2°50; luster vitreous to pearly; color violet-red. An analysis by Chr. Christensen gave: SiO, Al,O; Na,O H.O 58°74 17°73 19°91 4°19 = 100°57 This leads to the formula HNa,A1 (Si0,),.—Zs. A7vyst., liv, 120, 1914. FaRATSIHITE is a hydrated ferric silicate described by A. Lacroix, from Faratsiho, Madagascar. It occurs in masses of a yellow color, resembling nontronite. Under the microscope it shows a crystalline structure like that of kaolinite, being made up of aggre- gates of minute hexagonal scales. An analysis gave: SiO, Al,03 Fe,0; H.0(105°) H,O(red heat) X 41°60 22°68 15°22 o71 13°02 1:44 = 100°20 Here X = FeO 0:54, MgO 0°11, CaO 0:60, Na,O 0°16, K,O 0-22, Ti0, 0:13, P,O, 0:21. The formula deduced is H,(Al, Fe), Si,O,, and it may be regarded as a kaolinite in which part of aluminium is replaced by ferric iron.— Bull. Soc. Hr. Min., xxxvii, 231. SPEZIAITE is a kind of amphibole described by L. Colomba, from the pyroxenite of Riondello, Traversella. It occurs in fibrous and acicular forms with cleavage angle of 55° 30’ to 56°; color green, pleochroic; the angle cac = +238° to 24°. An analysis gave: SiO. Al,O: Fe.0; FeO MnO CaO MgO Na,zO K,20 H.O 26:21 079 34:57 3:56 067 10°58 787 4:08 0:93 0:50 = 99°71 The calculated formula is that of an orthosilicate 5Fe,(Si0,),.12 (Ca, Mg, Fe, Na,, H,) SiO,.— Atti Accad. Sci., Torino, xlix, March, 1914. Lus.initz is described by R. Lang as a monoclinic modification of calcium carbonate forming the soft earthy aggregate called 90 Scientific Intelligence. “rock milk”; the occurrence examined was from the Diessener valley near Horb on the Neckar. A similar form, also earthy and made up of microscopic needles, from the Adams valley near Briinn has been studied by O. Miigge. The conclusion reached by him is that there is no reason for assuming the existence of a new modification of CaCO,; the peculiar characters being prob- ably due to pseudomorphism after organic remains.—Lang, J. Jahrb. Min., Beil. Bd., xxxviii, 121, 1914; Miigge, Centraibdl. Min., 673, 1914 ; ; Lang, ibid., 298, 1915, HeweErrire, Meranewertire and Pascorre are hydrous cal- cium vanadates described by Hillebrand, Merwin and Wright. Hewettite was obtained by D. F. Hewitt, at the vanadium locality of Minasragra, Peru. It occurs in deep red lumpy aggregates of microscopic crystal needles ; it is derived from the oxidization of the sulphide patronite. Specific gravity 2°55; melts easily to a dark red liquid; slightly soluble in water. A similar red ore of vanadium, Wetahewettite, has been found in the Paradox valley, Montrose Co., Colo., and elsewhere over a wide area extending into Utah. For these two minerals the same composition, Ca0.3V,O,.9H,O is obtained, assuming the maximum content of water at room temperatures; but both are found to be very sensitive to atmospheric changes of humidity. The course of progressive dehydration over sulphuric acid has been minutely studied and the marked difference in this regard is the reason for the difference in name given. Pascoite, also from Minasragra, occurs in clusters of crystal- line (monoclinic ?) grains of a dark orange color. Hardness 2°5 ; specific gravity 2°457. It melts to a deep red liquid and dissolves easily in water. Analysis gave: V2.0; MoO; CaO H.O 100 = H.O + undet. 64:6 10°3 12°6 13°8 78 [0-9] = 100. The calculated formula is 2Ca0.3V,O,.11(?)H,O. — Proc. Am. Phil. Soc., liii, 31, 1914. Piyrapoire and UvanirE are new vanadium minerals from Utah described by Hess and Schaller. Pintadoite forms a thin green efflorescence on the sandstone cliffs of the Canyon Pintado, San Juan Co. An analysis (Schaller) gave: V2.0; CaO HO 42-4. 22°6 300 = 100. The calculated formula is 2CaO.V,O,.9H,O. Uvanite is similar to carnotite in appearance and occurrence, but has a brownish yellow color; it has been found only at Temple Rock, Emery Co. An analysis (Schaller) gave : V.0; P.O; As,O; UO; CaO MgO K.O 4,0 insol. 37°70 «= (006 «S005 83960 «1°78 S004 2S 0380) = 1828 =1:24=99-00. After the deduction of impurities, the formula 2U0,.3V,0,.15H,O is calculated.—_Jowr. Washington Acad. Sci., iv, 576, 1914. Botany. on 7. dn Amatew’s Introduction to Crystallography ; by Sir Witiiam Puirson Beare. Pp. vi, 220, figs. 126. London, 1915 (Longmans, Green and Co.).—This book, as its title indi- cates, is intended for the use of the non-scientific reader but is nevertheless quite scientific in its treatment. By a somewhat novel method the subjects of crystal axes, indices and symmetry are introduced and illustrated. A brief discussion of the differ- ent crystal classes follows. In appendices more detailed descrip- tions of the methods of crystal calculation and drawing are given. The book is well illustrated. W. E. F. 8. Die 32 kristallographischen Symmetrieklassen und ihre einfachen Formen; by E. A. Wi xrine. Pp. 48, figs. 260, pls. viii. Berlin, 1914 (Gebriider Borntraeger).—This is a second edition of the work, the first having been issued in 1895. The text, now added to the original work, includes brief discussions of crystal symmetry, the division of crystal forms into the thirty- two classes and of the simple forms characteristic of these classes. The tables, which are bound up separately in a small atlas, illus- trate diagrammatically the matter of the text. W. E. F. 9. Annual Tables of Constants and Numerical Data. Chem- ical, Physical and Technoiogical. Vol. iii. Chapter on Crystal- lography and Mineralogy, pp. 425-446 ; by L. J. Spencer. Pub- lished by Gauthier-Villars et Cie, Paris and the University of Chicago Press, Chicago, 1914.—This is a single chapter of an important scientific publication designed to summarize yearly the new data of physical science. The present volume, No. Ill, covers the year 1912. The mineralogical chapter gives in brief form new mineral analyses, new crystal forms and axial ratios, new determinations of optical constants, of specific gravi- ties, etc. It summarizes also the crystallographic and optical work done on artificial inorganic and organic compounds. W. EL F. Ill. Borany. 1. Transpiration and the Ascent of Sap in Plant; by Henry H. Dixon. Pp. viii, 216, 30 figs. London, 1914 (Mac- millan and Co.).—The valuable and original researches of Professor Dixon on the complex phenomena connected with the movement of sap in plants are here brought together in connected form. He looks upon transpiration as something more than a mere phys- ical process and considers that an active excretion of water by the green cells is involved. In this way a high osmotic pressure is developed in the cells, sometimes amounting to over 20 atmos- pheres. He considers further that a continuity is maintained between the liquids in the green cells and those in the absorbing cells of the root by means of the liquids in the tracheids and other conducting elements of the stem and root; and he lays a great deal of emphasis on the cohesive or tensile strength of this 92 Scientific Intelligence. continuous liquid column. In some cases, according to his estimates, this strength exceeds 200 atmospheres. The high pres- sure maintained in the green cells by transpiration is transmitted through this liquid column but is clearly insufficient to rupture it, and in this way the upward passage of the sap is assured. Some of the most interesting of the experiments described by the author are those connected with the determination of the osmotic pressures in cells. They are based on the relationship which exists between the freezing point of a solution and its osmotic pressure, and in the determination of the freezing point an exceedingly delicate thermo-electric method has been employed. Professor Dixon’s experiments and conclusions are of much import- ance and throw a great deal of light on one of the most difficult problems in plant physiology. A. W. E. 2. A Manual of Weeds; with descriptions of all of the most pernicious and troublesome Plants in the United States and Canada, their habits of growth and distribution, with methods of control ; by Apa EK. Ggoreia. Pp. xi, 593, 386 figs. New York, 1914 (The Macmillan Company).—The main purpose of the present work is to enable the growers of useful and ornamental plants to recognize and combat the numerous weeds which infest farms and gardens. The introductory chapters deal with general statements about weeds, about the financial loss which they cause, about the ways in which they disseminate themselves, and about the use of chemical herbicides. The body of the book, however, is devoted to full descriptions of individual weeds, and to definite methods of controlling them. The author gives in each case the botanical name of the weed, the English name or names, and tabulated information about the time of blooming, the time of seeding, the geographical distribution, and the hab- itat. The descriptions, although written in semi-popular language, are clear and accurate, and the numerous figures which accom- pany them should help make the determination of the weeds an easy task. The concluding pages give a bibliography, a list of poisonous plants, and a glossary of botanical terms. Miss Georgia’s volume is issued in the series of Rural Manuals edited by Professor L. H. Bailey, but it will be found useful not only to those for whom it was written but also to those interested in weeds from a botanical standpoint. A. W. E. 3. Plant-Breeding ; by L. H. Bartry. New edition revised by Artuur W. Gitgert. Pp. xvili, 474, 118 figs, New York, 1915 (The Macmillan Company).—As stated in the historical introduction the first edition of Plant-Breeding was published in 1895, and the present thoroughly revised edition is the fifth. The topics treated, which will give some idea of the scope of the volume, are the following: variation, mutation, hybridization, heredity, methods of crossing plants, and the forward movement of plant breeding. The discussion of variation and of the important part which it plays in originating new varieties is accompanied by a wealth of illustrative material and is especially Miscellaneous Intelligence. 93 to be recommended. The body of the work is followed by five appendixes, which include a glossary, a list of plant-breeding books, a list of periodicals, a bibliography of references related to plant-breeding, and a series of laboratory exercises. A. W. E. 4. Fundamentals of Plant- Breeding; by Joun M. Coutrzr, Pp. xiv, 347. New York 1914 (D. Appleton and Company).— The present book aims to give a thoroughly modern account of evolution and heredity as applied to plant-breeding and agricul- ture in general. It describes clearly the theories of variation, natural selection and mutation, it reviews the recent work done in genetics, and it gives in detail the methods to be employed in securing resistance to drought and to disease. Chapters on for- estry and on the work done by departments of agriculture both in this country and abroad are likewise included. The work pre- sents in a graphic way the remarkable advances in plant-breeding which have been made possible by the scientific study of heredity and related topics, and it outlines some of the results which may be expected in the future. A. W. E. 5. The Principles of Fruit-growing, with applications to practice; by L. H. Bairzny. 20th edition, completely revised. Pp. xiv, 432, 186 figs. New York. 1915 (The Macmillan Com- pany).—The original edition of this handbook was published in 1897, and the appearance of twenty editions within less than twenty years gives evidence of its great popularity. Although most of the book relates to the larger fruits of northern climates, especial attention being given to the laying out and care of orchards, the smaller fruits are by no means neglected. The information given is thoroughly practical in its nature and is designed primarily for those who raise fruit on a commercial basis. A. W. E. ITV. Mrscecuanrous Screntiric INTELLIGENCE. 1. The Carnegie Foundation for the Advancement of Teach- ing. Ninth Annual Report of the President, Henry S. Prir- cHETT, and of the Treasurer, Ropert A. Frank. Pp. vi, 154. New York City, October, 1914 (issued in June, 1915).—The total endowment of the Carnegie Foundation at the close of the last fiscal year amounted to $14,130,000, to which is to be added $1,250,000 specifically devoted to the division of Educational Enquiry. Of the general income of the year,—nearly $700,000,— $635,000 was devoted to retiring allowances and widows’ pen- sions, while $26,500 was carried to surplus. The Educational Division had an income of $50,350, of which all was expended in its work except some $3,320. During the year 29 retiring allow- ances and 15 widows’ pensions were granted, the average grant being $1,648. The total number of allowances now in force is 332, and of widows’ pensions 100. There have been granted 595 94 Scientific Intelligence. allowances since the beginning, the expenditure for this purpose being $3,551,000. The details in regard to the work of the Foundation are always interesting, but even more the discussions given to the special topics with which it is concerned. This is particularly true of the extended remarks by the President on the subject of pensions in general. The movement in the direction of teachers’ pensions has progressed rapidly and thirteen of the states now have such systems ; it is shown, however, that most of these are radically faulty in their provisions for the future, and only that of Massa- chusetts deserves particular commendation. Industrial pensions are also briefly considered, and further the scandalously extrava- gant system of federal war pensions. Much attention is given in the report to the subject of medical schools, to which the Founda- tion has already contributed so largely. The radical change accom- plished in the country since Dr. Abraham Flexner’s critical report was published in 1912 (vol. xxxiv, 96) proves what can be accom- plished by throwing the light of day into dark corners of the educational world. The most important recent work done by the division of Educa- tional Enquiry is that in the study of legal education, which has led to the publication of Bulletin No. 8 by Professor Redlich of Vienna, published some months since (see vol. xxxix, 611). The study of education in Vermont is also spoken of at length ; the bulletin on this subject (No. 7) was issued a year since (vol. . XXxvii, 564). 2, Publications of the Carnegie Institution of Washington. —- Recent publications of the Carnegie Institution are noted in the following list (continued from vol. XXXV1ll, p. 489): No. 203. A study of prolonged fasting ; by Francis G. Bewnepicr. Pp. 416; 5 pls., 47 figs. No. 204. The water-relation between plant and soil; by B. E. Livryesron and L. A. Hawxins. The water-supplying power of the soil as indicated by osmometers; by H. E. Purine and B. E. Livineston. Pp. 49-83; 2 figs., 13 tables. No. 205. Genetic studies on a Cavy species cross; by J. A. DsrLeFseN, with a preparatory note by W. E. Castix. Pp. 134; 10 pls. No. 210. The Absorption Spectra of Solutions as studied by ‘means of the Radiomicrometer; by Harry C. Jones and coL- LABORATORS. Pp. 202; 58 figs., 50 tables. 3. The Crocker Land Expedition. —A committee of the Ameri- can Museum of Natural History, Prof. Henry Fairfield Osborn chairman, calls (May 20) for contributions toward the expense involved in bringing home the staff of the Crocker Land—or George Borup Memorial—Expedition, and in clearing up all out- standing obligations connected with it. A sum of $16,000 is needed to accomplish these ends and the same degree of liber- ality shown in earlier contributions for this Expedition is to be looked for now. It will be remembered that the sudden death of Miscellaneous Intelligence. 95 the leader, Mr. George Borup, caused a postponement in its depar- ture till 1913, when it left New York under the leadership of Mr. Donald B. MacMillan. The grounding of their vessel in the Straits of Belle Isle, in July, 1913, involved an expense of some $11,000, which explains a considerable part of the large sum now needed. The “ George B. Cluett,” the hospital and supply auxili- ary schooner belonging to the Wilfred T. Grenfell Association, has been chartered and will proceed to Etah next summer and bring the members of the Expedition and their collections back to New York. Checks should be made payable to the American Museum of Natural History and sent to the chairman of the committee in charge. 4. Spencer Fullerton Baird. A Biography, including selec- tions from his correspondence with Audubon, Agassiz, Dana, and others; by Witt1am Hratey Datu. Pp. i-xvi, 1-462, with 19 illustrations. Philadelphia and London, 1915 (J. B. Lippin- cott Company).—Spencer Fullerton Baird (1823-1887), father of the U. S. National Museum, the U. 8. Commission of Fish and Fisheries, and the Marine Biological Laboratory at Wood’s Hole, received his B.A. degree from Dickinson College, Carlisle, Penn- sylvania, when he was seventeen years of age. Even at this time he was an industrious collector of birds and in correspondence with Audubon about a new species of flycatcher ; long before his call to the Smithsonian Institution he was known to all natural- ists in the United States. At the age of twenty-four, Baird was apprised by James D. Dana of the possibilities at Washington, and by him recommended to Henry, then secretary of the Smith- sonian Institution, as keeper of the natural history cabinet. However, because of difficulties in erecting the building and because of shortage of funds, the appointment of Baird as assist- ant secretary did not come until July 5, 1850. For thirty-seven years thereafter he was intensely active in upbuilding American natural history and in laying the foundations which led to the establishing of many of the scientific departments of the Govern- ment bureaus at Washington. The present biographer of Baird, Doctor Dall, became acquainted with him in 1862, and in 1865 was attached to the staff of the Smithsonian Institution, where he is still actively at work. He is, therefore, well qualified to be Baird’s historian and to tell us how he lived and worked, with glimpses of his relations to his contemporaries, to the promotion of science, and to great public services. We are here also introduced to nearly all of the pioneer American naturalists, and told how the Smithsonian Institution came to be through the munificence of the Englishman, James Smithson, and how Baird, appointed its assistant secretary at an annual salary of $1500, gradually developed the U.S. Na- tional Museum and the Commission of Fisheries. These great institutions were being built up with little money and scant sym- pathy from the Government, and we learn that the Commission 96 Scientific Intelligence. of Fish and Fisheries was started in 1870 with a grant of $100 and the use of the sloop yacht “‘Mazeppa,” loaned by the New Bedford Custom House. It is interesting to note that this biography of a great and sym- pathetic man is written by one who has himself given ba sea to the public service in science. 5. Publications of the British Museum of Natural Histone — The following volumes have recently been issued (see vol. xxxix, p. 325 and earlier). Catalogue of the Lepidoptera Phalene in the British Museum. Supplement. Volume I. Plates I-XLI.—This volume of plates, admirable in its execution both as to drawing and reproduction, belongs to the volume of text already noticed (I. ¢.). A Revision of the Ichneumonide with descriptions of new genera and species. Part IV. Tribes Joppides, Banchides and Alomyides; by CraupE Moriey. Pp. xii, 167.—This third part of the revision of the Ichneumonide by Mr. Morley embraces three additional families. A colored plate of Joppa nominator, by Mr. Robert Stenton, accompanies the volume. The Syrphide of the Ethiopian Region, with descriptions of new genera and species; by Professor Mario Buzzi. Pp. 146; 28 figs.—The collections of African Syrphide in the British Museum, received from the Imperial Bureau of Entomology, are remarkably rich and complete; they form the basis of this study by Prof. Bezzi. Some sixty new forms are described, making the whole number of species from this region 249; the family num- bers some 2300 species, distributed over all parts of the world. Report on Cetacea stranded on the British Coasts during 1914; by S. F. Harmer. Pp. 16, 4to; 1 text fig., 3 maps.—It is remark- able how many stranded whales are recorded from the shores of Great Britain. In 1914, upto August, some 43 had been noted; after that date the record was largely interrupted by the war. Of the total number of Cetaceans noted the larger part could be definitely determined as to species. Instructions for Collectors: No. 12. Worms. Pp. 23; 17 figs. —The instructions contained in this pamphlet have been drawn up by Mr. H. A. Baylis, assistant in the department of zoology. 6. The Rumford Medal of the American Academy of Arts and Sciences.—It has been recently announced that the Rumford medal of the American Academy has been awarded to Dr. Charles G. Abbot, director of the Astro-physical observatory of the Smith- sonian Institution. This medal, established through a donation from Count Rumford (Benjamin Thompson), in 1796, to the Academy, is annually given for researches in light and heat. OBITUARY. Sir ArtHur Hersert Cuvurcnu, the veteran English chemist and mineralogist, died on June 2 at the age of eighty-one years. Dr. Hueco Mtxter, president at one time of the London Chem- ical Society, died on May 23 at the age of eighty-one years. Dr. Axset S. Stren, director of the Norwegian Meteorological Institute, died on May 10 at the age of sixty-six years. Warns Natura Science Estas isHMent A Supply-House for Scientific Material. Founded 1862. Incorporated 1890. A few of our recent circulars in the various departments : Geology: J-3. Genetic Collection of Rocks and Rock- forming Minerals. Mineralogy: J-109. Blowpipe Collections. J-74. Meteor- ites, ; Paleontology: J-134. Complete Trilobites. J-115. Collec- tions.. J-140. Restorations of Extinct Arthropods. Entomology: J-30. Supplies. J-125. Life Histories. J-128. Live Pupae. Zoology: J-116. Material for Dissection. J-26. Compara- tive Osteology. J-94. Casts of Reptiles, etc. Microscope Slides: J-135. Bacteria Slides. Taxidermy: J-1388. Bird Skins. J-1389. Mammal Skins. Human Anatomy: J-16. Skeletons and Models. General: J-100. List of Catalogues and Circulars. Ward’s Natural Science Establishment 84-102 College Ave., Rochester, N. Y., U.S. A. EIMER& AMEND Complete Laboratory Furnishers Chemical Apparatus, Balances, etc. C. P. and T. P. Chemicals: and Reagents Best Hammered Platinum Ware, Blowpipe Outfits and Assay Goods WE CARRY A LARCE STOCK OF MINERALS FOR BLOWPIPE WORK, ETC. EST’B - 1851 - 203 -211- THIRD -AVE NEW-YORK- CITY. CONTENTS. Page Art. ].—Factors in Movements of the Strand Line and their Results in the Pleistocene and Post-Pleistocene ; by JeBARRELL:. 6. ess 2. Se Se er 1 II.—Heat of Formation and Polymerization of some Oxides and Determination of the Heat of Combination of Water. by Fusion with Sodium Peroxide ; by W. G. Mixter __ 23 III.—A Study of the Relations existing between the Chemi- cal, Optical and other Physical Properties of the Mem- bers of the Garnet Group ; by W. E. Porp..._.-.__.__ 338 1V.—The Lower Ordovician (Tetragraptus Zone) at St. John, New Brunswick, and the New Genus Protisto- graptus.; by F. H.cMcCLasRNe. 2: 2.53 22 oe ee 49 V.—A Study of the Recent Crinoids which are Congeneric with Fossil Species ; by A. Ht. Crark ___-../ 2 202223 o0 VI.—Relation between the Maximum and the Average Bathy- metric Range, etc., in the Subfamilies and Higher Groups of Recent Crinoids:;‘by Aci. Ciark 221205559) eeeee 67 VII.—Separation of Potassium and Sodium by the Use of Aniline Perchlorate and the Subsequent Estimation of the Sodium ; by D.U. darts. ie 75 SCIENTIFIC INTELLIGENCE. Chemistry and Physics—Prout’s Hypothesis, W. D. Harkins, 78.—Action of Chloroform upon Metallic Sulphates : Explosives, A. MarRsHALL : Chemical Technology and Analysis of Oils, Fats, and Waxes, J. Lewkowi1Tscu, 79.— Annual Reports on the Progress of Chemistry for 1914 ; X-Ray Band Spec- tra, H. Wacner, 80.—EHlements of Optics, G. W. PARKER: Dielectric Phe- nomena in High Voltage Engineering, F. W. Prrx, Jr., 82.—Radium- Uranium Ratio in Carnotites, S. C. Linp and C. F. Wuirremore, 88. Geology and Mineralogy—Climate and Evolution, W. D, Marramw, 83.— Publications of the U. S. Geological Survey, G. O. SmirH, 85.—U. S. Bureau of Mines: Canada Department of Mines, 87.—Lavas of Hawaii, W. Cross, 88.—Brief Notices of some Recently Described Minerals, 89.— Amateur’s Introduction to Crystallography, W. P. BnaLe: Die 32 kris- tallographischen Symmetrieklassen und ibre einfachen Formen, E. A. Wtuirinc: Annual Tables of Constants and Numerical Data, Chemical, Physical and Technological, 91. Botany—Transpiration and the Ascent of Sap in Plant, H. H. Drxon, 91.— Manual of Weeds, Apa E. Groreia: Plant-Breeding (Bailey), A. W. GiL- BERT, 92.—Fundamentals of Plant-Breeding, J. M. CouLTER: Principles of Fruit-growing, with Applications to Practice, L. H. BAILEy, 93. Miscellaneous Scientific Intelligence—The Carnegie Foundation for the Ad- vancement of Teaching; Ninth Annual Report, H. S. PritcHEeTT, 93.— Publications of the Carnegie Institution of Washington : Crocker Land Expedition, 94.—Spencer Fullerton Baird ; A Biography, ete., W. H. Dat, 95.—Publications of the British Museum of Natural History: Rumford Medal of the American Academy of Arts and Sciences, 96. Obituary—A. H, CourcH: H. Mturmr: A. S. Stein, 96. Library, U. S. Nat. Museum. | See? eaenien n Instita pm. ‘ Ss is AUG 2 1915 _ Eprror: EDWARD S. DANA. : ASSOCIATE EDITORS eae GEORGE L. GOODALE, JOHN TROWBRIDGE, Ww. G. FARLOW anp WM. M. DAVIS, or Camprincz, 'ROFESSORS ADDISON E. VERRILL,. HORACE L. WELLS, LOUIS V. PIRSSON, “HERBERT E. GREGORY AND HORACE S. UHLER, or New Haven, ~Proressor HENRY S. WILLIAMS, or IrHaoa, _Proressor JOSEPH S. AMES, or Battimorsg, Mr. J. S. DILLER, or Wasuinerton. FOURTH SERIES _ VOL. XL—[WHOLE NUMBER, CXC}. No, 236—AUGUST, 1915. NEW HAVEN, CONNECTICUT. 1 TITLE, MOREHOUSE & TAYLOR CO., PRINTERS, 123 TEMPLE STREET. Seeenehly. Six dollars per year, in advance. $6.40.to countries in the 5 $6.25 to Canada. Remittances should be made either by money orders, ters, or bank checks (preferably on New York banks). IMPORTANT TO COLLECTORS I take pleasure in announcing that the large collection of minerals recently received has been thoroughly gone over and properly labelled and is now ready for sale. This collection consists of over 1,000 specimens of excellent quality, some of them from old finds, and almost all very well crystallized. Let me know what you desire and I shall be pleased to send you a selection on approval. Indian Relics In addition to the minerals, a large assortment of Indian relics were also included :— arrow heads, tomahawks, spears, celts, ceremonials, pipes, — scrapers, pestles, implements, obsidian knives, etc., etc. Also an assort- ment of Mexican relics of all descriptions. Are You Interested in Gems If so, you will find my stock now richer than ever before in beautiful examples suitable for both jewelry and specimens. Among the synthetic gems in my stock, all of which are of the finest quality, I have the follow- ing: Rubies, pink, white, blue and yellow Sapphires, and the latest dis- covery, the alexandrite. I have an unusually large stock of common and rare Semi-Precious and Precious Stones, both cut and in the rough. Iam able to supply any gem desired, in best quality and all sizes. - New Minerals HODGKINSONITE:—I have been fortunate enough to secure the best specimens of this very rare mineral. It is from the celebrated Franklin Furnace Mines and is a rare compound, the formula of which is Mn(Zn OH).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. BETAFITE:—A member of a group of cubic minerals, niobo- tantalotitanites of uranium, ete., including also blomstrandite (of G. Lind- strom 1874) and samiresite (q. v.); they are closely allied to pyrochlore and hatchettolite, but differ from the former in containing titanium. Betafite is a hydrated niobate and titanate of uranium and occurs in pegmatite near Betafe, Madagascar. Named after the locality. Any of the above which may be desired for selection I shall be glad to send on approval to patrons and customers. Information and prices of individual specimens cheerfully furnished upon request. ALBERT H. PETEREIT 81-83 Fulton St, New York City THE Sena AMERICAN JOURNAL OF SCIENCE [FOURTH SERIES.] ——_—_ $4 — Arr. VIII.—The Lgneous Origin of the “ Glacial Deposits” on the Navajo Reservation, Arizona and Utah; by Herserr EH. Grecory. Introduction. Tue lowest recorded level reached by Pleistocene glaciers of Utah and Arizona south of Lat. 40° is 8500,* and it was there- fore a matter of considerable interest when glacial deposits were reported from the Chinle Valley in northeastern Arizona at an elevation of less than 5000 feet. The area in which the so-called glacial deposits occur is the home of the pyrope garnets or “Arizona rubies” exported from the Navajo Reservation, and a report by Sterrett on the pro- duction of precious stones in the United States calls attention to the unusual character of the material which overlies bed rock in the garnet fields. “The drift is over 100 feet thick and is composed of bowlders which vary from stones weighing many tons to cobble size, mixed through a matrix of pebbles and sand. The gravel and bowlders consist of biotite granite gneiss, porphyritic biotite granite gneiss, hornblende or diorite gneiss, partly epidotized trap and basaltic rocks, epidote hornstone, soapstone, tremolite asbestos, sugary quartz, and large blocks of light gray colored fossiliferous lime- stone of Carboniferous age. Just where the origin of this con- glomeration is to be sought is not known. The general appear- ance of the drift is that of a glacial deposit. Glaciation has taken place in the San Francisco Mountains of Coconino County, Arizona, and moraine deposits have been formed.t The latter are thought to be of rather recent age, probably Quaternary. Whether there has been glaciation in the slightly higher country * Dutton: Geology of the High Plateaus of Utah, 1880, p. 42. + Atwood, W. W.: Glaciation of the San Francisco Mountains, Arizona, Jour. Geol., vol. xiii, pp. 276-279, 1905. Am. Jour. Sc1.—Fourta Srrtes, Vou. XL, No. 236.—Aveust, 1915. - 98 Gregory—Lgneous Origin of the“ Glacial Deposits” west and northwest of the garnet deposits is not known. It is probable that the garnet-bearing drift deposits are of greater age than the glacial deposits of the San Francisco Mountains, for the. former are covered with a stratum of hard white sandstone, and are at almost as great an elevation as any of the surrounding region. ‘The presence of such quantities of crystalline and ancient rocks in the drift cannot be explained by very recent action, as these rocks do not outcrop near the locality.”* The belief in the probable glacial origin of the gravels asso- ciated with the garnets of the Navajo Reservation was strength- ened by the discovery of similar materials near the San Juan River at an elevation of 4800 feet. The deposits at this local- ity have been described by Woodruff. “Tn the southeastern part of the [San Juan Oil] field there is a small area covered by debris which is believed to be of glacial origin. ‘This material consists chiefly of fragments of sandstone, but includes also a considerable amount of shale and lesser amounts of interbedded limestone, conglomerate, and igneous rocks. The conglomerate is of two types—(1) well-rounded pebbles, in general similar to the Dakota conglomerate which is exposed to the northeast of the field, and (2) apparently meta- morphosed conglomerate. Neither type resembles any of the other rocks exposed in this field. The igneous rocks comprise schist and gneiss. The deposit is a heterogeneous mass which shows no evidence of bedding, though some of the constituent blocks show traces of their original stratification. Fragments vary in size. One large block of conglomerate was found to be more than 100 feet in length. ‘The mass rests in an old channel carved in the Moencopie and Dolores formations. The general trend of the channel is north and south, and it terminates at the south abruptly against a wall of shale and sandstone. Small gar- nets were found in the anthills on the top of the debris. These garnets are of interest in a study of the origin of the deposit, because similar ones were found scattered over the surface and in fragments of schist in the southeastern part of the field on the divide between Gypsum and Chin Lee creeks, where it is crossed by the wagon road. Sterrett suggested the glacial origin of garnet beds immediately south of the San Juan field, where rocks similar to the igneous rocks in this field are scattered over the surface.” + Granting that the deposits mentioned are glacial, the inter- est of the problem presented by the interpretations of Sterrett and of Woodruff is two-fold: ; 1. If the glaciation be assigned to Pleistocene times, these discoveries are remarkable in several respects. In the first place, a new record for the lower limit of the ice cap south of *Sterrett, D. B., Min. Res. U. S., for 1908, Pt. II, p. 826. + Woodruff, E. F.: Geology of the San Juan Oil Field, Utah, U. S. Geol. Survey, Bull. 471, pp. 85-86, 1912. on the Navajo feservation. 99 latitude 40° is established, viz.: 4800 feet. Again the effect of climatic conditions producing glaciation must have been peculiarly localized -along the Arizona—Utah boundary. The nearest known glacial materials are in the La Plata Mountains, about 100 miles to the northeast at an elevation of 8500 to 8800 feet and at San Francisco Mountain, 160 miles to the southwest where the terminal moraine of a glacier two miles in length rests at an elevation of 9200 feet. Carrizo Mountain, 40 miles east of the garnet fields, and 4620 feet higher in eleva. tion, and Skeleton Mesa, 35 miles west and about 3000 feet higher, are without evidences of glaciation. Moreover the association of materials composing the drift of the garnet fields is unlike that reported elsewhere from Arizona, Colorado, New Mexico, or Utah. (2) The statement of Sterrett, that “ the garnet-bearing drift deposits... are covered with a stratum of hard white sand- stone,”’* points to a pre-Quaternary period of glaciation not elsewhere recognized in Arizona or Utah. The erratics on the border of the glaciated (?) areas described by Sterrett and Woodruff were noted by the writer in 1910. In 1913 the outer edge of the “drift” at Garnet Ridge was mapped with the assistance of Mr. K. C. Heald, but the sear- city of water and the demands of the work in hand permitted no more than a superficial examination. During the past sea- son a desire to examine the geologic features of the lower Chinle Valley and to study the problem presented by the glacial (?) deposits was realized in consequence of a grant from the Dana Research Fund of Yale University. Geography. Three areas covered by erratics have been located, one south and two north of the Arizona—Utah boundary line in longitude 109° 45’ (see map, fig. 1). The southernmost field has a super- ficial extent of about 1°2 square miles, but the erratics are dominant only in a belt one-half mile long and one-fourth mile wide at the eastern end of Garnet Ridge,t and on the other hand isolated bowlders are to be found beyond the limits of the area mapped. The Mule Ear field contains about -25 square mile, in which bowlders are most abundant on the high ridge form. ing the west wall of Mule Ear Pass. The erratics of the Moses Rock field are strewn along a narrow, irregular belt six miles in length. The commanding topogr: aphic feature of the region is “the Comb,” a cuesta which forms the eastern boundary of Monument Valley. The face of the cuesta is a wall of massive * Loc. cit., p. 825. + The topographic and geologic terms appearing in this article are those adopted for use in for theoming - reports on the Geography and the Geology of the Navajo Reservation. 100 Gregory—Igneous Origin of the “ Glacial Deposits” Fie. 1. LATITUDE 37° 13’ APPROXIMATE LONGITUDE 109°40' APPROXIMATE 1) 5 S00 Soo, ARIZ-_ fies €72% Noe 25 »\ s Ss. en Stesrare Ab 1390 Mi 2 & Xo, ote. ae i) ° 1 2 Scale of Miles Contour interval 200 feet Fic. 1. Map of a portion of northern Arizona and southern Utah, show- ing the location of fields of erratic bowlders. Base compiled mainly from reconnaissance topographic sheets of the United States Geological Survey, and from Map of the San Juan Oil Field by E. G. Woodruff. The dissected ridge extending southward through Mules Har, Moses Rock and Garnet Ridge is the Comb Monocline; Monument Valley extends westward beyond Gypsum Creek, Since the preparation of this map the spelling of Chinli has been changed by official action to Chinle and of Kayenta to Tyende, on the Navajo Reservation. 101 red rock rising 300 to 500 feet above the valley at its western base and presenting slopes exceeding 60 degrees. The ridge is unbroken except where trenched by” the rock-walled canyon of the Chinle, and its crest is set with red sandstone teeth cul- minating in a massive projection locally known as Mule Ear. West of this primary cuesta are a series of hogbacks in parallel position formed of the upturned edges of eastward- dipping sandstone strata. The surface of the entire area is trenched by closely-spaced, deep-cut canyons with perpendicular walls; and innumerable mesas and buttes occupy the inter-canyon spaces. South of Mule Ear Pass the water channels are tributary to the Chinle, either directly down the back slope of the Comb Monocline or by devious routes among the ridges to the west. Drainage north of Mule Ear enters the San Juan, the master stream of the northeastern portion of the Navajo - Reservation. The Chinle canyon is deeply filled with alluvium into which an inner canyon 10-80 feet deep has been cut ; elsewhere a mantle of soil is lacking except for mconspicuous patches of recently disintegrated rock and piles of scattered dunes. Vegetation is therefore scanty except for isolated tufts of grass and hardy weeds which spring up rapidly following showers. Greasewood and yucca in widely-spaced groups are dominant in the lower lands; on the crest of the Comb, sage, pifion and cedar are able to maintain themselves. The absence of soil, the steep gradients of the canyons, the scant annual precipitation which is made up of sudden short- lived showers, and the high values of rock absorption and of evaporation give to the water courses in this region their typical desert character, viz.: well-formed channels admirably adapted for carrying water but functioning only for a few hours at a time. The fields of erratics may be reached from stations on the Denver and Rio Grande Railway at distances estimated ‘as follows: Farmington, N. M., 80 miles; Dolores, Colorado, 110 miles ; Thompson Springs, Utah, 160 miles. From the south a passable road, about 190 miles long, extends from Flagstaff, Arizona, via Tuba and Kayenta. From Gallup, N. M, the field may be reached, via Fort Defiance and Chinle, by a traverse of approximately 120 miles. The scarcity of water and of feed for stock, rather than distance, are the significant factors of travel in this region. The Moses Rock and Mule Kar fields may be studied with Moses Rock Spring as a base ; in the vicinity of Garnet Ridge no water is available except that remaining in rock pockets for a few days following heavy rains. Saddle horse, pack train and a competent guide are essential accompaniments of detailed work. 102 Gregory—Igneous Origin of the “ Glacial Deposits” General Geologic Relations. The members of the stratigraphic column represented in the area under discussion embrace the following: the Goodridge formation (Pennsylvanian); the Moenkopi formation and the De Chelly sandstone (Permian); the Shinarump Conglomerate (Upper Triassic); the Chinle formation (Upper Triassic); the La Plata Group (Jurassic); and the McElmo formation (Jurassic or Lower Cretaceous). The Cretaceous strata out- cropping at Carrizo Mountain and the Tertiary sediments of the Boundary Mountains to the east, which doubtless formerly extended over southern Utah and northeastern Arizona, are not present in the Chinle Valley. The deposits of erratic bowlders are associated with all the formations represented. In the Mule Ear area they are found upon and within the Moenkopi, the De Chelly, and the Chinle formations; in the Moses Rock field they occur upon and within the Moenkopi and rarely overlie Goodridge strata. At Garnet Ridge the Navajo sandstone, the upper member of the La Plata Group, is dotted here and there with i igneous erraties ; portions of the McElmo floor are thickly strewn with bowlders, and beds and lenses of “glacial conglomerate” are incorporated within the McElmo sediments. Structurally the area under discussion is part of the Monu- ment uplift, the eastern border of which is outlined by the Comb Monocline. In the Moses Rock field the Moenkopi strata dip eastward at angles of 4 to 20 degrees; at Mule Ear the beds are upturned at angles exceeding 50°; Garnet Ridge, on the back slope of the monocline, is composed of sedimentary beds whose eastward dip averages between 2° and 8°. A fault with slight displacement traverses the Moses Rock field, as noted by Woodruff, and minor displacements of strata at Mule Ear and Garnet Ridge, accompanied by landslides, have resulted in’ producing masses of jumbled rocks in greatly confused arrangement. DESORIPTION OF THE “ GLACIATED ” ARBAS. The Mule Har Field. As mapped by Woodruff, the “ glacial ” deneeee at Mule Ear Pass cover an area of about one-fourth of a square mile and rest “‘in an old channel carved in the Moencopie and Dol- ores formations.”* Approaching this field from the south, attention is attracted to the high ridge forming the west wall of the pass, and trending parallel with the dominating cuesta of Comb Monocline (fig. °2), Throughout its course, for many miles, this ridge is formed by upturned edges of massive De * Bull. 471, p. 86. on the Navajo Reservation. 108 ‘Chelly sandstone and Shinarump conglomerate in relations normal to the Navajo Reservation. At the pass, however, the usual wavy and serrate crest of the ridge disappears and its place is taken by a disorderly array of massive blocks of sandstone and limestone ; red or buff in general color, but in places painted black by a thick coat of desert varnish. Except for a distance of about one-fourth of a mile the sedimentary series is undis- turbed ; the underlying Moenkopi is followed in regular ascend- Fig. 2. Fic. 2. Mule Har Pass viewed from the south. The rock in the left fore- ground is De Chelly sandstone and Shinarump Conglomerate; the valley is cut im Chinle shales ; the massive rock walls in the right half of the picture, including the two points of Mule Har, are formed of La Plata sandstone. The erratics cover the broken ridge in the extreme upper left-hand corner of the view. : ing order by De Chelly sandstone, Shinarump Conglomerate, Chinle shales and the three members of the La Plata Group. In ascending the ridge it is found that erratics are thickly strewu over the slopes and extend into the valley carved from the soft Chinle shales. Fragments and bowlders of granite, granite gneiss, oarnetiferous diorite, slates, phyllites, ” schists and fossiliferous limestone are represented by numerous indi- viduals. On the crest of the widened ridge and along its upper flanks bowlders and blocks of these materials are found in abundance. One granite bowlder three feet in diameter 104 Gregory—Igneous Origin of the “ Glacial Deposits” was noted and angular blocks of Carboniferous limestone 20 to 100 feet in long diameter are not uncommon. This accumula- tion of igneous and metamorphic rock appears strangely out of place since no materials of these types are found 7m setw within the limits of the Navajo Reservation or beyond its borders for a distance of 100 miles from Mule Ear. Among the materials capping the ridge masses of congloree erate, consisting of the materials represented in the “ drift,” were noted, and where outcrops are favorably displayed lenses and stringers of this conglomerate appear wedged between the blocks and tightly plastered against limestone and sandstone fragments. At the immediate contact between the conglom- erate and the sedimentary rock, the limestone is in places dis- colored and partially er ystallized and the sandstone is baked or even altered to a vitreous quartzite. In the jumbled mass it was found impracticable, with the time at our disposal, to map the seattered outcrops or to determine their inter-relations. Across the top of the ridge dark green bands of decomposed rock were noted, which on examination proved to be composed of fragments of igneous, metamorphic, and sedimentary rock set in a groundmass of finer materials. These bands, 2 to 6 feet wide, intersect the strata at various angles, have distinct borders, and are sharply differentiated from other portions of the terrane by their color. At no point was the material com- posing these streaks consolidated at the surface or in the two- foot trenches dug to determine their character. On the east side of the ridge one of these bands, traversing the strata in a vertical direction, was found to consist of fragments of granite, gneiss, schist, quartzite, and limestone held together by a paste of basic igneous fragments. It will be seen from the above description that the conglomer- ate is intrusive in origin and that the green bands intersect- ing the upturned Mesozoic sediments are dikes of somewhat unusual aspect. The material exposed is rather a heterogeneous mass of sedimentary fragments intersected by stringers and lenses of conglomerate and paste rather than an igneous intru- sion carrying inclusions of foreign rock. The number of dikes and their mutual relations was not determined, but the field relations suggest a plexus of contemporaneous dikes associated in a general intrusion. It is possible that the intrusion par- takes of the nature of a volcanic neck or pipe, and that blocks of deep-seated rock were carried upward along a poorly-defined vent. The fact that some of the bowlders of limestone are of greater diameter than the width of any of the dikes observed is in harmony with this assumption. All the erraties of the Mule Ear field may be accounted for on the assumption of igneous intrusion. The blocks of lime- on the Navajo Leservation. . 105 stone carrying Productus cora, Spirifer rocky montanus, Pugnax Utah, and other characteristic Pennsylvanian fossils are identical with the strata of the Goodridge formation found in the San Juan Canyon, at a stratigraphic horizon several hundred feet lower than their position at Mule Ear. The igneous and metamorphic fragments are from a much greater depth. The Moses Rock Field. One of the sources of the “* Arizona rubies” (pyrope garnets), offered to traders by the Navajo Indians, is a broken line of rounded knobs about two miles west of Moses Rock, Utah, In this region,—an area which has a north-south extension of about six miles and a width varying from less than five feet to one and one-half miles,—the bed rock is mantled by drift made up of fragments of granite, basic igneous rocks, a variety of gneisses, schists, and slates, and blocks of limestone and sand- stone. On the west the field has a somewhat definite border ; on the east the erratics are distributed along the water courses or in mesas and ridges and mounds in an apparently capricious manner. In the canyons northwest of Moses Rock the erratics, at first sparingly distributed on the dry valley floors or dis- played on intercanyon mesas, increase in abundance westward to the crest of the ridge whose rounded and broken summit is buried in gravels, cobbles and bowlders, and square blocks from one inch to three feet in diameter. On the extreme sum- mit of the ridge blocks of limestone 10 to 20 feet in width are mingled with the finer materials. A reconnaissance of the field indicated that the erratics were concentrated along a nar- row irregular band with an average north-south trend and that the bowlders strewn over the eastern portion of the field owe their position to the storm water torrents which add their floods to the Chinle. The “drift ”-covered belt consists of three dis- tinet portions : 1. The northern portion, two miles in length, is a band 2-5 feet in width, consisting of an agglomerate of buff limestone and sandstone containing Pennsylvanian fossils. Included within this mass are fragments of red shale, sandstone, quartz- ite. minette, granite, and gneiss. Viewed from a distance this belt appears as a yellow-green streak crossing canyon and flat and ridge, sharply outlined against the dark red sandstone and shales of the Moenkopi formation. Its trend is N. 10° W., but numerous offsets of 10-50 feet give to the line a zigzag form. This conglomerate band is even with the surface throughout most of its extent, but on canyon walls it is re- placed by a trench and at a few places it stands above the sur- face as the core of small mesas. The band clearly cuts the 106 Gregory—Igneous Origin of the “ Glacial Deposits” country rock as a dike, and although no consolidated material was found either at the surface or by digging shallow trenches, the intrusive origin of the material is scarcely a matter of doubt. It differs from other dikes of the region only in the fact that the igneous paste constitutes probably less than five per cent of the mass. It is essentially a pudding of sedimentary fragments holding igneous plums. 2. The largest accumulation of erratics oceurs at a point about two miles northwest of Moses Rock. At this locality the dike-like band is replaced by mounds of broken strata deeply covered with erratics and occupying an area about 1000 feet long and 500 feet wide. Four hills, highly irregular in outline and arranged along a north-south axis, rise 100-200 feet above the surrounding surface and are capped by chunks of limestone and sandstone 10 x 30 x 6 feet to 2x 2x4 feet, arranged at various angles. Within this area both native and foreign rocks are piled in confusion, while on and among them are found the erratic gravels including fragments of rock of various types. All about this area of jumbled blocks the Moenkopi strata are undisturbed, displaying their normal east- erly dip. With the drift-covered knolls the N. 10° W. direc- tion of the belt of erratics abruptly ceases and the band of regular mounds and ridges made of broken strata tilted at various angles extends westward for about one-half mile. This second portion of the erratic belt is believed to have the same origin as the narrow band at the north, from which it differs mainly in width and in quantity and variety of foreign material. 3. The southernmost portion of the Moses Rock field extends southwest, south and southeast in a sweeping curve, nearly four miles in Jength. At the north end, and especially at the south end of this line, the erratics form a belt of low knobs, in places merging into the sand-covered plain. For a distance ot bout one mile the drift forms a ridge which stands 20 to 40 feet above the country at its base,—a ridge coated with ash- grey to green-erey gravels, contrasting strongly with the dark red Moenkopi strata upon which it appears to be resting. The strata both east and west of the grey-green ridge are seemingly in undisturbed position. The materials forming the crest of the ridge and strewn over its flanks include many of the erratics elsewhere noted in the Moses Rock area. Numerous fragments of chert and chalcedony are present and there is a large amount of mica-diabase or minette. Like other parts of the Moses Rock belt, this southern portion is thought to be located along a dike and the ridge is believed to owe its pres- ence to remnants of the country rock made resistant by local metamorphism. The presence of fragments of diabase or mi- — nette, in places constituting 15 per cent of the debris, suggests on the Navajo Reservation. 107 that this intrusion is similar in composition to that of the dikes and necks of Monument Valley 20 to 50 miles to the west. Along the axis of the Moses Rock field the Moenkopi strata are faulted as indicated by Woodruff.* The amount of displacement is difficult to determine because of the ab- sence of distinctive beds within the 500 or 600 feet of strata exposed. At the north end of the field the downthrow on the east appears to be less than ten feet; further south the displacement is somewhat greater but probably does not exceed 200 feet, the maximum assigned by Woodruff, and may be con- sider ably less. North of Moses Rock terraces along the Chinle and its west- ern tributaries are heavily floored with gravel consisting of monzonite, quartzite, and sandstone. These materials find their origin in the Carrizo Mountains, about fifty miles distant, ; as measured along the streams, and are unrelated to the “dr ift” under discussion, “from which ‘they differ both in composition and origin. The erraties of the Moses Rock field are believed to have no connection with Pleistocene or Mesozoic glaciation, but to owe their origin to a dike or possibly a group of dikes intruded into Pennsylvanian and Permian strata. Tue Gargnet RipGE Frexp. Character of the “drift.’—In the lower reaches of the Chinle Valley the Navajo sandstone is laid bare over wide areas along the eastern limb of the Comb Monocline. At one locality about three miles south of the Utah line overlying strata remain in the form of a terraced ridge which terminates in aseries of rounded buttes. This area is the principal source of the Arizona garnets of commerce.+ The remarkable feature of the district is the presence of fields of bowlders, hills capped by gravel of an unusual aspect, and beds and lenses of conglom- erate of a character not observed elsewhere. This erratic con- glomerate covers the east end of Garnet Ridge, forms a talus on its south and west flanks, mantles the adjoining buttes, and forms a thin disconnected cover of the wind-swept Navajo sandstone extending from Garnet Ridge southward toward Tyende Creek and “eastward nearly to the Chinle. The bowl- ders are most numerous on top of the ridge and buttes and along their sonthern slopes and upon the bare rock floor at their bases. At one point forty bowlders exceeding two feet in diameter are in view and smaller erraties may be counted *Loe, cit., pp. 88-89. + Stezrett, loc. cit. Gregory, H. E.: Garnet Deposits of the Navajo Reservation, to be pub- lished in Economic Geology. 108 Gregory—Ilgneous Origin of the “ Glacial Deposits” Fic. 3. Fic. 8. View of a part of the Garnet Ridge Field, showing distribution and size of bowlders. Note the saddle horse standing among the bowlders. Fic. 4. Fic. 4. Bowlders of granite, granite gneiss, limestone, and sandstone, Garnet Ridge. ; on the Navajo Reservation. 109 by hundreds. Two bowlders of porphyritic granite, 14 x 8 feet and 5 x 8 feet respectively, stand isolated on the dune- covered flat fully a mile from their nearest companions. The field entirely or in part mantled by these materials is about 1:2 square miles, but individual fragments are much more widely scattered. The size and mode of distribution of these erratics are illustrated in figures 3, 4 and 5. Specimens collected from the drift include the following: sandstone, shale, limestone, biotite granite, garnetiferous dio- Fie. 5. Fic. 5. Granite bowlders, Garnet Ridge. rite, diabase(?), minette, granite gneiss, porphyritic granite eneiss, garnetiferous diorite gneiss, muscovite schist, chlorite schist, slate, hornstone, quartzite, garnet, peridot, lustrous feld- spar, quartz, chalcedony, augite, diopside, epidote, soapstone, tremolite, asbestos. None of the igneous and metamorphic rocks included in the list are found in place within 100 miles of this area, and the nearest outcrop of Pennsylvanian lime- stone occurs in the bed of the San Juan river 20 miles north. Tt thus appears that this deposit is local and that its position and character are peculiar. Features suggesting glaciation.—The photographs (figs. 3 and 6) bear out the statement of Sterrett, ‘“The general 110 Gregory —Lgneous Origin of the * Glacial Deposits” appearance of the drift is that of a glacial deposit.” In fact several features ordinarily relied upon to establish the glacial origin of surface deposits are present at Garnet Ridge. The debris is accumulated in mounds or spread unevenly over the surface. It rests in places directly upon the smoothed surface of bed rock, and fragments of country rock are found in abundance among the pebbles. The deposit is heterogeneous; fragments of various materials, in sizes from that of the con- stituents of rock flour to bowlders several feet in diameter, are Fic. 6. Fic. 6. General view of a portion of the ‘‘ glacial field,” Garnet Ridge. Highteen varieties of rock fragments were collected within the area shown. contusedly intermingled ; and several of the larger bowlders are perched in an insecure position. The constituents of the drift are of a variety of shapes; subangular specimens are most abundant; some are rounded, many are angular. Cobbles with smoothed faces and angular edges, also soled and snubbed and polished fragments, are not uncommon, and a few striated pebbles were seen. The conglomerate interbedded with Juras- sic (#) strata is identical in texture and composition with the surficial drift. If the form and arrangement of pebbles be emphasized, the materials described may properly be classed on the Navajo Reservation. af fed with tillites, but as suggested elsewhere* the shape of pebbles is of little significance especially in small outerops. Disregarding the possibility of an igneous origin of the peculiar conglomerate interbedded with the sediments, the assumption of glaciation during Mesozoic time is directly opposed to the physiographic evidence. The color, composi- tion, texture and structure of the sediments indicate a warm, arid climate, and uninterrupted sedimentation. Under such conditions the presence of vigorous glaciers is not to be expected. It is believed by the writer that the demonstration of suitable climatic conditions is an essential feature in the recognition of ancient glacial deposits. Position of the “ glacial” conglomerate in the stratigraphic column.—After confirming the observation of Sterrett that the “olacial” materials not only covered the surface, but were also interstratified with undisturbed sandstones and shales, a section of Garnet Ridge was measured as follows: Section of Garnet Ridge, Arizona. Taken on N, 40° W. line Dip of strata, 40° EH. Z 3° Feet 1. Sandstone, grey to white ; fine, uniform grain, except for millet seed grains cemented on bedding planes and cross-bedding laminae; ripple marks and mud cracks present. Grains of clear, round quartz, white with rare red and black individuals, poorly cemented, cross-bedded at low angles rarely tangential. Inter- sected by seams of calcareous sand 1 /100-—2 inches in width. Prominent joints N. 40° E. and N.40° W. 25 2. Arenaceous and calcareous shales and sandstone. Light red, dark red and brown in alternating bands, 1 to 3 feet in thickness. Cross-bedding poorly defined, but bedding planes irregular and wavy and_ rock traversed by undulating streaks and minute faults ETN OMO Siete gests ee noe Ree tee oes TNS eee 30 3. Sandstone, grey, white and variegated; calcareous, friable, irregularly bedded, imbricated and cross lieolClaC he Gaaeee Ve TREE en Oe errata encom are 10 4. Shales, arenaceous and argillaceous, with variable amounts of sandstone; dark red, light red, green- white, ash grey or variegated in color. Within a few feet along strike shale in places replaced by lenses of calcareous sandstone built of overlapping short laminae==-- 2 ~~ eA eto ae. LD 5. Sandstone, light red, white on fresh fracture ; cal- careous, irregularly bedded, ripple marked ._-_. __-- 2 *Gregory: Note on the Shape of Pebbles, this Journal, xxxix, pp. 300- 304, 1915. 112 Gregory—IRagneous Origin of the “ Glacial Deposits” Feet 6. Shales, argillaceous, calcareous, arenaceous, with 2-14 inch beds and lenses of white and of light red sand- stone ; shales various tones of red and brown, also ash erey. and oreen2= =. 2 52.5). 322 9 = ee 7. Conglomerate, olive-green in tone, compact, but not firmly cemented. The pebbles in sizes up to 16 inches in long diameter are subangular with flattened and in many cases polished surfaces, and include white granite, granite porphyry, granite gneiss, diorite, garnetiferous diorite gneiss, mica schist, chlorite schist, serpentine, asbestos, quartzite, blue-grey marble, grey and buff limestone, red sandstone, red shale ; the matrix consists of tiny particles of rock types present as pebbles, and also flakes of biotite and muscovite, rare garnets aud peridots. Materials of the conglomerate identical except in size with those forming the surface “drift” of this area. The upper and lower contacts of the conglomerate some- what wavy and shales enclosing it are slightly dis- colored. Bed traced by excavation for 200 feet, terminates abruptly at the west---- - Per 2 4 8. Sandstone, fine, even-grained, light red, lenticular ; replaced along the strike by white, sandy shales____ 1 9. Shale, arenaceous ; banded white, green, ash gray, dark red, light red, chocolate ; colored bands in places continuous for several hundred feet; elsewhere distributed as discontinuous stripes or as dots ; beds paper-thin to ¢ inch in thickness ; uneven, wavy, and in places imbricated <== :222-- --sse5=— 49 10. Sandstone, red-brown, fine; well-rounded, quartz erains.; calcareous cement (225 5- -2 == 52 eee 5 11. Shale, arenaceous ; dark red to light brown, with scat- tering patches of black, white and green ; irregularly bedded ; where shale partings are not developed by weathering, strata assume appearance of thick bed- ded, variegated sandstone. 2232 (2-22 -== 25222 ees 12. Sandstone, dark red with white patches ; many curved beds ; massive or irregularly bedded ; friable ; fine quartz, even-grained ; weathers into knobs, hoodoos, stone babies: bobbins] 922-22... 25. se 2 ae eee Et 13. Sandstone, light red, massive, cross-bedded ; Navajo —— sandstone, the upper member of the La Plata 227 Group. The strata included in this section, Nos. 1-12, inclusive, are tentatively assigned to the McElmo formation, Jurassic (?) in age. A section of the southwest end of Garnet Ridge, measured with the assistance of Mr. Heald, is substantially the equiva- on the Navajo Leservation. 118 lent of the one given above, except that No. 7 is not repre- sented. As shown in the diagram (fig. 8), the conglomerate bed (No. 7) could be traced for less than 300 feet, but material identical in character, in an uneven stratum 2 to & feet in thickness (fig. 7), is found about 500 feet farther east beneath the cap of white sandstone (No. 1 of the section), and a bed of the same composition occurs near the top of a detached butte near Fie. 7. Fic. 7. Hast end of Garnet Ridge. The man in the figure is standing on the stratum of ‘‘ glacial” conglomerate. at hand. On following the conglomerate bed by excavations, it was found to turn abruptly upward, and after an offset of 20 feet to extend to the top of the ridge. To the north of the ridge the bed was again found on the floor of a natural ditch. Further examination led to the conclusion that the stratum of conglomerate in both vertical and horizontal position is of 1gne- ous origin—a dike and sheet of abnormal composition. Microscopie examination of the bed furnished corroborative testimony. The stratifieation of Garnet Ridge is undisturbed except for a distance of about 400 feet, within which the sediments are broken into blocks tilted at various angles and rendered more Am. Jour. Sci.—Fourts Series, Vou. XL, No. 236.—Aveust, 1915. 8 114 Gregory—Igqneous Origin of the “ Glacial Deposits” confused by sliding of block ov block and of parts of the mass on the shales beneath. Within this tangled mass of blocks the conglomerate is distributed as chunks and lenses and stringers and sur- ficial patches in a most capricious manner. It was found impracticable to deter- mine the distribution and continuity of the various ig- neous and sedimentary beds involved in this complex, but it is believed that these dis- turbed beds mark the loca- tion of the principal intru- sion. That the assumed intrusion is local, somewhat in the nature of a neck, is suggested by the fact that the beds near at hand, north, south and west of the dis- turbed area, are continuous and in their usual positions. It is possible, also, that a dike extends eastward through the heavily-mantled knolls and buttes, but no defi- nite trace of it was discoy- ered. The age of the intrusion is unknown beyond the fact that it is post-Jurassic, but the field evidence demands no date earlier than that as- signed to the other voleanic features of the Navajo Reser- vation, viz: Tertiary time. As in the case of the Mule Ear and the Moses Rock localities, the igneous origin of conglomerate at Garnet Ridge appears to be demon- strated. Corroborative evidence.— At a number of localities on of the interbedded stratum of conglomerate, The construction of the volcanic pipe is wholly tion, pp, 111-112. , B, and of the bowlders strewn over the surface. rammatie section of east end of Garnet Ridge, showing the positi The numbers on the beds correspond with those described in the sec of the area of disturbed strata fre. 8. Diag theoretical. A, on the Navajo Leservation. 115 within the boundaries of the Navajo Reservation igneous intru- sions in the form of dikes, sheets, laccoliths, and voleanic necks oceur. Few of these exhibit the features ordinarily associated with igneous masses. Many of them are vertical sheets of consoli- dated agglomerate in which massive igneous rock plays a minor role. A few dikes and necks studied consist of over 75 per cent of conglomerate, in which fragments of sedimentary rock exceed in bulk the broken bits of minette, shonkinite, and augi- tite. The most common foreign ingredients are blocks of sand- stone and shale through which the igneous material has made its way, but bits of igneous, metamorphic and sedimentary rocks from unknown deep sources are also in evidence. At Alham- bra Dike, granites and gneisses are embedded in the augite minette. DPlastered against the diabase of Church Rock two varieties of granite gneiss and fragments of chlorite schist, quartzite, chalcedony and augitite were found. Within one of the dikes cutting Comb Monocline garnetiferous gneiss, granite, slate and limestone are found among the inclusions. The igneous masses within Buell’s Park hold inclusions of horn- blende, gneiss, garnetiferous diorite gneiss, marble, slate, mica schist, and chert. It is probable that careful search for inclu- sions within the igneous masses of the Reservation would result in a collection containing nearly all the varieties of pebbles found in the “ drift ” of the garnet fields. Conclusion. Materials resembling glacial deposits on the Navajo Reserva- tion occur as strata interbedded with Mesozoic sediments and also as superficial drift. Their position in the stratigraphic column is believed to be due to igneous agencies ; their distri- bution over the present surface the result of erosion rather than of Pleistocene glaciation. Most of the bowlders are assigned to pre-Cambrian formations. Whether the “ glaci- ated ” forms of the pebbles have resulted from compression in ancient conglomerates, from a pre-Cambrian period of glacia tion or directly from igneous activity, has not been determined. 116 L. Page—Lnergy of « Moving Electron. Art. [X.—The Energy of a Moving Electron; by Leen Page. Ty a previous paper* it has been shown that the equations of the electromagnetic field can be derived in their entirety from two fundamental assumptions, to wit: (a) An ether exists which transmits strains In accordance with the postulate of the relativity of all systems moving with constant velocities ; (0) The elementary charge is a center of discrete, uniformly diverging, tubes of strain. The object of the present paper is to discuss and compare the expressions for the energy of a Lorentz electron moving with constant velocity as obtained by three distinct methods. The first of these methods consists in finding the electric and magnetic energies of the electron’s field. The second consists in subjecting an electron which is at rest relative to the observer to an infinitesimal mechanical foree for an infinite time. In this way a finite velocity is imparted to the electron without any finite radiation of energy. Hence the sum of the initial electrostatic energy of the electron, the work done by the mechanical force in accelerating the electron, and the work done by the ether pressure,—or whatever other force is pos- tulated in order to prevent the disruption of the electron,— in producing the progressive contraction of the electron as its velocity relative to the observer increases, gives the energy of the moving electron. The third method is analogous to that used in finding the electrostatic energy of a charged conductor. Starting with an uncharged moving electron, an infinite num- ber of shells of infinite radius are shrunk down to the surface of the electron. Each shell carries an infinitesimal charge and maintains throughout the process of contraction the same veloc- ity as the electron. The sum of the work done in contracting the shells and that done in maintaining their constant velocity against the retardation of the field gives the energy of the charged electron. . The value of the total energy is, as it must be, independent of the method used. The division into kinetic and potential energies is, however, not the same for the different methods. In the first method the kinetic energy is taken as the magnetic energy of the moving electron’s field. The second method gives for the kinetic energy the work done by the accelerating force. This is the expression peculiar to the dynamics of relativity, and does not agree with that obtained by the first *This Jour., xxxviii, p. 169, 1914. L. Page—Energy of « Moving Electron. 117 method. In the third method the kinetic energy is measured by the work done in maintaining unimpaired the velocity of the contracting shells. Its value is equal to that of the mag- netic energy of the moving electron’s field as given by the first method. Of the three methods the second would appear to be the least trustworthy on account of the necessity of dealing with an accelerated system. If the elementary charge or electron is a uniformly charged spherical shell of radius @ to an observer relative to whom it is at rest, it will appear as an oblate spheroid to an observer relative to whom it has a constant velocity v, the dimensions at right angles to the direction of motion being the same as when at rest,and the dimensions in the direction of motion ‘ : i ——,, v being shortened in the ratio 4/1—’:1 where ¢ = 3 = velocity of light. If a mechanical force be applied to such an electron, the electron’s own field will exert a force opposite and proportional to the acceleration produced. By means of this retarding force Lorentz* explains the inertia mass of a moving electron. For a quasi-stationary state of motion he finds the transverse mass to be given by e m, i —# (1) Mm, = ~ 6 rac’/1—B and the longitudinal mass by e m — = 2S 2 3 (2) érac’ (1— #*)?~ (1— 8)” If the acceleration is finite and in the direction of relative velocity, it must be remembered that the acceleration of the front of the electron is less than that of the rear, since the electron contracts progressively as its velocity relative to the observer increases. The author has shown, in the paper already referred to, that Lorentz’s expressions for the mass hold good for any acceleration which is constant relative to the electron’s own system, provided we take as the acceleration of the electron, not the acceleration of its geometric center, but that of the plane perpendicular to the direction of motion, which divides the electron into two parts having equal charges. The point where the axis of the electron cuts this plane may appropriately be called the center of acceleration. The restric- tion of constant acceleration relative to the electron’s own system is not serious, since a constant acceleration for a time comparable to that taken by light to travel a distance equal to m * Theory of Hlectrons, p. 212. 118 L. Page—Energy of a Moving Electron. the diameter of the electron is all that is necessary in order that the expressions given above shall hold. While the ex- pression for the longitudinal mass has not been verified experi- mentally, that for the transverse mass has been found to hold very exactly for high speed @ particles, first by Kaufmann* and Bucherer,t and more recently to a “high degree of preci- sion by G. Neumann.{ First Method. To calculate the energy of a moving electron by the custom- ary method use is made of the familiar energy equation which follows at once from the equations of the electro-magnetic field and the transformation equation for the force. If vectors be denoted by Gothic letters, the force equation, in Gibbs’ notation, has the form F=p| E+ vxH | (3) where F denotes the force per unit volume, due to the electric strain E and the magnetic strain H, on a ‘charge of density p moving with velocity v. The energy equation is als faa (EXH)- ds +/F- vdr=0 (4) The first term represents the rate of increase in the energy of the electron’s field, the second term the rate of radiation, and the third the rate at which work is done on the electron by its own field. The first term gives for the potential energy of the moving electron’s field — é are a (5) and for the kinetic energy T= = [Hd = Co oie ae (6) 2 12ra 4/1 — g° ) V1—B Hence the total energy of the moving electron is ic 8 Be ae eee w= We OS 72,€ Vink =e, Vink | (7) *Kaufmann, Ann. d. Physik., xix, p. 487, 1906. + Bucherer, Phys. Zeitschr., ix, p. 705, 1908. {Neumann, Ann. d. Physik., xly, p. 529, 1914. L. Page— Energy of a Moving Electron. 119 and the gain in energy due to the motion, obtained by sub- tracting from (7) the electrostatic energy of an electron at rest relative to the obser ver, is given by AW me| Ta —1 [. ++ Mee | (8) The rate of radiation of energy from an electron is given exactly by dh _ €¢ WE 6xc° (9) where ¢ is the acceleration of the center of strain* of the elec- tron relative to its own system. Second Method. Let an electron which is initially at rest relative to the ob- server be given a finite velocity v by imparting to it an infini- tesimal acceleration for an infinite time, and then let the electron be allowed to maintain this velocity forever after. Since the acceleration is infinitesimal, the radiation term will be neghgible compared to the other terms in (4). Hence we may write fel ae alo fet H') dr |+fFvar=o (10) To find the gain in energy we must calculate the work done by the constant mechanical force which produces the acceler- ation. This force is obviously given by K=_fRa= —s (11) where the acceleration, since it is infinitesimal, ean be consid- ered to be the same for all points on the electron. Integrating (11) we find for the work done by the mechan- ical force acting on the electron Ti mel ae | (12) which is not the same as (8). In fact, if v, is the velocity of the geometric center, the work done by the mechanieal force as calculated above is T= — fv.de [Fa *By ‘‘center of strain” is meant that point inside the electron at which its charge can be considered as concentrated without altering the external field. 120 L. Page—Energy of « Moving Electron. whereas the change in energy as given by (8) is aw=| 5 [+m )ar |—-| 5 [ews |=— f{fevarae which is not the same as 7, because for the Lorentz electron Lrvt feaaf [Evaed (13) This inequality means, physically, that in caleulating the work done we cannot replace the forces acting on the elements of the electron by a single force acting at the center or at any other specified point in the electron. This is due, of course, to the fact that the electron is deformable and consequently when it is accelerated, different points on its surface have dif- ferent velocities. In the case of a rigid electron (1. e., one which maintains the same size and shape whatever its velocity relative to the observer), such as Abraham’s, the velocities of all points would be the same, and (13) would be an equality. It is of interest to examine more closely the right-hand side of (13). If the origin be taken at the center of the electron, and the Z axis in the direction of relative velocity, =v E = Se (14) where ¢ is the infinitesimal, constant acceleration of the elec- tron relative to its own system, and primes refer to the - electron’s system at the instant considered.* Also es e a) 7 df= — (.-4"? ). "de (15) Multiplying (15) by v, and integrating over the surface of the electron we get »,f dF, = — 7 Uv, (16) for the rate at which work is done by the electromagnetic forces in resisting the change in motion. Changing the sign of this expression and integrating with respect to : the time, we obtain the work done by the mechanical force producing the acceleration as given by (12). Multiplying (15) by the second term in (14), we get upon integrating over the surface of the electron a7, a-p) fans - op (17) * See ‘‘ Relativity and the Ether,” this Journal, xxxviii, p. 169, 1914. L. Page—Energy of « Moving Electron. 121 which is obviously the rate at which work is done by the electro- magnetic forces in resisting the progressive contraction of the electron as its velocity increases. Changing signs and integrat- ing with respect to the time, we get for the total work done in contracting the electron the expression 4 U, =" [1-vi-# | (18) Adding this to (12), the expression for the work done by the mechanical force in producing the change in velocity, we get, as we should, for the total increase in energy the same expression (8) as obtained by subtracting the initial value of J 2 72 —f@ + Hl) dr from its final value. In accordance with the usual definitions we should consider (12) as representing the kinetic energy acquired, and (18) the increase in potential energy due to the contraction. It is to be noted, however, that the kinetic and potential energies so measured do not correspond to the magnetic and electric energies respectively of the electron’s field. In order to provide a mechanism for produems the contrac- tion of the electron Poincaré* assumed that its surface is subject to a constant hydrostatic pressure —————— 1 2 327° a ( ) This pressure is just sutticient, when the electron is moving with constant velocity, to counteract the electric forces tending to disrupt the electron, and when the acceleration is injinites- wmal, the work done in contracting the electron against the electromagnetic forces due to its own field can be accounted for by the work done by this hydrostatic pressure. Unfortunately, however, Poincaré’s stress loses its significance when an electron is accelerated by a finite mechanical force. If the force is constant, the electron’s field will be such that if the surface of the electron is assumed to coincide with a level surface + of the electromagnetic field—and it is hard to see * Poincaré, Rendiconti del Circolo Matematico di Palermo, xxi, p. 129, 1906. + By ‘‘level surface” is meant a surface everywhere perpendicular to the electromagnetic forces of the field. See ‘ Relativity and the Ether.” ‘This Journal, XXXViii, p. 184, 1914. In the paper referred to, instead of ‘‘ level surface” the term ‘‘equipotential surface” is used. The latter term is, however, objectionable, since the force in an accelerated system is not derivable from a scalar potential and hence there can be no meaning to ‘““ equipotential surface” in the sense of a surface all points of which have the same potential. 122 L. Page—Energy of « Moving Electron. how the charge could be in equilibrium otherwise,—the distribution of electricity on the surface of the electron will not be uniform, and consequently the electric forces tending to disrupt the electron cannot be counteracted by a hydrostatic pressure. If the pressure is not hydrostatic, but of such a magnitude at each point on the snrface as to balance the disruptive forces, it will exert a resultant force in the direction of motion that will exactly balance the retarding force due to the electron’s own field, and in order to satisfy the equation of motion for the electron it would be necessary to introduce a mechanical mass. The most obvious way to avoid this difficulty is to deny the existence of mechanical forces per se, and put everything on an electrodynamic basis, at least in so far as the motion of electronsis concerned. Then in dealing with the motion of an accelerated electron we could not eliminate the external electromagnetic field which was responsible for the acceleration. We should have to deal with two overlapping fields, which would render the problem more complicated. So far as mechanical forces are concerned, it would seem that a body which is subject to a mechanical force must have a mechanical mass. Third Method. We will now calculate the energy of an electron moving with a constant and finite velocity v by a method analogous to that used in finding the electrostatic energy of a stationary charged conductor and without causing the electron to pass through a series of states in which the velocity varies. To an observer relative to whom it is at rest, we have assumed the electron to be a uniformly charged spherical shell. Hence the kinematical transformations of relativity show that it will be an oblate spheroid to the observer relative to whom it has the constant velocity v. Ifthe origin of moving axes is taken at the center of the electron with the Z axis parallel to the direction of motion, the surface of the electron will be that formed by revolving the ellipse pi a 2 wv a a’({1 — B°) (20) about the Zaxis. If @ is the angle made by any radius vector of the ellipse with the Zaxis, the charge on the annular ring between the cones defined by @ and 6+ d@ is easily seen to be e (1—f’) sin 6d6 ee 9 (1—# sin* 6)3 (21) L. Page— Energy of « Moving Electron. 123 In order to find the energy of the moving electron we shall charge it by shrinking down from infinite size to its surface a series of moving shells. Each shell will carry an infinitesimal charge de distributed over its surface in such a way as to give rise to no field inside the shell, and will maintain throughout the process of contraction the same velocity as the electron. The work done in shrinking these shells against the forces exerted by the charge already on the electron will be equal to the energy of the electron in its final state. First it is necessary to show that the energy of m shells of radius # and charge de. where ZF is infinitely great and nde = é is finite, is negligible compared to that of an electron of charge ¢ and finite radius a. A consideration of dimensions alone shows that the energy of one of these shells must be (de)’ proportional to Re while that of the electron itself must be 2 proportional to < . Hence the energy of the m shells will be proportional to “e which is an infinitesimal of the second 2 € ° een : order compared to a This must be true quite irrespective of the velocities of the shells and of the electron, provided these velocities are less than the velocity of light. Hence the energy of the shells of infinite radius moving with velocity w is negligible compared to the work done in contracting them. Moreover in contracting a shell, the work done against the electromagnetic forces due to the shell itself will be of the order oe and hence negligible compared to the work done against the forces due to that part of the charge which is already on the surface of the electron. So in calculating the energy we need consider only the work done against the forces exerted by the charge already on the electron’s surface. Let O (fig. 1) be the center of the electron and P a point on the contracting shell. The electric and magnetic forces at P are given by the familiar expressions _ e (1-8) fe Arr” (1—* sin’ 6)2 (So HH 2 Bsin 6 (1—8’) (23) dar? (1— 8’ sin’ 6)3 In order that the contracting shell shall give rise to no field im its interior, it is obviously sufficient that relative to an 124 LI. Page—tnergy of a Moving Electron. observer moving with it the shell should be a uniformly charged sphere concentric with the electron. Let the infinites- - imal radial velocity of contraction relative to such an observer be denoted by w’ and the radius of the sphere by 7’. Then the Lorentz-Einstein kinematical transformations show that, to an observer relative to whom the electron and shell have the velocity uv, the shell will be an ellipsoid similar to the electron Fie. 1. itself, but with its center displaced relative to that of the elec- tron in the positive Z direction by an amount* _ OQ=0"" Byi-# (24) Nevertheless, the charge on the annular ring between the cones defined by @ and @ + dé (@ being, as before, the angle between the radius vector drawn from the center of the electron and the Z axis) will bear the same ratio to the total charge on the sheli as that on the corresponding ring of the electron’s surface does to the total charge on the electron; in other words, the dis- tribution of charge on the surface of the shell will be that given by (21). Consequently when the shell has contracted down to the surface of the electron and has imparted its charge - to the same, no redistribution of electricity over the surface of the electron will be necessitated. *TIf the shell were not contracting it would be concentrie with the elec- tron, The eccentricity of the contracting shell is due to the smaller velocity on the front, and greater velocity on the back of the shell. L. Page—Energy of « Moving Electron. 125 From (8) we find for the rate at which work is done on the charge contained on an element of the surface of the contract- ing shell dw pase | ae Eve: | (25) Putting Wa UZ. where dU. dr SS = — 9) dt ek lt 2) and aT, i di = ~--deE-v (27) we find, on substituting from (21) and (22) UES Wee) qu. ee BY _ fae a oe 8a 7” (L—B* sin’ @)* ° b ble 3 See ind Vy eee av —B /1—? sin? 6 © Integrating with respect to e¢ to find the work done in shrink- ing all the shells down to the surface of the electron, we find for the total potential energy e 3—° mic 3-6 C= 54 - Zoe = zee (29) ma 4/1 — 8° 4 V/ 13 which is the same as the electric energy of the electron’s field, i.e. 1 Ts ye dr (30) Subtracting from U, the electrostatic energy of a charged con- ducting sphere, we find for the increase in potential energy due to the electron’s motion mC 3— Gr SU = a —3 | (31) (28) It where 0 4 ies B Equation (27) indicates that there is a resultant electric force on the contracting shell in such a direction as to oppose 126 L. Page—Energy of « Moving Electron. its motion. The mechanical force which must be applied to counteract this resistance is easily found to be K=- [Bde _ edew' B 677""C dr’ = but e/g since dt’/=dt »/1— 8 The total work done by this force during the shrinking on of all the shells is seen to be : e 1Oin Meee URKE a ao Pee ae As this is the work done by the force applied to the shells in the direction of motion in maintaining their velocity unim- paired, it may properly be called the kinetic energy of the electron. It is equal in value to the magnetic energy of the electron’s field, i. e. T= 5 [wa (34) The sum of the potential and kinetic energies as determined above is, of course, equal to the corresponding total energy as measured by adding to the electrostatic energy of an electron at rest the work done in imparting to it, by means of an infini- tesimal mechanical force, a velocity v. It is to be noted, how- ever, that the division of the total energy into kinetic and potential is different in the two cases, as is seen by comparing (31) with (18), and (83) with (12). Denoting the transverse mass by m, (83) gives for the kinetic energy the familiar ex- pression (33) 1 2 Gis’ me - (85) instead of the expression (12) peculiar to Einstein’s mechanics. It is of interest to investigate the validity of the general- ized force equation for a contracting shell. If this equation holds as : dv ae since 7, = 0. The mass of the electron plus the mutual mass of electron and shell is e° ede 1 < peas | ee ¥ re | /1— ey) Hence ede _ B i ye tae aS (38) Which is a force in the same direction as that given by (82) but twice as great. However, while the force A, given by (82) is the only force in the direction of motion that does any work, it is not the total force that must be applied to the contracting shell in order to keep its velocity unimpaired. In fact (3) shows that the magnetic field exerts a resultant force on the shell in such a direction as to decrease the velocity. The mechanical force which must be applied to counteract this resistance is ] dr K=— sf % ap x (110) 63 62 49 Crystal No. 2—Crystal No. 2 (Case Collection) illustrates a prismatic type of development in the direction of the brachy axis (a). A drawing of this crystal is shown in fig. 3. The right-hand side of this crystal was more or less broken off, and a photograph of the opposite, or more perfect side, is shown in fie 1, with the basal pinacoid, c(001) at the top, and the macropinacoid, a(100) on the right. This specimen had been glued together at some time previous to its arrival for erystallographie orientation (the full and long dashed line on the faces ¢(001), A(203), p(223), and a(100) represents the glued contact). This specimen was found to contain 12 forms which was more than were observed on either of the other erystals as follows :—the pinacoids, (100), 6(010), c(OOL); the prisms, m(110), 7(810), (140), a(230); the domes, 2(201), A(203), (023); and the pyramids, p(223), and &(214). The forms were identified, and their relationship established on the basis of the following measurements : 150 Van Horn and Hunt—Bournonite Crystals. Contact Measurements Calculated Angles hea = (203) : (100) 57° 57° 297 era =) (201) 1100) 28 27 364 fae == (214))(01) 26° 30 27 50 nc — (223) < (00) 40 41 9 y:b = (028) : (010) 58 591g MeN — (NO) 2 LOO) 43 43 10 7. @ = (310) : (100) 18 17 22 &-b = (140) : (010) 15 14 56 a: 0: ==(230) e010) 36 35 24 ol), = (310) : (010) ALY 30 72 38 pip = (228) = (223) 98 97 42 Crystal No. 3.—Crystal No. 3 (Utah Collection) also shows a short prismatic habit similar to No. 1. A drawing of this crystal is shown in Fig. 4. The forms observed were all large, and well developed, especially the prisms, m and f, .and the basal pinacoid ¢. The 10 crystal forms present were the pinacoids, a(100), 6(010), c(001); the prisms, m(110), 2(320), (120), B40), (810); the dome, 0(101); and the pyramid, u(112). The angular measurements obtained are listed below : Contact Measurements Calculated Angles mu:m = (110) : (110) Cvis 86° 20! Cu = (820) = (10) 31 Be el Vag = (20) 20) 57 56 «68 WO, = (120) 8 (O10) 29 28 4 ®:m = (140) : (110) 62 61 46 CeO == ((OOW)3 (Gon) 43 43 48 Com == (OI) 3 (2) 33 33 15 9 0 == (OO) (010) 72 30’ (Bis) Crystals No. 1 and No. 3, in general, resemble those from Neudorf in the Harz, and Nagyag in Hungary, except that they are inclined to be somewhat more tabular, and that the macropinacoid is less developed. Crystal No. 2 seems to be of different habit than crystals from other localities in its distinct prismatic development in the direction of the braehy axis. The habit of all three crystais is quite different from certain specimens which the writer has recently seen from Cornwall, England; Horhausen, Prussia; and Pribram, Bohemia. Mineralogical Laboratories Case School of Applied Science and University of Michigan January, 1915. Udden— Castile Gypsum and Rustler Formation. 151 Arr. XIV.—The Age of the Castile Gypsum and the Rustler Springs Hormation ;* by J. A. Uppxn. In 1904, Richardson described two formations in Culberson County, Texas, the Castile Gypsum, known to be at least 300 feet thick, and the Rustler formation, some 200 feet thick. He found them extending in a continuous broad belt from the New Mexico border southward to within 15 miles of the Texas and Pacific Railway, the Rustler formation overlying the Castile. He found that the Castile formation is separated from the Delaware Mountain formation (Permian) by an un- conformity.t From the few fossil fragments which were secured from these formations, he was unable to draw definite conclusions as to their probable age, and he stated in his report that “the age of the Rustler formation is not known.” Neither could he make any definite statement as to the age of the Castile formation. His descriptions are the best we vet have. Although he searched for fossils in all exposures examined, only a few poorly preserved fossils were found. Dr. T. W. Stanton, who examined one fossil, found that it might be a Mytilus or a Myalina. Dr. F. H. Knowlton reported on some poorly preserved plant remains as probably being of Mesozoic age. On Willis’ geologic map of North America, the area of the outcrop of these formations is represented as Permian, and in his Index to the Stratigraphy of North America, he states that “recent work has shown that the Castile Gypsum and Rustler formation are parts of the group of red beds of Pecos Valley, which are of Permian Age.” He cites Richardson’s paper : “Stratigraphy of the upper Carboniferous in West Texas and Southeast New Mexico.”{ It appears that Richardson in this paper refers these two formations in question to the red beds, mainly on the basis of their lithologic character, and on field work by other geologists who have traced the red beds around the Staked Plains from New Mexico to Oklahoma and to the Pecos Valley in Texas. The Mesozoic as weil as the Paleozoic formations are poorly exposed in this region, and all who have worked here have recognized the difficulties attendant on making correct cor- relations between distant and limited outcrops, in beds that rarely yield any fossils. It seems pertinent to place on record the finding of some fossils in the Castile formation, which in the writer’s opinion * Published by permission of Dr. W. B, Phillips, Director of the Bureau of Heonomic Geology and Technology, University of Texas, Austin, Texas. + University of Texas Mineral Survey, Bulletin 9, pp. 48-45. ¢{ This Journal (4), vol. xxix, p. 325, 1910. 152 Udden—Castile Gypsum and Rustler Formation. indicate that the Castile Gypsum is not Permian, and show that the Castile Gypsum is not a stratigraphic unit, but contains considerable amounts of marly clay in separate beds. In the latter half of 1914 the Troxel Oil Company made a test hole for oil near the center of the south line of Survey 24, Block 110, on the Public School Lands in Culberson County. This lis about one mile east of Rustler Springs. Mr. C. R. Troxell has furnished the writer a set of cuttings from this exploration. A description of these cuttings is as follows: Description of Samples of Cuttings from Troxel Well No. 1, located near the south line of Survey No. 24, Block 110, Public School Lands, Culberson County, Texas. Depth in feet below surface. From To White gypsum. Slight effervescence with acid-__- 4 8 Dolomite, gray and yellow, of a fine, uniformly sized crystalline texture. In thin section it shows scattered porosities and minute yellow streaks or blotches believed to be bituminous. The sample contains some black, some gray, and some white quartz. Pyrite and gypsum are. present..2 22.2 Seep ee ae eee 42 58 Mostly yellow and in part dolomitic limestone. In thin section this is seen to contain fine sand and some other clear, transparent particles. Some is porous and contains black specks of pyrite. Pyrite and gypsum, some fibrous, are present. Some fragments of red and yellow sandy fy psuml Mote == 5 aes = eases 74 90 Gypsum, mostly coarse-grained, at__..---------- 105 Red shale, with some gypsum, fragments of lime- stone and some dark quartz pebbles_------_-- 115 123 Gray marly material, with many fragments of dark dolomite, some fine sandstone having a calca- reous cement, a yellow sandy limestone, gyp- sum, pyrite and quartz sand. The dark dolo- mite contains transparent angular quartz grains and some scattered needle-like crystals, also traces of fossils, and has faint reddish bitumi- nous (?) streaky blotches. The crystals in the dolomite are from 0:005 to 0°01 mm. in diame- ter. Among the finer fragments of the sample were found several foraminifera which are unlike Paleozoic forms, and resemble Zextularia glob- ulosa Khrenberg, Globigerina bulloides d’Or- bigny, Bulimina pupoides V@Orbigny, Anoma- lina ammonoides Reuss, Nodosaria sp. . - -- - -- 141 145 Udden— Castile Gypsum and Rustler Formation. 153 Dark gray to. yellow dolomite, much gypsum, some quartz sand and pyrite. Fumes of sulphur and bituminous matter were given off when heated in closed tube. The dolomite is compact, but in some fragments small round pores, suggest- ing organic origin, were seen. ‘There are scat- tered minute black specks, probably pyrite_.- Red and green shale, dolomitic limestone and gyp- sum. Some gypsum is fibrous, evidently from thin layers. The sample contains some small pebbles of vein quartz and a few grains of pyrite. There is also some light gray lime- stone. Heated in closed tube, fumes of sul- phur and bitumen are given off.__.._____--- Brownish gray shale and dark gray dolomite, some red or yellowish limestone, considerable gyp- sum, some quartz grains, some pyrite.” A small erystal of quartz was noted and a Foraminifer ike mAcuielonee, GLODULOSA. 5-2. Le See Boo Red shale, dark gray dolomite, and a few frag- ments of red limestone, gypsum, rounded quartz grains, and pyrite. One quartz crystal noted. Fumes of sulphur, bitumen and ammo- nia noted on heating one part of the sample in closed tube. Another sample of the dolomite gave much oil and gave strong fumes of ERUUNIAT ON cl eee oy Sy 5 SNe ee ey Scie ee SS Gray marly shale, with some fragments of dark dolomite in part impregnated with pyrite, some yellow limestone, gypsum, and a few grains of quartz. Forms like Globiyerina bulloides and Textularia globulosa quite abundant. Boliv- ina like punctata @Orbigny noted and a small fragment of a test like Textularia turris d’Or- DIGI asses Sle Ne SSE eels 2 oS eee eee Dark dolomite and yellow limestone, much gyp- sum, some quartz grains, and pyrite. Fumes of sulphur, bitumen and a trace of ammonia noted on heating in closed tube-------- tee Dark gray dolomite with a little gypsum, and a few grains of quartz and pyrite. Sulphur and bituminous fumes were given off when heated. The dolomite contains more or less angular sand and the dark color is due to pyrite, dis- seminated in very fine cubic crystals. The sample contains Textularia and Globigerina, perhaps from overlying shales__.-..--. ------ Depth in feet below surface. From 165 180 209 To 195 154. Udden—Castile Gypsum and Rustler Formation. Depth in feet below surface. From To Gypsum and dark gray dolomite. The sample con- tains a few scattered grains of quartz. Sul- phur and bituminous fumes were given off when heated in closed tube. -..-----2=2-2-22 275 Gypsum containing dark dolomite and a few scat- tered grains of quartz and pyrite. Sulphur and bituminous fumes were given off when heated =. 2.2 once be ee 296 Gray dolomite, with a few fragments of gypsum, quartz and pyrite. Fumes of sulphur and bitumen noted on heating in closed tube ---- - 327 335 White gypsum, some dolomite and a few quartz grains and scattered grains of pyrite, at-_---- 375 Gray dolomite, white gypsum, and a few grains of quartz. Fumes of sulphur noted on heating. Below _. 3228. 2e ee eee APE ae ee 400 The location of this boring is in the west slope of the Rustler Hills, where the general dip of the Rustler formation clearly Fres. 1-14. ~o~ > wry yr ae gem Nie Tos . IZ 5 10 Microns | Fie. 1. Foraminifera from marls interbedded with the Castile Gypsum in the Troxel oil test No. 1, Survey 24, Block 110, Public School Lands, Culber- son County, Tex. Tentative identifications, made by the author, are as follows : 1, 2, Globigerina bulloides d’Orbigny (?), depth 141-145 ft.; 3, same, depth 209-213 ft.; 4, 5, Textularia globulosa Ehrenberg (?), depth 141-145 ft. ; 6, 7, same, depth 209-213 ft.; 8, 9, 10, Bulimina pupoides d’Orbigny (2), depth 141-145 ft.; 11, Bolivina punctata d’Orbigny (?), depth 209-218 ft.; 12, Anomalina cammonoides Reuss (?), depth 141-145 ft.; 18, 14, Nodosaria, sp., depth 141-145 ft. Udden— Castile Gypsum and Rustler Formation. 155 is to the east. It is evident that the greater part of the ex- ploration is through the Castile Gypsum, ~ which is interbedded with red shale, yellow dolomite, and some gray marl. In this eray marl the foraminifera were found, recurring in at least three such beds, at 141-145, 180-195, and at 209-213 feet below the surface. The nature of the eround shown by, these samples corresponds with the descriptions i in the driller’s log, which is not available for publication in full. The upper 300 feet were described mostly as gypsum, red clay with some gray marl, and yellow limestone (dolomite), variously interbedded. The accompanying figure shows most of the foraminifera found in these samples. I have seen what appears to me to be identical forms in the basal Comanchean at other points in this part of this state. On the other hand, I have examined many foraminifera from the Carboniferous, including the Permian, without ever finding any assemblage of forms at all like the lot found in these gray marls.* It seems to me they are undoubtedly Mesozoic, demonstrating the post- Permian age of both the Rustler Spring formation and of the Castile Gypsum. Gypsum beds associated with red and gray marls are known to lie near the Jurassic east of the Finlay Mountains in El] Paso County, of this state, and I believe that the Castile Gypsum and the Rustler Springs formation will be found to be equivalents of these beds, whether Comanchean or of earlier age.t University of Texas, Austin, Tex. * Bulletin University of Texas, No. 363, pp. 72-81. + Not wishing to rely solely on my own identification of these fossils, some material was submitted to Dr. Joseph A. Cushman, who kindly examined the same. He says that it contains scattered foraminifera which are more likely to be Mesozoic than Paleozoic ; that they are almost entirely Globigerinidz, a family which is rare or wanting before the Mesozoic. ‘‘In this lot they are very clearly shown and I certainly snould say, Cretaceous.” Ihave also submitted a small sample of the foraminiter-bearing cuttings from the Troxel boring to Mr. G. B. Richardson, of the U. 8. Geological Survey, and he has kindly furnished for publication with this article a state- ment anent the questions involved in the correct stratigraphic determination of the Rustler and Castile formations. A desire to throw whatever little new light there may be on an obscure subject, is responsible for this paper. Mr. Richardson’s communication is as follows : Dear Doctor Udden: Replying to your request for a statement from me concerning the age of the Rustler formation and Castile gypsum : Clearly more direct evidence of the age of these formations is much to be desired, and I am interested in your discovery of foraminifera. But it seems to me the question may be raised whether present knowledge of minute Mesozoic and Paleozoic foraminifera is sufficient to warrant their use in crit- ical age determinations. I have thought such evidence as we have indicates the Permian age of these formations, They occur in the lower part of a distinctive group of rocks 156 Udden—Castile Gypsum and Rustler Formation. which outcrop in the valley of Pecos River in southeast New Mexico and west Texas, which have been referred to as the ‘‘red beds of Pecos Valley.” These red beds form part of an east-dipping series and are underlain on the west, on the flanks of the Guadalupe and Delaware Mountains, by the Gua- dalupe Group, and on the east, in the escarpment of the Staked Plains, are - overlain by the Dockum formation. The Dockum formation is assigned to the Triassic on the evidence of vertebrate bones. But the precise position in the upper Carboniferous of the Guadalupe Group, which contains a unique fauna, remains to be determined. It is assigned to the Permian by Girty. The ‘‘red beds of Pecos Valley” consist of vari-colored, chiefly red, sand- stone and shale and interbedded lenses of magnesian limestone and gypsum at least 1600 feet thick. The base of the red beds, considered as the lowest occurrence either of red strata or of gypsum, is not a constant horizon because the stratigraphic position of the red-colored rocks varies along the strike. Very few. fossils, shells, and determinable plant remains have been found in these rocks. Concerning the shells which I collected from lime- stone in the Rustler Hills, Texas, from a horizon presumably above the mouth of the well from which you obtained the foraminifera, G. H. Girty reports: ““Two forms are included in this collection, one of them suggesting by its shape a small Myalina, the other being perhaps a Schizodus and having the general shape of Schizodus harei.” 'T. W.Stanton reports that in his opinion these shells are Paleozoic. The best collection known to me from the “‘red beds of Pecos Valley” was made by Beede, who found shells in a limestone lens near Lakewood- north of Carlsbad, New Mexico, which he eorrelates with the fauna of the Quartermaster and Woodward formations, parts of the well-known Permian red beds of Oklahoma and north Texas. I hope you will succeed in collecting more fossils from these rocks. Yours very truly, G. B. RicHARDSON. Washington, Mareh 30, 1915. G. S. Jamieson—Determination of Lead as Sulphite. 157 Arr. XV.—On the Determination of Lead as Sulphite; by Grorce S. JAMIESON. Ir has been known for a long time that lead sulphite was a very insoluble compound, but it remained for V. N. Ivanov* to show that very small amounts of lead could be detected by precipitation as the sulphite. He found a 2 per cent solu- tion of acid sodium sulphite most suitable, because this reagent could be employed for the detection of lead in the presence of iron, nickel, silver, copper, calcium, magnesium, and aluminum. Barium must not be present as it is precipitated under the same conditions as lead. Since it appears that this reaction has never been used for the quantitative estimation of lead, the following investigation was made in order to determine the conditions under which lead could be precipitated and weighed as the sulphite. It was found that lead could be precipitated quantitatively from a slightly acid solution with sodium or ammonium bisulphite, or even with aqueous sulphurous acid, if the acidity of the solution is carefully controlled. When a solution contains much acid it should be almost neutralized with ammonium hydroxide before attempting to precipitate the lead. For convenience, a solution of lead acetate was prepared which contained 78320 grams of lead and 10 grams of actual acetic acid in 1000°. Measured quantities of this solution was taken in beakers and diluted to about 100°. To each solution an excess of a 2 per cent solution of sodium bisulphite was added. ‘The solutions were thoroughly stirred and allowed to settle for an hour or longer. The precipitates were filtered on Gooch crucibles and thoroughly washed with cold water. It is important for satisfactory washing uot to allow all the liquid to pass out of the crucible until the washing is completed. The crucibles containing the lead sulphite were dried at about 150° C. for an hour. The drying is rapid and the weight very constant at this temperature. The following results were obtained : No. eccofPbsol. Pbtaken Wt. of PbSO; Pb found Error 1 25°0 "1958 PMD 1954 — ‘0004 2 22°0 "1724 2394 1724 ‘0000 3 22°0 W724 +2390 lees — 0001 4 30°0 -2350 "3262 2352 +0002 5 30°0 "2350 3262 2302 +°0002 6 20°0 *1566 Gopleda WOH +:0001 Tl 19:95 1562 "2165 "1561 —'0001 * Chem. Ztg., xxxviii, 450, 1913. Zz ° an 158 G. S. Jamieson—Determination of Lead as Sulphite. Another series of experiments was made using measured quantities of a lead nitrate solution which contained 73950 erams of lead and 5 grams of actual nitric acid in 1000%. The lead was precipitated with a solution of sulphurous acid which contained 36°8 grams of sulphur dioxide per liter. In some of the experiments, sodium acetate was added after precipitating the lead sulphite, as indicated in the following table of analyses : ee of Pb Wt. of NaC.Hs02 ce of sol. Pb taken PbSO,;, Pb found Error sol. 10% SO, il F015 1489 2071 "1493 +:°0004 ee 18 2 15°0 "1108 "1530 “1106 — ‘0002 wisest 15 3 25°0 1850 *2550 1839 — ‘0011 Bets £ 20 4 20°05 "1482 "2056 "1483 +0001 10 15 5 25°10 lisa) AG "1858 — 0002 10 18 6 25:0 "1850 "2563 "1848 — ‘0002 Tae 18 ol 20:0 1478 °2045 "1475 — 0003 = ee 18 8 10°0 "0739 1027 "0740 +:0001 10 8 ) 101 0747 "10384 0746 —‘0001 10 16 0 10°0 "0739 OL ‘0740 - +0001 20 24 In the first three analyses given above the lead sulphite was filtered ten minutes after adding the sulphurous acid. In experi- ments 8 and 9 the precipitates were allowed to settle an hour before filtration, while the others were allowed to stand over night. In dealing with unknown quantities of lead and free acid, it is safer to add some sodium acetate. Also it should be observed that it is possible to have such an excess of sulphur- ous acid that it will exert a solvent action on the lead sulphite. Several experiments were made in order to determine the conditions under which it would be possible to precipitate lead sulphite quantitatively from solutions containing considerable free acid. It was found that it was only necessary to neutra- lize the free acid with ammonium hydroxide. However the solution should be left slightly acid so that when the lead is precipitated it will separate in a satisfactory condition for fil- tration. The following results were obtained by precipitating the lead with a solution of sodium bisulphite after the acid (10° HNO,) had been nearly neutralized : Wt. of Pb No. Pbhtaken SOs Pb found Error 1 sikitily "1554 “1120 +°0008 2 1479 "2049 1478 —°0001 3 1845 °2565 "1849 +:°0008 4 "1482 *2060 "1485 +°0008 5 1479 °2046 1476 —°0003 _G. 8. Jamieson—Deternunation of Lead as Sulphite. 159 In attempting to precipitate lead as sulphite in the presence of copper it was observed that high results were always obtained when following the directions given above. The error was apparently due to the precipitation of some lead sul- phate as the precipitates were found to be free from copper. After much experimentation it was found that a satisfactory separation of the lead could be obtained if care was taken to filter the lead sulphite within about an hour after precipitation. It was found best to filter the solution, leaving as much as possible of the precipitate in the beaker. After washing the precipitate once with about 10° of water and decanting again, it was heated for 5 minutes with 10-15° of a strong solution of ammonium sulphite. Then the precipitate was transferred to the crucible and washed thoroughly with water and dried. The ammonium sulphite was prepared by passing sulphur dioxide gas into a solution (1:1) of ammonium hydroxide until it was strongly acid and adding enough more of ammonia to neutralize the solution. The following results were obtained : No. Pb taken Cu taken Pb found Error 1 *1004 “1554 1014 +-°0010 2 SII} 2331 “SIL +°0004 3 °1505 ail: “LUZ, +°0007 4. -0100 3885 “0101 +:°0001 5 1405 "1941 *1405 ‘0000 6 *1405 "1941 “1413 +:0008 ie "1204 2381 *120] —*0003 8 “1204 3h) Ik *1200 —'0004 Four analyses were made by simply adding an excess of sodium bisulphite and allowing the precipitate to settle in most eases for several hours before filtration. No. Pb taken Cu taken Pb found Error 1 “1566 1504 "1598 +°0032 2 “1566 miles *1606 +0040 3) 0739 "3885 0752 +:0012 4. 1471 “2331 "1499 +°0020 An attempt was made to determine lead as sulphite in the presence of calcium, but calcium sulphite was found to be par- tially precipitated along with the lead, although a calcium solu- tion gave no precipitate with sodium bisulphite even after standing two days. It was found that lead could be readily separated from zine as shown by the following experiments: 160 G. S. Jamieson—Determination of Lead as Sulphite. No. Pb taken Zn taken Pb found - Error il 1183 “1791 STARS +0004 2 SA: "1194 ‘1170 —'0004 These two experiments would indicate that lead can be sepa~ rated by means of sodium bisulphite without special precau- tions from those metals (except the alkali earth metals) which are not reduced by the sulphite. It will be observed that this method could be applied with advantage to the precipitation of lead from solutions contain- ing acids other than sulphuric, because, as shown above, it is only necessary to nearly neutralize the acid with ammonia instead of removing it by evaporation with sulphuric acid, which is the usual procedure with the sulphate method. Several attempts were made to titrate lead sulphite with potassium iodate in the presence of strong hydroehloric acid, but as the results obtained were always somewhat low, it was found preferable to weigh the precipitate. Sheffield Chemical Laboratory, Yale University, May 20, 1915. NS NV. L. Bowen— Crystallization of Magmas. 161 Arr. X VI.—The Crystallization of Haplobasaltic, Haplodio- ritic and Related Magmas ; by N. L. Bowsn. INTRODUCTION. Basar may consist essentially of labradorite and monoclinic pyroxene and, if the nature of the plagioclase is regarded as the determining factor, a mixture of andesine and monoclinic pyroxene may be called diorite, though commonly diorites favor hornblende rather than pyroxene. In their essentials, then, the mixtures of diopside with the plagioclases, whose thermal behavior is treated in the present paper, may be regarded as basaltic, dioritic and so forth, according to the nature of the plagioclase. The mixtures are, however, basalt, diorite and so forth reduced to their simplest form, for the pyroxene is pure, theoretical diopside, and the plagioclase pure, lime-soda feldspar free from potash. I have referred to the mixtures, therefore, as haplobasaltic and haplodioritic,* pre- ferring to emphasize in the title the petrologic aspect of the investigation rather than its physico-chemical aspect. The arti- ficial mixtures are, moreover, believed to be sufficiently close to basaltic and other magmas to throw considerable light on the crystallization of these natural mixtures. Though five oxides, Si0,, Al,O,, MgO, CaO and Na,O, enter into the composition of the mixtures studied, yet all the phases appearing can be expressed quantitatively in terms of the three components albite, anorthite and diopside. From the phase- rule point of view, then, the mixtures studied constitute the ternary system : diopside-anorthite-albite. MerrnHop or WoRKING. In studying equilibrium in the various mixtures the quench- ing method was used exclusively. By this method of sudden chilling the phases present at the measured furnace-tempera- ture are fixed and examined under the microscope at leisure. Liquid becomes a glass, erystals remain as such and exhibit their distinctive properties, with the result that the phases present at all desired temperatures can be readily determined. Equilibrium is assured by making the time of holding at the constant furnace-temperature sufticiently long. The quenching system was calibrated against the following fixed points: gold—1062°5, Li,SiO,-1201°, diopside-1391°5°, anorthite-1550°. In making up the mixtures the following ingredients were used: Na,CO, dried at 300°, Al,O, purified by boiling with *From the Greek axAdoc = simple. Am, Jour. Sct.—Fourts Series, Vou. XL, No. 286.—Aveust, 1915. 1 162 WV. L. Bowen— Crystallization of Haplobasaltic, NH,C1 solution, MgO obtained by igniting the precipitated carbonate to constant weight, specially prepared OaCO, and silica as ground quartz purified by heating with HOI. OptTicaAL PROPERTIES OF THE CRYSTALLINE PHASES. The crystals obtained throughout the work were in nearly all cases very small and no precise measurements of optical properties were undertaken. It was necessary only to deter- mine such properties as permitted the distinguishing of diop- side from plagioclase. The diopside occurred usually in stout crystals with pris- matic elongation and strong birefringence. The plagioclases had a pronounced tabular development | 010, the plates being often so thin that when viewed in plan (rhomb sections) their birefringence was barely discernible, whereas viewed on edge (needle sections) the birefringence was marked. The refractive indices of the plagioclases were usually not very different from those of the glasses in which they were embedded, whereas the diopside indices were always con- spicuously higher. In the mixtures rich in albite the crystals were always very minute and the indices of the plagioclase erystals were considerably higher than the glasses. The plagio- clase crystals of these mixtures were, however, readily dis- tinguished from diopside by their low extinction angles and negative elongation, whereas diopside has extinction angles up to 38° and those sections with low extinction have positive elongation. TaBLeE I. Composition Tempera- Diopside Anorthite ture Time Result 80 20 1348 4 hr. Glass and diopside 80 20 1352 thr. Glass only 60 40 1275 3 hr. Glass and diopside 60 40 1280 3 hr. Glass only 60 40 1272 3hr. Glass and diopside 60 40 1268 x hr. Diopside and anorthite 50 50 1326 3 hr. Anorthite and glass 50 50 1330 3 hr. Glass only 50 “50 1268 3 hr. Diopside and anorthite 50 50 1272 +hr. Glass and anorthite 40 60 1388 hr. Glass and anorthite 40 60 1392 4 hr. Glass only 20 80 1483 thr. Glass and anorthite 20 80 1487 3 hr. Glass only Haplodioritie and elated Magmas. 163 Tasxe II. Composition Tempera- Diopside Albite ture Time Result 75 25 1337 lhr. Glass and diopside 75 25 1342 lhr. Glass only 50 50 1283 lhr. Glass and diopside 50 50 1287 lhr. Glass only 25 75 1208 lhr. Glass and diopside 25 75 1212 lhr. Glass only 10 90 1142 lhr. Glass and diopside 10 90 1147 lhr. Glass only THE Binary SYSTEMS. It was necessary to study first the three binary systems: anorthite-albite, anorthite-diopside, and albite-diopside, and these results will be presented first. Anorthite- Albite. The system anorthite-albite had been studied formerly and, for details, reference can be made to the original paper.* The equilibrium diagram is reproduced in fig. 1. Diopside-Anorthite. The system diopside-anorthite is an ordinary eutectic sys- tem with the eutectic at 1270° and 42 per cent anorthite. The system is sufficiently described by presenting the equilibrium diagram fig. 2. The quenching determinations are given in Table I. Diopside-Albite. The diopside-albite system is also a eutectic system. It exhibits in emphasized form the location of the eutectic close to the low-melting component, being in this respect analogous to the silver-lead system among the metals. The equilibrium diagram is given in fig. 3 and the quenching determinations in Table IT. The liquidus of diopside was determined as far as 90 per cent albite but in mixtures richer in albite the high viscosity prevented the attainment of equilibrium. The position of the eutectic can, therefore, only be estimated by extrapolation of the liquidus, but it is clear that it must lie close to albite, proba- bly at about 97 percent albite. The temperature of the eutec- tic is not far below the melting-point of albite, probably at *N. L. Bowen, this Journal (4), xxxv, p. 577, 1913. 164. WV. L. Bowen—Crystallization of Haplobasaliic, Fie. 1. 1300 1200 ALBITE WT. PER CENT ANORTHITE Fie. 2. 20 30 40 50 60 7O 80 90 DIOPSIDE . WT. PER CENT ANORTHITE Haplodioritic and [elated Magmas. 165 about 1085°, at which temperature apparent traces of melting were found in a finely-ground mixture of crystalline diopside and Amelia County albite. This location of the eutectic does not depend entirely on extrapolation of the diopside liquidus, for it is confirmed by the shape of the fusion surface in the ternary system. 10 9 DIOPSIDE WT. PER CENT ALBITE A rapid increase in the convexity of the diopside liquidus as it approaches albite is necessary in order to carry it below the melting point of albite. This increase has already begun at the last determined point (90 per cent albite) and is, theoreti- eally, a general property of any liquidus as. it approaches the other component.* We are accustomed to having the eutectic in a median position, so that the liquidus ends by meeting the other liquidus before passing into this region of increased curvature. Tur TERNARY SYSTEM. In the ternary system the expectation is that the composition triangle should be divided into two jrelds, one containing all *See Roozeboom, Heterogene Gleichgewichte II, p. 275, fie. 103. 166 WV. L. Bowen—Crystallization of Haplobasaltic, points representing the composition of liquids which can exist in equilibrium with plagioclase crystals, and the other all points indicating the composition of liquids which can exist in equilibrium with diopside. There should, presumably, be but a single boundary curve separating these fields and indicating the composition of all liquids which can exist in equilibrium with both plagioclase and diopside. This is the condition actually found and shown in fig. 4. Thus the field AED Fie. 4. A DIOPSIDE SUBITE < ; ANORTHITE © contains all points representing the composition of all liquids which can exist in equilibrium with (are saturated with) diop- side and DECB is the corresponding field for plagioclase. By means of isotherms we may indicate also the temperatures at which various liquids become just saturated with either solid phase (temperatures of beginning of crystallization), and this has been done in fig. 5. The quenching results on which the location of the isotherms and boundary curve rests are given in Table III. ; Figure 5 may be regarded as a contoured map of a solid model showing fusion surfaces which represent the tempera- ture of saturation (beginning of crystallization) for the various compositions, temperature being plotted vertically and compo- sition horizontally on triangular coérdinates. A vertical sec- tion of such a model along the direction of the boundary curve is shown in fig. 6. Haplodioritic and Related Magmas. 167 Fie. 5. DIOPSIDE R S = ALBITE 20 40 WT PER CENT 60 80 Fie. 5. Isotherms. Fig. 6. Fic. 6. Vertical section along boundary curve. NV. L. Bowen—Crystallization of Haplobasaltic, Diopside 168 Composition Anorthite Albite 74 16 74 16 56 20 56 20 37 18 37 18 34 16 34 16 44 4] 44 4] 42 33 42 33 36 34 36 34 32 30 32 30 26 24 26 24 10 12 10 12 26 49 26 49 24 46 24 46 21 39 21 39 16 25 16 25 25 62 25 62 17 65 17 65 15 60 15 60 14 76 14 76 10 75 10 75 9 7 9 75 5 85 5 85 4 85 4 85 TaBLeE IIT. Tempera- ture 1484 1479 1402 1405 1255 1259 1265 1269 1372 BST 1337 1341 1296 1300 1248 1253 12733 L277 1343 1347 1256 1261 1222 1226 1250 1254 1298 1302 1303 1307 1240 1244 1208 1212 1252 1255 1197 1202 1180 1183 1175 1180 1150 1155 ler (Sr |Sr fer ler Je- \sr ler s7 |S- er ler Bless ao Ge ugh ae cui ice tet isd fd 5D ed ler} n Fe i eee a Per rrr eres = . I — i=) i} il Ime. — jerisr er ler| it lope Goh pee gh tod Ged ed ied a eh eed tor} ° Lec Si sles cies tenure Result Glass only ' Glass and plagioclase Glass and plagioclase Glass only Glass and plagioclase Glass only Glass and diopside Glass only Glass and plagioclase Glass only Glass and plagioclase Glass only Glass and plagioclase Glass only Glass and plagioclase. Glass only Glass and diopside Glass only Glass and diopside Glass only Glass and plagioclase Glass only Glass and plagioclase Glass only Glass and diopside Glass only ; Glass and diopside Glass only Glass and piagioclase Glass only Glass and plagioclase Glass only Glass and diopside Glass only Glass and plagioclase Glass only Glass and plagioclase Glass only Glass, plagioclase and diopside Glass only Glass and plagioclase Glass only Glass, plagioclase and diopside Glass only Haplodioritic and Lelated Magmas. 169 Three-Phase Boundaries. In the simple system showing a ternary eutectic the location of the boundary curve and isotherms is sufficient to determine the crystallization of any mixture. The crystallization curves (curves indicating the change of composition of the liquid) are, within each field, straight lines radiating from the composition of the pure component (see fig. 7). The composition of the liquid passes along one of these straight lines until it meets a Ine), Lie Fie, 7. Crystallization curves in a eutectic system. boundary curve and then follows the boundary curve to the ternary eutectic where crystallization is completed. On the other hand, in any system in which, on account of the exist- ence of solid solution, there is no ternary eutectic the crystalli- zation curves in the solid solution field are not straight lines, and there is no single temperature common to all mixtures at which crystallization is completed. Evidently, then, further determinations besides the locating of boundary curves and iso- therms are necessary in the present system before the crystalli- zation of individual mixtures can be described. Equilibrium in a eutectic system at a certain temperature T° may be represented graphically in fig. 8. DEF is the isotherm for T°. The phases present within the various areas are as follows : 170 =. L. Bowen—Crystallization of Haplobasaltic, area DEF all liquid ie ADB liquid D, solid A and solid B cr DBE liquid D — E and solid B EBC liquid HE, solid B and solid C ie EFC liquid K — F and solid C ae FAC liquid F, solid C and solid A “a FAD liquid D — F and solid A Fic. 8. A \? Fic. 8. Eutectic system at T°. Evidently locating the isotherm DEF fixes all these rela- tions since all the other lines are simply straight lines joining D, E and F with the corners of the triangle. In a system where solid solution occurs the matter is not so simple. Equi- librium at 1230° in the investigated system is shown in fig. 9 by means of full lines. DFE is the isotherm of 1230°. The phases present in the various areas are as follows : area DFEB all liquid i EFG liquid E — F and plagioclase K — G ip ADF liquid D — F and diopside S AKG liquid F’, plagioclase G and diopside eS AGC plagioclase G — C and diopside Haplodioritie and Related Magmas. 171 The line FG which bounds the three-phase field AFG has been called a three-phase-boundary.* AF and AG likewise bound the three-phase field, but they are merely lines radi- ating from A, their position being completely determined by the points F and G. In order to be able to predict the phases present at 1230° in a given mixture it is obviously necessary to Hie. 9! B EVES K K’ G few c Fic. 9. Phase relations at 1230° (full lines) and 1250° (broken lines). know the position of the line FG, the three-phase-boundary for that temperature, or at 1250° the line F’G’, and so on for other temperatures. Determination of Three-Phase- Boundaries. The determination of three-phase-boundaries can be accom- plished in a number of ways the principles of which are read- ily understood by reference to fig. 9. For a given position ot the isotherm DFE, i. e., a known temperature, it is only neces- *N. L. Bowen, The Ternary System : Diopside-Forsterite-Silica, this Jour- nal (4), xxxyiii, p. 222, 1914. 172 WV. L. Bowen—Crystallization of Haplobasaltic, sary to note that G represents the composition of the plagioclase which at the temperature considered is in equilibrium with both diopside and liquid. If a mixture which gives all three phases is held at the desired temperature and the composition of the plagioclase is determined by optical methods, then the point G is determined and the figure for that temperature can be drawn. This may be called a composition method since it depends on the determination of the composition of the mix- erystal. The three-phase-boundaries may be located by starting with a mixture of known composition and determining the tempera- ture at which the three-phase area is entered either from above or below. Any point on the line AG, such as P, lies on the border of the three-phase area for 1230°, and at this tempera- ture, as at all lower temperatures, it consists entirely of plagio- clase and diopside, but if the temperature is raised very slightly the point enters the three-phase area for this higher temperature (note that the point P lies well within the three-phase area for 1250° AF’G’)i. e., liquid is added to the phases already present. A three-phase-boundary can, then, be located by determining the temperature of beginning of melting for any mixture. Thus if any mixture of diopside with plagioclase of composi- tion G is taken and the temperature of beginning of melting is determined it will be found at 1230°. Jf the isotherm DFE for 1230° as previously determined is drawn then the join FG is the three-phase-boundary for 1230°. We may determine three-phase-boundaries by entering the three-phase field fromabove. The point R lies within the field of plagioclase and liquid, E’F’G’ at 1250°, is on the border of the three-phase area for 1230°, and if the temperature is lowered a little it enters the three-phase area for this lower temperature, i.e., the liquid and plagioclase are joined by diopside. By determining the temperature (approached from above) at which plagioclase and liquid are joined by diopside in any mixture the three-phase-boundary passing through the point representing the composition of the mixture is thereby determined. Thus, if the mixture R is taken it is found that diopside first appears at 1230° and if the isotherm of 1230° DFE is drawn, FR joined and produced to G, then FG is the three-phase- boundary All three of these methods were applied to the determina- tion of three-phase-boundaries. Composition Method. —'The composition method (first method) depends on the ability to determine the composition of the plagioclase crystals. These occur always as minute tabular crystals embedded in glass and their properties are very difficult to determine. In soda-rich mixtures the index of the Haplodioritic and Related Magmas. 173 plagioclase crystals is higher than that of the glass, in the lime- rich mixtures the reverse is true. In one intermediate mixture it was found that the index a of the crystals nearly matched that of the glass so that by measuring the index of the glass by the immersion method the composition of the plagioclase was determined. It was in this manner that the three-phase- boundary of 1250°, F’G’ of fig. 9, was located. At 1250° the liquid in equilibrium with both plagioclase and diopside, F’, has a refractive index as glass at room temperature of 1°572 and a of the plagioclase crystals nearly matches the glass, but is appreciably less (1-571), which places the composition of the plagioclase at Ab,An,(G’). This was the only case in which the composition method was applicable, for, when the indices of the crystals differ considerably from the containing glass, the crystals are too small for the determination of their refrac- tive indices by direct comparison with immersion liquids. Temperature Methods. — Determination of three-phase- boundaries by the method of finding the temperature of begin- ning of melting (second method) depends on the ability to detect the first trace of glass. This is a comparatively easy matter in a mixture otherwise homogeneous, so that the tempera- ture of beginning of melting of plagioclase mix-crystals, say, can be accurately located. But in a mixture of plagioclase and diopside the first trace of glass is not easily detected and there would appear to be a tendency for this method to give a some- what high result. Fortunately, however, this tendency is not very strong since the amount of glass increases very rapidly with a small rise of temperature for most compositions. (Note extent of flatter portion of boundary curve, fig. 6. Determination of three-phase-boundaries by this method was carried out by grinding together crystalline diopside and erys- talline plagioclase of definite composition and determining the temperature at which the first liquid was formed. The results obtained are given in Table IV. TaBLeE IV. Temperature of beginning of Mixture melting Ab,An, + diopside 1176° Ab, An, + diopside 1200° Ab,An, + diopside 1219° An,An, + diopside 1240° Determination of three-phase-boundaries by approaching the three-phase area from above (third method) was carried out with only two mixtures. In these plagioclase begins to separate VEON.L. Bowen— Crystallization of Haplobasaltic, from the liquid first and is later joined by diopside, the tem- perature at which diopside appears marking the border of the three-phase area. In Ab,An, 85 per cent—diopside 15 per cent, diopside first appears at 1216°; in Ab,,An,, 75 per cent—diop- side 25 per cent, at 1230°. Fie. 10. DIOPSIDE ALBITE ; ANORTHITE Fie. 10. Three-phase-boundaries. In fig. 10 the determined three-phase-boundaries are drawn in full lines and numbered according to the method of deter- mination. In dotted lines the theoretical general direction of others is shown. A three-phase-boundary intermediate between two that are determined may be found by interpola- tion. Crystallization of Mixtures in the Diopside Field. Three-phase-boundaries being located, the course of erystal- lization (with perfect equilibrium) of any mixture whose com- position is represented by a point in the diopside field can now be quantitatively described. Thus in fig. 11 the mixture F 4 Haplodioritic and Related Magmas. 175 (Ab,An, 50 per cent-diopside 50 per cent) begins to crystallize at 1275°, diopside separating and the liquid changing along the straight line AFG towards G. At 1235°, when the liquid has the composition G, plagioclase of composition H (Ab,An,) begins. to crystallize, the point H being determined by the three-phase-boundary (GH) through G. As the temperature is lowered the composition of the liquid follows the boundary Fig. 11. DIOPSIDE ALBITE he Be AGT CWS ANORTHITE Fic. 11. Crystallization of mixtures in diopside field. curve towards M. At 1218°, when the liquid has the composi- tion K, the plagioclase has changed in composition from H(Ab,An,) to L(Ab,An,). Finally at 1200° the liquid is used up, the last minute quantity having the composition M, the plagioclase having now changed in composition to Ab, An,(Z). MZ is the three-phase-boundary through the composition Ab,An, and fixes at the point M the temperature of final con- solidation of F or of any other mixture of Ab,An, and diop- side. The mixture X(Ab,An, 60 per cent—diopside 40 per cent) begins to crystallize at 1252° with the separation of diopside, 176 «WV. L. Bowen— Crystallization of Haplobasaltic, and the liquid changes along the straight line AXK. At 1218°, when the liquid has the composition K, plagioclase of composi- tion L(Ab,An,) begins to crystallize, KL being a three-phase- boundary. On further lowering of temperature the composi- tion of the liquid moves along the boundary curve towards §, plagioclase increases in amount and changes continually in com- position, diopside also increases in amount, until at 1176° the liquid is finally used up. The last of the liquid has the com- position S and the feldspar has changed in composition to T, ST being a three-phase-boundary. These examples make clear the necessity of determining three-phase-boundaries in order that the composition of plagio- clase at any temperature may be known. Crystallization of Mixtures in the Plagioclase Field. For all mixtures in the diopside field the change of composi- tion of the liquid until the boundary curve is attained is repre- - sented by a straight line. (Note AFG and AXK.) For all mixtures in the plagioclase field, however, the liquid follows a curved course in reaching the boundary curve. The crystallization of any of these mixtures can not, therefore, be quantitatively described unless these crystallization curves in the plagioclase field are determined, and for this reason these curves were determined for two representative mixtures. Though applied to only two mixtures the method may perhaps prove useful in other more or less similar investigations and will therefore be described in full. Determination of the Composition of Liquid and of Miz- crystals in a Two-Phase Mixture.—In order to find the com- position of liquid in equilibrium with crystals at any tempera- ture in a binary mixture, it is necessary only to hold a mixture at the desired temperature, quench it and determine the refrac- tive index of the glass.* In a ternary mixture, however, the measurement of the refractive index of the glass is not sufli- cient to fix its composition. The composition can, neverthe- less, be located as lying on the curve joining the composition of all glasses having that measured refractive index. Such curves will be referred to as isofracts.+ But it is known also that the composition of the liquid must lie on the isotherm of the temperature at which the liquid was held. It must, there- fore, lie at the pomt of intersection of the isofract and the *N. L. Bowen: The Melting Phenomena of the Plagioclase Feldspars, this Journal (4), xxxv, p. 585, 1913. ’ + Objection to this term, based on its mixed derivation, seems to me to be outweighed by the fact that the prefix iso is that commonly accepted in this sense and therefore preferable to, say, equi while fract is mnemonic of refrac- tive index. : Haplodioritic and Related Magmas. 177 isotherm. In order to apply this method, then, it is necessary to determine isotherms and isofracts, to hold the desired mix- ture at a measured temperature, quench and determine the refractive index of the glass. The point of intersection of the isotherm of the measured temperature and the isofract of the determined refractive index represents the composition of the liquid. TABLE V. Composition of glass Diopside Plagioclase Refractive Index 100 0 1607 50 50 Ab 1°548 0 100 Ab 1°489 17°5 82°5 Ab,An, 1°528 60 40 Ab,An, 1°571 40 60 Ab,An, 1°553 25 75 Ab,An, 1539 0 100 Ab,An, W517, 50 50 Ab, An, 1°569 30 70 Ab,An, 1°553 0 100 Ab, An, 1°531 45 55 Ab, An, 1°573 25 75 Ab,An, 1°560 0 100 Ab, An, 1°545 60 40 An 1°594 0 100 An 1°575 The measurements of refractive indices, on which the loca- tion of isofracts is based, are given in Table V and the isofracts are drawn in fig. 12. The refractive indices were determined on glasses of known composition, the glass being compared with immersion liquids until a liquid whose index matched the glass was obtained. The index of the liquid was then deter- mined on the refractometer. The probable error is usually not more than -001 but an error of :002 is possible in some cases. If the mixture Ab,An, 85 per cent-—diopside 15 per cent (D) is held at 1300° and quenched, the refractive index of the glass is found to be 1-539. Its composition is therefore fixed at the point P, fig. 12. The mixture, Ab,,An,, 90 per cent—diopside 10 per cent (E), held at 1400° gives a glass of refractive index 1-561, the composition being, therefore, that of the point R, fig. 12. rt is important to note also that this determination of the composition of the liquid fixes the composition of the plagio- clase crystals at the same time. Thus the composition of the erystals in the former case is given by joining PD and produc- ing it to G which represents the composition of the plagioclase Am. Jour. Sct.—Fourts Serizs, Vou, XL, No. 236.—Aveust, 1915. 12 “ALBITE S 20 ¥ 40 WT.PER CENT : 60 © \S 178 WV. L. Bowen—Crystallization of Haplobasaltic, crystals. Similarly the point K on the straight line REK represents the composition of the plagioclase in the latter case. This method is the only one applicable to the determination of the composition of the plagioclase crystals in equilibrium with any liquid in the plagioclase field (i. e., not on the boundary Fie. 12. DIOPSIDE 60 ' REFRACTIVE INDEX X_ DETERMINED POINTS me ° Fie. 12. Isotherms and isofracts. curve, in which case the three-phase-boundaries fix the eom- position of the plagioclase) for the crystals themselves are too small for precise optical determination. Crystallization Curves in the Plagioclase Field.—W ith the aid of the foregoing determinations of the composition of the liquid two representative crystallization curves in the plagio- clase field can be drawn and the erystallization of the mixtures Haplodioritic and Related Magmas. 179 discussed. The mixture, Ab,An, 85 per cent—diopside 15 per cent (D, fig. 13), begins to crystallize at 1875° with the sepa- ration of plagioclase of composition Ab,An, As the tem- perature falls the plagioclase increases in amount and changes in composition until at 1300° the liquid has the composition P and plagioclase the composition Ab, An, (G of fig. 12). When the temperature has fallen to 1216° diopside begins to crystallize, Fig. 13. A Fic. 13. Crystallization of mixtures in the plagioclase field. the liquid then having the composition M (fig. 13) and plagi- oclase the composition S (Ab,An,), SM being the three-phase- boundary through D. With further lowering of temperature the liquid follows the boundary curve, both diopside and plagioclase crystallizing, and at 1200° the liquid is all used up. The composition of the plagioclase is now Ab,An, (F), FH being a three-phase-boundary. In the case of the liquid E (Ab,,An,, 90 per cent— diopside 10 per cent) crystallization begins at 1480° with the separation of Ab,An,, and the composition of the liquid follows the curve ERN. At 1245° diopside begins to crystallize. The com- 180 WV. L. Bowen— Crystallization of Haplobasattic, position of the plagioclase is now T (Ab,,An,,), NT being a three-phase-boundary. As the temperature is lowered both plagioclase and diopside erystallize until at 1237° all the liquid is used up. The composition of the final liquid is O and the feldspar has attained the composition L (Ab,,An,,.) It should be noted that the crystallization curves DPM and ERN apply to the liquids D and EK respectively and to no other liquids. Thus the crystallization curve of the liquid P is not the curve PM but the new curve PL, i. e., if we start with a liquid P free from erystals the composition of the liquid follows the course PL. Only when the liquid P econ- tains in it the crystals formed during the change from D to P does the further course of the liquid coincide with PM. More- over, the liquid P when originally free from crystals becomes on cooling completely crystalline, not at 1200° (H), as before, but at a somewhat lower temperature. Crystallization with Zoning. Throughout the foregoing discussion of crystallization per- feet equilibrium is assumed. The conditions are supposed to be such that crystals of plagioclase can change their composi- tion through and through in response to the demands of equi- librium. It may be considered, however, that crystallization takes place in a quite different manner. When plagioclase of a certain composition has separated it may remain as’such and become surrounded by layers of different composition deposited by the continually changing liquid. The liquid is in equi- librium at any instant only with the material crystallizing at that instant and not with erystals already formed.* yaa eee il 522 KCl sino Mie SOs aaa aes 1 *229 KBr so SOF saee sees 1 077 KGS mo MES ©) eee 1 —'152 Sodium hydroxide -:-._---- il "346 Potassium hydroxide -_-_-_--- il "348 Sodium thiosulphate -.----- 5 —'147 Potassium cyanide.....---- 1 (about) —611 *The anomalous behavior of solutions of copper and mercury will be dis- cussed later. J. H. Reedy—Anodic Potentials of Silver. 289 Influence of Concentration on Reaction Potentials.—In the preceding table, it is noticeable that silver shows the same reaction potential (about 521 volts) with a large number of anions. In such cases the reaction potential is independent of the concentration of the anion (see Table II), which seems to show that the role of the latter is negligible, or at least of secondary importance. TasiE II, Reaction Potentials of Various Concentrations of Sulphuric Acid. Molar Reaction concentration potential 1:0 °521 volts 205) "521 "25 "522 ‘J "521 On the other hand, with anions which showed reaction potentials lower than °521 volts, concentration effects were marked. TasLE III. Concentration Effects in Solutions of Low Reaction Potentials. Concentration KClin 5M KBrin°d5M Klin ‘dM NaOH in M in moles H2SO4 H.SO, H.SO, NaNO; 1 "229 ‘O77 —°152 °346 volts ail "285 °139 —'087 “400 ‘Ol °350 *200 —*027 "462 001 "413 "259 ‘035 "484 ‘0001 467 °319 “099 *500 00001 “510 “391 "183 508 All of the above data were found to be definite and repro- ducible within a limit of one millivolt, excepting those for the -00001 molar concentration, where the experimental error is necessarily relatively great, and the results are consequently slightly uncertain. Deposition Potential.—The low reaction potentials of silver with the halogen ions, the cyanide ion, the thiosulphate ion, and the hydroxyl ion might be explained by the assumption that these ions give up their electric charges to silver very readily, and are deposited on the silver anode. From this point of view, the potential necessary to deposit an ion on an electrode is regarded as its deposition potential and varies with the concentration of the ion and the metal of the electrode. This must be distinguished from the discharge potential on unattacked electrodes, where the discharged ions combine with one another, or react with the solvent or other materials 290 J. H. Reedy—Anodie Potentials of Silver. present. Deposition potentials on attacked electrodes lie below the discharge potentials on inert electrodes, since they are less by an amount which must correspond in some degree to the energy changes involved in the formation of the compound. It is very noteworthy that in the case of silver the substances formed by deposition of anions are the ones which furnish low Ag--ion concentrations, owing to low solubility or to the formation of complexes. Further, the deposition potentials and ionic concentrations are in the same order. This suggests, to say the least, that the ionic concentration of the silver may be the determining influence in these reaction potentials. Ac- cording to the deposition theory, it must be assumed that, in the case of the thiosulphate and cyanide ions, these anions are first deposited on the silver, and the silver salt thus formed later reacts with the excess of the electrolyte to form the final complexes. Significance of the Potential -521 Volts——The reaction which takes place at the potential °521 volts,-which is the high- est anode potential observed in solutions containing no Ag—ions, amounts in practice to the solution of silver. This action seems to take place irrespective of the anions present or their concentration, provided they do not belong to the group which gives low reaction potentials, as mentioned above. If there is present at this potential a small concentration of the ions of an insoluble silver compound, as OH’—ion or Cl’—ion, a precipita- tion occurs, not only upon the electrode, but in the solution. In view of this behavior, this maximum potential may be called the solution potential of silver. The mechanism of this reaction can be explained from three different points of view: (a) Anions of the solution are discharged on the silver, a large number of anions haying the same deposition potential, 521 volts. (b) Hydroxyl ions from the water are discharged, and the silver hydroxide thus formed is instantly neutralized by the acid formed at the same moment. (c) Silver sends ions into solution at this potential. The first theory seems unlikely, since such a coincidence of discharge potentials for several anions would be remarkable indeed. And the additional fact that the potential is not influ- enced by change of concentration (see the case of sulphuric acid, page 282) is a further argument against it. A third objection is found in the fact that an electrolyte may show two different reaction potentials, and yet contain only a single kind of anion (not counting the OH’—-ion from the water.) Such a case was found in 0°5 molar potassium bromate. (See fig. 4.) An anodic reaction potential first appears at °398 volts, and the J. H. Reedy—Anodic Potentials of Silver. 291 flexure at 521 volts indicates that at that potential a second reaction has been superposed upon the first. The lower poten- tial may in all probability be taken to mean the deposition of BrO,’—ions, exactly analogous to the deposition of the halides and other anions which form silver compounds of low solubility. The second potential, 521 volts, therefore must mean either the discharge of the hydroxyl ions of the water (hypothesis 3), or the direct formation of silver ions (hypothesis c). Hypothesis 6 assumes that hydroxyl ions from the water are discharged on silver at °521 volts, forming AgOH (or Ag,O), and that this reacts with the hydrogen ion, also formed by the dissociation of water, to form water and silver ions,— AgOH + H'—>H,0+ Ag:. This hypothesis is open to the following objections: (i) The reaction is uninfluenced by acidity since acids and their neu- tral salts give the same reaction potentials. (See Table I, page 288.) If this hypothesis were correct, the theory of equi- librium between the ions and the electrode would require that the potential for acids should be much higher than for neutral salts, since in the acids the concentration of the hydroxyl ion must be very low. This should mean a difference of -4 volts between the potential in an acid of normal concentration and that of one of its neutral salts. (ii) Neutral solutions were electro- lyzed on silver anodes, and no acidity was developed at the anode, as would be expected if hydroxyl ions were precipitated. Nor was there any evidence of the formation of silver oxide, either by its dark color or by loss in weight on ignition. The following considerations appear to favor the theory that the silver goes directly into solution in the form of silver ions (hypothesis c): (i) If a very dilute solution of hydrochloric acid is electrolyzed at an anode potential of less than ‘521 volts, silver chloride is formed on the anode, and no turbidity appears in the solution. But if the potential is raised above °521 volts the solution at once shows the characteristic opalescence. This effect is made all the more striking by avoiding agitation of the electrolyte. In this case a white cloud soon envelops the anode and slowly diffuses outward into the liquid. Within this region of precipitation Ag—ions seem to be in excess, and as they diffuse outward are precipitated by the Cl’/-ions. A similar phenomenon was produced with alkaline solutions. (ii) The theory of the direct formation of Ag-—ions is made more probable by thermal data. According to the theory of the primary electrolysis of water (see }, page 290), the solution of silver would take place in stages, which, with their heats of reaction, would be represented by the following equations : 292 J. H. Reedy—Anodic Potentials of Silver. 20H'—>430, + H,O— 108,800 cal.* 2Ac¢+40,—>Ag,0 +5,900 cal.f Ag,0O+H,SO,.Aqgq—>Ag;,SO,.Aq+H,0O+14,490 cal.t 20H’ + 2Ag+H,SO,.Aqg—> Ag,SO,.Aq +2H,O0—88,410 cal. Compared with the energy required to effect in this way the solution of two gram equivalents of silver, the heat required to form two gram ions of silver by direct ionization is only 46,600 ealories.t From this it appears that silver most probably passes into solution by ionization, and not by the discharge of the hydroxyl ions of the water. (iii) The direct ionization of silver must be postulated to explain the equilibrium between a silver electrode and Ag--ions, as in the case of all other elec- trodes of the first class. The necessity in that case should at least count for probability in the case of solution potentials. The relation between the “solution potential” and the “electrolytic solution pressure ” of silver is yet undetermined. Without doubt, other influences, such as electrostatic attraction, are operative in determining the solution potential. In this connection, the energy relations of solutions which show two reaction potentials come up for discussion. In addition to potassium bromate, this behavior was found in dilute solutions of chlorides, bromides, iodides and hydroxides. In the case of the bromate curve, for example, the question presents itself: Since in both reactions the same substance is formed (i.e., silver bromate), why should the potential be different, when the latter is understood to be a measure of the chemical work done? The answer is, that the lower potential doubtless represents the work done in depositing a bromate ion of that concentra- tion on silver, while the higher potential represents the work done in forming a silver ion. To be sure, the silver ion is at once precipitated by a bromate ion of the solution, forming the same AgBrO, as before. It then follows that the total change in free energy is the same in both cases; but in the latter case the combination of the silver radical and the bromate radical occurs 7 solution, and not on the electrode. Hence the difference in potential may be explained upon the assumption that the energy of ionic combination appears as heat, and not as electrical energy. Concentration Effects in the Light of the Nernst Formula. —Figure 5 shows the influence of concentration on the anodic * Ostwald, Grund. allg. Chem., 1909, p. 309. + Thomsen, Thermoch. Untersuch., III, p. 381. { Ostwald, Grund. allg. Chem. 1909, p. 309. J. H. Reedy— Anodic Potentials of Silver. 293 reaction potentials of silver with the halogen and hydroxyl ions. The potentials are plotted as ordinates and the logarithms of the dilutions of the ions as abscisse. For dilutions up to 1000 liters the ionic concentration of the anions was calculated from conductivity data; for dilutions of over 1000 liters the ionization is assumed to be complete. These potentials were Fic. 5. Reaction Potentials H/ Pi FARA Wl eo eee 10000 100000 Logarithms of Dilutions mS Fic. 5. Effect of concentration on reaction potentials on silver anodes. obtained on silver anodes coated electrolytically with the silver compound corresponding to the anion of the solution. The lines for bromine and iodine ions are apparently straight ; that is, the potential is a logarithmic function of the dilution, although the slope is greater than would be expected from the Nernst formula. Instead of :0585 (for 22°, the average temperature of the experiments), the slope appears to be -0646—about 10 per cent higher than the theoretical value. No explanation of this deviation is offered. At ordinary concentrations (1 to ‘001 molar) the lines for chlorine and hydroxy] ions are straight, just as in the case of bromine and iodine ions. But at higher dilutions the increase of potential 294 J. H. Reedy—Anodic Potentials of Silver. Fie. 6. Reaction Potentials in Volts 00/ 000! SAafI7 Ul SUOLLNIIG 000/01 00000! 000'0001 ee ee ~_HLe-- = =- —= - ~ te =] Me - ~ 00000000! 0000000! Fie. 6. Solubilities of silver compounds. J. Z. Lteedy—Anodie Potentials of Silver. 295 with dilution falls off very markedly, so that their lines become almost horizontal. They seem to approach the solution potential of silver (represented by the dotted line) asymptotically and the deviation from the simpler behavior of the other two anions seems to be due in some way to the proximity of the potentials to this limiting value. Presumably, the bromine and iodine lines would show a similar curvature if it were possible to obtain satisfactory data for the extreme dilutions that would be required. Solubility of the Oxide and Lalides of Silver.—Figure 6 represents in a comprehensive way the effects of the concen- tration of five different ions upon the reaction potential of a silver electrode. The line AK shows the potentials of silver in contact with a solution containing Ag~ions ; that is, when it functions as an electrode of the first class. The three lower lines show the behavior of AgCl, AgBr and AglI electrodes,— electrodes of the second class. By extending the lines for the anions according to a logarithmic formula, they cut the upper lineat F, G, H, and K. These intersections are the points where the silver may be regarded as an electrode of either class, since they lie on both loci. Further, at these points the cation and the anion have the same concentration. That is, they represent saturated solutions of the silver compounds in question. At the high dilutions here represented, ionization may be considered to be complete. In brief, assuming that the potential is a logarithmic function of the dilution, the solubility of the silver compound in question may be estimated by noting the ionic concentrations corresponding to these intersections. In the case of the hydroxyl and chlorine ions, these intersections were actually realized by determining the cathodic reaction potentials in solutions saturated with silver oxide and silver chloride, respectively. For bromine and iodine ions the intersections were obtained by extrapolation. Below are the values of the solubilities found in this way compared with those by other methods: TABLEIV. Solubilities of Silver Compounds. Silver Graphic Nernst E. M. F. Conductivity compound method formula method method Ag,O WK 1O= HSNO = a a Oa HO) 5105, It will be noticed that the results obtained in this way (see col- umn above, headed “ Graphic Method’’) agree very well with those by the other methods. The somewhat higher values may be due to the presence of the solvent sulphuric acid. 296 J. H. Reedy—Anodic Potentials of Silver. Liffect of Deposits of Silver Compounds on the Anode.—In general, the presence of a coating of a silver compound on a silver anode raises the reaction potential for solutions contain- ing the corresponding anion. ‘This is shown in the following table of reaction potentials for the halogen ions: TABLE V. Effect of Coating on Anodes, Solution Bright silver Coated Elevation anode silver anode M. KCl in 5M. H,SO, 934 999, 019, volts M. KBr in -5M. H,SO, 028 077 "049 M. KI in 5M. H,SO, —1'75. | / 52) ianoea Evidently the effect of the coating of silver halide is to oppose the discharge of that particular halogen ion—a behavior iden- tical in effect with the so-called electrolytic solution pressure of the anion of the insoluble compound in electrodes of the second class. Boettger,* in a recent paper, reports that he found the same elevation, though he does not appear to have isolated and measured it. The amount of the deposit—provided it exceeds an exceed- ingly thin layer—seems to have no influence on the potential. The approximate thickness of the layer necessary to make the electrode function normally as an electrode of the second class was determined in the following way: A silver electrode, 35 <38™™", was made the anode in a molar solution of potassium chloride until it gave the same potential as a coated one. From the gain in weight (assuming that the density of silver chloride is 5°56) the average thickness of the layer was calculated to be 24x10~"". Below this limit the potential is indefinite, and lies between the values for bright and coated electrodes. Influence of Free Halogens.—Free halogens exert a strong depressing action of the reaction potential of silver, as is shown in the following way: Solutions of bromine and iodine were made by shaking the free halogens with 0°5 molar sulphuric acid. These may be considered approximately saturated. Using in these solutions the coated silver anodes as in the pre- vious experiments, reaction potentiais of -259 volts and -357 volts, respectively, were obtained. Compared with the reac- tion potential of 0°5 molar sulphurie acid (*521 volts), these represent lowerings of -262 and -164 volts. Undoubtedly the reaction in such cases is to a large degree molecular, as for example, 2Ag+Br,—~>2AgBr ; and as such may be assumed to have no effect on the potential. The lowering of the poten- * Verh. Ges. deutsch. Aertze, 1913 (1914), IT, 361. J. H. Reedy— Anodic Potentials of Silver. 297 tial is to be attributed to the halogen ions, produced by the hydrolysis of the free halogen, according to a reaction of the type, br, + H,OS—>H-+Br +HOBr. The same assumptions explain very satisfactorily the rise in potential which was found to take place with silver electrodes in acidified iodide solutions upon standing in contact with the air. Part of the hydriodic acid in the sojution is oxidized to free iodine, which combines with silver to form silver iodide. Besides the reduction of the iodine content of the solution in this way, the I’-ion concentration is further diminished by the fact that some of the iodine passes into the form of free halo- gen and hypoiodous acid, both of which have no effect on the potential. In brief, oxidation amounts to a decrease in I’-ion concentration. Influence of Cations on Reaction Potentials.—As is seen in the table of reaction potentials on page 288, all sulphates have the same reaction potential of °521 volts except those of cop- per and mercury. With solutions of these metals the results were unexpectedly high. Moreover, silver anodes in these solutions behaved abnormally in that their potentials were in- definite, and showed a tendency to rise, even during the course of a determination. Upon examination of the silver electrodes after such experiments 1t was found by means of chemical tests that some of the less noble metal had been deposited on the silver. This points to the following reactions taking place, at least to a small degree: 2Ag + Cur z™ 2Ag: + Cu. Ag +Hg- w™ Ag’ + Hg. The progressive increase in the reaction potential may be accounted for in some degree by the appearance of silver ions in the electrolyte. That this small increment of the Ag-—ion concentration should cause such large changes in the reaction potential seems very surprising. It is much more likely that this elevation is mainly due to the formation of solid solutions of copper and of mercury, respectively, in silver. As has been pointed out by Foerster,* the solution of a small amount of a less noble metal (here copper and mercury) in a more noble metal may diminish the electrolytic solution pressure of the latter, just as the presence of a small amount of dissolved sub- stance reduces the vapor pressure of the solvent, even though the solute in the pure state may have the higher vapor pressure. For example, a small amount of ether dissolved in a large amount of water may raise the boiling point of the latter. In a perfectly analogous way, the potential of silver may be raised * Hlecktrochemie, 1905, p. 208. 298 J. H. Reedy—Anodic Potentials of Silver. by the presence of very small amounts of the baser metals in the form of solid solutions. Outside of such cases as copper and mercury, the cations of metals less noble than silver seem to have no effect upon the reaction potentials of silver. Finally, it was found that the presence of the nobler metals, as platinum and palladium, in contact with the silver anode exerted no perceptible influence on the potential. Summary. 1. The term “reaction potential” is used to designate the potential difference that must exist between a metal and a solution for a definite action to begin. This potential is best determined by the third electrode method. 2. For polarizable silver electrodes, where the products of the reaction are soluble, the point of greatest curvature on the current-potential graph may be taken as the reaction potential. 3. A large number of electrolytes show the same reaction potential (521 volts) with silver anodes, regardless of their concentration. This value is interpreted as the “solution (or ionization) potential” of silver. 4, Reaction potentials below °521 volts appear in cases where the silver compound formed is insoluble or contains the silver largely in the form of complex ions. These “deposition potentials,” except in the neighborhood of the solution poten- tial of silver, are logarithmic functions of the dilutions, as would he expected from the Nernst formula for electromotive force. However, the increase in the reaction potential with dilution was found to be somewhat greater than would be ex- pected from theoretical considerations. 5. Certain electrolytes were made to show two reaction potentials, the lower one due to the deposition of the anion, the other (521 volts) due to the solution of the silver. 6. Electrodes coated with the insoluble silver compound corresponding to the anion of the solution show higher deposi- tion potentials than do electrodes of bright silver. 7. Asarnule, the presence of cations of other metals was found to have no effect on the reaction potentials of silver. In the case of the cations of copper and mercury, however, silver shows abnormally high solution potentials. It is suggested that these metals may form solid solutions with silver, with a resultant lowering of its electrolytic solution pressure. C. Barus— Use of Compensators. 299 Arr. XXIIIl.—The Use of Compensators, Bounded by Curved Surfaces, in Displacement Interferometry; by Cart Barvs.* 1. LIntroduction.—The method of increasing the sensitive- ness of the displacement interferometer by increasing the dis- persion of the grating readily suggests itself. Unfortunately the interference pattern loses sharpness in the same ratio and ultimately becomes too diffuse for practical purposes. Similar sensitiveness is secured when the glass and the air paths of the component beams of light are respectively identical, with the same inadequacy in the huge mobile figures, for the purpose of adjustment. In fact, if for simplicity we consider the inci- dence normal (/ = & = 0, linear interferometer), the sensitive- ness becomes d0/dn = r*/{2eD cos 6. ((u + 26/A*) — WV)} where @ is the angle of diffraction for the wave length A, e the thickness of the plate of the grating, w its index of refraction, D the grating space, m the order of the fringe and 6, WV con- stants. Hence other things being equal, @@/dn increases as D and e¢ grow smaller, where e = 0 is obtained by a compensator counteracting the thickness of the plate of the grating. It occurred to me that the difficulty of diffuse interference patterns might be overcome, in part, by the use of compensators with curved faces, when the case would become similar to the conversion of the usual interference colors of thin plates into Newton’s rings. Naturally a cylindric lens with its elements normal to the slit is chiefly in question, though an ordinary lens also presents cases of interest because of the easy conver- sion of elliptic into hyperbolic patterns and the lens is more easily obtained. Other methods were tried. For instance in using a Fresnel biprism with its blunt edge normal to the slit, two sets of inter- ference patterns, one above the other in the spectrum, are obtained. When the blunt edge is parallel to the slit, either side of the prism gives its own interferences, but they cannot be made clearly visible at the same time. A doubly reflecting plate or a thin sheet of mica covering one half of the beam will produce two intersecting patterns, but these also are of little use for measurement. A very promising method, how- ever, consists in the use of compensators of equivalent thick- ness, but of different dispersive powers, crown and flint glass, for instance. These experiments are in progress. * Abridged from a forthcoming Report to the Carnegie Institution, of Washington, D. C. 300 C. Barus— Use of Compensators. 2. Lens systems.—lf but a single compensator is to be used, i. e. compensation in one of the component beams only, the lens in question must be of a very small focal power; other- wise the adjustment will be impossible, as the two direct images of the slit will be in very different focal plains. More- over the focal power should be variable. All this makes it necessary to use a doublet, preferably consisting of lenses of the same focal power, respectively convex and concave. If these lenses are themselves weak, say one meter in focal dis- tance, both slit images may easily be seen in the telescope and be sufficiently sharp for adjustment. If the lens first struck by light is convex and the second coneave, their focal distances Ff, and f, respectively, and their distances apart D, the focal power of the combination used is V/P= DIPS, = DS (1) since f =f, f,. The position of the equivalent lens is d = DF/f, = f, and it lies on the same side of the doublet as the convex lens. In the actual experiment, however, the rays go through the lens system twice. In this case it is perhaps best to compute the distances directly. Of the two adjustments, the one with the concave lens toward the grating and the convex lens toward the mirror has much the greater range of focus relative to the displacement D. Supposing the mirror appreciably in contact with a convex lens therefore, if 6 is its prineipal focal distance measured from the concave lens, 6+ D = MM its prinei- pal focal distance from the convex lens or mirror, is Wi VAG ae 2D) ee (2) MeN j= WAGE se ID) where f, is the (numerical) focal distance of the concave and J, that of the convex lens. If we now write b= Bil — D(2/f, —1/(f, + DP) (3) equation (2) is easily converted into 1 1 1 D 2B © Jy 1d eee so that the usual value of the principal focal distance has been halved relatively to the new position of the equivalent lens. Lh, Sy = 2B =f'/D;b = (4) ca J? — 2D : = ii Sta. 2D fi Dp. >%=¢= 2 — aye Thus if D increases from 2 to 25 em., JZ decreases from 2450 to 165 cm., 26 from 2500 to 200 em., d from 49 to 35 em. As C. Barus— Use of Compensators. 301 b is smaller than B by equation (3), the equivalent lens is on the side of the convex lens and at a distance B- M=(f?—2D)/2(f + D) behind the mirror, or B-b=f(f + 2D)/2(f + D) behind the coneave lens. If the system is reversed, 7, and.f, are to be replaced by — f, and — f, whereas D remains positive. Hence the equiva- lent lens has the same focal distance as before, but it is now placed in front of the system, at a greater distance than it was formerly behind it. The total displacement of the equivalent lens on reversal is about one meter. ~ 3. The effective thickness of the lenticular compensator.— The compensator with curved faces may change the interfer- ence pattern in two ways; viz., by changing the angle of incidence and refraction of the rays at the grating and by changing the path difference of successive rays passing through it. Both conditions are virtually the same, or at least occur simultaneously. If there is but one compensator, as above, the two effects must be small, since the rays reflected from each of the opaque mirrors, JZ and J, of the interferometer, must eventually enter the telescope, to unite in two nearly identical images of the slit. It was rather unexpected to observe that the interferences are still obtained, even when the two slit images are quite appreciably different in size. They are then confined to a single plane, however, as will be shown in $6. Since the beam of light coming out of the colimator and traversing the grating is a vertical ribbon of light, several cms. high, vertically, but very thin in comparison (a few mms.) hori- zontally, it is relative to the vertical plane that the marked effect must be expected. If the beam consists merely of the axial pencil, the distortion of pattern due to the introduction of the doublet is slight for any value of the distance apart lenses, . The two lenses are practically equivalent to a plate. If a broad beam-is in question and the rays retrace their path, the same is still true. Butif on changing J the rays do not retrace their path, so that the equivalent lens is convergent or divergent, then the rays after leaving J re-impinge on the grating at different angles than before and the interference pattern is correspondingly changed, principally in its vertical relations. Thus it is the lens system which changes the obliquity of rays lying in a vertical plane and passing through the grating 302 C. Barus— Use of Compensators. to the effect that the axial rays may represent a case of either . maximum or minimum path difference. The latter will be the case when the divergent pencil which usually traverses the grating becomes convergent in consequence of a sufficiently large value of the D of the convergent lens system. 4. Observations largely with weak lenses and short inter- Serometer.—The film grating used (Wallace, 14500 lines to the inch) was cemented with Canada balsam to a thick piece of plate glass, so that the total thickness of plate at the grating was 1:734°". This introduces a large excess of path in one of Fini NY, ee O° DK == D=5 o> D = 15 10 5 the component beams; but it is generally necessary if the sta- tionary interferences, due to the reflection at the two faces of the plate of the grating, are to be obviated and if the ellipses produced are to be reasonably large for adjustiment (ef. §6). The lens doublet was to be added on the same side as the glass specified, so that the excess of glass thickness on one side was further increased by about -19°™ on the average. Under these circumstances the ellipses were strong, but in view of the large dispersion with inconveniently long horizontal axes. On inserting the doublet (convex and concave lens, each 1 meter in focal distance) with its concave lens at the mirror and gradually increasing the distance ) by moving the convex lens toward the grating, a series of forms were obtained which passed from the initial horizontal long ellipse, through cireles, vertically long ellipses, vertical lines, into hyperbolic forms of increasing excentricity, as recorded in fig. 1. On reversing the system, keeping the convex lens fixed near the mirror and increasing the distance D by moving the other lens toward the grating, the original ellipse usually flattened out further, as shown in fig. 2. Moving the lenses sideways, a i C. Barus— Use of Compensators. 303 parallel to themselves, had no definite effect. Moving them fore and aft together (D constant), produced results similar to the above. The vertical lines of fig. 1 are liable to be sinuous, or to resemble the grain of wood around a knot. If corresponding to fig. 1, the convex lens is kept fixed near the grating and the concave lens gradually moved up to it, the order of forms is reversed but not quite completely. They usually terminate in long vertical ellipses, before reaching which the wood-grained forms are sometimes passed. The same is similarly true for the case of fig. 2. With cylindrical lenses (respectively convex and concave, each one meter in focal distance), very little effect was observed when the axes of the cylinders were parallel to the slit. With the axes perpendicular to the slit the effects of spherical lenses were virtually reproduced, except that the central fields par- took of a more rectangular character. A variety of experiments were also made with strong lenses, with similar results, the interferences being seen most clearly in the principal focal plane of the telescope. 5. Remarks.—A few explanatory observations may here be inserted. The occurrence of the elliptic or oval and the hyperbolic type of fringes may be most easily exhibited, by laying off the order of the fringe in terms of the distance (in arbitrary units) above and below the center of the image of the slit. If we call the latter y and consider the allied colors of thin plates, for instance, 1=2eucos7r/X or more generally n= (eu/r)f (yr), (where e is the thickness of the plate, u its index of refraction X the wave length of light in case of a dark locus of the order 7) is to be expressed in terms of y, 7 being the angle of refraction at the plate of the grating. The pheno- menon will thus be coarser for red light than for violet light since » decreases when A increases and for the present purposes any two curves 7 and 2, fig. 3, may be assumed as the loci of the equation in question. If now horizontal lines be drawn for n = 1, 2,3, ete., they will determine the number of dark bands in the spectrum for any value of y. If the central ray is also a line of symmetry and intersects the grating normally, it must correspond to a maximum or a minimum in 2. ‘These conditions are shown in the diagram at M/, where the maximum number of points or bands occurs, and at m, where the reverse is true. The question is thus referred to two sets of loci v7’ and wv’, or 77” and v’v’, ete. In the former case ecosr varies with y in the same sense as #/X ; in the latter in the opposite sense and is preponderating in amount. Both may vary at the same rates in the transi- tional case, in which therefore the two curves 7 and v are at the same distance apart for all values of y. Am. Jour. Sct.—FourtH SErizs, Vou. XL, No. 237.—SErremBer, 1915. 804 0. Barus— Use of Compensators. Suppose furthermore the same phenomenon is exhibited in terms of wave length ), as in the lower part of the diagram, the spectrum being equally wide for all values of y, while at any given y, the upper diagram still shows the number of dark points or bands between 7 and v. If now we suppose that under any conditions these dark points are grouped symmetric- ally with reference to any given color (which is probable, for a maximum or a minimum of any value of y will be so for all Fie. 3. values) and that the successive dark points have been connected by a curve, the interference pattern will be of the elliptic type in case of aa’, wa’, and of the hyperbolic in the ease of a’a’’. The other features of the phenomenon are secondary and therefore left out of the diagram. Thus, for instance, the dis- tance apart of the bands shrinks from red to violet and the ovals, ete., are only appreciably symmetric because they occupy so small a part of the spectrum. Whether the long axes of the ellipses are horizontal or vertical depends upon the slope of the lines 7 and v. Maxima and minima will not, as a rule, occur close together, though in certain wood-grain shaped patterns this seems to be the case. C. Barus— Use of Compensators. 305 In conclusion, therefore, the main feature in modifying the type of interference pattern is the varying thickness of the compensator. For oval types the preponderating lens is con- vex, for the hyperbolic type it is concave. Neither of these lenses are here appreciably affected in modifying the horizontal distribution of path difference because the dispersion of the grating requires a horizontally parallel system of rays. 6. Lelescopic interferences—If interference patterns of small angular extent are to be obtained, it is essential that the rate at which obliquity increases from ray to ray be made as large as practicable. Probably, therefore, an opportunity for realizing these conditions may be found within the telescope, i. e. after the rays pass the objective. The endeavor would, therefore, be directed to bringing two spectra, focussed in two planes one of which is behind the other and consequently of different sizes, both vertically and horizontally, to eventual in- terference. The experiment was made on the long interferometer, fig. 4, the distances between mirror J/ and grating G and from the latter to the mirror JV being nearly two meters each. JZ is the lenticular compensator, consisting of two lenses, respectively concave and convex, each having the same focal distance #=+50™. The distances apart, D, of the lenses may be varied. The glass plate C’, which is revolvable about the ver- tical, is thick enough to exactly counterbalance, if necessary, the thickness of the glass plate of the grating and of the lens system Z. -A sharp wedge sliding transversely may also be used. It is best to replace C’ by two plates of glass, one thick and the other thin, so that the latter may be removed. The telescope directed along the axis / will, therefore, in general see two white slit images, A and A’, fig. 5, not both in focus at once; A coming from J/ being larger, A’ from V (parallel rays) smaller. The focal plane of A’ will be towards the grating as compared with A and A’ is larger than A, in proportion as the distance apart of the lenses, Z, is larger. Similarly the two spectra are observed along the axis, D, not in focus at once and of different areas. To obtain the interferences, the slit image A must be placed anywhere within A’ and they will occur at the top of the spectrum if @ and @’ are vertically in coincidence; in the middle if 6 and 0’ coincide; ete. The plane of the new interferences is no longer the prin- cipal focal plane, containing the Fraunhofer lines, but lies in front of it; i.e. towards the eye of the observer and away from the grating. This distance, measured along D for the given small telescope used, was fully 1™. The focal planes of the 306 C. Barus— Use of Compensators. two spectra are usually not so far apart. -A’ corresponds to a virtual object behind the observer. If the vertical plane at which the interferences lie be taken as the image, the object would be situated about 3 meters beyond the objective of the telescope used. This would place it about 30™ in front of the mirror J/ or J, where there is but Fies. 4-7. a single beam in each case. In fact, the telescope may be brought quite up to the grating. Hence interference is pro- duced in the telescope itself, where rays are relatively very divergent, a condition which accounts for the smallness of the interference pattern. This understanding of the case is tenta- tively shown in fig. 6, where O is the objective of the telescope, If the larger image from the mirror with the lens compensator and JV the image from the other mirror (parallel rays). If the C. Barus— Use of Compensators. 307 corresponding rays be drawn through the extremity of J/ and JV, their fields of interference / and /’’ would begin in the plane Z/’. For axial rays it would be at z. Thus the locus as a whole would not be a plane, and this seems to be the case. If the telescope moves towards the grating, /7’ moves toward the right in the figure, as though the virtual object beyond the grating were fixed in position. At all events, the problem is to find the interference diagram of two symmetrical, plane, parallel spectra of different areas and placed at definite dis- tance apart. The appearance of the fringes is indicated in fig. 7 where S is the height of the spectrum, usually quite out of focus. There are many more lines than could be drawn in the sketch. The ends a@ and } seem to surround small ellipses, but these are not quite closed on the outer edge. The center of symmetry is at c. The demarcations are stronger and broader vertically, if the distance apart of the doublet Z, fig. 4, is small; fainter but nevertheless clear and narrower if this distance is large. Horizontally the fine lines thread the spectrum. The best results were obtained when the lenses Z are less than 1™ apart, the middle band being about half as high as the spectrum. Two contiguous lenses gave a design which nearly filled the spectrum. For practical purposes the lens compensator Z is to be attached to the mirror, J/, just in front of and moving with it. It makes little difference, here, whether the concave lens or the convex lens of the doublet Z is foremost. _ If the micrometer J/ is moved, or if the telescope is slid to the right or left, or forward, so as to take in other parts of the spectrum, the nearly closed lines at a and 6 become finer and finer crescent shaped lines, always open outward, till they pass beyond the range of vision. The whole phenomenon remains on the same level of the spectrum. On moving the telescope towards G, fig. 4, the ocular has to be drawn outward (towards the eye) till it is fully 2°" beyond the position of the principal focal plane. The whole spectrum is now seen with the interfer- ences from red to violet (no ellipses), but having the same relative position as before. The central horizontal band measures about 1/5 the height of the spectrum, while the fine parallel horizon- tal lines extend to the upper and lower edges. The appearance is now curiously like a blunt wedge; the band is nearest the eye and the lines running abreast extend towards the rear. This impression is probably an illusion due to the shading. The lines grow finer and are more crowded toward the bottom and top of the spectrum. The illusion of a reéntrant wedge is thus not possible. . To use this interference pattern for measurement, the cross hair is supposed to pass through the region ¢, fig. 7, symmetric- 308 C. Barus— Use of Compensators. ally. Very slight motion of the micrometer mirror J/ then throws c¢ either to the right or the left of the cross hair. In this case the lens doublet at Z is attached to the mirror and moves with it, as stated. To obtain the extreme of sensitive- ness, the path difference of VG and GJ/ must be all but zero ; i. e. the grating plate @ and the lens doublet Z, fig. 4, must be all but compensated for equal air distances, by the compensa- tor ('. In this case of full compensation the interferometer pattern in the absence of a doublet Z, would be enormous and diffuse, seen preferably in the principal plane of the telescope, but useless for adjustment. ‘The introduction of a lenticular compensator balanced by a compensator in GJ, transforms the huge pattern into the small interference fringes in question, with the advantage that the high mobility of the course design has been retained. In other words, an index suitable for measurements has been found, compatible with extreme sensi- tiveness. In fact, it is difficult to place the micrometer mirror MM so that the region ¢, fig. 7, is exactly bisected. As the plane in which these interferences are seen most distinctly is 1°" or more anterior to the principal focal plane, the Fraun- hofer lines are unfortunately blurred and a cross hair is needed as a line of reference. Brown University, Providence, R. I., July, 1915. R. R. Ramsey—Radioactivity of Spring Water. 309 Arr. XXIV.—Radioactivity of Spring Water; by R. R. Ramsey. Tue springs tested are flowing springs which issue from the ground at the base of or on the side of a hill. A great number are those which were used by the early settlers as a source of water supply. Some are still the main water supply of the farm. Most of these springs are the so-called “ never-failing ” springs which flow the entire year. All are more or less affected by the rainfall. The wells are dug, driven, or drilled wells whose depth varies from 15 feet to 150 feet. Pumps must be used to raise the water to the surface of the ground. The method of measurement is the “shaking method” as used by Schmidt* in which a known volume of water is shaken vigorously with a known volume of air for two minutes, and then the air is pumped through the chamber of the electroscope by means of a rubber bulb pump until the emanation is thoroughly mixed with the air in the electroscope and the air in the shaking can. Then the following equation holds: Tee (=) (te : V, Ve Vi, Where, V, = the volume of water in the shaking can. V — « 6c (3 air (a4 66 (3 2 Nee oe ens « «bulb pump and connect- ing tubes. Vase 84 Gia en 6 “« « jonization chamber. a = the absorption coefficient of water for radium emanation. é = the amount of emanation per liter in V,, the elec- troscope. EK = the amount of emanation per unit of volume of water. Hlectroscope.—The electroscope used at first was one made in the shop of sheet tin. Sulphur was used as an insulator. Later a Schmidt electroscope made by Spindler and Hoyer, Gottingen, was used. Both electroscopes were calibrated by using Duane and Laborde’s formula,+ Cis * = 6-31 X 10° (1 — 0°572 S/V) * Phys. Zeitschr, vol. vi, p. 561, 1905. +Le Radium, vol. xi, p. 5, 1914; Ann. der Phys., vol. xxxviii, p. 959, eae Compt. Rendus, vol. cl, p. 1421, 1910; Jour. de Phys., vol. iv, p. 605, 5. curies 310 &. Rk. Ramsey—Ladiouctivity of Spring Water. Where e = amount of emanation in electroscope. dmax= Maximum current, at end of three hours, in E. S. units. V = volume of ionization chamber. S = surface of ionization chamber. This formula applies to a cylindrical chamber whose volume is about one liter and whose height is one to three times the diameter. Accuracy of the formule.—aAs a test of the Duane formula the following test made by two students will serve. The two electroscopes were connected together and filled with emana- tion, care being taken to see that the emanation was the same density in each electroscope. This experiment was carried out under favorable conditions in the laboratory. Hlectroscope: 3 essa ei Schmidt Obsetver< cso e ec Gel W.D.S. Diameter of chamber---- 10°8 7°8 cm. Height of chamber_----. 12°71 20°3 cm. Volume of chamber -----_- 1102- 968° cu. cm. Surface of chamber... .--- 594° 586°6 sq. cm. Capacity of electroscope.- - igs 6°3 cm. Amount ef emanation__-_- 206000: x 10—-" 200000: x 10—” curies Another test made in the field at the spring by the same observers will give an idea of the accuracy under unfavorable conditions. Illinois Central Spring, July 24th, 2 p. m. Water temperature, 12°°4C.... Air, hot and Temperature - windy. 104° F. in shade. Observer’): c. 2302 = eee B..G.T. W.D.S. Klectroscope -23s. ee eee aA bi ates Schmidt. Curies per liter ef water -__--- 360°X10="7 1) S007 x Apr. 15, 1915 Curies Temp. per liter tea G3 Qe Ome 5° 45° Ney 70° 19° 45° 26° 70° ay 95° ae 70° 13 200° 12 185 bss 420° 16 560 = 100° 15°8 250 19°5 300 Wy 610 19°5 140 17 350 17 350 12°5 600 16° 355 10°3 430° 11°5 170 11°5 660° 12°2 265 ile fl 12°3 175 13 650 12 1150 ae! 425 12°5 2150 = 500° 12 750 ou 1640: Sis 1920° LR. &. Ramsey—Ladiouctivity of Spring Water. 318 INDIANA SPRINGS. 39— 91 86—30°3 June 4, 11°8 WACO S) k ; (e) Lightning creek ; Stanton and Knowlton’s sketch sec- tion of 1897 (p. 131) is reproduced in Figure 5. Stanton’s section (c), 1909.—In 1910 (p. 184) Doctor Stanton gives an addi- tional section in this area, taken on the south side of the Cheyenne River at the mouth of Lance Creek, and extending up the creek a mile and a half or two miles. It is as follows: White cross-bedded sandstone with irregular brown indurated bands, masses, and concretions __.. -. 50 Soft sandy shale with bands of lig- nitic shale. Fragments of dino- saur bones were found on the surface here’) 2.2. eee 50 FEET ) Fresh-water Miocene resting unconformably on Fort Pierre ; -water Laramie fossils ; (d) Lance cree ES 5 ou 2 “G = Sandy shale full of Corbicula = 43s cytheriformis? and Corbicula Ae subelliptica var. moreauensis._._. 4 to 1 guiG More or less carbonaceous shale __- 15 £3 Soft massive gray sandstone with B Cis many brown concretions --_-. .-- 25 O8s Gray sandstone and sandy shale ons with bands of sandstone contain- ae ing Fox Hills fossils, about...._ 150 Sue g Cross-bedded, ripple marked, red- pat dish brown sandstone with irreg- aie ular base... 2.2 en 8 to10 Of; Massive soft buff sandstone with aaa many large concretions and in- BO durated masses and an abundant Ou Ss Fox Hills fauna - 222 22> seeeeee 100 2 oF Pierre shale with only the top ex- Ss 2 posed Boe Stanton says further of this section: wae “When studying the section it was S33 believed that the upper four members 345 belong to the Lance formation, but after- * < sections at the south end of the field it seemed more probable that all the beds R.S. Lull— Mammals and Horned Dinosaurs. 339 examined here belong to the Fox Hills. The higher unques- tioned Lance formation was not studied at this place.” Knowlton (1911, p. 372), however, says of the section unde consideration : 5 “The section made on the south side of the Cheyenne River at the mouth of Lance Creek shows a thickness of 405 feet of Fox Hills above the Pierre, but the highest point in the section at which marine Fox Hills invertebrates were found is over 109 feet below the top. It further appears from this section that the upper four members, aggregating 115 feet in thickness, contain carbonaceous and lignitic shales as well as fragments of dinosaur bone and brackish-water invertebrates, certain of which are the same as those found in, and there said to indicate the Laramie age of, the 400 feet of beds already mentioned as reported by Stanton and Knowlton above the typical marine Fox Hiils. To the writer | Knowlton] it seems altogether more probable that the four upper members of this section belong to the Lance forma- tion and not to the Fox Hills... If this portion of the section is placed in the Lance formation, where it certainly appears to belong, the thickness of the Fox Hills in the section is reduced to 285 feet, or but little more than half of the maximum thickness assigned to beds of this age in the Converse County region. While this evidence may not be considered conclusive, it must at least be admitted that it strongly suggests the possibility that even here, as in the areas already discussed in the Dakotas and Montana, the Fox Hills is of variable thickness, due to the erosion of the upper portions before the deposition of the Lance formation.” Doneghy’s section, 1914.—Mr. Doneghy’s section, prepared in the snmmer of 1914, begins in very nearly the same place where Stanton’s section (b) and that of Hatcher in 1893 were taken, that is, in the canyon of Buck Creek near the mouth of the draw mentioned by Hatcher and a short distance north of the Buck Creek pens. While it was occasionally necessary to depart to the right or left of a straight line in order to follow the exposures, the general trend of the section was N. 80° W. It was continued up the slope across the successive outcrops until the summit of the divide between Lance and Buck creeks was reached. ‘Thence west to Lance Creek the general slope of the land surface, which is about 10°, approximately coincides with the dip, so that practically nothing further can be learned. Mr. Doneghy did not go west of Lance Creek. This section is more detailed than any of the others and hence is hard to com- pare with them. Our supposition was that the lignite beds were necessarily Lance, so that the horizon ‘“*” about 200 feet above the bottom of the section was taken as the possible line of demarcation between the Fox Hills and the Ceratops 340 R. S. Lull—Mammals and Horned Dinosaurs. beds. Productive ant hills and the remains of dinosaurs were found at a level some 300 feet higher, which is in agreement with the statement by Stanton (wide supra, p. 337). Doneghy’s section follows : Gray sandstone, poor or no outcrops, surface slope approximating that of beds + 10° W.-- Massive hard gray sandstone, often capped by huge concretions, productive ant hills, prob- able level of Quarry No. 2 (Skull No. 1).---- Lignite, dinosaur bones at top..-..-.-----.--- Buff concretionary sandstone, ? possible limit of Fox Hills... 24). =2o)45 5502 pee ae oe White concretionary sandstone -----_--------- White sandstone with interbedded lenses of shale and lignite, massive in appearance but weathering brings out extremely irregular cross-bedding, concretions scarce, poor out- erops, ant) hillsibarrense se! 9-2-2 eee eneeae Hard red sandstone, cross-bedded, jointed, slightly. flexed. oes. Sore sees oe Massive gray sandstone, bedding like above---- Talus of soft sandstone, shale and lignite, no out- OTOP y= = cei epee oa eee a bedded...) =. j55 Rp eee oe tee eee eee Soft sandstone, poor outcrop ---------.-------- Hard red bed, massive in appearance, but seems to weather into thin dirty brown slabs. Un- weathered layers show no sign of bedding, about, ...20 40.4) eee Se eee eee Red beds, irregular bedding. -..---------.---- Hard buff massive sandstone---_-------------- Thin-bedded, even-bedded, hard sandstone- .- -- Sandstone)... 2h. = eer see es ee ee Alternating soft shale and lignite_.....-----.- Massive ibuit ‘sandstones == = eee Hard buff concretionary sandstone-.-__--------- Alternating soft buff sandstone and red shale, with plant and bone remains, beds average 1 foot muthickness; aboutus. eee. aoe eee icnitesbed, vabout. 2) S27 2a= a aeeeeeee eee Resistant pink concretionary sandstone - - -- ---- *W hite to gray massive sandstone.--._----.--- Hard medisand stone.) ue 5 - Dae ee See ere Sandstone in alternating beds, 6 inches to 10 feet, occasional hard and concretionary layers.---- Red sandstone, waboute.2- jas. eo ee ae Thick-bedded sandstone ____.-_---__.-__------ Soft alternating shales and sandstones in beds 1 to 5 inches thick, varicolored..........--- FEET 125 to 9 to 8 to 75 to 45 to 20 to 15 to 75 to 15 to 150 10 18 1 9 100 50 ~T 0 600 500 400 300 100 R.S. Lull—Mammals and Horned Dinosaurs. 341 mSolleanaiatahuigu: £5. 2 234s Sa aS eee 8 Resistant white concretionary sandstone, a ledge RAINS epee he eB AEE Spey het eee ea ay ibsiay 2 Hard white massive sandstone, regularly bedded - 15 Alternating white to brown sandstone, in beds 1 inch to 1 foot thick, regularly bedded- ---- 10 Very hard brown sandstone, concretionary, shark MeapeAbOutie a) eos a ane eee ee See 14 Resistant white sandstone..-.._..------------ 14 50 SON LIEGE pa I le oe ee Aj Massive white sandstone. ....--_-.--.----.--- Gto 7's Sill GOEL ag ej sma A Seg he RU A ete a 17 Massive irregularly bedded buff sandstone- ---- 8 Ceratopsian Localities Geologic sequence.—As a result of the labors of Mr. Hatcher and his aids, no fewer than thirty-two ceratopsian skulls in varying degrees of preservation were brought to light, in addi- tion to several partial skeletons. This material was studied in part by Professor Marsh and later, in preparation for the Cera- topsia monograph, by Hatcher and Lull. The collection, there- fore, includes every Lance ceratopsian type and nearly all of the figured material, so that its importance can not be over- estimated. The material is now about equally divided between the United States National Museum and that of Yale. The collection of the former institution is entirely prepared, while upon the Yale material there is still much to be done. The skulls were given a series of numbers, 1-19, and 19A-3L, mak- ing thirty-two all told, while the skeletons were indicated by letter, skeleton © and skull 26 constituting the composite mounted Triceratops at the National Museum. Professor Marsh and Mr. Hatcher naturally chose the best specimens for description, with the exception of the type of Zriceratops horridus, which is Skull No. 1, and one or two others. Cir- cumstances were such that the skulls of the upper levels were _ the best preserved, so that with the exception of skulls Nos. 1 and 9 (Triceratops obtusus), which are the lowest in the series, most of the types come from the upper portion and are all quite near one another stratigraphically, while between No. 1 and the next higher known form there are a number of un- identified and indeterminable specimens, which may or may not have their representatives in the higher levels. I have arranged the skulls in their stratigraphic sequence, based upon all the data available at present, but taken very largely from a study of the map, and the tables will show not alone the ceratopsian sequence but that of each adjacent mammal quarry as well. I place rather less value upon an evolutionary sequence of ceratopsian species within the Lance 342 R.S. Lull—Mammals and Horned Dinosaurs. than I did some years ago (1912, p. 774) as there seem to be more lines of descent than were recognized at that time. Evolutionary sequence.—Lambe (1915)* recognizes three main phyla of Belly River Ceratopsia, two of which lead into the two main groups of Lance forms, the third extending perhaps into the Edmonton but not as yet recognized in the Lance. He dismissed for the present the old Judith River genera Monoclonius and Ceratops as being insufficiently characterized. Brown (1914B, p. 550), however, considers Lambe’s genus Centrosaurus to be the equivalent of the former. According to Lambe, the separation of these three phyla based upon the main characteristics of the horn-cores and neck frill, appears to be as follows: ** Hoceratops to + Triceratops [also +Diceratops| Large brow-horn increasing in size. Nasal horn persistently small. Squamosal broadly triangular. Parietal fontanelle disappearing (closing). ** Centrosaurus [ = Monoclonius ?], ** Styraco- saurus, and ** Brachyceratops. } Brow-horn persistently small. Nasal horn persistently large. Squamosal continuing small. Parietal fontanelle diminishing. ** Chasmosaurus to + Torosaurus Brow-horn increasing. Nasal horn decreasing. Squamosal lengthening. Parietal fontanelle diminishing. In the Lance, therefore, there are two distinct series (Lull 1912, p. 774), the Zvriceratops-Diceratops race and that of Torosaurus. The latter is extremely rare and contains but two species, which may prove identical, as they are very similar and trom nearly the same locality, though separated by a stratigraphic interval of about 200 feet. Diceratops, with the obsolete nasal horn, represents an aber- rant race of the Zviceratops series, while of the genus Triceratops itself I can recognize two well-defined phyla and some additional species difficult to place. Of these the first phylum is represented by Z?riceratops prorsus, characterized by a well-developed nasal horn pointing forward, and moderate brow-horns. The race includes 7. prorsus, the closely allied * Geol. Surv. Canada, Mus. Bull. 12, p. 15. + Lance. +Two genera described by Brown from the Edmonton seem to belong to this group: Anchiceratops, showing affinities with Styracosaurus; and Leptoceratops, with Brachyceratops. ** Belly River. ~R.S. Lull—Mammals and Horned Dinosaurs. 343 TaBLeE 1. Phyla U. S. N. M. skulls, other than types, identified by Gilmore. cr Yale skulls, other than types, identified by Lull. Rae 4 EES rs (S) | Cy lane Level a Mus. nae | Type Genus and species E g = = zs = s : High | il 19 | Y.M. | 1830 | Holo.| Torosaurus latus x 2 2 31 | Y.M. | 1838 Triceratops sp. indet. 3 19A | Y.M. | 1831 | Holo. | Torosaurus gladius x 4 24 | Y.M. | 1828 Triceratops ‘‘ingens” MS. | x | 12 5 30 | Y.M. | 1837 | Triceratops sp. indet. | 6 18 | Y.M. | 1829 | Triceratops 2elatus x 7 27 | N.M. | 5740 Triceratops sp. indet. 8 21 | Y.M. | 1882 | Ples. | Triceratops ?brevicornus Be 9* | 22 | Y.M. | 1834 | Holo.| Triceratops brevicornus pre bia 19 10 2 | Y.M. | 1821 | Holo.| Triceratops flabellatus 28 ie ee 11 25 |N.M. | 2412 | Holo.| Diceratops hatcheri x 12 16 | N.M. | 1201 | Holo. | Triceratops elatus Ex 8 13 20 | Y.M. | 18838 Triceratops (suggests Diceritops) ox me 14 oe Y.M. | 1836 Triceratops? sp. indet. 15 26 | N.M. | 2100 | Ples. | Triceratops 2prorsus or elatus ox 16 5 |N.M. | 4276 | Holo.| Triceratops sulcatus AN x | 5 yy 29 | N.M. | 4928 | Holo. | Triceratops calicornis as, x at 18 4 | Y.M. | 1828 | Holo.| Triceratops serratus 19 3 | Y.M. | 1822 | Holo. | Triceratops prorsus S 20 14 14 | N.M. | 7239 Triceratops sp. indet. 26 21 |l15 | N.M. | 1208 Triceratops sulcatus x 22 28 | N.M. | 6679 Triceratops sp. indet. 23 10 | N.M. |} 5741 Triceratops elatus x 24 i ee INEM 1205 Triceratops prorsus x 25 11 | N.M. | 4708 Triceratops elatus x 26 12 | N.M. | 4286 Triceratops sulcatus x 27 138 | N.M. | 2124 Triceratops sp. indet. 28 8 | N.M. | 5788 Triceratops sp. indet. al 29 6 |N.M. | 2416 Triceratops serratus x 30 | NEM Cannot find eels 31 1 | Y.M. | 1820 | Holo. | Triceratops horridus x 2 Ee 9 | N.M. | 4720 | Holo.| Triceratops obtusus x x| 4 * Stanton says not much higher than skulls Nos. 3, 4, 5. ** P, Peterson’s mammal quarry; B, Beecher’s; the rest were numbered. 344 PR. S. Lull—Mammals and Horned Dinosaurs. if not identical Z. brevicornus, probably TZ. serratus and T. horridus. All are of moderate size with the exception of the last and their range in the stratigraphic column is low, especially as Skull 22 and possibly 21 in the table may be placed entirely too high. Of the second Zvriceratops phylum T. elatus is typical. In this race the nasal horn is very small and set well back from the terminus of the beak; the brow-horns, on the contrary, are very large. To this group belong 7. calicornis, closely related to if not identical with 7. elatus ; probably also the immature though large 7. flabellatus, and the gigantic “Skull 24” to which Marsh gave the manuscript name of 77iceratops ingens. The members of this second phylum are all large and range through the middle and upper portion of the Lance. To derive them from the first phylum would necessitate the retrogression of the nasal horn to which certain critics (Gidley in Peale, 1912, p. 751) object, although the principle is certainly well established in evolution. I am willing, however, to accept Mr. Gidley’s main contention of the improbability of the reduction of a highly developed nasal horn while the brow- horns were being developed to be the principal ones. Mammalian Localities The table of ceratopsian distribution also shows the nearest mammal localities stratigraphically to each skull, and has enabled me to arrange them in an approximate sequence which, however, is open to the same chance for error as that of the skulls in that it is derived largely from the study of the map and the sections. It should, therefore, be veritied by - further accurate field work. One thing at least is certain, that instead of being confined to any one level, mammalian remains are found throughout nearly the entire Ceratopsia-bearing beds. From Peterson’s quarry, Beecher’s quarry, and Quarries Nos. 1-10 we have a recorded mammalian fauna. From Quarries 11-27 there is no recorded material at Yale or at Washington, so that at present I am unable to give a list of genera and species from them. This is unfortunate, as the four quarries highest in the list are included within the number and a record of their mammalian contents would be of great interest. There are also a number of species of which the types bear no quarry number but which are in part duplicated by unnamed material the source of which is indicated. The following table gives Marsh’s original genera and species and in the adjacent columns their distribution in the several quarries is indicated. I have also identified in so far as possible all the additional material from these quarries, with TABLE 2. High <— Strat. sequence ——> Low Mammal-bearing quarries T=holotype X=occurrence a 3 £ P=figured specimen g e S EF She eee |e eae | Sc) lle (8 s\élelslelé|ale jeje 2s Suborder Trituberculata Stagodon nitor “ae ee rey tumidus € Tt galeparl| E validus Pee plo Pen |e ae Sp. indet. DX@ xe aa Xe Platacodon nanus: T P sp. indet. 5 Py os ere eel a Pediomys elegans a GaP ENS: aa sp. indet. x ary on Baty Batodon tenuis Beniips. ea P Pane Didelphodon comptus pxe bx RP 2P | 2X ferox x Ae ay vorax TP Poa sp. indet. GD cae ipxe| [Exe Cimolestes curtus XE 7 | aa pi ea Mees. imeisus Sa@Er re x en a ea eee sp. indet. eam lGxe PERT |pexe | xX Telacodon levis T hae BEF prestans poly FSO dead a Dryolestes tenax a | ay Genus novum fale ap | Suborder Allotheria (Multituberculata) A Allacodon fortis t lentus Xe TP) Xa SP. x pumilis x | T me rarus al at sp. ax aa Cimolomys agilis Ay bellus aa axe i eles digona xX 2x | T Paley) | (Halodon) formosus* KE Le lee x x x gracilis TP (Wanomyops) minutus 1 1p Woe sp. aa xe |X nitidus Je |an ie le parvus T P aes (Halodon) serratus* T x |x 12 sp. x x xe | x Meniscoéssus (Selenacodon) brevis aX T x Eas (Tripriodon) caperatus. Av (ey Xe ceelatus iE. CF Ge (Selenacodon) fragilis x ale ae (Dipriodon) lunatus x iNe x alle te robustus SX /EXe PN EP xX T XG | (Halodon) sculptus P TP) PAE, x P Wics sp. Bes | xa) eos Oracodon anceps BXGiI Xe P ies conulus At Camptomus amplus TP aw * Referred to Ptilodus by Gidley, 1909, pp. 622-623. 346 R.S. Lull—Mammals and Horned Dinosaurs. the exception of Quarry No. 1 and Peterson’s quarry. The former particularly was an immensely prolific locality and there are yet at least eighty-five vials containing unnamed specimens at Yaleand more at Washington from these localities. These additions have served very largely to amplify the table, especially from Quarries Nos. 10, 9, 5, 7, and 3. Comparing this table with that showmg the distribution of ceratopsian species, it will be seen that no mammals are recorded above the level of TZ7iceratops brevicornus type, so that those which were associated with the huge Z. “ingens” and with Torosaurus, if there were any such, are unknown. Turning again to the table, the great number of species from Quarry 1 must be due in part to a happy accident of preser- vation; nevertheless some species range from the level of Quarry No. 1 upward, others downward, while some of the commoner species like Ieniscoéssus robustus and I. sculptus continue throughout nearly the entire series. One rather remarkable feature is brought out, however, in that the Tritu- berculates are rarely found below the level of Peterson’s quarry; Quarry No. 7, which is very productive, and a lone specimen in Quarry 38 constituting the record. The Multituberculates, on the other hand, are as plentiful below the level of Quarry No. las above. This may be significant, though the greater resistance to destruction offered by the more massive multituber- culate molars may in part explain it. The material the quarry of which is unrecorded probably came very largely from either Peterson’s quarry or Quarry No. 1, if one may judge from the date of shipment as compared with similar dates in Hatcher’s field diary, in which the work done is meagerly recorded. Quarry No. 8 I can not locate, but it may be one of two unnumbered mammal localities, one of which would bring it not far from the level of Beecher’s quarry, the other between Quarries No. 5 and No.10. The last would be more in keeping with the “tritubereulate ” distribution. Summary It will be seen that the results of the above study are inconclusive in some important details, which only serves to emphasize the need of exact field work covering the entire area. ‘This should include a carefully prepared topographic map whereupon the sections and fossil localities could be plotted with great accuracy, while the paleontological task should be to search minutely the whole region from south to north, keeping a careful record of the exact occurrence of any mammal deposits stratigraphically. There is reason to believe that such exhaustive search would be rewarded by R.S. Lull— Mammals and Horned Dinosaurs. 847 much more perfect material than any yet preserved. Possibly more complete jaws or even skulls and associated skeletal remains may be found, which alone will serve to put our knowledge of Lance mammals upon a more exact basis. The main conclusions reached in the present paper are as follows : 1. That the mammals, instead of being confined to any one horizon, are pretty uniformly distributed throughout the entire Cer atopsia-bearing beds. 2. That the mammals are apt to be not far removed from lignitic deposits and are found in association with often worn seales and teeth of fresh-water fishes, generally in a bone conglomerate at the base of invading sands. This would seem to imply some transportation from the actual living habitat. The associated plant remains, both with the mammals and the Ceratopsia skulls, imply an abundance of vegetation, possibly forested conditions, in their respective homes, whether they differed or not. 3. That the more conservative and older Multituberculates range throughout the entire Lance, while the Trituberculates thus far found are absent from the lowermost quarries. This seems to be significant, for Multituberculates are known from the Jurassic and become extinct in the Paleocene, while descendants of the Trituberculates may still exist. Bibliography Brown, Barnum. 1907. The Hell Creek beds of the Upper Cretaceous of Montana. Bull. Amer. Mus. Nat. Hist., xxiii, pp. 823-845, figs. 1-8. 19144. Cretaceous Eocene correlation in New Mexico, Wyoming, Mon- tana, Alberta. Bull. Geol. Soc. America, xxv, pp. 355-380, figs. 1, 2. 1914B. A complete skull of Monoclonius, from the Belly River Oreta- ceous of Alberta. Bull. Amer. Mus. Nat. Hist., xxxiii, pp. 549-558, pls. 38-40, text figs. 1, 2. Douglass, Earl. 1902. A Cretaceous and Lower Tertiary section in south central Montana. Proe. Amer. Philos. Soc., xli, pp. 207-224, pl. 29. 1908. Vertebrate fossils from the Fort Union beds. Ann. Carnegie Mus., v, pp. 11-26, pls. 1, 2. Gidley, J. W. : 1909. Notes on the fossil mammalian genus Ptilodus, with descrip- tions of new species. Proc. U. 8. Nat. Mus., xxxvi, pp. 611-626, pl. 70, text figs 1-9. 1915. An extinct marsupial from the Fort Union with notes on the Myrmecobide and other families of this group. Ibid., xlviii, pp. 895-402, pl. 238. Hatcher, J. B. 1893. The Ceratops beds of Converse County, Wyoming. This Journal (8), xlv, pp. 185-144. 1896. Some localities for Laramie mammals and horned dinosaurs. Amer. Nat., xxx, pp. 112-120, pl. 3 (map). 348 R.S. Lull—Mammals and Horned Dinosaurs. 1908. Relative age of the Lance Creek (Ceratops) beds of Converse County, Wyoming, the Judith River beds of Montana and the Belly River beds of Canada. Amer. Geol., xxxi, pp. 369-375. 1904. An attempt to correlate the marine with the non-marine forma- tions of the Middle West [with note by T. W. Stanton]. Proc. Amer. Philos. Soc., xliii, pp. 341-365, figs. 1-3. Hatcher. J. B., Marsh, O. C., and Lull, RB. 8. 1907. The Ceratopsia. Monograph U.S. Geol. Sury., xlix. Hay, O. P. 902. Bibliography and catalogue of the fossil Vertebrata of North America. Bull. U. S. Geol. Sury., No. 179. Knowlton, F. H. 1909. The stratigraphic relations and paleontology of the ‘‘ Hell Creek beds,” ‘‘ Ceratops beds” and equivalents, and their reference to the Fort Union formation. Proc, Wash. Acad. Sci., xi, pp. 179-238. 1911. Further data on the stratigraphic position of the Lance forma- tion (‘‘Ceratops beds”). Jour. Geol., xix, pp. 308-376, figs. 1-3. 1914. Grtncsode Tentiany boundary in the Rocky Mountain region. Bull. Geol. Soc. America, xxv, pp. 325-340. Lull, R. 8. 1912. The evolution of the Ceratopsia. Proc. Seventh Internat. Zool. Cong., 1910, pp. 771-777, 1 fig. Advance print, 1910. Marsh, O. C. 1889A. Discovery of Cretaceous Mammalia. This Journal (8), xxxvili, pp. 81-92, pls. 2-5. 1889B. Discovery of Cretaceovs Mammalia, Part II. Ibid., pp. 177—- 180, pls 7, 8. 1892. Discovery of Cretaceous Mammalia, Part III. Ibid., xliil, pp. 249-263, pls. 5-11. Matthew, W. D. 1914. Evidence of the Paleocene vertebrate fauna on the Cretaceous- Tertiary problem. Bull. Geol. Soc. America, xxv, pp. 381- 402, figs. 1-3. Osborn, H. F. 1910. The age of mammals. Peale, A. C. 1912. On the stratigraphic position and age of the Judith River formation. Jour. Geol., xx, pp. 530-549, 640-652, 738-757. Stanton, T. W. 1909. The age and stratigraphic relations of the, **« Ceratops beds” of Wyoming and Montana. Proc. Wash. Acad. Sci. ., X1, pp. 239-298. 1910. Fox Hills sandstone and Lance formation (‘‘ Ceratops beds”) in South Dakota, North Dakota and eastern Wyoming. This Journal (4), xxx, pp. 172-188. 1914. Boundary between Cretaceous and Tertiary in North America as indicated by stratigraphy and invertebrate faunas. Bull. Geol. Soc. America, xxy, pp. 841-354. Stanton, T. W., and Hatcher, J. B. 1905. Geology and paleontology of the Judith River beds, with a chapter on the fossil plants by F. H. Knowlton. ‘Bull. U.S. Geol. Sury., No. 257. Stanton, T. W., and Knowlton, la alals 1897. Stratigraphy and ‘paleontology of the Laramie and relaeeel formations in Wyoming. Bull. Geol. Soc. America, viii, pp. 127-156, figs. 1, 2. Browning—Detection and Separation of Platinum, ete. 349 Arr. XX VII.—A Note on the Qualitative Detection and Sepa- ration of Platinum, Arsenic, Gold, Selenium, Tellurvum and Molybdenum; by Putire EK. Brownine. [Contributions from the Kent Chemical Laboratory of Yale Univ.—cclxx. ] Propasiy the best working scheme for the qualitative sep- aration and detection of platinum, gold, arsenic, selenium, tellu- rium and molybdenum is that of Noyes and Bray.* Briefly stated the method is as follows: the platinum is first separated as potassium chloroplatinate by evaporation with a potassium salt, and the arsenic in the filtrate is removed as ammonium magnesian arsenate by precipitation with magnesian chloride mixture in ammoniacal solution. The filtrate is then made acid with oxalic acid and warmed to precipitate the gold. The filtrate from the gold separation is evaporated almost to dryness and treated with strong hydrochloric acid and sodium sulphite, which removes the selenium, and on dilution of the filtrate and treatment with potassium iodide and sodium sul- phite the tellurium is thrown down. After the removal of the tellurium and the addition of zine and potassium sulphocyanide to the filtrate the red molybdenum sulphocyanide is obtained. Recent work by a class of about forty in the application of this method to the analysis of a common solution showed a marked variation in results. With a considerable proportion of the class the tests for gold and arsenic were unsatisfactory and in some cases the tests for selenium and tellurium were doubtful. An investigation of the causes of these results revealed the following facts. First, that a solution of tellu- rous chloride when treated with a sufficient amount of ammonia to form the soluble ammonium tellurite and then with magne- sium chloride mixture will give a flocky precipitate of magne- sium and ammonium tellurite not easily soluble in an excess of ammonia. This reaction has been noticed by Heberleint and use has been made of it by him in the purification of crude tellurium. Second, selenious acid when treated with ammonia and magnesium chloride mixture tends to give, especially on warming, a crystalline precipitate of magnesium and ammonium selenite also quite insoluble under the condi- tions. Both this compound and that of tellurium have been described by Hilger.t Third, a solution of gold chloride when treated with ammonia gives a precipitate of fulminating * Noyes and Bray, Jour. Amer. Chem. Soc., xxix, 137. + Heberlein, Inaug. Dissert. Basel, 1898, 37. Gmelin-Kraut, 7th Edition, TI (2), 859. t Hilger, Ztschr. anal. Chem., xiii, 132. 350 Browning—Detection and Separation of Platinum, ete. gold; and after filtering and washing, gold may be detected both in the filtrate and in the precipitate. From these facts it would seem probable that in the attempt to precipitate the arsenic acid as the ammonium magnesium arsenate one might under certain conditions not always easy to avoid, precipitate some of the compounds of gold, tellurium and selenium, and not only vitiate the test for arsenic but also partly destroy the delicacy of the tests fur gold, selenium and tellurium. To avoid these difficulties, one may leave the treatment with magnesium chloride mixture until after the gold, selenium and tellurium have been removed and detected. On evaporat- ing with bromine or nitric acid, the filtrate from the tellurium, which contains hydrochloric, sulphurous and hydriodie acids, oxidation of the arsenic and molybdenum readily takes place, and the precipitation of the arsenic by magnesium mixture can be satisfactorily made. The molybdenum sulphoeyanide is easily obtained in the filtrate by acidifying with hydrochloric acid and adding zine and a sulphocyanide. As a substitute for oxalic acid, hydrogen dioxide in alka- line solution has been found very satisfactory for the precipita- tion of gold. In conclusion, it may be stated that the above indicated modification of the original method gave much more satisfac- tory results in the hands of the class. June, 1915. O. Andersen—Aventurine Feldspar. 351 Art. XX VIII.—On Aventurine Feldspar ; by Otar ANDER- spn. With Plates I-III. CONTENTS. BION ONTO Nes ne a ee 351 PREWVAOUS WORK. Soe Pete oe eee ee eee oe eee eee 302 I, METHODS OF EXAMINATION AND GENERAL RESULTS._.. 354 DETERMINATION OF THE ORIENTATION OF THE LAMELLA__-_------ 354 IRIGNeSNOfROnientahion 2 list i Le ee ee ceek sees 305 Measurements with the microscope _____._.._------------- 355 Remarks on the optics of aventurine feldspars._-.-----_- 306 Goniomeirice measurements ==-2 2-2-5 228222223 -8 2 lee 361 Resultsvofpmeasunements= 2245 ace 0 0-2 cscs sees ental 363 . Orientation of the edges of the lamellw____.....------------- 367 THE PROPERTIES OF THE REFLECTING LAMELLM_-------...---___- 369 SIZE ECTHYSUML OWLNES= me nae me seeeee elke ease eee ee 369 Inter ‘ference Colors) MihiChnesse nee = ean. eee ee eee 369 PADSOMILO I MCOLOGS aH ete be aber ese Sa ey. wee eee 370 Doublen mes mac lO tees ese ean ne ne A eae e esos 374 Ghemicalttestse sere sae ne oes on se oa sosiaeeosee kos 374 Some observations on hematite._...._..........-.------ 375 Sunmary of the properties of the lamelle____.._-_.---- 375 PIPED HEU NU AMT RIAU 2 ors ie, UA ern ye yas win Sin aka ay eaid amie ama Sine 376 ORIGIN OF THE HEMATITE LAMELLZ..-_---..-.-..-.---.----..-- 379 II. DESCRIPTION OF THE SPECIMENS -.._-...-.-.-----_---_-- 380 = Albite from msher Hill Mines22. 2222222552225. . -2hs li. 2 381 = ALLULEN NOM NEAMIVMEORNG an soe 2 22. 2hi Sa sce le uke ses. osde 382 — Oligoclase from Statesville.._.....-.------------------------ 384 a OR GOClase frony KRAGEnOs =.= sesso) 4 ene sacs ean ee oSs 384. Ohogoclasefron Puedestrands.- & =. 22-2222 22_ 2) Besse ee 386 — Oligoclase from Aamland ..........--.--..------.---------- 389 idoradortte: from LaOrador = 92 252-2 eo ene ose eons 6 es a O90 Wicrnocline pertiite from berth. — eee oa Shwe oe 391 — —_Microcline perthite from Mineral Hills.-----....------------ 393 Microcline perthite from Ndskilen..-.....-....--.---------- 893 Microcline perthite from Stene.._.-..----------- ----------- 395 Eig SCeLLOMmeOWS OCCULT ONCESH Hehe See ee aoe leo cee Sas sae 396. SION LANNY Sey pel RS ee eS SS Se San eS re 397 INTRODUCTION. DirreRent varieties of feldspar show a more or less distinct metallic schiller, aventurization (sunstone schiller), when light rays fall in certain directions on cleavage faces or artificially polished faces. This schiller is caused by oriented lamellar inclusions which reflect the light with great intensity. The “fire” of the schiller is due to the brilliant interference colors produced by the thin film action of the reflecting lamelle. Aventurization may be defined as a play of light and colors caused by strong reflections from thin oriented lamellae of visible size included in the feldspar. _ The terms aventurine feldspar and sunstone have been used interchangeably by previous authors. It seems advisable to Am. Jour. Sci.—Fourt# Srerizs, Vou. XL, No. 2388.—Octoser, 1915, 23 352 O. Andersen—Aventurine Feldspar. make some distinction between them by using aventurine feldspar as the general term, embracing all feldspars which show aventurization, without regard to the intensity of the phenomenon. Sunstone should then be the special term for those varieties which have intense aventurization and there- fore sometimes find use as gem stones. Aventurine feldspars (sunstones) should be sharply dis- tinguished from the other group of color-playing feldspars known as moonstones, murchisonites and labradorites. These feldspars are characterized by a rather subdued, generally blu- ish or greenish play of colors (“glaukisiren”*) which is not caused by any visible lamelle but perhaps by submicro- scopic inclusions. The colors are probably due to scattering of light by particles smaller than the wave length of light,+ and can not be explained as ordinary interference colors of thin films. A survey of the literature shows that the conceptions of the problems connected with aventurine feldspars have been rather diverse. A general treatment of the subject based on thorough examinations of different aventurine feldspars has never been attempted. It, therefore, seemed of considerable interest to subject these problems to a somewhat closer study. I had at my disposal good material from a number of occur- rences. Specimens from Norwegian localities were obtained from the Mineralogical Museum of Kristiania University through the kindness of Professor Dr. W. C. Brégger and Mr. Jacob Schetelig. During visits to some of these localities I have also had the opportunity to collect specimens and to study their occurrences. From American localities | obtained good specimens from U. S. National Museum through the courtesy of Dr. G. P. Merrill and Dr. E. T. Wherry, who placed at my disposal (among other specimens) the entire feldspar collection of Isaac Lea, containing the type specimens for the paper referred to below. PREVIOUS WORK.t{ Aventurine feldspars or sunstones are mentioned in some of the earliest systematic works on mineralogy,§ in which they are described as varieties of feldspars without explanation of the aventurization. *C, Viola, Zs. Kryst., xxxiv, 171, 1901. + Cf. explanation of the blue of the sky, C. Viola, loc. cit., p. 188. } The following review is not intended to be complete, the most important work only being mentioned. Some additional references will be found in different parts of the present paper. §e. g. Delametherie, Théorie de la Terre, vol, II, p. 201, 1797, where the term heliolithe is used. [ R. Jameson, A System of Mineralogy, 1820, vol. II, p. 17, where aven- turine feldspars from The White Sea and Archangel are mentioned. O. Andersen—Aventurine Feldspar. 353 Very little was known about the exact localities or the’ mineral associations of the sunstones. The first description of an occurrence was given by K. G. Fiedler,* who discovered a locality near Werchne Udinsk, Siberia. Fiedler does not deseribe the feldspar and the explanation of the aventuriza- tion is disposed of in the following remarks: “ Ueber diesen Feldspath ist noch zu bemerken, dasz er seinen Goldschimmer der Vulkanitiit verdankt in welcher er entstand.” The well known occurrence at Tvedestrand, Norway, was discovered by Weiby in 1844. This locality has furnished a large quantity of good specimens, the first of which were care- fully examined by Th. Scheerer,t who analyzed the feldspar (oligoclase) and described and gave drawings of the reflecting inclusions, which he determined as hematite. He found these inclusions to be oriented parallel to (001), (010), (221) and a vertical prism and summarized his explanation of the origin of the sunstone as follows :t¢ ‘‘ Man musz also annehmen, dasz Oli- goklas und Hisenglanz die Producte eines eleichzeitiven Krys- tallizationsactes sind, und dasz beide, in ihrer regelmissigen Werwachsung ein dem Schriftgranite ahvliches Gemenge dar- stellen.” A. Kenngott§ discussed the qualities of the reflecting lamel- le, and concluded that they were goethite (“ pyrrhosiderit’’). E. Reusch|| discussed, in connection with a careful study of _ moonstones, the problem of reflections and refractions in bodies like sunstones and moonstones and made a few observations on sunstone from Tvedestrand, correcting and supplementing some of Scheerer’s measurements. Isaac Lea*] made microscopic examinations of sunstones from different (mostly American) localities and described the reflect- ing inclusions, which were considered goethite, but did not attempt to determine their orientation, thinking that “they usually lie parallel with the principal cleavage of the feldspar.” ** A. Schrauf++ studied aventurization on labradorites and also examined the inclusions in sunstone from Tvedestrand. It was shown that the inclusions from the sunstone were identi- eal with those found in ecarnallite, the latter determined to be hematite. The lamelle causing aventurization on (010) of cer- tain labradorites (which also showed labradorization) were found to be oriented approximately parallel to (180) and (170). J. W. Judd,tt in discussions on the schiller of minerals, *Poge. Ann., xlvi, 189, 1839. + Pogg. Ann., lxiv, 153, 1845. t Loe. cit., vp. 161. § Sitz.-Ber. Akad. Wien, x, 179, 1853. || Poge. Ann., exvi, 396, 1862. “| Proc. Acad. Nat. Sc. Philad., 1866, 110. ** Loc. cit., p. 111. tt Sitz. -Ber, math. naturw. Cl. Ak. Wien, Ix, I, 1024; 1869. 4¢ Quart. Journ. Geol. Soc., xli, 374, 1885. Min. Mag., vii, No. 33, 81, 1886. 354 -O. Andersen—Aventurine Feldspar. embracing aventurine feldspars, came to the following conelu- sion concerning the origin of the inclusions :* “These enclos- ures are of the nature of negative crystals which are more or less completely filled with products of decomposition of the mineral.” Judd considered these products of decomposition as chiefly consisting of amorphous hydrates of silica and ferric oxide. - H. Tertscht examined the sunstone from Tvedestrand and found the reflecting lamellee oriented parallel to (538) and (417), which forms were considered boundary positions (“Grenzla- gen’) of the simpler form (218). A. Johnsen determined the inclusions in carnallitet and can- crinite§-as hematite, and found them identical with those con- tained in aventurine feldspars. The presence of the inclusions in earnallite and cancrinite was explained as due to secondary reactions and unmixing in the solid state. It was intimated that the inclusions in aventurine feldspars might be explained in the same way. I. METHODS OF EXAMINATION AND GENERAL RESULTS, The study of the specimens embraced three groups of exam- inations: (1) Microscopic examinations with the object of determining the feldspars. (2) Determination of the orientation of the reflecting lamelle. (8) Examination of the properties of the reflecting lamelle. The results of the first group will be given in the section in which the specimens are described and details of all measure- ments are given. The present section contains a brief review of the general results of the last two groups in connection with descriptions of the methods applied and discussions of the simple optical problems involved. DETERMINATION OF THE ORIENTATION OF THE LAMELLA. For a complete determination of the orientation of the lamellee we must know both the codrdinates of the different planes parallel to which the lamelle are oriented and the diree- tions of certain edges of the lamelle. All orientation must be referred to some crystallographic planes or axes of the feld- spar. * Quart. Journ. Geol. Soc.. xli, 384, 1885. + Tsch. Min. Petr. Mitt., xxi, 248, 1902. } Centralbl. Min., 1909, 168. - § Centralbl. Min., 1911, 369. O. Andersen—Aventurine Feldspar. 355 Planes of Orientation. As the specimens consisted of cleavage pieces, in general showing no other faces than the cleavage faces (001) and (010), all measurements had to be referred to the elements (001) (010) and the a-axis. In determining the planes of orientation of the lamelle we have thus to deal with the following angle codrdinates: pp = angle between P (001) and lamelle. pu = angle between M (010) and lamelle. op = angle between line of intersection : lamellz A P (001) and a-axis. $y = angle between line of intersection : lamellee , M (010) and a-axis. A combination of two of these angles determines the plane of orientation of a set of lamelle. To be able to refer the measurements to the proper octants in the axial system of the feldspar we must know the approxi- mate direction of the c-axis (or the direction of the positive or negative a-axis) for each cleavage piece examined. In the plagioclases the difference between the obtuse and the acute edges of the a-axis must also be noticed. On specimens of microcline the perthite striation on (010) will generally indi- cate the approximate direction of the c-axis. In the plagio- clases it is necessary to look for indications of the third cleav- age, parallel to (110), or else to rely on the determination of extinction directions on small cleavage pieces chipped off from the larger. The angles ¢ could be measured either with the microscope on oriented sections after (001) and (010), or with the goni- ometer on cleavage pieces. ‘The angles p were always measured with the goniometer. Measurements with the microscope. The measurements of the angles ¢ on thin sections consisted in determining the angle between the cleavage lines and the lines of intersection of the lamelle with the surface of the sec- tion. The latter lines we will call the section lines of. the lamelle. _ As there were always more than one set of lamellz to be measured, the difficulty in determining the orientation of each set by measuring the two angles @ would consist in a correct correlation of the measurements from the two different sec- tions. For each set, the angle ¢ of which had been determined in one section, it was necessary to estimate the angle p in the same section in order to get an idea of the orientation of the 356 0. Andersen—Aventurine Feldspar. lamelle in space. It was then possible to identify the same set of lamelle in the second section on which the other angle @ was to be determined. } A more detailed description of the microscopic measurements is unnecessary, especially since the method was only used in the preliminary work and in cases where the angles p of the lamellze were too large to be measured with the goniometer. Fie. 1. n 1 1 1 1 ! | ! 1 ! U 3 Some of the measurements will be recorded in connection with the description of the specimens. Remarks on the Optics of Aventurine Feldspars. Before describing the measurements with the goniometer a brief general discussion of the course of light rays in aven- turine feldspars will be in place. We consider only rays in the main reflection plane, the plane perpendicular to the sec- tion lines of the lamellee, as only such rays are used in the measurements. In fig. 1, A B represents the surface of a cleavage piece con- taining a reflecting lamella H / G H. The plane of the drawing is perpendicular to the section line of the lamella. The lamella, drawn at an angle p of approximately 21°, has O. Andersen—Aventurine Feldspar. B57 the actual position of the main set of lamelle that cause the aventurization on (001) of the feldspars. The rays are con- structed at the approximately true angles obtaining in an oligo- clase (x = 1°54). A very high refractive index (n = about 3) is assumed for the construction of the rays inside the lamella. In the discussions we disregard (as we do in the figures) the double refraction of the feldspar, and also an eventual double refraction of the lamelle. This will simplify the problems very considerably without changing their general character. The course of the different rays originating from the inci- dent ray a p isshown in fig. 1. Itis supposed that the lamella is thin enough to allow a part of the light to penetrate to its lower surface G H/ where one ray o s proceeds into the feldspar and the other is reflected in the direction 0 k. There will then be opportunity for an interference between the two rays h p’ and & p” with a path difference equal to 40 & or approximately the double thickness of the lamella (as the angle A o & is always very small). In white light we will, therefore, see interference colors in the direction p’ a (p” a”) of the light rays which pass out through the surface A B. In the following discussion we use the symbol for the angle a p n (fig. 1), the angle of incidence of rays falling on a surface (A B) and the symbol 7 for the angle a’ p’ n’, the angle at which the same rays emerge from the same surface after the reflection from the lamelle. Fig. 2 shows the relations between the angles 2 and 7 for a number of rays constructed on the basis of the same properties as fig. 1. (Feldspar with m = 1°54; lamelle with angle p = 21°.) The angles 7 and? will differ somewhat for different aventurine feldspars and will especially depend on the angles 358 O. Andersen—Aventurine Feldspar. p. As will be shown further on, however, all the lamella that cause aventurization on (001) have practically the same angle pp (about 21°) and those producing aventurization on (010) have an angle py only a little smaller (around 19°). The differ- ences in the angles 7 and 7 will therefore be small. Fig. 2 may then be considered a fair representation of the general relations between the angles 7 and 7 in aventurine feldspars where the aventurization is observed on the cleavage faces Fie. 3. (001) and (010). The incident rays are marked with small letters and the corresponding reflected rays with the same let- ters distinguished by prime signs. It is seen that rays @ p of angle 7 = 90° are reflected in the direction p a’ at a very small angle 7, For an angle p of 21° all the incident and reflected rays fall on the same side of the face normal p 7. For smaller angles p they may, in part, fall on opposite sides. The direc- tion ep represents the ray which after the refraction at the surface A B will coincide with the normal to the reflecting lamellee. Incident rays along the line e p will, therefore, emerge along the same line, that is, the angles 7 and 7 for these rays are equal and e p represents, ina way, the axis of incidence for the whole reflecting system feldspar-lamelle. Rays falling on the surface A £B in any direction between p a’ and p g will not pass out through the same surface after the reflection from the lamellee, because they will be totally reflected at the surface. O. Andersen—Aventurine Feldspar. 359 By considering fig. 3 we infer easily how the angle p is cal- culated from the angles ¢ and 7 of any rays.* a+r 20) = SIME ot sine p= 5 Ghd SS == 5 SUN eS 2 n n n is the mean refractive.index of the feldspar. In order that rays falling on a certain surface after the reflection from the lamelle shall emerge through the same sur- face, the angle p of the lamella must not exceed the angle of total reflection for feldspar against air. This is easily seen in fig. 4, which represents a feldspar containing a lamella of Fie, 4. qoccctctrce s v a { 1 1 1 1 1 * 1 1 WV Ss angle p= 40°, approximately the angle of total reflection for oligoclase. For lamelle parallel to the surface of the feldspar, the re- flected rays will, of course, coincide with the rays reflected directly from the surface. Such lamelle do not, therefore, produce the same brilliant aventurization as lamellee of medium angles p, because the colored light reflected from the lamelle will be blurred by the white light reflected directly from the surface of the feldspar. We may now consider a case where the light rays pass in through one face and after the reflection from the lamelle pass out through another face. In the case illustrated in fig. 5, AB and A Crepresent the two cleavage faces (001) and (010) of an oligoclase and #’a lamella oriented parallel to (021); that is *See EH. Reusch, loc. cit., p. 401, where many of the optical problems are discussed in detail. - 360 O. Andersen—Aventurine Feldspar. the lamella and the two faces of the feldspar lie in the same zone. ays in the plane perpendicular to the a-axis of the feldspar will, therefore, remain in this plane throughout and the calculations of the relations between the different angles are simple. We find easily the following formule : _Atv—2r ae aap a UT? sine sin 7’ = — ; sin7’= it) 0 => x =. q, The meanings of the different symbols are indicated in fig. 5. Fie. 5. C These formulee enable us to calculate the angle p of any set of lamellz lying in the zone of two faces when the angle » between the faces is known and the angles 7 (or 2,) and 7 of any light ray can be measured. In the preceding discussion we have, in general, tacitly assumed that the light was homogeneous. In fig. 6 the incident ray @ p is supposed to consist of white light. After O. Andersen—Aventurine Feldspar. 361 the refraction at the surface A B the red rays follow the course p hh’ p” a” and the violet rays the course php’ a’. The reflection from the lamelle of aventurine feldspars will thus, in general, be accompanied by a color dispersion of the light. The magnitude of the dispersion will depend on the angle 7 of the incident rays, the angle p of the lamelle and the specific power of dispersion of the feldspar. As this power of disper- sion is low with feldspars the actual color dispersion in aven- Fic. 6. ‘’ : a E turine feldspars will always be insignificant. For lamellee parallel to the surface (ep = 0) there will be no dispersion. Goniometric Measurements. For the goniometric measurements cleavage pieces with smooth faces of the forms (001) and (010) and showing distinet aventurization on both faces were selected. The faces gener- ally measured from 2 x 2™™ to 10 x 10™". The measurements were made with a Goldschmidt two- circle goniometer in the following way: The cleavage face to be used as equator was first adjusted parallel to the vertical circle of the goniometer. The cleavage piece could then be rotated round the horizontal axis without changing the vertical position of the face. In cleavage pieces composed of polysynthetic twins the single twinning lamellee were sometimes broad enough for in- dividual adjustment and contained a sufficient number of 362 O, Andersen—Aventurine Feldspar. reflecting lamellz for measurement. As a rule, however, they were too narrow and the whole set of parallel faces had to be adjusted as one face without regard to the intervening twin- ning faces. By selecting the broadest set this could be done without difficulty and after the adjustment the signals belong- ing to the adjusted faces could be easily distinguished from signals belonging to the other twinning faces. ‘The determination of the angle ¢ of any set of lamellz that reflected light through the adjusted cleavage face would now Fic. ie Collimator 2 consist in measuring the angle between the following two zone axes, both lying in the vertically adjusted face: (1) The zone axis of the two cleavage faces (a-axis). (2) The zone axis of the section lines of the Jamelle. Each zone axis was in turn ad- justed parallel to the vertical axis of the goniometer and the position for each read on the vertical circle. The angles p were determined by measuring the angles 7 and 7 in the following way: The cleavage piece wasadjusted with one face parallel to the vertical circle as before, and the section lines of the set of lamella to be measured were set parallel to the vertical axis of the goniometer. The vertical cleavage face was then fixed in a position which gave a suitable angle of incidence to the rays from the collimator and readings were made on the horizontal circle with the telescope in the following three positions (see fig. 7): (1) Z;, position of reflections from lamellae. (2) Z3, posi- tion of direct reflection from the cleavage face. (3) Zs, posi- tion of direct signal from the collimator. How the angles 2_ a O. Andersen—Aventurine Feldspar. 363 and 7 are computed from the readings is easily seen. The formule for the calculation of p from 2 and 7 have been given before (p. 359). Direct measurements of the angles p could be made occeasion- ally when lamelle were exposed on fracture faces. These measurements were, of course, in no way different from ordi- nary goniometric angle measurements. In determining the angles p by measuring the angles 2 andr we should theoretically obtain two values owing to the double refraction of the feldspar. The deviation between these values is, however, always so small as to be negligible even if a high degree of accuracy were desired. It was, in fact, in most cases, impossible to distinguish two signals in the goni- ometer telescope. The color dispersion of the light observed in the same measurements (see p. 361) was also insignifieant and it was without noticeable influence on the accuracy of the results when white light was used instead of monochromatic. The relative accuracy of the goniometric measurements will, of course, depend on the variable qualities of the cleavage faces and the reflecting lamella. Owing to the poor cleavage faces after (010) compared with those after (001) (especially in the plagioclases) signals reflected from or passing through (001) were much sharper than signals influenced by (010). It should be noticed that the error in the determination of an angle p is much smaller than the actual errors of measure- ments of the angles 7 and 7 from which p is caleulated, pro- vided that the refractive index is approximately correct. This is plainly inferred from the formule p. 359. We find that the error in the angle determination in general will increase in the order pp, py, dp, dy. In the best deter- minations of the angles p the probable error of single measure- ments did not exceed that of ordinary goniometric measurements of medium sharpness (2’ — 3’). In the poorest determinations of the angles ¢ it would reach 1/2° or more. (See Tables 2-12.) In most of the measurements the goniometer telescope was used with a reducing attachment; in a few cases with a lense system of low magnifying power. Each angle value listed in the tables (Tables 2-12) is the average of several (generally 5) readings. Resulis of Measurements. The results of the microscopic and goniometric measure- ments of the planes of orientation may now be briefly sum- marized. 364 O. Andersen— Aventure Feldspar. It was found that the lamelle, in all the different varieties examined, were oriented parallel to definite, crystallographic faces of the feldspar, and all these faces had ‘rather simple indices. The lamelle causing the aventurization on the cleay- age faces were oriented along the same erystal faces in all the varieties. Altogether it was found that the planes of orientation of the - lamellae were parallel to faces of the following forms: (112), (112), (118), (150), (150), (110), (110, (021), (010), (001). Of these (601), (010), (110) and (110) are known as actual faces in practically all feldspar crystals; (021) is also a fairly common form, whereas (112), (112), (113), (150) and (150) belong to the very rare forms that have been observed only occasionally and with insignificant faces. Table 1 gives the angles pp, py, dp and $y of the forms men- tioned, for albite, anorthite and orthoclase calculated from the known axial ratios of these minerals.* Baye For the sake of completeness the angles of the forms (118) and (021) are also given. All the angles are given with posi- tive and acute values. The direction in which the angle should be counted and whether the positive or the negative q@-axis should be used as base is always easily inferred. For comparison with the measurements the calculated angles of the aventurine plagioclases were simply computed by inter- polation between the angles of albite and anorthite on the basis of the known compositions of the plagioclases (see Tables 2-8). This method may not be strictly correct, but it is accurate enough for our purpose and, in fact, probably the most aceu- rate method available,t as the axial ratios of the different plagioclases are but imperfectly known. The angles measured on microclines (microcline perthites) are compared with the calculated angles of orthoclase for the reason that the microcline, as usual, was always so finely twin- ned that the orientation of the reflecting lamelle could not be measured in relation to single twinning lamellee (as in the case * Albite: a; b: c= 0°6367 : 1 : 0-5598 a= 94° 15'; B= 116° 37; y= 87 41’. (A = 86° 24’; u= 63° 28" v = 90° 287.) C. Dreyer und V. Goldschmidt, Meddeleieer om Groénland, xxxiy, 1907. (In the table of elements, p. 48, is given A= 86° 42’ instead of 86° 24’.) Anorthite: a: b: c= 06347: 1 : 0°5501 a = 98° 13’; B= 115° 56’; y = 91° 12’. (A°= 85° 50'; w= 68° 56’: v—87° 6). V. Goldschmidt, Winkeltabellen, (1897, p. 141). Orthoclase: a: b; c; = 0°6585 : 1 : 0:5554 B= 116° 3’. V. Goldschmidt, Winkeltabellen (1897, p. 148). {See V. Goldschmidt, Winkeltabellen (1897, p. 404). 365 O. Andersen—Aventurine Feldspar. . y £9 | G8 6E/ 98 LAL | Te 8 j| & $9} ST GI} EL LL | 4G F8 OST 4G €9 | Ih 8f | PS OF | GS T8 = = yb $9 | 8E 61 | 8s LL) 8S 6h || & E9 | To GE | 8& L4T)cG 8h OST b £9) 26 69 | L6 8G] 06 69 || & &9) 9G 6S | GG Go| OL 69 Olt 4g $9 | S&& 6E | 8& 9G | 4h L9 b *9/|F 88) oh 9¢| SS GO|) & £9 | 8& 09 | OF 68)8 9 OTT 0 0 1 4H) 0 O | 9F 9F 1/0 O | ty 9F)}0 O | GF OF 120 () 0) y GP iO - 99 bP 0 0 | cr 8/0 O | 88 Gil0O O | It &/0 O | & &F T0 VE 91 | 8t §8| ch 9G | 4h GI || 9F OF | 9 E8 | OL 6S | 0€ GT Sit b OF | OF 64 | 8& 9F | @ GT : FE 91 | SE 94) 46 8G | 4G BL || 9F OF | 8E GL} oS Go| FF BT SIL 8 81/1 8] 246 8¢ | 9T Te |) OF 8F} Te 18) co Gc) Ig Te cas nia ohh OV 08h, het o9G ihG 0G Tae) 8 Sf | 1h .Ph | SF 9G |S TS || OL BF | OL 94 | OF .69 | 8E 0c rae We Wd dy dg We Ng dg dg Wg Wd dg dg OLO OF [BOON 100 07 [BULON OFO OF Teuton TOO NOP TE UATON OTO 0} [RULION LOO OBO N || omco : 910g + [Og : e[Og : e[0g : a[0g e0g HaSV TOOHLYO ALIBLYONV alla lV [ wiav iy kk eo) ine) 1915. S, Vou. XL, No. 238.—Ocrozmr, 368 O. Andersen—Aventurme Feldspar. of the plagioclases), but could only be referred to the appar- ently monoclinic elements of the entire twinned cleavage pieces. Moreover, the difference in angles between orthoclase and microcline is insignificant. The faces representing the planes of orientation of the re- flecting lamelle are plotted in stereographie projections in figs. 8-11. Figs. 8 and 9 show the faces observed in micro- clines and figs. 10 and 11 those observed in plagioclases. In figs. 8 and 10 the plane of projection is (001), in figs. 9 and 11 it is (010). The most important planes of orientation are marked with large dots; the less important with smaller dots. The most numerous lamelle are those which produce the aventurization on the cleavage faces. In the plagioclases the aventurization on (001) is chiefly caused by lamelle after (112) and to a less extent by lamelle after (112). In the microclines these two forms are equal. The aventurization on (010) is always caused by lamelle after (150) and (150), both of about equal development, in plagioclases as well as in microclines. Sometimes we observe numerous lamelle parallel to one or both of the cleavage faces (001) and (010), but as a rule these faces contain very few lamelle and often none. The same holds true for the faces (110) and (110). Parallel to (021) there were seen a few lamelle in two of the plagio- clase varieties examined. Lamellee parallel to (113) were only observed once, also on a plagioclase. Orientation of the Edges of the Lamelle. In cases where the lamellze showed definite crystal outlines it was plainly seen that there were certain directions along which the edges were more frequently oriented than along others. This orientation was, however, not so regular as the orientation along the planes. Simple crystallographic relations of the directions of the edges were only found exceptionally. The angles varied considerably even on the same specimen, and on different specimens the orientation was often quite dif- Lerenie) ee It should be emphasized that we do not know the crystallo- ° graphic relation between any of the edges and the crystal axes of the lamelle, even if we take it for granted that the lamellee are hematite crystals. In the six-sided lamelle, for instance, we have no means of deciding whether the edges are formed by faces of the first or the second order ; and in the distorted lamel- le with eight- or ten-sided outlines the identification of the edges is still more uncertain.» Moreover, many of the lamellze have no regular outlines. We are, therefore, in general notin a O. Andersen—Aventurine Feldspar. 369 position to establish the mutual orientation between the crystal axes of the feldspar and those of the lamellze. In view of these facts, it was of minor interest to undertake extensive measurements of the orientation of the edges. Measurements were only made, on sections after (001) and (010) of the angles between the a-axis (cleavage lines) of the feldspar and the projections of the edges of the lamelle on the respective cleavage face. These measurements can be most conveniently included in the special descriptions of the speci- mens. Tur PROPERTIES OF THE REFLECTING LAMELLA. The lamellz were too thin to be separated mechanically from the feldspar. They were, therefore, chiefly studied under the microscope in cleavage pieces or sections of the feldspar. The different aventurine feldspars showed considerable variations as to shape and size of the lamellee, but no distine- tion between different varieties could be made on the basis of other qualities. There is, therefore, no reason to consider that the lamellee consist of more than one mineral species. In the following descriptions we refer to the general qualities of lamelle from all the specimens examined. Size ; crystal outlines. | The smallest lamellee were hardly visible under the micro- scope; the largest could be easily seen with unaided eye. Between these there were all transitions. The lamelle of the microcline were generally smaller than those of the plagioclase, the former seldom measuring more than 0°2™™ in diameter, the latter often measuring as much as 3™™. As observed by Scheerer* and others who studied aven- turine feldspars under the microscope, the lamelle sometimes form nearly regular hexagons but more often they show distorted, six-sided or rhomb-shaped outlines. Unsymmetrical eight- or ten-sided outlines are also often seen, and narrow rectilinear strips several times as long as wide are very com- mon. lLamelle with rounded or irregularly curved outlines are frequently observed. When parallel to one of the cleavage faces the lamellz often showed more regular hexagonal or rhombic outlines than in other positions (see Plates I-III). Interference colors ; thickness. The intense colors displayed by most of the lamelle in reflected light were explained by Sheerert as colors of thin films. The correctness of this explanation could be readily proved in the course of the present microscopic study. The * Loe. cit., p. 156. + Loe. cit., p. 157. 370 O. Andersen—Aventurine Feldspar. light reflected from the opaque lamelle was always of the same grayish color, whereas the reflections from the trans- parent lamellze showed vivid colors varying with the absorp- tion colors, that is, with the thickness of the lamelle. It was, for instance, observed that the absorption colors of some lamellee changed gradually from very light yellow near the edges to light brown red towards the middle. The interfer- ence colors in reflected light changed in the same lamelle from dark gray of first order at the edges to the brilliant colors of second order near the middle. In the thin lamelle of yellowish or light brown tints (in transmitted light) the interference colors were not noticeably modified by the absorption of the one of the interfering rays that passed through the lamelle. As the lamelle became thicker, however, the influence of the absorption was more pronounced and in the deep brown red lamelle the absorption was so strong as to suppress the interference colors. The reflected light of these lamellee was therefore grayish like that of the opaque lamelle. We know that the thickness of the lamelle is about 1/2 of the path difference between the interfering rays reflected from them. It is obvious, then, how the thickness of the thinner lamelle that show distinct interference colors can be approx- imately determined. The very thinnest of the lamelle showed the interference colors gray and white of first order and their thickness could accordingly be estimated at from 50 to 100 mp. Thicker lamellee showed interference colors from yellow of first to red of second order corresponding to thickness of from 150 to 500 up. The majority of lamelle had a thickness of between 100 and 400mm. In sections where the lamellz were cut approximately perpendicular to their planes they appeared as almost invisible streaks entirely too thin to be measured with the ordinary devices of the microscope. During the observation of the interference colors it was noticed that light reflected from lamelle with small angles p never showed any detectable polarization. Absorption colors. The color of the reflecting lamelle in transmitted light varied with their thickness from very light yellow and reddish brown to deep brown red or blood red. Lamellee forming small angles p with the plane of the section showed no pleochroism. Lamelle of large angles p, on the other hand, appeared at first sight to be strongly pleochroic. In ordinary thin sections of the feldspar we observe the follow- ing absorption colors of such lamellz (e. g. lamelle parallel to (112) seen in sections after (010) or lamelz parallel to (150) in O. Andersen—Aventurine Feldspar. 371 sections after (001) ): (1) In vibration directions perpendicular to the section lines of the lamellee—colors varying from yellow to strong brown red or blood red (depending on the thickness of the lamelle). (2) In vibration directions parallel to the section lines of the lamelles—colors dark brown of nearly the same tinge in all lamellee (independent of their thickness). The change is thus apparently stronger in the thinner than in the thicker lamelle. There is no distinct change in the quality of the colors but rather a change in the tints of the same brownish color from dark in one direction to light in the other. This change in tints has been observed by previous authors and explained as pleochroism.* As the reflecting lamelle have been considered hematite in tabular crystals after the base a strong pleochroism > e (@ dark brown, e light brown) has been generally adopted as one of the characteristic qualities of hematite. A somewhat closer consideration of the observa- tions described will, however, show that this explanation is not correct. If the lamelle are hematite crystals in plates after the base, those parallel to the surface of the feldspar section, or forming small angles (p) with the same, must show the absorption color of the vibration direction . According to the observations on such lamellz their absorption colors vary between yellow and brown red. According to the observations on lamelle of large angles p, on the other hand, the color of the vibration direction (direction parallel to the section lines of the lamelle) should be dark brown with very little variation, whereas the colors of ¢ (or strictly a direction between e and ) should vary between yellow and brown red. In other words the absorption colors of @ in the lamelle with small angles p correspond to the colors of ¢ in the lamelle with large angles p. This indicates that the lamelle have only a weak pleochroism, if any, and the dark absorption colors for vibration directions parallel to the section lines of the lamelle of large angles p must be explained in the manner outlined below. Fig. 12 shows the light rays passing through a cleavage piece A BCD containing a lamella HG H of angle p= 75°. The light of the incident ray a p / is supposed to be unpolar- ized. By the influence of the lamella the refracted and reflected rays become polarizedt with the reflected ray hp’ a/ *F. Rinne, Neues Jahrb. Min., 1890, i, p. 183. + The phenomenon of polarization by reflection and refraction is too well- known to need any explanation. It may only be recalled that the polariza- tion, in the case of metallic substances like the lamellz here considered, is never complete either in the reflected or in the refracted ray, but reaches a maximum for a certain (always large) angle, the main angle of incidence, and becomes insignificant for small angles of incidence. C. Forsterling (Neues Jahrb. Min., B. B., xxv, 360, 1908) determined the main angle of incidence for hematite at 71°-73° for rays of medium wave lengths. 372 0. Aiidlby sepi=oseh turine Feldspar. vibrating perpendicular to the plane of incidence and the trans- mitted ray o p’”’ a’” vibrating in the same plane. Now if the incident ray (ap), therefore, consists of polarized light the transmitted ray (p’” a’’’) will have its maximum of intensity for the vibration direction perpendicular to the section line of the lamella and the minimum for the direction parallel to this line. This is exactly the apparent pleochroism (@ > e) observed on lamelle of large angles p. The correctness of the above explanation was proved by the fact that the light reflected from lamelle of large angles p HrGeloe a” a (rays p’ a’, fig. 12) was found to be strongly polarized with vibration direction parallel to the section lines of the lamelle. This was observed in a number of sections with lamelle of angles p around 75°. It is then a necessary conclusion that the transmitted rays (p’” @’’’) must be polarized with vibration direction perpendicular to the section lines. The apparent pleochroism in the lamelle of aventurine feldspars when observed at large angles p is not, therefore, due to any strong difference in absorption between the vibra- tion directions @ and e (if we suppose that the lamelle consist of hematite tables), but is explained, as outlined above, by the polarization of rays when they fall on the lamelle at the — appropriate angles of incidence. O. Andersen—Aventurine Feldspar. 373 In thick sections or cleavage pieces the lamelle of large angles p show in polarized light (with only one nicol) dis- tinct interference spectra with bands parallel to the section lines. ‘These lamelle show the same apparent pleochroism as those in thin sections. The thicker the part of the feldspar penetrated by the lamelle the larger is the number of bands in the spectra. In lamelle penetrating only a thin part of the section we see, as in ordinary thin sections, no extended spectra but sometimes single, colored stripes. By rotating the micro- scope stage all spectra disappear in two positions at right angles with each other coinciding with the extinction direc- tions of the feldspar between crossed nicols. This phenomenon is explained by the fact that the lamellee (of large angles p) are included in the doubly refracting feldspar. In fig. 13 4 Fie. 13. indicates a lamella included in a feldspar section A B CD at an angle p large enough to make the rays transmitted through it become noticeably polarized. If we use the lower nicol of the microscope the lamella will form a somewhat imperfect analyzer for the wedge-shaped part of # /’ fe of the feldspar that lies below it. If we use the upper nicol the lamella will be polarizer for the wedge-shaped part # /'f’ e’ of the feldspar that lies above it. In either case we shall obtain an interfer- ence spectrum corresponding to the part of the feldspar wedge that lies between the lamella and the surface of the section, (the lower surface in one ease, the upper in the other). It is obvious that the bands of the spectra must always be parallel to the section lines of the lamelle. In such a stauroscopic system, where one of the nicols is replaced by lamellee included in the section to be observed, we can not change the angles between the ‘“nicols” and at the same time the relative posi- tion between the vibration directions of the “ nicols”’ and those of the section at will. If the “nicols” are crossed the feldspar 374 O. Andersen—Aventurine Feldspar. section may not be in a favorable position for showing inter- ference colors and when the feldspar is in the most favorable position the vibration directions of the “nicols’** may form only a small angle with each other. Remembering this the behavior of lamelle of large angles p on rotating the micro- scope stage presents no difficulty. Double Refraction. Lamelle parallel to (or forming small angles p with) the sec- tions extinguished simultaneously with the feldspar between crossed nicols. Lamelle of large angles p, on the other hand, remained light when the feldspar extinguished and thus proved to be anisotropic. This is the relation to be expected if we consider the lamellz uniaxial crystals in plates after the base. Attempts to discover axial figures in the isotropic lamelle failed, but this is not surprisimg when we remember the ex- treme thinness of the lamellz, most of which were actually no thicker than about 1/100 of the thinnest rock sections (see p- 370), and also consider the disturbing influence of the bire- fringent feldspar in which the lamellee were enclosed. The course of the rays through lamelle showing double refraction’ is seen in fig. 12 (aphop’’ a”). The rays are never transmitted through the lamelle in a direction near the plane of the lamelle. If the lamelle are considered to be hematite or other uniaxial crystals in basal plates, we see that the transmitted rays will never have the vibration direction e. Consequently we shall in no case obtain the maximum double refraction and the path difference of the transmitted rays will depend on the thickness of the lamellee and on their angle with the section (angle p). Chemical Tests. Scheerer* found that the originally red powder of sunstone from Tvedestrand turned white on heating with hydrochloric acid, the filtrate containing iron oxide. In the course of the present study similar tests were made on samples of various aventurine feldspars with the same result. Microscopic ex- aminations of the powder showed that the red (pink) color was due to the presence of the thin reflecting lamelle described and the discoloring by treatment with HCl was due to the solution of these lamelle. These tests, therefore, show that the lamelle of the aventurine feldspars contain iron oxide. * Pogg. Ann., lxiy, 160, 1845. O. Andersen—Aventurine Feldspar. 315 Some Observations on Hematite. For comparison with the reflecting lamelle of aventurine feldspar several samples of hematite were examined as to trans- parency, absorption colors and pleochroism. Fine powder of the samples was imbedded m liquid Canada balsam and examined under the microscope. The degree of transparency could be estimated by selecting tabular grains lying on the flat surface, noting the color and then tilting the grains on edge (by moving the cover glass) for measurement of the thickness. The transparency varies within wide limits, some hematites being practically opaque even in the finest powder, others being transparent. For a sample of medium transparency (micaceous hematite from Montgomery County, Pa.) the fol- lowing observations were made: Absorption color Thickness Perfectly opaque 0:0 22 Very dark blood red, almost opaque 0:003™™ Deep blood red 0-001=™ Grains of various orientation were seen in the sections. Some were parallel to the base and were perfectly isotropic ; others were oriented at angles with the base and showed a strong double refraction with definite extinction directions. If the hematite possessed a strong pleochroism this should be observed in the grains showing double refraction. In spite of careful observation, however, no distinct change in color or tint could be seen in any of the grains examined. We, there- fore, conclude that the hematite has little or no pleochroism. The absorption colors of the transparent varieties of hematite are very characteristic, blood red, brown, red, etc., according to variety and thickness and are distinctly different from those of goethite, which are much lighter and always a purer brown or yellowish. It will be seen that these properties of hematite agree well with the corresponding properties of the reflecting lamelle of aventurine feldspars. Summary of the Properties of the Lamelle. We may now summarize the data bearing on the identifica- tion of the reflecting lamelle. (1) The hexagonal outlines often shown by the lamelle, taken in connection with the optical properties which agree with uniaxial crystals im plates after the base, point to a hex- agonal or trigonal symmetry. (2) The absorption colors are the same in the lamelle as in hematite and the lack of distinct pleochroism is also character- istic for both. 376 O. Andersen—Aventurine Feldspar. (8) Chemical tests show that the lamelle contain iron oxide.* (4) Thermal experiments (to be described below) show that the lamellee do not undergo any essential change even by a prolonged heating of the aventurine feldspar at temperatures around 1050°. If the lamelle consisted of goethite or other hydrated iron oxides we should expect a considerable change due to the decomposition of such hydrates by heating. All these data lead to the conclusion that the reflecting lamellae of the aventurine feldspars consist of hematite in tabular crystals after the base, as first suggested by Scheerer. THERMAL Dara. In order to obtain, if possible, some information on the stability relations between the feldspar and the hematite in- clusions, a number of heating experiments were carried out.t+ Fresh, transparent cleavage pieces containing hematite lamellz of various size and thickness were selected for the experi- ments. The outlines of the cleavage pieces and their included lamellee were drawn with camera lucida and the colors of the different lamelle in transmitted light were noted. After appropriate heating (in the electric resistance furnace) the cleay- age pieces were removed to the air, examined under the micro- scope and compared with the drawings. A brief record of the experiments is given below. (1) Cleavage piece with numerous transparent hematite lamellae heated for one hour at 1260°: The feldspar remained birefringent throughout with outlines sharp as before the cheating; colorless; transparent, but somewhat dim. All hematite inclusions disappeared. (2) Cleavage piece with few light-colored hematite lamelle of sharp outlines heated for one hour at 1150°: No visible change. (8) Same piece heated for one-half hour at 1200°: No change. (4) Same piece heated for one-half hour at 1230°: Feldspar unchanged. Some hematite lamellze disappeared; others be- came lighter and were corroded at the edges. (5) Piece from exp. (1) heated for twenty-four hours at about 1050°: Feldspar milk white, dull, full of very fine black dust. (6) Cleavage piece with numerous hematite lamellae heated for one-half hour at 1235°; Feldspar birefringent ; colorless ; somewhat dim, but still transparent. All hematite disappeared. *In this connection it should be noticed that lamelle of similar qualities forming inclusions in carnallite have been analyzed separately and found to consist of Fe.O3 without H.O (O. Ruff, ‘“‘ Kali”, i, 81, 1907). {Specimens from Aamland, Séndeled, Norway, yielded the most favorable material for these experiments on account of the freshness of the feldspar and the large size of the hematite lamelle. : O. Andersen—Aventurine Heldspar. 317 (7) Cleavage piece with some transparent and some opaque lamellee heated for one hour at 1230°: Feldspar practically unchanged. Nearly all hematite lamelle disappeared. (8) Piece from exp. (7) heated for eighteen hours at about 1050°: Feldspar white, dull, only little transparent. The opaque lamellze reappeared in the same positions and with the same outlines as before the heating (of exp. 7). They did not reflect the light as before, however, and were evidently made up of a fine aggregate of an opaque or very dark brown sub- stance. The originally transparent lamelle did not reappear, but at the places they had occupied a dense crowding of black dust was seen. (9) Piece from exp. (8) heated for one hour at 1235°: The opaque lamelle again disappeared. (10) Same piece heated for 40 hours at about 1050°: Opaque lamelle reappeared as in exp. (8). The feldspar opaque, full of black dust. (11) Cleavage piece with opaque and transparent lamelle heated for one hour at 1235°: Feldspar practically unchanged, only a little dim. Both the opaque and the transparent hema- tite lamelle disappeared. (12) Same piece heated for 45 hours at about 1050°: Feld- spar dim and full of black dust. Opaque lamelle reappeared in the same positions as before. The substance of the lamelle now a dark aggregate as in exp. (8). (13) Cleavage piece with opaque and transparent hematite heated for 22 days at about 1050°. The piece was examined and replaced after 1, 2, 7 and 14 days of heating. After 1 day some of the opaque lamelle had become transparent with a deep brown color; others remained opaque. The originally transparent lamellae were apparently unchanged. After 2 days no further change visible. After 7 and 14 days all lamelle had become visibly lighter, some of the originally opaque lamellz now being reddish brown, others being deep red. After 22 days the same relations persisted. The feldspar remained perfectly clear throughout (without formation of dark dust as in the cases where the heating had been first earried up to 1230°-1260°). The reflections from the lamellee were just as intense as before the heating. The result of these experiments may be summarized as follows: By heating fresh cleavage pieces of aventurine feldspar at temperatures below 1230° (around 1050°) the hematite lamellee undergo a slow change. The opaque lamellz become more or less transparent and the transparent ones generally grow a little lighter. The change from opaque to transparent with brown red color seems to take place rather quickly, whereas 378 O, Andersen—Aventurine Feldspar. the change from darker to lighter brown red or to yellowish brown is very slow and in the lightest lamelle a change is hardly detectible, even after a long heating. The opaque lamellze are seldom homogeneous, but generally contain a num- ber of irregular spots of transparent substance which shows the same colors as the entirely transparent hematite lamelle. By a short heating these transparent parts undergo practically no change while the opaque parts become transparent and soon acquire the same color as the transparent parts so that the lamellee become homogeneous. It seems as if the changes take place with retarded velocity and gradually cease when a certain stage is reached. These changes might be explained as being due to a direct solution of the hematite into the feldspar (in the solid state) whereby the hematite lamelle grew thinner and lighter. In that case we should expect the most conspicuous change in the thinnest lamellee, some of which ought to disappear if the heat- ing were continued for a sufficiently long time. As the ex- periments show, however, the thinnest lamellee were evidently the most persistent ones, none of them disappearing and most of them undergoing no visible change even after 22 days’ heating. The change may be explained by assuming that there is a transition from a darker to a lighter modification of the hematite. The darker form is perhaps a secondary product and the original, lighter form is restored by heating. An ex- planation of this nature seems to account for the actual behavior of the different hematite lamelle on heating. As the experimental data are few and only suggestive we have no basis for a detailed discussion of these problems. The sudden disappearance of the hematite lamelle at 1235° is most reasonably interpreted as due to a simultaneous melting of hematite together with a portion of the feldspar surround- ing it; perhaps a eutectic melting or possibly a melting witha reaction between the feldspar and the hematite. Owing to the extreme thinness of the lamelle the amount of feldspar neces- sary for such a melting must in any case be small. The liquid (glass) resulting from the melting will therefore occupy only very thin films in the place of the lamelle and will escape detection under the microscope. It will look as if the lamellee disappeared without leaving any trace, while the surrounding feldspar was unchanged and showed no sign of melting. The fact that the opaque lamelle reappear in the same places by heating at a lower temperature proves that their substance can not have travelled far. The substance of these reappearing lamellze is evidently different from that of the original lamelle, and the iron oxide must therefore have undergone some change i> al O. Andersen—Aventurine Feldspar. 379 by the melting and recrystallization. We have no means of deciding of what nature this change may have been. It is not unlikely, however, that at least a part of the hematite in melting has been reduced to magnetite. ORIGIN OF THE HEMATITE LAMELLZ. The investigations above described show that the aventurine feldspars must be considered oriented intergrowths between feldspar and hematite. ‘The essential features of the inter- growths are these: The hematite crystals form exceedingly thin plates after the base and the plates are oriented parallel to some simple crystal faces of the feldspar with the edges of the hematite crystals also, in general, definitely oriented. These facts should be borne in mind when we seek an adequate explanation of the origin of the hematite lamelle. Of the different possible theories there are two that seem to account well for the oriented intergrowth, viz.: The theory of simultaneous crystallization suggested by Scheerer and the theory of unmixing in the solid state intimated by Johnsen (see review of literature). In discussing the origin of the hematite Jamellz in carnellite and cancrinite Johnsen points out the reasons for preferring the theory of unmixing to the theory of simultaneous crystallization in the-cases of the two minerals mentioned. Similar reasons are evidently valid also for the aventurine feldspars. It is obvious that the planes of growth of the feldspar have not, in general, coincided with the principal planes of orienta- tion of the hematite lamelle, as the latter planes represent extremely rare forms with the feldspars. It is highly probable, for example, that the faces 112 and 150 have never existed as erystal faces (faces of growth) in any of the specimens con- sidered, and still they are the most important planes of orien- tation of the lamelle in all aventurine feldspars. If the aventurine feldspars were considered products of simultaneous crystallization of feldspar and hematite, therefore, we would have to assume that the majority of the extremely thin hema- tite lamelle during their growth formed angles with the prin- cipal faces of growth of the feldspars. This is improbable according to the common experience with crystalintergrowths, and the orientation of the hematite lamellze thus forms a strong objection against the theory of simultaneous crystalliza- tion; a theory which otherwise would seem very reasonable. The formation of aventurine feldspars by unmixing in the solid state may be conceived as follows: The feldspar crystals were, at the time of their separation, wholly or in part homo- geneous and contained small amounts of Fe,O, in solid solution, 380 O. Andersen— Aventurine Feldspar. either as hematite or as a constituent of a ferric compound. By a change in the exterior conditions prevailing at the time of formation, e. g. change in temperature, the equilibrium of the solid solution may be disturbed in such a way that Fe,O, can no longer be held in solution, but must separate in individual crystals. The hematite molecules will then move towards the centers of crystallization (that is the locations of the hematite lamellze) and feldspar molecules must move in the oppo- site directions. From the extreme thinness of the lamellae we conclude that practically all these movements have taken place in the planes of orientation of the lamelle. These planes, there- fore, seem to represent definite structural planes in the feldspar, perhaps translation planes,* along which the molecules can move relatively easily. In such planes there will again be certain direc- tions, translation directions,* of maximum mobility of the mole- _ cules, and these may account for the distortions of the hema- tite lamelle. It should be noticed that some of the planes of orientation actually are important structure planes of the feldspar. Thus (001) and (010) are both cleavage planes and twinning planes. Of (110) and (110) one or both are cleavage planes and (021), is a twinning plane. It is, therefore, reason- able to consider the other planes of orientation, especially (112), (112), (150) and (150), which are observed in all aven- turine feldspars, as definite structural planes, planes of transla- tion, as suggested. We may summarize the conclusions as to the origin of the hematite lamellee as follows: The aventurine feldspars have been formed by unmixing of an originally homogeneous solid solution of the feldspar and hematite (or a ferric compound) in such a manner that thin hematite lamella have separated along structural planes (translation planes) of the feldspar. II. DESCRIPTION OF THE SPECIMENS. The present section contains brief descriptions, including tabulations of measurements, of all the specimens examined. The optical properties of the feldspars were only determined to the extent necessary for an identification of the species.+ Extinction angles on (001) and (010) were determined on thin sections or cleavage pieces. Refractive indices were deter- mined on powdered material by the immersion method. As a rule only approximate determinations of the mean refractive index 8 were made in white light. Exceptionally the refractive *See A Johnsen, Fortschritte der Mineralogie, vol. iii, p. 93, 1913. + The graphical plot devised by F. E. Wright, this Journal (4), xxxvi, 540, 1918, was used in the determinations of the plagioclase, O. Andersen—Aventurine Feldspar. 381 index of the feldspar glass, produced by melting the powdered feldspar, was determined.* In the descriptions of the hematite lamelle only the features most characteristic for each specimen are given. Qualities like thickness, absorption colors, etc., have already been described, and as they were practically the same for all varie- ties they will be mentioned only exceptionally in the follow- ing descriptions. Albite from Fisher Hill Mine, Mineville, Essex County, New Yo ork. The feldspar.—The cleavage pieces were rather fresh, trans- parent, of a strong red color, with patches of a green substance not identified. Extinction angle on (001) = + 3° “ 6c 6c (010) — a6 IIS)" B = 1535 Composition: Ab,,An,, a comparatively pure albite. Polysynthetic twinning after the albite law with the one set of lameilz comparatively broad and the other very narrow. The twinning striation on (001) was accordingly very fine. The hematite lamelle.—The aventurization was rather sub- dued, silky, produced by a great number of very small lamelle. Most of the lamellz form very narrow, linear strips of maxi- mal dimensions 0°3 xX 0:01"; some form larger flakes with more equal diameters (maximum 0°3™"). The narrow lamelle are generally irregularly rounded at the ends, seldom showing edges that indicate six-sided outlines. The larger flakes are sometimes approximately hexagonal, but more often irregularly rounded or tongued. Figs. 1 and 2, Pl. I show the most char- acteristic shape of the lamellee. Orientation of the lamelle.—The goniometriec measurements are given in Table 2._ The planes of orientation were: (001), (010), (112), (112), (021), ( (150), (150). Most prominent were (112), (150), and (150), all of which contained numerous lamelle. After (010) there were also many lamelle, often larger than the others ; after (001), (112) and (021) there were few. The lamelle after (112) and (112) were generally oriented with the projections of the elongated edges on (001) parallel to the a-axis. Most of these lamelle were of the type of nar- row strips described above. Other lamellz after (112) were oriented with the projections of their elongated edges on (001) * This method for determining plagioclase feldspars is very convenient, when a high temperature furnace is available, and probably as accurate as ‘the best of the other optical methods. See E. S. Larsen, this Journal (4), XXvili, 265, 1909. {Specimens from U. 8. National Museum, No. 47773. 382 O. Andersen—Aventurine Feldspar. TABLE 2 AbooAns n= 1°535 Measured p @ Form Calculated Calculated | Calculated i r from from Measured | from measurements | axial ratios | axial ratios Pole: Normal to 001 (pp and ¢p) 62° 18’ | 10° 21’ | 20° 58’ 58° 30’ 112 20° 39’ _- eS 59° 0' 62 50 | 10 10 215 0 57 22 62 43 | 12 14 21 40 56 30 | 112 21 49 | |e on é 62 52 | 12 22 21 438 57. OO | 021 {|166 52 | 27 24 46 23 46 49 0 0 OF e0 Pole: Normal to 010 (py and ¢y,) 150 4454 | 17 15 19 16 19 24 |] 68 24 Se 3 150 45 7% | 16 35 19 6 19 16 64 22 Oo forming +58° or +63° with the a-axis. They formed com- paratively broad strips and penetrated a number of the twin- ning lamellee of the feldspar. Lamelle after (021) were often oriented with their elongated edges approximately parallel to the a-axis. Most of the lamelle after (150) and (150) formed narrow linear strips parallel to the c-axis; the angle between these strips and the a-axis was measured at about 63°5°. The lamelle after (010) formed the larger flakes observed in sec- tions after (010). Albite from near Media, Delaware County, Pennsylvania.* The feldspar.—The cleavage pieces were fresh, transparent, grayish or colorless. Extinction angle on (001) = + 1° 6 73 66 (010) = 2b 16° B = 1°535 Composition : Ab,,An,, albite. The feldspar consisted of single individuals, sometimes with- out any twinning but generally with very narrow twinning lamellee after the albite law inserted at regular intervals. * Specimens from U. S. National Museum, No. 79828. O. Andersen—Aventurine Feldspar. 383 The hematite lamelle.—Large portions of the feldspar were perfectly free from lamellz ; others contained many and showed a strong aventurization. Most of the lamelle were elongated, often without regular terminal edges. Sometimes they were also approximately six- _ sided or rhomb-shaped. They were very small, seldom more than 0°2 x 0-1". Orientation of the lamelle.—The goniometric measurements are given in Table 3. The forms observed as planes of orienta- tion were : (001), (010), (112), (112), (150), (150). Of these the cleavage faces (001) and (010) contained very few lamelle, often none. The majority of lamelle were orientated after (112), fewer after (112), while (150) and (150) both contained a considerable number. Projections of the elongated edges of lamelle after (112) on (001) formed often about +73°, sometimes —14°, with the a-axis. Other measurements did not seem to represent general orientations. TABLE 3 AbgiANg ;n= 1-585 Measured p @ Form Calculated Calculated Calculated i r from from Measured from measurements | axial ratios axial ratios Pole: Normal to 001 (pp and ¢p) 47° 18’ | 20° 26’ 20° 53’ 58° 41’ 112 58 45 | 12 42 21 2 20° 40’ 58 37 58° 57’ 59 39 | 11 36 20 82 58 55 55 608] «17 «O87 21 56 56 24 112 58 389 | 15 12 21 49 21 48 55 «49 566 59 55 | 14 20 21 58 56 24 Pole: Normal to 010 (py, and ¢,, ) 150 54 21 10 54 | 19 31 19 24 63 31 & 47 40 | 15 12 iG) 118) 63 20 63 27 150 ig) ile =e 54 45 9 49 19 16 63 58 — Am. Jour. Sct.—Fourts Smrizs, Vou. XL, No. 238.—Ocrossr, 1915. 25 384 O. Andersen—Aventurine Feldspar. Oligoclase from near Statesville, Iredell County, North Carolina.* The feldspar.—The cleavage pieces had a grayish-brown color, and were not quite fresh, being clear in spots only. Extinction angle on (001) = + 1/2° (45 66 [73 (010) = + g° B = 1:54 Composition : Ab,,An,,, oligoclase.t The cleavage pieces form polysynthetic twins after the albite law, the lamelle of the one individual being compara- tively broad, those of the other extremely narrow. The hematite lamelle.—In the clear spots of the feldspar there were numerous lamelle which produced a strong aven- turization. The lamelle formed sometimes long narrow strips and sometimes distorted six-sided or rhomb-shaped plates. The largest dimensions of the strips were 0°8 x 0:03™"; the plates were seldom more than 0°3 xX 0:17". Orientation of the lamelle.—The goniometric measurements are given in Table 4. The planes of orientation were: (001), (010), (112), (112), (118), (021), (150), (150). The most prominent set of lamelle were parallel to (112), (150) and (150). After (112) there were fewer, and after the other planes, (113), (021), (001) and (010), very few. é The projections of the long edges of lamelle after (112) 0 (001) often formed about +-74° with the a-axis. These lamellee were elongated, strip-shaped, and traversed a great number of twinning lamelle. Six-sided lamelle after (112) were oriented with the projections of one of the edges on (001) forming —80° with the a-axis. The lamelle after (112) were often elongated — with the projections of the long edges on (001) approximately parallel to the a-axis. Lamelle after (150) and (150) had dis- torted hexagonal or rhombic outlines. The projections of their elongated edges on (010) sometimes formed +11° with the a-axis. Oligoclase from Krageré, Norway. t The feldspar.—The cleavage pieces have a strong, red color, and are usually perfectly fresh and transparent. Extinction angle on (001) = +1° “c “ (3 (010) = se B = 1°543 Composition : Ab, An >» Oligoclase. * Specimens from U. S. Nationai Museum, No. 80324. +G. F. Kunz has described orthoclase sunstone from Statesville, N. OC. (History of the Gems found in North Carolina, p. 27). This may be the same sunstone as the one described here. Owing to the very fine twinning striation the feldspar might have been mistaken for orthoclase, by a macro- scopic examination. {Specimens from U.S, National Museum, No. 44776. O. Andersen—Aventurine Feldspar. 385 TABLE 4 Abg.Anj3 ;vu= 1:540 Measured p @ Form Calculated | Calculated Calculated t r from from Measured from measurements | axial ratios axial ratios Pole: Normal to 001 (pp and ¢p ) 46° 40’ | 20° 53’ 20° 47’ DOs 20° 48) ft age aye 72 40 6 5 Pile wt 59 11 46 5d 23 20 21 36 569. 21 44 =e 56 23 72 2&2 7 48 21 42 ail) ts} 46 56 Viele lied 9 49) 19 35 569 56 23 162 24 31 29 46 53 0 0 269 20 20 4 46 47 46 48 0 O OW) 369 42 25 27 46 37 0 0 Pole: Normal to 010 (py and ¢y ) 48 58 14 50 19 22 64 2 Die 8 50 19 24 19 28 63 51 58 4] 8 58 19 38 63 21 63 31 48 29 13 55 19 2 63 538 19 18 _ 56 48 8 32 19 14 63 38 12 = 93° 37’ 2}, = 98° 55’ 32 = 98° 57’ The feldspar consisted of large, single individuals, or was sometimes composed of broad twinning lamell after the albite law. The hematite lamelle.—As the lamellz were large and pres- ent in great numbers the aventurization was exceedingly strong, especially on (001). Large lamellz, measuring 2 X 1™™ or more, were numerous, but all sizes down to the very smallest were seen. The outlines were sometimes six-sided and often rhomb- shaped or elongated. Light-sided or quite irregular outlines were also frequently seen. (Plate II, fig. 1.) Opaque hexagonal lamella were sometimes arranged in regu- lar groups with the edges of each lamella parallel to the six- sided outlines of the groups. 386 O. Andersen—Aventurine Feldspar. — Orientation of the lamelle.—The goniometric measurements are given in Table 5. The planes of orientation were: (001), | (010), (112), (112), (150), (150), (110). TABLE 5 AbegoADoo >n= 1°5438 Measured p @ Form Calculated Caleulated Calculated ~ 4 r from from Measured from measurements | axial ratios axial ratios Pole: Normal to 001 ( pp and ¢p) 59° 35’ | 12° 14’ 20° 57’ 57° 54! 112 nel 17) TON 52 20 50 20° 43’ 58 (0 58° 41’ 1 20 53 58 8 a 60 39 | 14 28 21 50 56 36 } 112 a1 44 ———— | : Gl BO |) Wes 21 34 By 8 Pole: Normal to 010 (py and $y ) 150 1 19 30 19 28 | 68 38 150 ae 19 18 63 55 63 31 1 Direct measurement. Lamellee after (110) were identified in sections after (001) and (010) by the measurements: dp = 58°; dy = 64°. There were numerous lamellee after (112) and comparatively few after (112). After (150) and (150) there were also many ; after (001), (010), and (110) very few. For some of the lamellee after (112) the projections of the longest edges on (001) were parallel to the a-axis. For other lamelle in the same plane the edges formed +60° or +47° with the a-axis. Oligoclase from Tvedestrand, Norway.* The occurrence at Tvedestrand has been described by Weibyt as consisting of veins in gneiss, the essential minerals of the veins being oligoclase (sunstone) and quartz ; accessory minerals, apatite, hematite, cordierite, hornblende and zircon. *Specimens from the Mineralogical Museum of Kristiania University. | Cited by Th. Scheerer, loc. cit., p. 154. O. Andersen—Aventurine Feldspar. 387 The feldspar.—The cleavage pieces were generally fresh and clear, of a strong, red color. Extinction angle on (001) = +1° (19 73 (13 (010) — SG 05s B = 1°545 Composition: Ab,,An,,, oligoclase in agreement with Scheerer’s analysis.* The oligoclase was generally twinned both after the albite law and the pericline law, most of the twinning lamelle being broad (often more than 1™). The pericline striation on (010) formed +10° with the aaxis. Sometimes broad cleavage pieces, without any twinning, were seen. The hematite lamelle.—In some specimens the hematite lamellee were very scarce; in others they were densely crowded and of large size, producing the most brilliant aventurization. Lamelle measuring 3 X 2™™ or even more were frequently seen. The outlines were hexagonal, rhomb-shaped or irregular. The lamellze after (010) were often larger and more irregu- larly outlined than the others. Some few lamellae were opaque, but most of them were transparent with the usual absorption colors. Orientation of the lamelle.—The determinations of the planes of orientation made by Scheerer (see p. 353) and Tertsch (see p. 354) have already been mentioned. Of these only Scheerer’s have been partly confirmed by the present study (viz., the faces (001), (010) and (110)). The planes given by Tertsch, (538) and (417), could be found neither on the Tvede- strand sunstone nor on any of the other varieties examined. That Scheerer’s determination of (221) as a plane of orienta- tion is wrong was intimated by E. Reuscht and further proved by Tertsch.t According to my measurements (of which Table 6 contains those made with goniometer) the following forms were planes of orientation : (001), (010), (112), (112), (150), (150), (110). Of these (112) contained the largest number of lamelle which cause the brilliant aventurization on (001); (112) contained but few. After (150) and (150) there were numerous lamelle pro- ducing a strong aventurization on (010). Along the other faces (001), (010) and (110) there were comparatively few, none causing aventurization on the cleavage faces. The face (110) was identified as plane of orientation by microscopic measurements: dp = 56° — 57°; dy about 62°. The projections on (001) of the elongated edges of lamellee after (112) were often parallel to the a-axis ; some formed +72° *Loc. cit., p. 155. + Pogg. Ann., exvi, 396, 1862. ¢ Loe. cit., p. 248. 388 O. Andersen—Aventurine Feldspar. TABLE 6 Ab7zsANeo; n = 1°545 Measured p ) Form | Calculated | Caleulated | Calculated i r from from Measured from | measurements | axial ratios || | axial ratios Pole: Normal to 001 (pp and ¢p) 45° 4’ | 22° 11’ 20° 43’ 57° 55! 50 s2| 18 2 || 20 46 | 58 17 54 32 | 15 2 20 45 58 26 112 62 40 | 10 1 20 47 20° 44 || 58 22 58° 34’ 67 26 7 24 20 44 57 «449 68 40 7 3 20 49 58 24 1 20 50 58 27 48 30 | 22 31 21 42 56 48 * 2) 16 |) 19) 21 21 35 56CO7 112 21 42 al OMES 62 40 | 12 12 21 29 56 34 66 14 | 10 55 21 41 56 29 Pole: Normial to 010 (py, and $4, ) CeLOmme a 63 22 150 19 30 1 19 40 63 16 45 53 | 17 6 19 20 || 63 55 63 33 150 46 12 | 16 39 19 16 1 19) 63 26 oe LO any, 63 29 = Direct measurement. with the a-axis. Projections of the long edges of lamelle after (150) and (150) formed —77° or —84° with the a-axis. Lamelle after (110) were generally elongated approximately parallel to (001). O. Andersen—Aventurine Feldspar. 389 Oligoclase from Aamland, Séndeled, Norway.* The occurrence is much like the one at Tvedestrand (accord- ing to Weiby’ s description of the latter). At the place where the specimens were found the prevailing gneiss included a great number of irregular pegmatite veins varying in thickness from 0:5 m. down toa few mm. The main minerals of the veins were oligoclase (sunstone), quartz and cordierite ; accessory minerals, hornblende, biotite, apatite and magnetite. The feldsp ar.—The oligoclase is generally very fresh, in places strong red from the hematite inclusion, sometimes gray- ish or colorless. Eetpetion angle on (001) = +1° — 2° “ (010) = +5° B = 1°545 Refractive index of glass, ny, = 1°507 + 0°001 Composition : Ab,,An,,, oligoclase, practically the same as the sunstone from Tvedestrand. The crystals are often twinned with coarse lamelle, after the albite and pericline laws, thus showing striation on (001) and (010). Very often, also, large pieces without twinning are observed. The aventurization was of a variable intensity and seldom uniform over large pieces ; many specimens contained no reflect- ing lamelle; others contained many and showed a beautiful aventurization, not inferior to that of the best Tvedestrand specimens. The hematite lamelle.—The lamelle were large as in the Tvedestrand sunstone, up to 2 X 38™™ or sometimes more. Most of them were not six-sided, but showed distinct, unsymmetrical eight or ten-sided outlines. Often they were quite irregular. Fig. 2, Pl. Il, and fig. 1, Pl. ILI, show the characteristic shape of lamellee in sections after (001). Orientation of the lamellew.—The angles of the goniometric measurements are given in Table 7. The planes of ee were : (001), (010), ( 112), (112), (150), (150), (110), (110). these (112) contained the larger number of lamelle ; (9) only a few; (150) and (150) contained a considerable number, the other forms very few. The determinations of the planes (110) and (110) were based chiefly on the microscopic measurements; for (110): dp = 57°10’; $y = 63° 30’; for (110); de = BT? 35! ; 5 dy = 63° 58%. In one case direct voniometri i¢ measurements could be made on exposed lamella after (110): py = 61° 5’; dy, = 68° 55’. * This locality was recently discovered by Mr. Térje Térjesen, Risér. The specimens for the present description were collected by the author at the locality. 390 O. Andersen—Aventurine Feldspar. TABLE 7 AbreAnas 5 n = 1°545 2 Measured | p Ci) Forni Calculated | Calculated | Calculated a r from from Measured from measurements | axial ratios | axial ratios Pole : Normal to 001 ( pp and ¢p) ANG? Path |i Pasko al 20° 47 58° 3 112 49 58 | 18 28 20 48 20° 44’ o7 56 58° 34’ 57 26 | 138 13 20 47 58 19 46 27 | 238 9 21 22 56 (57 112 49 58 | 21 7 21 36 . 21 42 a7 «6 56 18 57 26 | 15 46 21 36 56 44 Pole: Normal to 010 (py and $y, ) 46 32 | 17 10 19 31 63 43 150 119 34 19 30 63 34 119 28 63 50 ee 46 34 | 16 40 19 22 63 17 150 47 35 | 16 0 19° 25 19 19 63 42 119 20 63 29 1 Direct measurement. Projections on (001) of the longer edges of lamelle after (112) formed often + 72° with the a-axis. Lamelle after (150) and (150) were oriented with the projections of the long edges on (010), forming about —79° with the a-axis. . Labradorite from Labrador.* The feldspar.—The cleavage pieces were fresh, of gray oon and showed a beautiful labradorization on cleavage faces of (010). Extinction angle on (001) = 5°5° 2 one MONO) =—alee B= 12558 * Specimens from the collection of the Geophysical Laboratory. O. Andersen—Aventurine Feldspar. 391 Composition: Ab,,An,,, a labradorite very near andesine. The cleavage pieces were twinned with coarse lamelle after the albite law. : The hematite lamelle.—There were but few reflecting lamelle and the aventurization was accordingly very indistinct. Most of the lamellee were opaque, but some were transparent with the characteristic colors of hematite. They were often linear, but sometimes approximately six-sided or irregularly rounded. The largest measured 0:03™" in diameter. Orientation of the lamelle.—The goniometric measurements are given in Table 8. The planes of orientation were: (112), (112), (150), (150) each with about the same number of lamelle. The linear, strip-shaped lamelle after (112) and (112) were generally oriented with the projections of the long edges on (001) parallel to the a-axis. The long edges of lamellee after (150) and (150) were frequently parallel to the c-axis. TABLE 8 AbsjADs1 ; n= 1:558 Measured p @ Form Calculated Calculated Calculated i r from from Measured trom measurements | axial ratios axial ratios Pole: Normal to 001 (pp and ¢p ) 112 d7° 20’ | 14° 12’ 20° 53’ 20° 51’ o7° 53! 57° 58’ 112 57 14 16 24 Pil B83 21 34 57 23 57 «610 Pole: Normal to 010 (py and $y, ) | 55 30 | 11 32 19 39 63 54 150 Pe Ee See ee 19 39 See 57 39 10 37 19 48 — oe 63 48 i 54 49 Oa 18 44 63 46 150 | 19 23 — 95 48 8 23 18 48 63 55d It should be noticed that the plane of labradorization on (010) does not coincide with any of the planes of orientation of the aventurizing lamelle. Microcline Perthite from Perth, Ontario, Canada.* The feldspar.—The aventurine feldspar from Perth was described by Des Cloizeauxt as an orthoclase perthite. The * Specimens from U. 8. National Museum ; Lea collection, No. missing. + Ann. Chim. Phys. (5), ix, 465, 1875. 392 0. Andersen—Aventurine Feldspar. specimens examined by me consisted of microcline perthite made up of copper red microcline, and colorless albite. The microcline showed an exceedingly fine cross hatching invisible with low magnification and in places hardly de- tectable even with a high magnifying power of the micro- scope. The albite forms coarse inclusions extended approxi- mately parallel to the c-axis but of very irregular cross sections. The hematite lamelle.— The reflecting lamelle were restricted to the microcline, the albite never containing any- They were very small, seldom more than 0:05™™ in diameter, of rather regular outlines, hexagonal or rhomb-shaped, or some- times forming linear strips. Orientation oy the lamellew.—The goniometric measurements are stated in Table 9. The following faces were planes of orientation: (001), (010), (112), (112), (150), (150).* After (112, (112), (150) and (150) there were numerous lamellz caus- ing a distinct aventurization on the cleavage faces; after (001) and (010) there were very few. TABLE 9 Microcline: n = 1°5238 Measured p @ Form Calculated Calculated Calculated 0 r from from Measured from measurements | axial ratios axial ratios Pole: Normal to 001 (pp and ¢p ) 49° 9’ | 18° 50’ 20° 59’ 56° 15' 112 || 20° 87’ ||_——_|_ 56° 38’ 54 22 | 15 10 on 56 32 Pole: Normal to 010 (py and $y, ) AO, 390 te) ford] eta 64 27 150 | | 18 41 |e 56 11 | 6 48 || 18 46 64 10 Many of the lamelle after (112) and (112) were oriented with the projection of their elongated edges on (001) approxi- mately parallel to the a-axis. The elongation of the lamellz after (150) and (150) was often approximately parallel to the c-axis. *For the sake of comparison with the plagioclase, the positive and nega- tive faces of forms like (112), that is 112 and 112, are put down as distinct forms, although their angles are referred to the monoclinic axes of ortho- clase (see p. 364 and Table 1) and therefore are the same. — O. Andersen—Aventurine Feldspar. 393 Microcline Perthite from Mineral Hills, Middletown, Delaware County, Pennsylvania.* | The feldspar.—The aventurine feldspar from Mineral Hills was described by Des Cloizeaux first as orthoclaset and later as microcline perthite.t The albite inclusions are regularly extended in the direction of the c-axis, but have very irregular cross sections, often as much as 1™™ broad. On (001) we therefore see very irregular patches or bands of albite in the microcline, on (010) we see regular alternate stripes of microcline and albite approximately parallel to the c-axis. The microcline is greenish-gray with red spots; under the microscope it shows the ordinary coarse cross hatching. The albite is colorless and shows the usual polysynthetic twinning after the albite law. The hematite lamelle.—Aventurization was only observed in the red spots of the microcline; the other parts of the microcline and the albite contained no hematite lamelle. The lamellz are distorted six-sided, rhomb-shaped or form elongated strips or irregular patches. The elongated strips measure at most 0°2 x 0°1™™, the more irregular lamelle about One x02" ™, Orientation of the lamelle.—The goniometric measurements are contained in Table 10. The lamelle were oriented after the faces: (001), (010), (112), (112), (150), (150). Projections on (001) of elongated edges of lamelle after (112) and (112) form often 20° with the a-axis. For other lamelle the cor- responding angle is 76°. The angle between the projections of the elongated edges of lamelle after (150) and (150) on (010) and the a-axis is often about — 80°. Microcline Perthite from Néskilen, Arendal, Norway. § The feldspar.—The cleavage pieces were rather fresh, of a brownish-gray color. The cleavage faces were often curved. The feldspar was a microperthite with very fine, rod-shaped albite inclusions of elliptical cross sections and with the elongated direction approximately parallel to the c-axis. The microcline showed a very fine cross hatching. “The hematite lamelle.—The aventurization was distinct, produced by numerous small lamelle which never measured more than 0°2™™ in diameter. The lamellee often formed linear strips, but were also sometimes approximately hexagonal or rhomb-shaped. * Specimens from U. S. National Museum ; No. 78700. + (Orthose aventuriné) Nouv. Rech., 1867, pp. 153 and 206. $ Ann. Chim. Phys. (5), ix, pp. 534, 460, 463, 1876. § Specimens from the Mineralogical Museum of Kristiania University. 394 O. Andersen—Aventurine Feldspar. TABLE 10 Microcline ; n = 1°523 Measured p o Form Calculated Calculated Calculated a Tr from from Measured from measurements | axial ratios axial ratios Pole: Normal to 001 (pp and ¢p) 51° 27’ | 16° 55! 20° 58’ 56° 16’ || 51 50 | 16 18 20 51 56 24 112 20° 517°, -||_ eee eames 64 50 | 8 19 20 57 56 23 Gon 4e liesrned 20 54 56 46 Pole: Normal to 010 (py, and ¢y ) | P55 gel Mtekcsge 18 28 63 47 150 || 55 31 | 6 48 18 37 18 41 63 57 | 68 57 59 13 | 4 46 18 44 63 20 Orientation of the lamelle.—Table 11 contains the gonio- metric measurements. The following faces were planes of orientation: (001), (010), (112), (112), (150), (150). Of these only (112), (112), (150) and (150) contained a large number of lamelle. After (001) and (010) there were very few. TABLE 11 Microcline; n = 1°523 Measured p @ Form Calculated Calculated Calculated a r from from Measured from measurements | axial ratios axial ratios Pole: Normal to 001 (pp and ¢p ) 112 Ont, atom edt: Pe 20° 57’ 57° 21’ 56° 38’ Pole: Normal to 010 (py and $y, ) 150 ayy i tk Oy |) 18 57 18 41 64 38 638 57 — O. Andersen—Aventurine Feldspar. 395 Projections on (001) of the long edges of lamelle after (112) and (112) formed frequently 15° ‘with the a- -axis ; others were approximately parallel to the a-axis. Projections on (010) of the elongated edges of lamelle after (150) and (150) formed often —83° with the a-axis. Microcline Perthite from Stene, Sannokedal, Norway.* The occurrence is an ordinary granite pegmatite dike (feld- spar quarry) containing microcline perthite, plagioclase (oligo- clase), quartz and biotite as main minerals and a number of other minerals in smaller quantities. Different varieties of graphic granite are abundant. The feldspar.—The perthite structure was in part coarse with visible lamellz of albite (after 110), in part a very fine microperthite structure, with extremely thin rod-shaped in- clusions of albite oriented. approximately parallel to the c-axis. There were all transitions between these two structures, both often being found in the same little cleavage piece. The coarse perthite often formed isolated patches in the micro- perthite. In the coarse perthite the microcline was developed with the ordinary cross hatching. In the microperthite the microcline structure was very fine and there was no regular cross hatch- ing. "The hematite lamelle.—The aventurization was restricted to the microperthitic parts of the feldspar and appeared with medium intensity about equally distinct on either of the two cleavage faces. The lamelle measured at most 0-1™™ in diameter and were very variable as to shape. Some showed very regular hexa- gonal outlines, others were rhomb-shaped or elongated and still others perfectly irregular. The lamellee parallel to (001) were often collected in groups with regular hexagonal or rhomb-shaped outlines. The single lamellee were small, mostly irregular, but sometimes hexagonal or rhomb-shaped and with the edges parallel to the outlines of the groups. The groups often measured 1 — 2™" in diameter and their outlines were definitely oriented. Orientation of the lamelle.—The goniometric measurements are given in Table 12. The planes of orientation were the following faces: (001), (010), (112), (112), (150), (150), (110), (110). _Most of the lamelle were parallel to (112), (112), (150) and (150) with about equally many after each ; a considerable number were also oriented after (001), (010), (11 0) and (110). *This locality was discovered some years ago by Mr. Peder P. Tangen, Kragero. The specimens examined were partly obtained from the Minera- logical Museum of Kristiania University, partly collected by the author at the locality. 396 O. Andersen—Aventurine Feldspar. TABLE 12 Microcline ; n = 1° 528 Measured p @ Form Calculated Calculated | Calculated a r from from Measured from measurements | axial ratios | axial ratios Pole: Normal to 001 (pp and ¢p) 47° 40’ | 19° 20’ 20° 47’ | 56° 19’ 112 50 17 | 18 14 S1 eo 20° 57’ 56 42 56° 38’ 51 24 | 16 34 20 50 56 55 Pole: Normal to 010 (py and ¢y 46 21 14 5 18 47 64 20 150 18 41 63 57 5y- a) 7 59 18 13 50 | The form (110) (embracing 110 and 110) was established by the microscopic measurements: dp = 56°2°; dy = 64°3°. The edges of the lamellee after (001) were sometimes parallel to the a-axis and also occasionally parallel to the d-axis. More frequently, however, they formed oblique angles with the a-axis. Angles of 40° and 70° between the a-axis and the elongated edges of lamelle (or the outlines of the groups of lamellee described) were measured. Lamelle after (010) were often oriented with one of the edges perpendicular to the a-axis more seldom forming +50° with the a-axis, The lamelle after (112), (112), (150) and (150) were generally so small and irregular that no detinite measurements of the orientation of their edges could be made. Miscellaneous Occurrences. Besides the varieties described in the preceding pages, a number of other specimens were examined more superficially.* All these specimens showed a very weak ayenturization and the measurements were only approximate. The results for all were that the aventurization on (001) was due to lamelle *Jn the literature we find mentioned a considerable number of occur- rences besides those described or referred to in this paper. No actual descriptions of these aventurine feldspars have been given, however, and it is therefore of minor interest to list the references, most of which can be easily found in standard handbooks of mineralogy or treatises on gems. See for instance : Max Bauer, Edelsteinskunde; G. F. Kunz, Gems and Precious ~ Stones of North America, O. Andersen—Aventurine Feldspar. 397 oriented after (112) or (112) and that on (010) was due to Jamelle after (150) and (150), that is, the same as in the other specimens examined. The measurements were made with the microscope. The localities are given below, together with brief characteristics of the different specimens. Mérefjer, Arendal, Norway.t—Microcline, microperthite of normal structure. Hematite lamelle few, transparent. Rosaas, Iveland, Norway. +—Microcline perthite partly with visible albite lamellae, partly a microperthite. Strongest aventurization on (010). Hematite lamelle sometimes very regular, six-sided or rhomb-shaped, transparent. Numerous lamellee of mica after faces ot (110). Hiltveit, Iveland, Norway.*—Microcline microperthite with very fine microcline structure and thin rod-shaped albite inter- growths. Strongest aventurization on (010). Hematite lamellee sometimes elongated and sometimes rather regular, six-sided ; transparent. Renfrew, Canada.t—A wicrocline perthite of very coarse structure. Few opaque hematite lamelle. SUMMARY. \ A number of varieties of aventurine feldspars were examined. Orientation angles of the reflecting lamelle were measured, chiefly with the goniometer, and the properties of the lamelle were determined under the microscope. Brief discussions of the optical problems are included in the record of these exami- nations. The reflecting lamelle are always oriented after simple erystal forms of which (112), (112), (150) and (150) occur as planes of orientation in all varieties, the first two causing aven- turization on (001), the last two on (010). The forms (001), (010), (110) and (110) also frequently contain reflecting lamelle. Exceptionally (021) and (118) are planes of orientation. The orientation of the edges of the lamelle is evidently regular but simple crystallographic relations could not, in general, be found. The reflecting lamelle were determined as hematite. They vary widely from one variety to another as to shape and size, showing hexagonal, eight- or ten-sided, rhomb-shaped, strip- shaped or irregular outlines. The largest measured 3°5™™ in one direction, the smallest were of submicroscopic size. The absorption colors are those characteristic of hematite. It was shown that the colors in reflected light are interference colors of thin films. By means of these colors the thickness of the transparent lamella could be approximately determined. It * Specimens from the Mineralogical Museum of Kristiania University. + Specimens from U. 8. National Museum; No. 83218, 398 O. Andersen—Aventurine Feldspar. was found to vary between 504m and 5004p. The lamelle were shown to possess no appreciable pleochroism. The ap- parent pleochroism observed in lamelle forming large angles with the section was explained as due to the effect of polariza- tion by reflection and refraction at the surface of the lamelle. The appearance of interference spectra in these lamelle was explained as due to the action of the lamelle as polarizers or analyzers for the wedge-shaped parts of the feldspar that lie above or below them in the sections. Thermal experiments with one of the varieties showed that the hematite lamelle persist up to about 1285°. At this temperature they disappeared, presumably by melting together with a small part of the surrounding feldspar to thin, invisible glass films. The feldspar remained otherwise unchanged (crystallized). By heating at lower temperatures some of the lamellee (originally opaque ones) reappeared in the same places and with the same outlines as before. By a long heating at temperatures around 1050° (of cleavage pieces not previously heated) the vpaque lamelle generally became transparent and the others became a little lighter. : The origin of the hematite lamellee was explained as due to unmixing of an originally homogeneous feldspar which con- tained iron oxides in solid solution. Thin lamella of hematite then separated along certain structural planes of the feldspar. In the concluding section all the specimens examined are described and the measurements tabulated. The Geophysical Laboratory of the Carnegie Institution of Washington, Washington, D. C., July 16, 1915. Plate |. Amer. Jour. Sci., Vol. XL, October, 1915. Fie. 1. Fic. 2. Amer. Jour. Sci., Vol. XL, October, 1915. Plate Il. Fie. 1. Amer. Jour. Sci., Vol. XL, October, 1915. Plate Ill. O. Andersen—Aventurine Feldspar. 399 EXPLANATION OF PLATES. PrArnoly Fie. 1. Albite (aventurine feldspar) from Fisher Hill Mine. Thick sec- tion after (001). Ordinary light. Magnified about 30 diameters. The vertically elongated patches are hematite lamelle arranged parallel to (112), the long edges being projections on (001) parallel to a-axis. The streaks running in the direction from lower left to upper right quadrant are also hematite lamelle parallel to (112). Fic. 2. Albite (aventurine feldspar) from Fisher Hill Mine. Thick sec- tion after (010). Ordinary light. Magnified about.30 diameters. The large patches are hematite lamella parallel to (010); horizontal streaks lamellz parallel to (021). Streaks running from upper left to lower right quadrant at small angle with horizontal are lamelle parallel to (112) and (112). - Streaks from upper left to lower right at large angle with the horizontal are lamelle parallel to (150) and (150) with the elongated edges parallel to the c-axis. ee PrRarE Til Fie. 1. Oligoclase (aventurine feldspar) from Kvrageréd. Thick section after (001). Ordinary light. Magnified about 30 diameters. Elongated patches running from upper left to lower right quadrant are hematite lamelle parallel to (150). Other patches are chiefly lamellz parallel to (112). Fic. 2. Oligoclase (aventurine feldspar) from Aamland. Thick section after (001). Ordinary light. Magnified about 30 diameters. Chiefly hematite lamelle parallel to (112); some parallel to (112). The variations in the outlines of the lamelle are distinctly seen. PuLate III. Fie. 1. Oligoclase (aventurine feldspar) from Aamland. Thick section after (001). Ordinary light. Magnified about 50 diameters. Hematite lamelle parallel to (112) showing various outlines. Near the center two lamellz parallel to (150) running steeply from lower left to upper right. Am. Jour. Sct.—FourtH SERIES, Vou. XL, No. 238.—OctozeEr, 1915. 26 400 J. H. Reedy—Anodic Potentials of Silver. Arr. XXIX.— Anodic Potentials of Silver: If. Their Réle in the Electrolytic Estimation of the Halogens; by Joun Henry Reepy. (Contributions from the Kent Chemical Laboratory of Yale Univ.—cclxxi.) Tue first attempt at the estimation of the halogens by electrolytic fixation on silver electrodes was made by Vort- mann.* The halides in alkaline solution were electrolyzed between a copper cathode and a silver anode, using a potential of about two volts. At the end of the electrolysis the anode was ignited to decompose the silver oxide formed during the process, and the deposited halogen was determined by gain in weight. Even in alkaline soiution small amounts of silver were transported to the cathode. The deposit was not very adherent, and Vortmann sought to overcome this by adding sodium-potassium tartrate to the solution. The problem of separating the halogens was not considered. _ The quantitative separation and estimation of the halogens was first attempted by Specketer.| The halides were dissolved in 0-5 molar sulphuric acid, and during the electrolysis a stream of hydrogen was constantly passed through the solution to provide stirring, and at the same time to make the platinum cathode function as a hydrogen electrode. Considering the anode reversible with respect to the Ag--ion, Specketer assumed that the maximum anode potential required in any determination would be the potential of a silver electrode in a saturated solution of the silver halide in question, and that this value could be calculated from solubility data, by means of the Nernst formula for electromotive force. In this way he calculated the potentials necessary for the precipitation of the halogens to be —-06 voltst for iodine, +°15 volts for bromine and +°50 volts for chlorine; that is, a potential of —-06 volts should deposit all the iodine free from chlorine and bromine, and +°15 volts all the bromine free from chlorine. In practice somewhat larger potentials were found necessary. In order to make this separation satisfactorily, Specketer found that the following precautions must be observed: (1) A constant volt- age must be employed; (2) a definite acidity of solution must be maintained,—for example, 0°5 molar sulphurie acid: (8) atmospheric oxygen must be excluded, since it has a marked effect on the potential of the hydrogen electrode. Specketer found it impossible to determine chlorine on a silver anode, * Monatsh. Chem., xv, 280; xvi, 674. + Zeitschr. Elektroch., iv, 542. enya to the hydrogen electrode in 0°5 molar sulphurie acids as 0 Volts. J. H. Reedy—Anodic Potentials of Silver. 401 since some silver always went into solution. He therefore recommended estimating chlorine volumetrically after the iodine and bromine had been precipitated. Noattention seems to have been given by Specketer to the possibility that, for certain concentratious, halogens might be deposited simultane- ously. EF. Smith and his collaborators* have studied extensively the electrolytic precipitation of the halogens on silver anodes. In the same laboratory the so-called “double-cup” was developed by Hildebrand, by means of which both metal and halogen are determined in one electrolysis. Very accurate results were reported. Gooch and Read,t and Peters,t however, found the estimation of chlorine by deposition on silver anodes to give results that were “irregular and always low,” owing to a tendency of the silver to be dissolved from the anode, and to be transported to the cathode, or precipitated as silver chloride in the solution. Very recently Boettger and Kelly§ have proposed a modification of the Hildebrand “double-cup” method. It consists in adding an excess of a salt whose cation is more readily deposited than the ions of the alkali metals, such as “cadmium sulphate. The mercury dissolves the deposited metal, and owing to the high over-voltage of the hydrogen on the amalgam cathode, no hydrogen is evolved, and the solution remains neutral. No report was made of estimating chlorine in this way. Presumably the authors found that it was not practicable. The determination of the decomposition potentials of solutions of potassium iodide, potassium bromide and potas- sium chloride of arbitrary concentration led Boettger and Kelly to the conclusion that the iodine could be separated from bro- mine and chlorine, but not bromine from chlorine. Satisfactory results were obtained. The methods used were more or less empirical, and no attention was given to the effect of varying the concentration of the halide solutions. Experimental Part. Preliminary Experiments.—Solutions of potassium chloride were dissolved in various dilute acids (0°5 molar sulphuric, 0-33 molar phosphorie, 0°5 molar oxalic, and 1-0 molar formic) and electrolyzed between a platinum cathode and a rotating silver anode, using very low tensions,—50 to 200 millivolts. In all eases it was found that shortly before the precipitation was * Withrow, Jour. Amer. Chem. Soc., xxviii, 1356; Hildebrand, ibid., xxix, 447; McCutcheon, ibid., xxix, 1445, 1460; Lukens, ibid., xxix, 1455 ; Goldbaum and Smith, ibid., xxx, 1705 ; xxxii, 1468. + This Journal, xxviii, 435. t Ibid., xxxii, 365. § Verh. Ges. deutsch. Aertze, 1913 (1914), II, 361. 402 J. H. Reedy—Anodic Potentials of Silver. complete the silver began to dissolve. The weight of the chlorine found was less than the theory by amounts varying from 1 to 6 milligrams. To ascertain if the solution of the silver could be due to oxidation at the anode, reducing agents like urea, formic acid and acetaldehyde were added to the electrolyte. Such expedi- ents did not prevent the loss of silver from the anode, as minute flecks of the metal or its salts always appeared in the solution sooner or later. Determinations in neutral solutions were also tried, and the results obtained in this case were likewise always too low. Isolation of the Anode Potential.—In all the above experi- ments the voltage measured was the potential fall across the cell. Since, however, the fixation of the halogens is strictly an anodic process, it is obviously better to determine the anode potential apart from the cathode potential, which, as shown by Specketer,* is strongly influenced by conditions of acidity and contact with air. The third electrode meets this require- ment satisfactorily and the anodic potentials in the following experiments were all determined in this way. Nature of the Hectrolytic Mediwm.—In addition to the halide to be determined, especially when its amount is small, it was found best to have some other electrolyte present in the solution in order to facilitate the passage of the current. As has been shown by Foerster + and others, alkaline solutions favor the formation of oxidized halogen compounds, such as iodates. This would introduce an error into the determination, making the results too low, since the silver salts of the oxygen- halogen acids are appreciably soluble. In neutral solutions, the solution will become basic unless an amount of metal ions equivalent to the anions discharged is also separated on the cathode, as would be the case when zine or cadmium salts constitute the electrolyte. This would increase the difticulty. of estimating any silver transported to the cathode, since it would necessitate in every case an analysis of the deposit on this electrode. It is much better to be able to estimate the silver directly by gain in weight of the cathode, without having to make a chemical separation. Consequently acid solution appears to be the most favorable medium for the electrolysis. Since the deposition potentials of the halogens increase with the dilution, we would seek as a solvent that solution which has the highest reaction potential with silver; for in this way the largest margin for the complete separation of the halogens would be available. But most acids show the same reaction potential,—'521 volts. (See Table I.) * Zeitschr. Elecktroch., iv, 542. + Elektrochemie, 1905, p. 348 et seq. _ {Referred to the hydrogen normal electrode as 0 volts. All potential — measurements in this investigation have been referred to this standard. J. H. Reedy—Anodic Potentials of Silver. 403 TaBLE I. Reaction Potentials of Acids. Acids Reaction Potentials Sulphuric acid, *5 molar _.-._.----- °521 volts INiinieraeidh Ie0 mola aes ee eee, = 520 Phosphoric acid, 33 molar .-_------- 521 @Oxalic acids -ommolane 22252 225s O20 Xceticracidy 1: Ommolai == 42 ee "529 TENG BOl oh imoleyy, 22 ee ae 521 This particular potential has been investigated in my preced- ing paper* and was found to be the potential at which silver begins to dissolve, irrespective of the nature of the anions present. For this reason it was given the name “solution potential of silver.’ Evidently there is a free choice among the acids in Table I, so far as margin for separation is concerned. However, 0°5 molar sulphuric acid was chosen, on account of its good conductivity and the fact that no intermediate solution would be necessary between the cell and the mercurous sulphate electrode. Stirring—To prevent concentration polarization and to hasten the electrolysis, it is desirable that the electrolyte should be well stirred. Since, as shown in my previous paper, t the deposition potentials of the halogens appear to be indepen- dent of the nature of the stirring, whether by a current of gas or by other mechanical means, a rotating anode was adopted, as it affords the most effective stirring of the anode solution. Inasmuch as iodides in acid solution are readily oxidized by the air to free iodine, some of which (as shown by experiment) tends to be lost by volatilization, it is evident that the presence of air in the cell must be avoided. On account of its high density, carbon dioxide was used for this purpose. During all experiments a steady streain of the gas was introduced by a tube reaching almost to the surface of the solution, and allowed to escape through the space around the anode stem in the lid of the cell. In this way the space above the solution was kept filled with this gas. Deposition Potentials of the Halogens.—Fig. 1 shows the relation between the concentration of the halogen ions and their deposition potentials. It is reproduced from fig. 5 of the preceding article, which has been so modified as to show only those lines which bear on the problem now being con- sidered. As explained before, these lines are the “reaction potential curves” of the halogens, and were obtained by plot- ting the logarithms of the dilutions as abscisse and the reaction potentials as ordinates. The figure accordingly shows in a * This Journal [4], xl, 281. + Ibid. 404 J. H. Reedy— Anodic Potentials of Silver. comprehensive way, not only the behavior of the anode poten- tial during an electrolysis, but also the value of the maximum potentials | necessary for the separation and estimation of the halogens. In no case may the anode potential rise above *521 volts, which has been defined as the “solution potential” of silver, for at that potential silver begins to dissolve. The End Point in a Precipitation.—Specketer calculated by means of the Nernst formula the maximum potential that Fie. 1. Reaction Potentials Logarithms of Dilutions. es Fic. 1. Influence of concentration on the deposition potentials of the halogens. would be required for the complete separation of any halogen on the assumption that the final concentration of the halogen would be that of a saturated solution of the corresponding silver halide. Results obtained in this way are only approxi- mate, because (1) the Nernst formula (as shown in the pre- ceding article) does not exactly express the relation between electromotive force and concentration of the ion in the case of silver electrodes; and (2) the values of the solubilities of the silver halides are more or less uncertain. Anodic reaction potentials of a saturated solution of a silver halide on silver electrodes coated with that halide are likewise J. H. Reedy—Anodic Potentials of Silver. 405 more or less uncertain, owing to the difficulty of obtaining concordant results at the high dilutions involved. The method used in this investigation for fixing the end point in a precipitation was to arbitrarily choose those points on the reaction potential curves (extrapolating when necessary) which correspond to 0°1 milligram—the ordinary limit of accu- racy in analytical work. (See fig. 1.) This method promises greater accuracy than those mentioned above, since it depends only upon the probable assimption that the relation which is found to exist at low dilutions also exists at high dilutions. Only in the case of chlorine is the limit value of the anode potential doubtful, owing to the fact that its graph is not a straight line, as in the case with the other two halogens. This method gives the following results for volumes of 200° : Maximum potential for ehlorine ---. ---- ‘505 volts Maximum potential for bromine ._-_- ___- -400 volts Maximum potential for iodine --------- "190 volts . Apparatus.—Excepting the electrolytic cell, the apparatus was the same as that used in the determination of reaction potentials, as described in the preceding paper. The electro- lytic cell (see fig. 2) was a glass cylinder 11°" high and 8-7" in diameter. It was closed with a lid of wood fiber, through which the various connections were introduced. All electrical connections were insulated by glass tubing. The anode (A in the figure) consisted of a frame of heavy silver wire whose cylindrical outer surface was covered with fine silver gauze. The diameter of this anode was 4:7°", and the height (of the cylindrical portion) was 2-7". It was rotated with a speed of about 200 revolutions per minute. The anode before using was heavily coated with silver by making it the cathode in an ammoniacal silver solution to which am- monium oxalate had been added, as recommended by Gooch and Feiser.* It was then thoroughly washed and heated to about 500° in an electric oven, in order to make the deposit firm and adherent. The cathode C was a piece of bright sheet platinum, 21 x 50™™, provided with a stem of platinum wire ending in a loop, for convenient attachment to the cathode connection. ‘To prevent swinging of the cathode by the rotating electrolyte, it was held in place by a glass rod whose lower end was drawn out into a fiber which was bent so as to form a hook, as shown in the figure. The cathode was weighed before and after each deter- mination and any silver transported was estimated directly by difference. * This Journal, xxxi, 109. 406 J. H. Reedy—Anodic Potentials of Silver. Experimental Procedure.—200° of 0°5 molar sulphuric acid was introduced into the cell, and to this was added by means of a pipette a definite amount of potassium halide solution, Fie. 2. CO, ban To third electrode Fie. 2. Diagram of the electrolytic cell. whose halogen value had been carefully determined by pre- cipitation with silver. The silver anode and the platinum cathode were weighed and put in place. After starting the motor which operated the rotating anode, the electrolytic eur- J. H. Reedy— Anodic Potentials of Silver. 407 rent was turned on, the voltage at all times being adjusted so that the anode potential would conform to the limits mentioned before. In order to expedite the precipitation, the anode poten- tial was brought at the start toa value just below the maxi- mum. This potential requires constant attention and must be adjusted very frequently since it tends to rise very rapidly with the decrease of the halogen ions of the solution. The completion of the precipitation is indicated by the lack of deflection of the needle of a sensitive galvanometer. Con- ducted in this way, twenty to forty minutes are required for an electrolysis. The electrodes were then removed from the cell, carefully washed, and finally dried in an electric oven at about 500°. The gain in weight of the anode represents the halogen deposited. Before making another determination the silver halide was reduced to metallic silver by making it the cathode in a dilute sodium hydroxide solution, using a tension of about two volts, and allowing the current to pass for a period of 50 to 60 min- utes. The anode was then washed in dilute sulphuric acid, carefully rinsed in distilled water, and finally heated to about 500° in the electric oven for an hour or more. The silver after reduction is black in color and not very adherent, so that the anode must be handled very carefully to prevent loss. After heating it is white and firmly sintered together, so that there is no risk of mechanical loss in the subsequent manipula- tions. Nature of the Haloid Deposit.—As long as the solution potential of silver is not exceeded, the haloid deposits on gauze anodes are quite adherent, and may be washed without fear of loss. If, however, this critical potential is exceeded, the silver halide seems to be formed in part out of actual contact with the metal, and for this reason does not adhere firmly. No such expedient as the alkaline tartrates used by Vortmann is necessary to make the deposit more adherent, if the proper attention is given to the anode potential. The precipitate is free from silver oxide, and hence does not require to be heated to a high temperature to effect the decomposition of this com- pound, as Vortmann, Gooch and Read, and Peters found nec- essary in their researches. Influence of Current Strength—Under the description of the procedure in an electrolysis, reference was made to the fact that the anodic potential must be watched continually, since it rises rapidly as the electrolysis progresses. To keep this potential from exceeding its critical value, the voltage of the electrolyzing current must be reduced from time to time, and this entails a corresponding diminution of current strength. Moreover, the current strength must be reduced, not in an 408 J. H. Reedy—Anodic Potentials of Silver. arbitrary manner, but in the way the behavior of the anodie potential demands. In a general way, the relation between current strength and time is represented i in fig. 8, which is drawn from data obtained Fie. 3. 200 160 Y @ o 120 Q. S Ss Ss epee ~ — gy Mee we SS 1S) 40 (@) Time in Minutes —Se Fic. 3. Change of current strength during an electrolysis. in a chlorine determination. The anodic potential was held at its maximum value by constant adjustment of the voltage, and from time to time readings of the current strength were made. It is seen that the current falls off very rapidly at first, but much more slowly after the halogen concentration has become low. This variation of the current strength with time depends — J. H. Reedy—Anodic Potentials of Silver. 409 on other influences, such as the concentration of the halogen anion, the resistance of the cell and external circuit, and the cathode potential. From this it is evident that any analytical method, in which it is sought to control the course of the elec- trolysis by regulation of the current strength without regard to the anodic potential,—as has been attempted by various inves- tigators,— is, to say the least, inaccurate and uncertain. Estimation of the Halogens.—Table II contains a summary of the results obtained in the estimation of the halogens. It should be stated that this table contains the results of all the determinations made in which the theoretical conditions were maintained. Ina number of experiments the anodic potential, TaBLE II. Estimation of the Halogens. (Volume of Solution=205°.) Halogen Halogen Maximum Number taken found Error anode poten’! of Exp. grams grams grams volts Iodine 1 0620 0618 — ‘0002 "205 EKstimations 2 °0620 0623 +°0003 "190 3 *0620 0617 —'0003 “190 4 0620 +0620 "0000 “190 5 0620 "0617 — "0003 °190 Bromine 6 0396 0386 —‘0010 440 Estimations El "0396 0393 —‘0003 *400 8 0396 ‘0396 “0000 -400 ) "0396 0393 — ‘0003 “400 10 "03896 ~ :0395 —‘0001 -400 Chlorine 11 ONG 0078 —‘0099 520 Estimations 12 0177 0166 —-0011 520 13 0177 0168 — ‘0009 "520 14 0177 “0164 —'0013 520 15 0177 "0125 — ‘0052 520 16 SOT 0162 —°0015 520 through lack of attention, rose too high, with the result that sil- ver went into solution. Such experiments have been excluded from the table. The tendency towards low results in the iodine and bromine determinations is probably only accidental, and the results con- firm the conclusions developed in the study of the conditions involved in the general process. But in the case of the chlorine estimations, the deficiencies are not to be attributed to experimental error. In every ease except one it was found that silver had gone into solution, and amounts varying from -0008 to -0010 grams in weight had been transported to the cathode. The exception was Exp. No. 15, in which the process was interrupted early, while the galva- 410 J. H. Reedy—Anodie Potentials of Silver. nometer still showed a minute current ; and although no silver had dissolved, a considerable amount of chlorine remained in the solution. By reference to fig. 1 it will be seen that just such errors as these found in the chlorine estimations should be expected. Upon progressive dilution the deposition poten- tial of chlorine approximates °521 volts—the solution potential of silver,—and at that potential both the deposition of ehlo- rine and the solution of silver will proceed simultaneously. That is, the anode potential of silver can not be raised high enough to deposit all of the chlorine without silver beginning to dissolve. For this reason, the estimation of chlorine on sil- ver anodes can not be an accurate and trustworthy method. Separation of the Halogens.—By reference to fig. 1 it will be seen that the halogens can not be separated electrolytically regardless of their concentrations. For example: For a volume of 200°, an anodic potential of -190 volts is necessary to completely precipitate iodine on silver anodes. The same potential corresponds to a concentration of about -0115 molar bromine. That is, if the bromine ions are present in greater: concentration than this, there is a certain range of potential for which both iodine and bromine will be deposited simul-_ taneously. For similar reasons chlorine must not be present in greater concentration than -0014 molar if bromine is to be pre- cipitated free from chlorine. In view of this small concentra- tion, it is not surprising that Boettger and Kelly,* working more or less empirically, reported that bromine could not be separated from chlorine electrolytically. In the separation of iodine from chlorine, however, the concentration of the chlorine may be even higher than molar without its deposition potential overlapping the maximum potential for iodine. Evidently, the approximate bromine and chlorine concentra- tions must be known before a separation may be attempted. If they exceed the above limits, proper dilution must be made. It should be remembered, however, that dilution is always made at the expense of accuracy in the determination of the halogen of low concentration. Analytical Separations.—The conclusions drawn above as to the necessary conditions for the separation of the halogens were applied experimentally in a number of determinations, with the results indicated in Table III. In Exp. No. 23 the anode potential rose too high and some bromine was deposited, which accounts for the positive error. In Exp. No. 32, the chlorine concentration was 0024 molar —considerably higher than -0014 molar, which was taken to be the maximum chlorine concentration allowable for a complete * Verh. Ges. deutsch. Aertze, 1913 (1914), ii, 361. J. H. Reedy—Anodic Potentials of Silver. 411 TasBLeE III. Separation of the Halogens. (Volume of Solution = 210°.) Iodine from Chlorine. Number Chlorine Maximum Iodine Iodine of cono. anode pot. taken found Error Exp. moles volts grams grams grams 7 0046 “190 "0620 0617 —°0003 18 “0046 "190 *0620 "0619 —'0001 19 ‘0046 ‘190 "0620 °0619 —‘0001 20 "0046 190 -0620 "0618 —-0002 21 *0046 "190 “0620 “0621 +0001 Iodine from Bromine. Numbezx Bromine Maximum Todine Iodine of conc. anode pot. taken found Error Exp. moles volts grams germs grams 22 "0046 *190 ‘0620 0623 +:°00063 23 °0046 213 “0620 "0628 +0008 24 "0046 "190 0620 ‘0621 +°0001 25 ‘0046 "190 "0620 "0620 0000 26 0046 ‘190 “0620 "0619 — 0001 Bromine from Chlorine. Number Chlorine Maximum Bromine Bromine of cone. anode pot. taken found Error Exp. moles volts grams grams grams 27 “0010 400 "0396 ‘0401 +°0005 28 “0010 “400 0396 0389 —‘0007 29 ‘0010 “400 "0396 "0394 —°0002 30 ‘0010 “400 "0396 "0395 —0001 31 ‘0010 “400 "0396 "0393 —'00038 32 "0024 400 "0396 "0469 +°0073 separation. Consequently, some chlorine was deposited along with bromine in the later stages of the electrolysis. It is inter- esting to note that the positive error of ‘0073 grams agrees closely with the error that might be expected from theoretical considerations.. The excess of the chlorine concentration over ‘0014 molar, the limit found above, was ‘001 molar. This, for 210°, the volume of the solution in this experiment, is equivalent to 0075 grams, which is practically identical with the experimental error. These results confirm in a forceful way the conclusions drawn above as to the effects of concentration in the electrolytic separation of the halogens. Analytical Applications.—The processes just described are not offered as analytical methods. If has been the purpose of the writer throughout this investigation merely to determine 412 J. H. Reedy—Anodic Potentials of Silver. the conditions which must be observed in the electrolytic deposition of the halogens on silver, and not to perfect any analytical procedure. Furthermore, the methods which have been used in this work will hardly commend themselves to the practical analyst. The apparatus is too elaborate, and the manipulation is unsatisfactory in that it demands thorough familiarity with this unusual kind of work, not to mention the fact that the constant adjusting of the anode potential precludes the carrying on of other experiments at the same time. How- ever, it is hoped that this work will make clear what is possible and what is impossible in the quantitative separation and estimation of the halogens by means of a silver anode. Summary. The conclusions reached in this study of the determination of the halogens by means of silver anodes may be briefly stated as follows: (1) The anodic potential must be known at every stage of the electrolysis, so that the maximum value of the potential for that particular process may not be exceeded. (2) In no case may the anodic potential rise above °521 volts, since silver dissolves at the potential. (3) The course of the electrolysis can not be properly con- trolled by the regulation of the current, independent of the anode potential. (4) Chlorine can not be satisfactorily estimated by electro- lytic precipitation on silver, since the anodic potential required for its complete deposition is approximately the same as that at which silver dissolves. (5) Quantitative separation of the halogens is possible only within certain definite limits of concentration. W. G. Foye—Nephelite Syenites of Ontario. 413 Arr. XXX.—Wephelite Syenites of Haliburton County, Ontario; by Wireur G. Foye. Part I. Two differentiated Laccoliths of Nephelite Syenite in Glamor- gan and Monmouth Townships. Introduction. A. Gooderham Laccolith. General Geology. Petrography: 1. Red syenite. 2. Schistose canadite. 3. Nephelite pegmatite. 4. Kaolinized contact rock. B. Crescentic Laccolith near Tory Hill. General Geology. Petrography: 1. Garnet-pyroxene contact rock. 2. Hornblende-nephelite rock. 3. Monmouthite. 4, Biotite-nephelite rock. 5. Pegmatitic nephelite syenite. Part II. Suggestion concerning the Origin of the Nephelite Syenites. Smyth’s Theory. Confirmatory Evidence. Pneumatolytic Metamorphism of the Limestone to Amphibolite. Nature of the Solutions Introduced into the Limestones. Quantitative Importance of this Alteration. Origin of the Fluids Producing the Amphibolites. a. Segregation by juvenile gases. b. Interaction of the granite magma and limestone. Emplacement of the Laurentian Granite and Origin of Lit-par-lit Structure. Genetic Relations of Amphibolite and Nephelite Syenite. Field Facts Showing Desilication of the Granite and Origin of Soda-rich minerals. Conclusions. PARE I, INTRODUCTION. The laccoliths of nephelite syenite, described in this paper, are located in Haliburton County, Central Ontario. The are reached by way of the Central Ontario Railroad from Trenton and the Irondale, Bancroft and Ottawa Railroad, which connects with the Central Ontario Railroad at Bancroft. One of the laccoliths lies southeast of the village of Gooder- ham; the other lies southwest of Tory Hill. Messrs. F. D. Adams and A. E. Barlow have described the general geolog of this area in their memoir, “The Geology of the Halibur- ton-Baneroft Area, Ontario.”* *F. D. Adams and A. E. Barlow, ‘‘ The Geology of the Haliburton-Ban- croft Area, Ontario,’’ Memoir No. 6, Canada Geol. Sury., 1910. W. G. Foye—Nephelite Syenttes of Ontario. 414 Fie. 1. ‘orqqen=—/ ‘ey1uexordg =a ‘euojseuty o1tjoqiydmy =p ‘auojsoulry—=o ‘oytueIpn=q ‘oytueds oytfeydeN=» ‘YyI[O00R, oTjUeDseEAQ—=—q ‘YAI[O00¥, Ueyepoogn—y ‘oleyuO ‘AqunoH uojrnqrey ‘diysumoy, uesiomeyy ‘ueqtepooH ynoqz yourysiq oy} jo dey youeyG *T “Sty W. G. Foye—Nephelite Syenites of Ontario. 415 A. THE GooDERHAM LACCOLITH. The nephelite syenite laccolith indicated on the accompany- ing sketch map by the letter A is situated in Glamorgan town- ship near the village of Gooderham. It is one of a number of nephelite syenite bodies lying in a syncline of limestone between large areas of Laurentian gneiss. General Geology. The accompanying sketch map (fig. 1) shows the general geology of the area immediately adjacent to the Gooderham nephelite syenite laccolith. Limestone surrounds the body on all sides. This limestone is cut by many granite pegmatite dikes and has lenses of amphibolite associated with it. The nephelite syenite body consists of three principal rock types. These are: (1) red syenite (umptekite), (2) nephelite Hie. 2: s Pry P ie CLZZA om Z ‘Ze ZZ 72.694 4 2.662 KIER ZEPINSS 6 39 2 1 6 2h te) 6 i) 1500 feet Fie. 2. Cross-section of the Gooderham Laccolith. 1=Red syenite, 2=Schistose canadite, 3=Nephelite pegmatite, 4=Amphi- bolite, 5=Granite pegmatite, 6=Limestone, 7=Amphibolitic limestone. (The figures within the sketch give the specific gravities of the several rock types.) syenite (canadite), and (3) nephelite pegmatite. A north-south cross section of the body shows the rock relationships (fig. 2). The specific gravities of the several rock types are indicated by numbers on the section. The mass of nephelite pegmatite is unduly large for the size of the laccolith. It cannot be doubted that the syenite magma was surcharged with pneu- matolytic gases at the time of its emplacement. The activity of the gases is held to be largely responsible for the differ- entiation of this body. The nephelite pegmatite rests on the denser canadite but the red syenite at the base has a lower den- sity than the overlying canadite. At other localities the red syenite cuts the canadite and again it is transitional to this rock. The contact in this instance is covered by drift. It is quite possible that the red syenite was intruded after the nephelite syenite had differentiated into nephelite pegma- tite and canadite. This fact may account for the seeming lack of gravitative control in the differentiation of this laccolith. However, pneumatolytic transfer of the elements of the peg- matite towards the roof of the body was the prime factor in the differentiation. Am. Jour. Sct.—FourtH SErins, Vou. XL, No. 288.—OcropeEr, 1915. 416 W. G. Hoye—Nephelite Syenites of Ontario. The limestone, north of the laccolith, has been eroded from beneath the red syenite. The latter rock has broken down and has formed a talus-slope which conceals the contact of the two rocks. Drift likewise covers the contact of the canadite with the nephelite pegmatite. The contact of the nephelite pegmatite with the overlying limestone is marked by a kaolinized zone a foot or two wide. Coarsely crystalline calcite is included in blebs in the kaolinized Hic, 3. Fic. 3. Distribution of the several rock types in the Gooderham Laccolith. 1=Red syenite, 2=Schistose canadite, 3=Nephelite pegmatite, 4=Am- phibolite, 5=Granite pegmatite. material. Scapolite, apple-green apatite, and some pyrrhotite are also associated with this rock. With increase of ealcite the kaolinized zone passes into the normal limestone. Petrography of the Gooderham Laccolith. Figure 3 shows the general distribution of the rock types associated with the Gooderham laccolith. 1. Red Syenite. Macroscopic description.—This rock is reddish gray in color, rather fine in grain and even-textured. It is composed pre- dominantly of pink feldspar in granules 1 mm. in diameter. Occasional crystals of white feldspar are likewise present. Flecks of black mica and some hornblende are evenly dis- seminated through the rock and are arranged in such a way as to produce a vague schistosity. The rock is very fresh, breaks with an irregular fracture, and is of average weight. W. G. Foye—Nephelite Syenites of Ontario. 417 Microscopic description.—The red syenite resembles very closely the rock described by Adams and Barlow* as umptek- ite. A rough estimate of the mineral composition of the rock is: Onhioclasag ys tise suesa ey ete te 20 per cent. Wine RO ein eyes eet ee) Sel ee eh 10 Alibite 2s 222.2 ah Tie ee eee 20 Microcline-albite microperthite _ 33 JBSIOIUS Wes sy ASE a Ts eB ame NY Honniblen die es 2 es Sn 0°75 CHIKCIIGS gees aes ema Cae ig” Ed ZAC OM earner a 5 oa "25 100°00 ‘The dzotzte is olive-green in color. The hornblende is also olive-green and extinguishes at a small angle. The albite has the approximate composition Ab,An,,. Calcite is present as irregular grains of primary origin. The rock has an allotriomorphic granular texture. 2. WSchistose Canadite. Macroscopic description—The typical nephelite syenite of the Haliburton area, of which this rock is an example, has been called by Quensel,t canadite. It is distinguished from the usual nephelite syenite by the predominance of albite instead of a potash feldspar. It is a light gray rock, occasion- ally dotted with pmk. Jt is medium and even-grained. Biotite and feldspar are about equal in amount and form the larger portion of the rock. The gray platy feldspar is sometimes replaced by a pink variety. The rock breaks with an irregular fracture and appears quite fresh to the unaided eye. Lee description.—A thin section of this rock shows that it is composed of calcite, biotite, microperthite, and nephelite. These minerals crystallized approximately in the order named. They are present in the following proportions by weight : Albi tee Soe en aes ae aaa gol oe 79°34 per cent MiiGro pen thi eye sear ese 3°03 Nepheliteees meneame sess. 3217 BOLIC teres seer a eee ee oN 11°94 Calcite, eee me pyaar os 2°42 100-00 *F. D. Adams and A. EH. Barlow, Memoir No. 6, Canada Geol. Surv., 1910, p. 821, +P. D. Quensel, Bull. Geol. Inst. of Upsala, vol. xii, p. 135, 1914, 418 W. G. Foye—Nephelite Syenites of Ontario. The biotite is nearly black in color. The albite corresponds to the mixture (Ab,,An,). The calcite occurs as rounded grains between the other minerals. It is evidently primary in origin. The mzcrocline is closely associated with the nephelite and occurs also as resorbed blebs in the albite. ‘The evidence points to the fact that the earlier found microcline was unstable and that the increasing amount of soda in the magma allowed albite to take its place. The rock is hypidiomorphic, sometimes poikilitic, in texture. Chemical composition.—No chemical analysis of the par- ticular rock described above is available. This rock is, how- ever, practically the same as one described by Adams and Barlow* from’ the adjacent township of Monmouth. The analysis of the latter rock is as follows :— SiO, 51°58 per cent The calculated norm is :— TiO, "35 Orthoclase 25°02 per cent Al,O, 19°40 Albite 38°84 Fe,O, 4°26 Anorthite 6°67 FeO 5°29 Nephelite 15°50 MgO “49 Diopside 90 MnO "20 Olivine 5°05 CaO 3°64 Ilmenite “73 Na,O 7:49 Magnetite 6°15 K,O 4°23 Apatite "34 Jer 15 Calcite 3°45 Co, 1°53 Water 1°02 H, 1-02 = = 99°67 99°59 The rock occupies_the following position in the “norm” classification : Class Tot ee. anaes Dosalane Onder-6" (aoe ees eee Nogare Rangy2'. eee ee Essexase Sub Tang 4 see See Essexose 3. Nephelite Pegmatite. Macroscopic description.—Nephelite in irreguiar, anhedral crystals 4-5 em. long and 3 em. wide composes the lar ger portion of the nephelite pegmatite. Masses of calcite which show part- ing on —$ R are found between the nephelite crystals. A few *F. D. Adams and A. E. Barlow, Memoir No. 6, Canada Geol. Sury. 1910, pp. 262-264, s W. G. Foye—Nephelite Syenites of Ontario. 419 masses of scapolite, 2-3 cm. wide, are irregularly distributed through the rock, and, here and there, a bit of pyrrhotite is seen. A delicate pinkish white mica is also present and areas of sodalite which fade gradually into the nephelite. The rock is allotriomorphi¢ granular in structure, breaks with an irregular, subconchoidal fracture and is very fresh in appearance. Microscopic description.—The microscope shows that neph- elite, sodalite and calcite predominate in the pegmatite. The nephelite is in amoeboid forms extending into the albite. The evidence seems to show that whereas albite replaced microcline in the canadite, due to the increasing amounts of soda in the syenite magma, in this case the albite was replaced by nephelite for a like reason. Sodalite, also, replaces the nephelite and wraps about the grains of the latter mineral in irregular, amoeboid granules. The albite is practically free from lime. It is present in anhedral granules with threads of calcite intergrown along its twinning planes. The order of crystallization is muscovite, calcite, albite, neph- elite, sodalite. 4, Kaolinized Contact Rock. Macroscopic description.—There is a gradual transition from the nephelite pegmatite to the limestone along the upper con- tact of the nephelite syenite laccolith. The nephelite peg- matite is gradually altered to kaolin and calcite becomes more abundant until it completely replaces the nephelite rock. Biotite and apatite with minute red garnets appear, together with large masses of granular seapolite and albite. Twenty per cent of the rock is nephelite. This mineral is altering to kaolin, which forms fifty per cent of the rock. Cal- cite occurs in resorped grains in the nephelite and kaolin. Apatite occurs as prismatic crystals 1-2 em. long and a cen- timeter wide. The mineral is a light apple-green in color. The kaolinized and the fresh pegmatite are equally exposed to the processes of weathering. The kaolin is, therefore, prob- ably not due to secular alteration. The presence of fresh cal- cite within the kaolinized rock, also, would indicate that secular weathering has not progressed very far and that the solutions which effected the kaolinization did not contain acids capable of corroding the calcite. It is believed that the kaolinization was produced by carbonated waters during the latter part of the period of pneumatolytic action which produced the nephel- ite pegmatite. 420 W. G. Foye—Nephelite Syenites of Ontario. Microscopic description.—The thin section of the kaolinized rock shows only a few features not discernible in the hand specimen. The calcite is present in large cleavage sections set in a matrix of fine, fibrous, decomposition products consisting of kaolin and muscovite. The muscovite often shows rosette arrangement. Fig. 4. 500 Fic. 4. Distribution of the several rock types in the Crescentie Laccolith of Tory Hill. a=Garnet-pyroxene rock, b-=Hornblende-nephelite rock, e=Monmouth- ite, d=Pegmatitic nephelite syenite, e=Biotite-nephelite rock, f=Granite pegmatite. B. THE Crescentic LACCOLITH NEAR TORY HILL. General Geology. This laccolith is indicated on the accompanying sketch map (fig. 1) by the letter B; fig. 4 shows the general distribution of the rocks associated with the body. There are two parts to the laccolith. These are separated by limestone. Limestone surrounds the mass on all sides and dips away conformably to the contact. A massive, granite- peeaue dike forms a hill 200 yards across, south of the ody. A basic nephelite rock, consisting predominantly of garnet and pyroxene, is found at the lower or northern contact of the laccolith. The contact occurs at the top of a steep hill. The garnet-pyroxene rock forms a zone from four to eight feet in thickness. It grades into another type, still basic, but contain- ing nodules of nephelite and some hornblende. The latter W. G. Foye—Nephelite Syenites of Ontario. 421 rock is transitional to monmouthite, which was first described from this locality.* A valley intervenes between the northern and southern por- tions of the crescentic body. Masses of graphite occur within this valley. They are sometimes found in the limestone but more often are associated with the nephelite syenite. The rocks of the southern portion of the body occur in the same order as those of the northern portion. A coarse-grained, almost pegmatitic variety not found in the northern portion, forms, however, a considerable area south of the Monck Road. The dips and strikes of the surrounding limestones and the fact that the limestone passes beneath the northern border of this body are clear indications of its laccolithic character. Fie. 5. \Se- Kh S= SAAS BEESESS BEES BESS AEE 1600 feet Fig. 5. ‘ Cross-section of the Crescentic Laccolith near Tory Hill. a=Garnet-pyroxene rock, b =Hornblende-nephelite rock, c=Monmouthite, d=Pegmatitic nephelite syenite, g— Limestone. Figure 5 is a north-south section of the laccolith. The figures within the section indicate the specific gravities of the rock at 20° C. These figures show that gravity controls the differentiation of this mass. The rocks are all nephelite types and red syenite is not associated with them. This fact strengthens the opinion that the red syenite of the Gooderham laccolith was intruded after the nephelite types had differen- tiated, as indicated in the description of that body. Petrography of the Crescentie Laccolith. Figure 4 shows the distribution of the several rock types associated with the crescentic laccolith. ‘1. Pyroxene-Garnet Contact Rock. Macroscopic description.—The pyroxene-garnet contact rock is transitional to the overlying hornblende-nephelite rock. It is reddish-black in color and is composed of greenish-black *F. D. Adams and A, E. Barlow, Mem. No. 6, Canada Geol. Surv., 1910, pp. 274-277. 429 W. G. Foye—WNephelite Syenites of Ontario. pyroxene granules a half a millimeter in diameter, associated with red garnet in granules of the same size. Occasionally, large crystals of pyroxene, from 1 to 5 em. long and a centimeter or more wide, are present. The rock is mottled with grayish-white areas of calcite and albite which have a border of red garnet against the pyroxene. Microscopic description.—The following minerals in approxi- mately the given proportions are found in the contact rock :— Calcite gays.) = Re eae 25 per cent Pyroxene 223) = eaes 34 Gernete 2) oe eee ee 37 A libite pe Oe Ces ener emma 100 Calcite, besides forming with pyroxene and garnet a ground- mass of mosaic texture, occurs as minute blebs in the larger pyroxene crystals. The garnet is often found as euhedral crystals. The pyrorene is apple-green in color and extinguishes 36-37° on (110). The texture of the rock is allotriomorphie granular. Chemical description.—The following analysis is quoted from Adams and Barlow.* The garnet-pyroxene rock just de- scribed is more basic than the rock described by the above writers astypel. Thelatter rock is intermediate in composition between the garnet-pyroxene and the hornblende-nephelite rock. The hornblende-nephelite rock corresponds closely to their type II. Analysis of type I of Adams and Barlow. Si0, 43°67 per cent The calculated norm is :— TiO, 0°78 Orthoclase 12°79 per cent Al,O, 20°91 Albite 29°01 Fe,O, 3°54 Anorthite 20°29 FeO 8:01 Nephelite 19°03 MgO 1°46 Olivine 10°58 MnO 0°05 Ilmenite 1°52 CaO 7:37 Maegnetite 5:10 Na,O 6°73 Apatite “34 K,O 2°25 Calcite 5-4] PO! 0-11 a CO, 2°37 97°07 H, 2°52 Water 2°52 99°T% 99°59 *F. D. Adams and A. E, Barlow, Memoir No. 6, Canada Geol. Surv. 1910, — pp. 269-270. W. G. Foye—Nephelite Syenites of Ontario. 423 Its position in the norm classification is therefore :— Glass dil es eaie =: Dosalane Order'é* 228... Norgare Rang 2 ........-.- Essexase (near Salemase) nub-rang4: e220... Hssexose 2. Hornblende-Nephelite Rock. Macroscopic description.—The hornblende-nephelite rock is mottled black and white and contains occasional grains of garnet from 1 to 5 mm. in diameter. The jet black hornblende granules are arranged in irregular masses or lines with bits of feldspar or nephelite lying between them. The nephelite has the creamy white color it assumes when slightly decomposed. Microscopic description.—The thin section cut from this rock does uot correctly represent the relative amounts of its constituent minerals. A comparison of the section and the hand specimen gave the following approximate percentage composition :— Nephelite. 2.52 - 30°0 per cent ICE ses 2. ake 70 Hornblende _ =. —+<. 30°0 PAyGOKGNO =e on geen 20°0 Gannetees -a hanes: 75 @alewtie sso 5 2.02% 5°0 Apa titel. 2 2 tea oa 100°0 The nephelite occurs as large irregular grains. The albite is in minute rounded erystals, within or adjacent to the nephelite. The hornblende is deeply pleochroic, dark bluish green to brown. The pyroxene is apple-green in color. It is much shattered and the cracks are healed by an epidote, or, sometimes, garnet. The rock as a whole shows granulations under pressure. Chemical description —Adams and Barlow give the follow- ing analysis for their type II, which is approximately the same as this rock :— 424 SiO, 42°72 per cent NCO 0°38 Al,O 25-08 Fe,O, 2°00 FeO 4°36 MgO 97 MnO 16 CaO 6°92 Na,O 11:02 K,O 2°69 JE) 218) CO, 2°99 H, 88 100°36 Rang 2 8. Monmoutiite. W. G. Foye—Nephelite Syenites of Ontario. The calculated norm is :— — Orthoclase Albite Anorthite Nephelite Diopside Olivine Ilmenite Magnetite Apatite Calcite Italare Dosalane Vulturase V ulturose 15°57 per cent 7°34 Macroscopic description.—Monmouthite was first described from the laccolith now under diseussion.* The rock which is about to be described resembles the type monmouthite but does not contain hornblende (hastingsite) or cancrinite. original type is very rare in this locality. The rock is bluish-gray in color, medium-grained and some- what schistose. It is flecked with minute bits of black mica (1-2 mm. in diameter) which are arranged in sub-parallel layers. Microscopic description.—The minerals found in this rock, arranged according to their proportions by weight, are :— Nephelite_..._.-.---- 56°68 per cent Scapolite seen == ae 10°18 Biotite hc «aus aoe 22°50 Al bitie': 78> a7 Hanser 9°25 Koolings ape ve ene 1:02 Caleitierer 220. 2) a “3 100°00 * F. D, Adams, this Journal, (4), vol. xvii, pp. 272-276, 1904. The W. G. Foye—Nephelite Syenites of Ontario. 425, A smail amount of zoisite is also present. Adams states that the specific gravity of the scapolite is 2-711, which shows that it is near the meionite end of the scapolite series. Biotite is closely associated with the scapolite. When biotite appears, nephelite disappears and scapolite or albite takes its place. The kaolin is largely the decomposition product of the nephelite. | Chemical description.—The following is the analysis of the type monmouthite as given by Adams and Barlow: * SiO, 39°74 per cent The calculated norm is :— iO, 0°18 Anorthite 12°51 per cent Al,O, 30°59 Nephelite 67°72 Fe,O, “44 Leucite 8°28 FeO 2°19 Olivine 3°70 MgO 60 Akermanite “40 MnO ‘03 Magnetite “70 CaO 5°75 Ilmenite "30 Na,O 13°25 Pyrite 14 K,O 3°88 Calcite 4-92 CO, 2°17 = so, tr 98-07 Cl 0°02 Water 1:00 s 0:07 === HO 1-00 99°07 99°86 (QUETSeyidl lie sat acta ae a aa Persalane Order spe ree ee Ouarae Ianger2e 2. 220s 2a) 2 22 23) Vs Monimonthase NSO) DCH 1S we a ees eek Cem Monmouthose 4, Schistose Nepheiite-Biotite Rock. Macroscopic description.—The eastern end of the northern portion of the crescentic laccolith is composed of a bluish-black, medium-grained and very schistose rock which is largely made up of biotite. The biotite is in crenulated layers which form _ wavy lines across the rock. Between these layers are lenses of gray and pink feldspar and nephelite from a centimeter to a centimeter and a half in length and from 8 to 10 mm. wide. Microscopic description.—An estimate of the percentage amounts of the several minerals composing the rock, made by *F.D. Adams and A. E. Barlow, Memoir No. 6, Canada Geol. Surv., 1910, p. 276. 426 W. G. Foye—Nephelite Syenites of Ontario. comparing the thin-section with the hand specimen, gave the following result:— Allbit@ ze. steps ates eye ere 45 per cent Nephelite. 0c ee eel Diotite ie < sae Saud eee ee 30 Orthoclase 221-3. eee 10 Pyirhotite 2 feo... ah. ee 2 Caleite’. = 2i>26. te 28 See 3 100 The twin lamelle of the albite are much contorted. The compressive stresses were able to fracture the minerals of the rock but reerystallization is not found. Nephelite occurs as amoeboid grains between the albite erys- tals. It is, also, included in the albite. The orthoclase is in allotriomorphiec crystals and is usually associated with biotite. No chemical analysis of this rock type is available. 5. Pegmatite-Nephelite Syenite. Macroscopic description—The pegmatitic facies of the erescentic laccolith is coarse-grained but is not as rich in nephelite as the nephelite pegmatite of the Gooderham body. It is light bluish-gray in color and is mottled with patches of black. The light-colored minerals are feldspar and nephelite. They oceur in allotriomorphic crystals from 2 to 3 em. long and from 1 to 2 em. wide. Grains of hornblende are grouped in masses from 2 to 8 cm. in diameter. These sometimes show a vague schistose arrangement. Titanite occurs in crystals 2 em. long and a centimeter wide. Apatite is present in long slender needles which are light-green in color. Bits of pyrrhotite are disseminated through the rock and sodalite may occasionally be seen. ; Microscopic description.—The following minerals compose the rock in the given proportions by weight :— Micropexthite =< 2722. easee 59°02 per cent AD IGG Sone poten isk cops pte aaa 12°24 INephelite ss. 5... 5.5 5 hee ama FHormblenden 23. ere Calcite Seekiee So 8 OS Ee er AMtaniteweme cece. hull 2s eeyaeWers 52 4 « te | W. G. Foye—WNephelite Syenites of Ontario. + bo =I Sodalite was not shown in the slide examined. The nephelite often surrounds the albite and has apparently replaced that mineral. A border of albite sometimes inter- venes between the nephelite and microperthite. The hornblende resembles hastingsite. The calcite is primary and occurs as rounded grains in the hornblende or between the hornblende and nephelite crystals. No analysis of this type is obtainable. PART II. SUGGESTION CONCERNING THE ORIGIN OF NEPHELITE SYENITES. Professor C. H. Smyth, Jr.,* has suggested that the nephelite syenites are segregated from a primary magma by means of pneumatolytic gases. The pneumatolytic origin of nephelite and sodalite has been advocated by a number of writers. Mier- isecht and Lacroixt found that these minerals were sublimation products within the cavities of limestone bombs from Monte Somma. Lacroix states :—‘ The veinlets of orthoclase, soda- lite, ete., traversing a doleritic leucite-tephroite, recall in a striking manner veinlets of nephelite aplite which traverse the essexites and nephelite gabbros of Madagascar, Brazil, etc.” Brégger,§ in his study of the nephelite pegmatites of Norway, found that nephelite was frequently altered to sodalite and infers that the latter mineral was the result of the action of hot sodium chloride solutions. Lacroix| intimates that chlorine and fluorine were influential in the formation of sodalite in the nephelite syenites of the Los Archipelago. F. W. Clarke considered the sodalite of the nephelite syenite type, litehfield- ite, to have been formed at the expense of nephelite. J. Lemberg** produced sodalite from nephelite powder by allowing it to stand six months on a water bath at 100° C. in a 20 per cent solution of sodium chloride. The accompanying black and white sketch (fig. 6) shows the yein-like aspect of the sodalite in a nephelite syenite from Ice River, British Columbia.t+ It would be difficult to disprove the pneumatolytic origin of the mineral in this case. * ©. H. Smyth, Jr., this Journal, (4), vol. xxxvi, pp. 33-46, 1913. + Bruno Mierisch, Tscherm. Mitt., vol. viii, p. 188, 1887-88. t A. Lacroix, Mus. d’Hist. Nat. (Paris), Nouv. Arch., (4), vol. ix, p. 101- 102, 1907. SW. C. Brogger, Zs. Kryst., vol. xvi, p. 167, 1890. || A. Lacroix, Mus. d’Hist. Nat., (5), vol. iii, p. 98. 4] F. W. Clarke, this Journal, (3), vol. xxxi, p. 268. ** S. Lemberg, Zs. d.d. Geol. Gesel., vol. xxxv, p. 582, 1883. ++ This slab may be seen in the geological laboratory of Harvard Uni- versity. 428 W. G. Foye—Nephelite Syenites of Ontario. The close association of nephelite syenites with granite peg- matites, in the Haliburton area, indicates that they originated during the pneumatolytic stage of the intrusion of the Lauren- tian granite. The parent magma of the Gooderham nephelite syenite laccolith must have been especially charged with pneu- Fie. 6. Fic. 6. Tracing of the sodalite areas of a polished slab of sodalite-syenite from the Ice River district, British Columbia. matolytic gases since the amount of nephelite pegmatite associ- ated with the body is unduly large. Many instances have been given in the discussion of the Gooderham and Tory Hill laccoliths of the replacement of microcline by albite, of albite by nephelite, and in certain cases of nephelite by sodalite. The accompanying microphotographs illustrate this point (fig. 7). It is believed that these replace- ments show that a magma, from which microcline was first capable of crystallizing, was gradually enriched in soda, until first albite, then nephelite, and, finally in the extreme cases of W. G. Foye—Nephelite Syenites of Onturio. 429 Fic. 7a. Albite and microcline associated with nephelite. The micro- cline has a border of albite against the nephelite (x 24), Fic. 7b. Fic. 76. Microcline resorped by nephelite (x 48). 430 W. G. Foye—Nephelite Syenites of Ontario. the nephelite pegmatites, sodalite was the stable product of erystallization. How these soda solutions were, introduced into the magma is not definitely known. Nevertheless, it is reasonable to sup- pose that the so-called pneumatolytic elements now associated with the nephelite syenites were the effective means of trans- portation. The minerals of the nephelite pegmatite of the Gooderham laccolith are nephelite, albite, sodalite, scapolite, apatite, calcite, titanite, and pyrrhotite. Chlorine is associated Fic. ‘ic. Fic. 7c. Albite resorped by nephelite (x 48), Fies. 7a, b, c. Microphotographs of the nephelite pegmatite from the crescentic body near Tory Hill. with three of these minerals, carbon dioxide with one, phos- phorus with one, and sulphur with one. These elements are believed to have been active in producing the nephelite syen- ites. The writer is firmly of the opinion that Professor Smyth is correct and that-many lines of evidence lead to the conclusion that the nephelite and sodalite rocks are associated with the pneumatolytic state of igneous activity and are produced by gaseous transfer of their elements. Recently while studying the chemical composition of the amphibolites of Haliburton County, which Adams and Barlow* have shown were derived from limestone, it was impressed on *F,D, Adams and A. E. Barlow, Memoir No. 6, Canada Geol. Sury. 1910, pp. 108-111; also F. D. Adams, Journ. of Geol., vol. xvii, p. 1-16, 1909. W. G. Foye—Nephelite Syenites of Ontario. 431 the writer that the origin of these rocks was perhaps closely associated with the origin of the nephelite syenites. Column 1 of Table I gives the composition of a typical amphibolite de- rived from limestone. Column 2 is the analysis of an average limestone from the same locality. TABLE I. il 2 SUG a ala 50°00 Pi iC i aa ae 0°82 ion EOD Wor rege te 126 CRS ete a Oa 13s JON Seeger aes aaa Bedi By et MirsO v2 2028 et 4°63 3°20 VETO NS 2s 2 \ os ee 0°08 ese O@2O8e 86 sss0us 5% 10°65 49°88 Na,O toes ee eee 4°48 See UO pe hg Gs 1118 ee Co, fe eee ee ea 0°10 APIS NS)» cet aire eee are 0°03 Rae (Ooh SE eee a 0°10 spared H,O Ae eee 1:00 ee as 99°97 99°39 It will be seen that the limestone lost in the process of its transformation to amphibolite 39-2 per cent of lime and 42°6 per cent of carbon dioxide. On the other hand, it gained 47-7 per cent of silica, 19 per cent of alumina, 6:8 per cent of iron oxide, 1-4 per cent of magnesia, 4°5 per cent of soda, and 1:2 per cent of potash, beside other minor elements. Table Il shows the added oxides and the same oxides from the type canadite of the Haliburton area (see page 418) recalcu- lated to 100 per cent. TaBLe II. 1 2 Amphibolite from Nephelite limestone syenite SiO sh oe eee 60°4 56°2 UN Orme eure ber 2 22°8 20°6 He O, & FeO... - 8°8 10:8 NEO Uae og 1-7 0-5 Na Oreo ee eae 5:0 76 KG Oh ate eae 1°3 4°3 100-0 100°0 Am. Jour. Scr.—Fourts SERIES, Vou. XL, No. 238.—Octosmr, 1915. 28 432 W. G. Foye—Nephelite Syenites of Ontario. If the alkalies are neglected, the difference between the per- centages of the other elements are not greater than those reg- istered by the varying types of the nephelite syenites. Column 1 of Table IL shows beyond doubt that solutions high in alumina and relatively high in soda were introduced into the limestone and that, to a degree, these solutions were similar to the material which went to form the nephelite syenites. To one who has studied the Haliburton district closely, it appears very clear that the processes which Adams and Barlow described have actually taken place and that amphibolites have been derived on a large scale from limestone. All steps in this process are continually met with. The limestone is first impregnated with diopside crystals. Then scapolite and diop- side appear side by side in a matrix of calcite. Later diopside and feldspar are developed, and finally the diopside is replaced by hornblende and the transformation is complete. The evidence of this transformation was accepted by the International Committee on Pre-Cambrian nomenclature.* The following statement is quoted froin their report : “The committee are of the opinion that the evidence is indis- putable that the granite bathyliths in this region change the invaded limestone into a dark-gray amphibolite, which, together with fragments of the interbanded amphibolites found every- where in the limestone series, occur scattered throughout the granite mass in the form of included fr agments.” The amphibolites are vastly more important areally than the nephelite syenites, and hence the origin of the solutions which left the granite magma and transformed the Grenville limestones to amphibolites is of prime importance. The evi- dence presented in this article shows a close genetic relation between the amphibolites and the nephelite syenites. A study of the map published by Adams and Barlow and an intimate knowledge of a portion of the Haliburton area war- rant the following statements : (1) The area embraces some 3456 square miles, a fourth of which is underlain by limestone and amphibolite and the re- mainder by Laurentian gneiss and amphibolite. (2) A conservative estimate would show that the amphibolites formed from 30 to 40 per cent of the limestone areas and from 10 to 15 per cent of the gneissic areas. (8) The contacts of large oval masses of gneiss with the sur- rounding limestones usually are characterized by a zone of amphibolite mingled with limestone. The average zone is one- quarter to one- half a mile in width and represents a thickness of from one to two thousand feet. = * Report on Correlation of Pre-Cambrian Rocks, Jour. Geol., vol. xv, p. 204, 1907. W. G. Foye—Nephelite Syenites of Ontario. 433 (4) A liberal estimate of the nephelite syenite bodies mapped in the Haliburton region would place their total area below sixty square miles, or less than ten per cent of the area of the amphibolites. (5) The amphibolites derived from impure lenses in the limestone and those derived from gabbro by dynamic meta- morphism are insignificant in amount when compared with the amphibolites derived from the limestone by contact meta- morphism. In view of these facts, it is desirable to know whether the solutions which transformed the limestones to amphibolites were segregated from the granite by juvenile gases, as Smyth has suggested, or whether the interaction of the limestone and the granite caused these solutions to be given off during the formation of the amphibolites. It is clear from the following analyses that the amphibolites did not result from a simple reaction between definite propor- tions of the granite magma and limestone. Soda, potash, alu- mina, and silica are present in the amphibolites in proportions very different from those in the granite. TasBie III. Analysis of Analysis of Typical Red Amphibolite Gneiss SHOE Nae eee 76°99 per cent 50:00 per cent SUG) bs ag) eres eee ace 0°82 PRO ei ooh: 12°45 18°84 Vs OO eee elena 1:03 2°57 HeOyes sake aa 0°49 5°51 Wie OF .2¢ Goes.) 0-21 4:63 ING ee ee tr. tr CiOp ers ae -se 0°98 10°65 INTRO) Sseeh 3 Stace ea 3°46 4°46 OPS ae eae 4:99 1°18 Oe or Ske 0°26 1:00 Cle eee Seatac . NONE 0°10 Sot eer ae 0:03 COME se 2c ee os = none 0°10 100°16 99°97 Before discussing this matter, the writer wishes to state a theory concerning the method of emplacement of the Lauren- tian granite and the origin of lit-par-lit structure in the Hali- burton area. It is his purpose to treat the subject in more detail at a later time. The Grenville series is very thick; how thick, no one wishes to state. Granite pegmatite penetrates this series in every 434 W. G. Foye—Nephelite Syenites of Ontario. direction. One of the striking facts concerning this area is the predominance of pegmatitic rather than normal granite, show- ing that the Pre-Cambrian granites which intruded the Gren- ville series were abundantly supplied with pneumatolytic gases. ~ It is believed that the presence of these gases made possible the lit-par-lit structure of the Haliburton area.* They were capable of penetrating the limestones along planes of weakness, induced probably by static metamorphism,t long before the liquid magma could be intruded. The reaction of the lime- stones with the gases opened passageways through which liquid magma entered. Gradually, in this way, an immense number of alternating layers of granite magma and limestone were formed. The materials of this gigantic steam pack reacted with each other and produced finally a lit-par-lit strueture of granite and amphibolite with a few residual layers of lime- stone. The nephelite syenites were developed during the period of intrusion of the granites. Their total mass is far less than one per cent of the total mass of the granites and they are transi- tional to granite pegmatite through the intermediate stage of red syenite. The nephelite syenite laccoliths occur as small isolated bodies in the limestone or as border facies of the gran- ite. It has been objected to the syntectic theory of the origin of nephelite syenites that the amount of limestone capable of reacting with the granite magma was entirely insuflicient to alter the nature of that magma. The method of emplacement of the granite just described and the large scale production of amphibolites from limestones most effectively removes this objection. If from ten to twenty per cent of the Haliburton area is underlain by amphibolites derived from limestone, then some drastic effects must be looked for in the granite magma which gave rise to the solution producing these amphibolites. Returning to the question of the origin of the fluids which reacted with the limestone to produce amphibolites, the writer wishes to state very positively that he believes with Smyth that magmatic gases were the effective means of transporting the fluids from the granite into the adjacent rocks. He believes, however, that the soda solutions were given off by the granite magma because lime was capable of replacing soda at high temperatures. The magmatic gases, therefore, were only a means of transportation and were not active in releasing soda solutions from the magma. If it is believed that the nephelite rocks were segregated independently of the action of limestone, the close association *Cf. C. N. Fenner, Jour. Geol.. vol. xxii, pp. 594-612, 1914. fot ee e A. Daly, Guide Book, No, 8, Inter, Geol. Cong. Can., Part II, p. 28, 19138. W. G. Foye—Nephelite Syenites of Ontario. 435 ™ of the granite pegmatites and the nephelite syenites would lead one to suppose that the same agents, 1. e. magmatic gases, segregated two very dissimilar magmas at the same time from the granite. It is difficult to conceive how granite, a magma high in silica, and nephelite syenite, low in silica, could both be transported through the primary magma chamber simulta- neously, unless one believes in Rosenbusch’s theory that the foyaite and granite kerns are immiscible. To the writer it appears more reasonable to suppose that soda solutions given off in response to the interaction of the limestone with the granite magma euriched small, confined portions of the granite pegmatite in those elements character- istic of nephelite syenites and so produced the alkaline types. Such a local desilication of the granite was often observed in the field. Near Bancroft, Ontario, large masses of augite, a foot or more in length, are found near the contact of the granite with the limestone. Wherever these occur, the granite is altered to red syenite and often contains lenses of nephelite syenite. Daly* has shown that lime in the formation of the pyroxene molecule binds to itself 2°5 its own weight of silica. That the granite should be desilicated under these circum- stances is not to be wondered at! At the limestone contact and within the main mass of that rock the granite has produced amphibolite instead of augite from the limestone. The formation of amphibolite binds still more silica than the formation of augite and is, therefore, a more effective means of desilicating the granite. It has already been stated that the contact of the am phibo- lite with the limestone is characterized by a scapolitic zone. Seapolite is a mineral high in alumina and often high in soda. It oceurs in large masses associated with the corundum syenite of Craigmont, Ontario, and is an essential constituent of most of the nephelite types described from Gooderham and Tory Hill (Part I). The formation of scapolite is decidedly a case in which soda-rich solutions transported by pneumatolytic gases are given off, due to the interaction of limestone with granite magma. V.M. Goldschmidtt+ has shown this in his study of the Christiania region. The constant association of this min- eral with nephelite in many localities cannot be without significance. Meionite and nephelite frequently line cavities in the bombs ejected from Monte Somma. Mierischt and Lacroix§ have shown that they were deposited as sublimations from a gaseous *R. A. Daly, ‘‘Igneous Rocks and their Origin”, New York, 1915, p. 431. +V.M. Goldschmidt, ‘‘Die Kontaktmetamorphose im Kristianiagebiet,”’ Kristiania, 1911, pp. 28 and 320. ¢ Bruno Mierisch, Tscherm. Mitth., vol. viii, p. 188, 1888. $A. Lacroix, Mus. d’Hist. Nat., Nouy. Arch., (4), vol. ix, pp. 101-2, 1907. 436 W. G. Foye—Nephelite Syenites of Ontario. state. R. Brauns* has described the seapolite bombs of the Laacher See district. They, likewise, contain nephelite. P. D. Quensel+ has found vesuvianite, a mineral related to scapolite, as a primary mineral in the canadites of Almunge, Sweden. He concludes, “Hence the possibility must be con- sidered that the vesuvianite is the remnant of a former assimi- lation process.” V. M. Goldschmidtt has recently described the enrichment in soda of a sandstone having numerous calcite lenses. The calcite lenses show an increase in their soda content of 2°58 per cent. The lime was taken into solution and transported to nearby slash veins, where it was redeposited as prehnite, clino- zoisite, diopside, and scapolite. Geological literature has numberless examples of the enrich- ment of caleareous rocks in soda. Some ascribe this to the susceptibility of limestone to such réplacement, but this is but another way of saying that lime and soda easily are inter- changed at high temperatures within an igneous magma, or, for that matter, within the laboratory, as those who use the J. Lawrence Smith method for the determination of the alka- lies know. If this is true, the emanation of solutions high in soda from a syntectic formed during the intrusion of a gran- ite magma into limestone is to be expected. That such solu- tion did enter the limestones of Haliburton County and transform them to amphibolites has been shown by Adams and Barlow. Conclusions—The close association of granite pegmatite with nephelite syenite indicates that they originated from the primary granite magma at approximately the same time. If they were segregated without the aid of foreign material, it would seem that the low percentage of silica in one magma and its presence in excess in the other should have been adjusted by an interaction between the two. To the writer, it appears more reasonable to suppose that the solutions which gave rise to the nephelite syenites had their origin near the surface, that they were produced by the reaction of limestone with granite magma, and that these solutions enriched certain confined portions of the granite magma in the elements characteristic of nephelite syenites. If the invading granite magma reacted with the various lenses of limestone between which it was intruded, there would be a possible source of soda solutions. That such solutions were produced is shown by the differential transfer of soda into the amphibolites. Harvard University, Cambridge, Mass. on Beran Neues Jahrb, Min., Beil. Bd. xxxiv, p. 85, 1912; also xxxix, p. 124, 1914 +P. D. Quensel, Centralbl. Min., No. 7, p. 205, 1915. $V. M. Goldschmidt., Neues Jahrb. Min., Beil. Bd. xxxix, p. 198, 1914. H. W. Shimer—Post-Glacial History of Boston. 437 | Art. XXXI.—Post-Glacial History of Boston; by H. W. SHIMER. CONTENTS, Introduction. Typical section through the post- glacial sediment of Boston. List of species. Conclusions. Bibliography. Introduction. Tue recently completed excavations of the Boston Elevated in the construction of the Boylston Street subway give many new sections down through the mud to the glacial clay. The resulting new exposures of the post-glacial shells through the middle of Back Bay proved to be so interesting that it was decided that a record of them should be kept. The foilowing is a brief summary of this work: Typical section through the post- Pleistocene sediment. The following section from the Boylston Street subway ex- eavation at Exeter Street is typical of all the Back Bay sections, and fairly so of all the others from Brookline to City Point. iil im blottme-out the Back Bay..-.--2.22.2------ 16 feet This fill has taken place mostly since 1868. 2. Gravelly black silt. Few fossils......--..---.------ 6 ceniuer Diack silt. » Many fossils. 2 222. 520-2222 12 5 The middle portion is very full of fossils ; the upper- most two feet and the lowest foot contain but few. This silt is a dark-gray (when dry) argillaceous sand with a considerable number of mica scales. The compound microscope shows that very minute sand particles make up fully nine-tenths of the mass ; there is merely sufficient clay and carbon particles to give consistency and a dark-gray color to the sediment. ssandy.tresh-water peato. J.s22205 022.22 35s se ek 5 . Blue sandy clay with some peat.-....-.--------.--- 1 foot This is the upper edge of the glacial deposit. oo The surface of the street here is 16°5 feet above mean low tide, while the bottom of the peat bed is 15:5 feet below mean low tide. List of species. The following list includes all post-glacial fossils noted, or listed by previous observers, from Brookline through Back Bay to City Point. Appended to the list are notes upon the peat and the oyster. 438 HH. W. Shimer—Post-Glacial History of Boston. 1. Plants (unrecognizable frag- ments) FoRAMINIFERA* . Polystomella striatopunctata ee SD. . Trochammina inflata m CO bo SPONGES 5. Cliona sulphurea BRYOZOA 6. Membranipora pilosa PELECYPODA 7. Yoldia limatula 8, Ostrea virginica 9. Pecten gibbus borealis . Pecten magellanicus . Anomia simplex . Mytilus edulis . Modiolus modiolus . Modiolus demissus var. plica- tulus . Clidiophora trilineata . Arctica islandica . Cyclocardia borealis . Astarte undata . Astarte elliptica . Lucina filosa . Kellia planulata . Laevicardium mortoni . Venus mercenaria . Gemma gemma . Petricola pholadiformis . Macoma balthica . Tagelus gibbus . Ensis directus . Angulus tener . Mactra solidissima . Mulinia lateralis . Mya arenaria 69. GASTROPODS . Acmaea testudinalis . A. testudinalis var. alveus . Vitrinella shimeri Clapp . Turbonilla winkleyi . Odostomia trifida . Odostomia fusca . O. bisuturalis . Littorina rudis . L. rudis tenebrosa . Littorina palliata . Littorina littoreat . Lacuna vincta. . Crepidula fornicata . Crepidula plana . Crepidula convexa . Polinices heros . Polinices triseriata . Neverita duplicata . Paludestrina minuta . Bittium alternatum . Triforis nigrocinctus . Columbella lunata . Cingula carinata . Buccinum undatum . Nassa trivittata . Ilyanassa obsoleta . Urosalpinx cinereus . Thais lapillus . Anachis avara . Tornatina canaliculata . Melampus lineatus CRUSTACEA . A copepod . Balanus balanoides . B. crenatus . B. poreatus . Crab claws VERTEBRATA Fish After an examination of the peat from the subway at Exeter Street, Mr. G. B. Reed of Harvard University writes: * Identified by J. A. Cushman. + This species was doubtless introduced into the fossil shells by the dredge, since nowhere else upon the American coast is this shell reported earlier than 1855 (Gulf of St. Lawrence). Europe. It is apparently a late migrant from HH. W. Shimer—Post-Glacial History of Boston. 439 “] find no plants or remains of plants such as now grow on salt marshes or anywhere below high tide level. But what species have entered into the formation of the peat I can not determine beyond the presence of grasses and sedges, probably both tops and roots, woody roots probably of some Ericaceous plants, and fragments of wood. A large part, however, is made up of much decomposed material now unrecognizable. It has apparently, too, undergone considerable compression as all the stems are flattened.” Ostrea virginica.—This, our only species of oyster, is very rare at Exeter Street, being represented by but three speci- mens, the largest of which is only 85"™™ long by 70™™ broad. At Charles River this shell is exceedingly abundant including both the long, narrow or so-called “current” form and the short, broad ‘‘ quiet-water” form. The most usual size of the former is 230" in length and 55™™ in breadth; of the latter the length is 130™™ and the breadth 70™™. At Berkeley Street the very large current form is common at a depth of 27 to 31 feet. A valve of one of these, an old individual, with a length of 140™™ has a maximum thickness of 50™". At City Point the specimens are similar in size and abundance to those from Charles River. Miss Bryant figures one from here 265™ by 80™™. This oyster, as native, is now absent from Massachusetts Bay; during early colonial days it occurred only locally and then, on account of the cold air at such depths as to be exposed only at the low spring tides. A large oyster-bank was situated at the mouth of the Charles River, another at the mouth of the Mystic and probably one on the Noddles Island, now East Boston, flats. That the large current forms flourished in Back Bay as late as the middle of the seventeenth century is shown by the fol- lowing quotations: “The Oisters be great ones in forme of a shoo horne, some be a foote long, these breed on certain bankes that are bare every spring tide. This fish without the shell is so big that it must admit of a devision before you can well get it mto your mouth.”* . . . . “Towards the southwest in the middle of this Bay” (i. e., Back Bay, at mouth of Charles River) “is a great Oyster-banke” . .. . “The Oyster-bankes” (referring to the same) ‘doe barre out the bigger ships.” In the first edition (1841) of the * Invertebrata of Massa- chusetts,” Dr. Gould says (p. 357) “old men relate that they were accustomed to go up Mystic River and Charles River, and gather oysters of great size, before it was the custom to _ *New England’s Prospect, etc.” : William Wood, London, 1634. Prince Society Edition, 1865, pp. 39-44. s 440 HH. W. Shimer—Post-Glacial History of Boston. bring them from New York. And even now individuals of enormous size are occasionally brought from both these places, and probably might be found by special search, at any time.” The cause of this great numerical reduction since colonial days is said to be a very severe cold spell about 1780 in which the sea bottom was covered with icc, thus preventing the oysters from getting air. Another factor which aided in the destruction of some of these species, especially the oyster, from the Back Bay region was the gradual obliteration of Boston as an island by the formation of a neck uniting it with the mainland to the south. Even during late colonial days heavy seas washed over this neck into the Back Bay. Oysters need a clean substratum, such as gravel, or other shells, to which the young, the spat, may attach themselves, otherwise they will perish; and the opening across Boston neck would give the tidal currents extra strength with which to cleanse this partially enclosed region from the river muds ; but that this was never so exposed to the action of waves as at City Point is shown hy the occurrenee of the surf-clam (Mactra solidissima) at the latter place*only. Many plantings of the oyster spat in its old home in the Charles River during recent days have resulted merely in the death of the spat. Conclusions. The deposition of the blue clay took place probably in a body of fresh water, since no remains of animal life are appar- ent init. The clay itself, derived from a near-by melting gla- cier, is the so-called glacial flour,—the material ground from the rocky floor by the stones held firmly in the advancing ice. A few unidentitied pieces of wood were noted in this clay. After the glacier had melted away from this region, the area was subjected to erosion by running water as evidenced by the gullies in the surface of the clay as well as in the sand plains (fossil deltas deposited by glacial streams). At the Longwood Bridge, Brookline, the sand-plain was eroded to a depth of 37 feet. During and subsequent to this erosive period, or at least the latter part of it, fresh-water peat was broadly devel- oped. Theland was at this time sinking with reference to sea- level, and continued its downward motion until the sea invaded the area under discussion. The evidence that the peat in Back Bay furnishes in regard to this downward movement is as follows :— The bottom of the peat at Fairfield Street is 23 feet below low tide, at Exeter Street 15:5 feet, at Church Street 33 feet, _ and at Charles Street it is 27 feet. With a height of tide of 10 feet, as it was in the Charles River before the construction of the tide-water dam, it would HH. W. Shimer—Post-Glacial History of Boston. 441 mean a submergence of this region of at least 33 plus 10, or 43 feet ; andif the peat was formed far above sea-level it would mean aso much greater submergence. Following the submergence occurred the deposition of the black silt, with which were enclosed shells and other records of the life then living in these waters. This record is of pri- mary interest because of the evidence it furnishes as to a warmer period between the presence of glaciers here and the present. That the climate of Boston has become slightly colder since the time of the maximum development of this fauna is shown in the change in the present distribution, especially of the mollusks, for all are still living. Of the sixty some species noted from the localities given, about half no longer occur north of Cape Cod, or only rarely in sheltered areas, but find their perfect environment farther south. These forms belong to the Virginian fauna, which is typically developed from Cape Cod to Cape Hatteras, though some of the species also thrive northwards in a few protected places. Ganong* men- tions nine such areas, including the Gulf of St. Lawrence, Oak Bay, New Brunswick, Casco Bay, Maine, and Massachusetts Bay. Between the retreat of the ice from this coast and the present time a period must have occurred during which the waters were as warm as those from Cape Cod to Cape Hatteras, and during which this Virginian fauna migrated northward. This was followed by a gradual refrigeration of the waters sufficient to prevent the breeding of many of the species except within a few areas protected enough to raise the temperature of the air and water sufficiently during the summer, or breed- ing season, for the development of the young. The following are some of the principal post-glacial fossils of Boston oceur- ring thus rarely or not at all to the north of Cape Cod,— Ostrea virginica, Venus mercenaria, Pecten gibbus borealis, Laevi- cardium mortoni, Triforis nigrocinctus, Mulinia lateralis, and Vitrinella. That this refrigeration continued during early colonial days is shown by the later disappearance from the vicinity of Boston of such species as the then abundant oyster, and the great reduction of many others both in number and in size. Most sections through the black silt in the Back Bay area show a more fossiliferous lower portion and an upper portion with comparatively few fossils. This difference may be corre- lated with the partial closure of Back Bay by the tidal build- ing of Boston Neck; the consequent reduction in tidal scour would then cause a more rapid accumulation of sediment within the Bay. *W. F. Ganong, Trans. Roy. Soc. Can., vol. viii, Sec. 4, pp. 167-185, 1890. 442 HH. W. Shimer—Post-Glacial History or Boston. A small portion just beneath the “fill” may be due to the presence of dam walls built in the early part of the 19th cen- tury. In 1814 a corporationn—“The Boston and Roxbury Mill Corporation,” led by Uriah Coting, obtained a charter from the General Court empowering them to build adam from the end of Beacon Street (at Charles St.) to Sewell’s Point in the uplands of Brookline, with a cross dam to Gravelly Point in Roxbury ; also to make a roadway of each dam and to levy tolls for its use. It could confine tide water within this area and run mills by the water power thus created. At this time there was nothing but water and salt marsh from the foot of the Common to the uplands of Brookline. The mill dam was finished in 1821. But the tidal power, rather insufficient at the beginning for the running of the mills, was soon encroached upon, first, by the owners of bordering property filling in their Jand, thus restricting the area of the dam; and especially, secondly by the buiiding of the Boston and Providence and the Boston and Worcester Railroad across the water basin (incor- porated in 1831). With this restriction of the tide and the increase in population this basin soon became a public nuisance, and in 1852 a commission of the state legislature reeommended that the property be abandoned for mill dam purposes and be filled in for building purposes. This was finally done, giving as a result the topmost 15 to 20 feet in the above Back Bay sections. Bibliography. Bryant, D. L. A Study of the Most Recent Geological History of the Tide- Water Region of Charles River. An unpublished thesis in the library of the Geological Department of the Massachusetts Institute of Technology. 1891. Crosby, W.O. A Study of the Geology of the Charles River Estuary and Boston Harbor, with special Reference to the Building of the proposed Dam Across the Tidal Portion of the River. Tech- nology Quarterly, vol. xvi, pp. 64-92, 1903. Upham, Warren. Recent Fossils of the Harborand Back Bay, Boston. Proc. Bos, Soc. Nat. Hist., vol. xxv, pp. 305-316, 1891. Windsor, Justin. The Memorial History of Boston. Massachusetts Institute of Technology, July 3, 1915. Chemistry and Physics. 443 SCIENTIFIC INTELLIGENCE. I. Cxemisrry anp Puysics. 1. The Improvement of Hiyh Boiling Petrolewm Oils.—For a long time it bas been known that a sufficient heating of high boiling oils, a process called cracking, will cause them to break down into lower boiling oils or even gases with the deposition of carbon. A difficulty in the practical application of this cracking process lies in the fact that in breaking down the high boiling hydrocarbons into simpler ones there is not enough hydrogen to saturate the newly formed bodies, in spite of the deposition of carbon, so that the products contain large amounts of unsaturated hydrocarbons, which give them undesirable qualities. Such pro- ducts can, of course, be refined somewhat with sulphuric acid, but there must be too much acid used and too much oil lost to permit in practice a thorough treatment with the acid. A. M. McAresz has applied Friedel and Craft’s reaction, that is, treat- ment with anhydrous aluminium chloride during the distillation, for the purpose of converting high boiling into low boiling oils, and he has found that with proper control of the vapors leaving the distilling system and entering the final condenser, and with sufficient time given the aluminium chloride, there is a complete transformation, and no matter how unsaturated the high boiling hydrocarbons may be, the low boiling oils produced from them are sweet smelling, water white and saturated. The reaction gives little gas, and only about the right amount of carbon is de- posited to allow for the production of saturated products. In carrying out the process in practice, the crude oil is first heated in the still to render it anhydrous, and also distil off any gasoline and kerosene that it may contain originally. In many of our crude oils, especially some of those from Texas, California, and Mexico, there is practically no gasoline and very little kerosene present. Anhydrous aluminium chloride is then added to the extent, ap- parently, of 5 to 74 per cent, the mixture is stirred and heated to boiling, usually at a temperature of about 500°F. The still is provided with large, air cooled, trap condensers which cause the condensation and the return to the boiler of the higher boiling products and of the aluminium chloride and its compounds. ‘When the vapors are allowed to pass into the final condenser at a temperature of about 300° F. the product is gasoline alone, which is ready for market when washed with an alkaline solution. The operation is carried on until the least valuable, so-called gas oil, portion of the oil has been converted into the low boiling distil- late, then the operation is stopped, and the more valuable higher boiling part is drawn off from the coky residue containing the aluminium chloride, in order that it may be worked up for lubricating oil and paraffine products. The latter products are 444 Scientific Intelligence. much improved by the aluminium chloride treatment, as they are thus saturated, and the resinous and asphaltic constituents are eliminated. The aluminium chloride is recovered for further use, best by distillation in an atmosphere of chlorine.—Jour. Indust. and Lng. Chem, vii, 737. H. L. W. 2. Arsenious Oxide as an Alkalimetric Standard.—Atan W. C. Menzixzs and F. N. McCarruy have found that arsenious oxide may be employed with accuracy, and without too complicated manipulation, as a primary standard in alkalimetry. A good quality of the oxide is purified by subliming once in a 6 X 1 inch test tube which has been drawn down to one-fourth its bore at a point about 2 inches from the closed end. The product is dried by heating and bottled hot. To prepare 500° of 1, N. solution a quantity of about 2°47 g. is weighed accurately in a 75°¢ conical flask. This is then treated with 5°° of pure, concentrated nitric acid followed by 5° of water. Solution and oxidation are effected by careful heating with a suitable trap attached to the mouth of the flask to avoid loss by spattering, and 5°° more of concentrated nitric acid are added. ‘The nitric acid is now removed by evaporation, best by the aid of a jet of cotton-filtered, ammonia-free air. ‘The heating can go as high as 230° C., but it is necessary to take up the residue in water and evaporate to dry- ness twice in order to remove all of the nitric acid. It is then only necessary to dilute to the required weight or volume of solution. To a measured quantity of about 30 or 40° of the solution are added phenol-phthalein as indicator and 3 or 4° of saturated barium chloride solution. The alkali is then run in until the amorpbous white precipitate, formed locally, is rather slow in re-dissolving, then the glass surface beneath the liquid is scratched until the lustrous, silky, crystalline precipitate of BaHAsO, begins to form, when the titration is completed in the usual way. Carbonate-free sodium hydroxide, containing barium hydroxide, may be used as the alkaline solution. The authors compared this method of standardization with several other accurate methods and obtained practically identical results in all cases.— Jour. Amer. Chem. Soc., xxxvii, 2021. + AER TSS Nie 38. A New Method for the Qualitative Separation and Detec- tion of Arsenic, Antimony and Tin.—F¥. L. Haun has proposed the following method, which he considers preferable in the hands of students to the methods usually employed. Starting with the mixture of the higher sulphides of the three elements mixed with much sulphur, this is extracted in the cold with a5 per cent solu- tion of Na,S. The sulphides go into solution very easily, while the sulphur remains behind. To the filtered liquid 10 per cent NaOH is added to the extent of about double the Na,S solution used. Then an excess of hydrogen peroxide is added and the mixture is boiled. In the presence of antimony a crystalline precipitate of Na,H,Sb,O, soon begins to separate, and the pre- cipitation of the antimony is made complete by cooling and add- ing about 4 the volume of alcohol. The precipitate is then Chemistry and Physics. 445 filtered off, the alcohol is boiled off, ammonium nitrate, best in the solid state, is added. This precipitates stannic hydroxide, and after the ammonia has been boiled off the precipitation is complete. From the filtered liquid NH,MgAsO, may now be precipitated as usual. It is stated that in this manner 1 mg. of arsenic, 1 mg. of antimony, and 2 mg. of tin can be detected with certainty in 1 g. of copper or lead.—Zettschr. anorgan. u. allgem. Chem, xcii, 168. H. L. W. 4, An Alleged Allotropic Form of Lead.—Hans HEtier has found that when thin strips of lead are placed for 3 days or more in a solution containing 1000 cc. of water, 400 g. of lead acetate, and 100 ec. of nitric acid of 1:16 s. g., the metal loses completely its original solidity and ductility and falls into small gray to gray- ish-black particles which can be rubbed to a powder between the fingers. It was found that the presence of a lead salt in solution was necessary to produce this change, and the conclusion was reached that the transformation is analogous to the well-known change of white tin into gray tin. It was found impossible, how- ever, to change ordinary lead into the gray product by inocula- tion with the latter, so that it is evident that the analogy to tin is not complete. Although it is stated that the change is not a chemical one but an actual transformation of lead into a new modification, this does not seem entirely: plausible, and it appears that a further study of the matter is desirable.—Zeztschr. physi- kal. Chem., \xxxix, 761. H. L. W. 5. Reflection of Gas Molecules.—Certain phenomena asso- ciated with the flow of gases have led to the assumptions that, on collision with a solid wall, the gas molecules are reflected at angles which are independent of the angle of incidence, and that the number of molecules reflected in any given direction is pro- portional to the cosine of the angle which this direction makes with the normal to the reflecting surface. An attempt to verify these hypotheses by direct experiment has been recently made by R. W. Woop. The first experiments related primarily to the production of a one-dimensional flow of gas. ‘The apparatus consisted essentially of a vertical glass tube with a small bulb at the lower end and a constriction just above the bulb. A short segment of the upper portion of the tube was bent through an angle of about 130° so as to form a convenient reservoir. In short, the apparatus had the general appearance of the ordinary form of cryophorus. A globule of mercury was first introduced into the bulb and then the apparatus was exhausted to as high a degree as possible. During the process of evacuation the tube was heated with a Bunsen flame to drive out the occluded gases from the glass walls. After the apparatus had cooled sufficiently the bulb alone was heated, thus causing the mercury to condense in the upper por- tion of the tube. The pumping was continued until all of the mercury had been driven out of the bulb and the lateral reservoir had been sealed by fusion. By simple manipulation the mercury 446 Scientific Intelligence. was collected as a droplet in this reservoir. Finally, the bulb and all of the straight, vertical segment of the tube were im- mersed in liquid air. At the expiration of twelve hours a faint circular deposit of mercury, having the same diameter as the lower constriction in the tube, was observed at the bottom of the bulb. By winding a heating-coil around the lateral reservoir and causing the mercury to distil at a temperature appreciably above that of the room, the deposit in the bulb was established in a minute or two. In the original paper is given a clear photo- graph of the mercury which had condensed on a glass plate placed in the bulb obliquely with respect to the long axis of the tube. The phenomenon is explained very clearly in the following sentences. ‘The mercury vapour enters the upper portion of the tube, and all of the molecules which are moving sideways are condensed on the wall, the deposit being very heavy at the top of the tube, and gradually thinning away to nothing a few cen- timetres below the surface of the liquid air. Below this point the molecules are moving all in the same direction, like bullets from a machine gun, and no further deposit is found until the con- stricted portion of the tube is reached. Here the molecular stream strikes the sloping walls of the constriction, and a heavy deposit of the metal occurs. Passing through the small opening, the gas traverses the exhausted bulb in the form of a jet which shows no tendency to spread out laterally, and deposits on the wall in the form of a small circular patch with very sharply defined edges.” In order to study the reflection of the beam of molecules a large drop of glass, at the extremity of a thin glass stem, was ground down and polished plane at an angle of about 45° with the axis of the stem. The reflecting surface was mounted con- centrically with the evacuated bulb by sealing the end of the stem to the inside of the lower, or drawn-out, portion of the bulb. The insulating stem and the surrounding vacuum enabled the reflecting surface to maintain its initial temperature, which was well above the point of condensation of mercury vapor, for an interval of time much longer than was necessary for the com- pletion of the observations. Under these conditions, the deposit began to appear in about three minutes after the heating-current had been started, and in ten minutes it had become entirely opaque to ordinary light. The photograph shows that the deposit covers the upper oblique half of the bulb with the excep- tion of a narrow circular zone just above the plane of the refiect- ing surface. ‘The results appear to show that the cosine law is approximately followed, for the density is greatest on the line of the normal, falling off gradually as the angle increases. At angles greater than about 80 degrees there appears to be no reflexion, at least in the case of the two surfaces which I have examined.” ‘The second surface referred to was a sheet of mica freshly split and mounted on a glass plate. The occurrence of the blank zone was also found to be independent of the size of Chemistry and Physics. 447 the bulb and of the angle between the reflecting surface and the axis of the incident beam of molecules. As yet no satisfactory explanation of the cause of the superior limit of the angles of reflection has been published.— Phil. Mag., xxx, p. 300, August, 1915. H. S. U. 6. Lhe Stark Effect for Solids.—The question as to whether excitation by canal-ray bombardment is a necessary condition for the electric resolution of spectral lines is suggested by the fact that in the single case where a different mode of excitation was employed, the Stark effect could not be detected. During the past year this problem has been attacked by C. E. MenpENHALL and R. W. Woop, and negative results were invariably obtained. This fact is of no little importance at the present time and hence a brief account of the experimental conditions may be of general interest. Most of the data were obtained photographically with grating spectrographs, and the electric fields were applied by means of a motor-driven Wimshurst machine. In order to have the spectral lines as narrow as possible the solids studied were usually main- tained at the temperature of liquid air. The strong, sharp fluor- escent line at X 5736, which was radiated by a certain specimen of Weardale fluorite, was the first to be examined. by RS. lear XXVII.—A Note on the Qualitative Detection and cepa ration of Platinum, Arsenic, Gold, Selenium, Tellurium and Molybdenum; by P. KE. Brownine XXVUI.—On Aventurine Feldspar; by O. ANDERSEN. (With Plates I-III) XXIX.—Anodic Potentials of Silver: II. Their Role in the Electrolytic Estimation of. the Halogens; by J. H. XXX.—Nephelite Syenites of Haliburton County, Ontario ; by W. G.. Foyvn 2.2252" ee ee XXXI.—Post-Glacial History of Boston; by H. W. Samer 437 SCIENTIFIC INTELLIGENCE. Chemistry and Physics—Improyement of High Boiling Petroleum Oils, A. M. McArsn, 443.—Arsenious Oxide as an Alkalimetric Standard, A. W. C. Menzies and F, N. McCartuy; A New Method for the Qualitative Sepa- ration and Detection of Arsenic, Antimony and Tin, F. L. Hany, 444,— An Alleged Allotropic Form of Lead, H. Hexuer: Reflection of Gas Molecules, R. W. Woop, 445.—The Stark Effect for Solids, C. E. Men- DENHALL and R. W. Woon, 447. Obituary—F. W. Putnam: J. H, Van Amrinece: K. E. Gurae; P, HaRiicH; J. VON Payer, 448, TEN-VOLUME INDEX. An extra number, completing volume XL and containing a full Index to volumes XXXI-XL of the Fourth Series will be ready in December. Sent only to those who specially order it, PRICE ONE DOLLAR. No free copies, ‘Library, U. S. Nat. Museum. one Eat 605, 13 a ioe ae 5 -- — ee VOL. XL: NOVEMBER, 1915. Established by BENJAMIN SILLIMAN in 1818. THE AMERICAN JOURNAL OF SCIENCE. Epiror: EDWARD S. DANA. ASSOCIATE EDITORS ) Proressors GEORGE L. GOODALE, JOHN TROWBRIDGE, W. G. FARLOW anp WM. M. DAVIS, or Camsruwnce, Proressors ADDISON E. VERRILL, HORACE L. WELLS, LOUIS VY. PIRSSON, HERBERT E. GREGORY anp HORACE S. UHLER, or New Haven, Proressor HENRY S. WILLIAMS, or Irnaca, Proressor JOSEPH S. AMES, or Battimore, Mr. J. S. DILLER, oF Wasutneron. FOURTH SERIES VOL. XL—[W HOLE NUMBER, CXC]. No. 239 NOVEMBER, 1915 Cainsonian Insite: “Naw a 916 || | NEW HAVEN, CONNEOTIOUI i y * Honal vi yet re i L915. —— THE TUTTLE, MOREHOUSE & TAYLOR CO., PRINTERS, 123 TEMPLE STREET. ne 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). New and Rare Minerals EUXENITE, near Tritriva, Antsirabe, Madagasear. Sharply developed ~ crystals; small and large. ‘75c¢ to $5.00. : AMPANGABEITE, nr, Tritriva, Antsirabe, Madagascar. Small and large crystals; new mineral. $1.00 to $8.00. BETAFITE, Betafe, Ambatofotsy and Antsirabe, Madagascar. ‘oc to $5.00. MANGANOSITE, Franklin Furnace, New Jersey. New discovery. $1.00 to $2.00. HODGKINSONITH, Franklin Furnace, New Jersey. $1.00 to $5.00. Just received a large collection of excellent specimens suitable for work- ing material. This is a fine opportunity to secure rare minerals at low figures. Kindly send for list. Prices range from 10c to $2.00. ; Important to Collectors If you want to sell your collection or some of your specimens, you will now have an excellent opportunity. We are making a specialty of selling collections on a fair commission basis. We have the best facilities and are always in contact with the greatest collectors, who are willing to pay a good price for choice minerals and gems, but they must be first class. The min- erals, your property, are just as safe with us as at your place. We are lo- cated in the financial district in a fire-proof building and have vaults for the valuable articles. Desire to state that the specimens and collections we have sold in the past always gave perfect satisfaction. We are always in the market for fine showy crystallized minerals and pay the highest prices for them, The minerals must be sent on approval and when not accepted will be returned promptly. Letters must accom- pany shipments, giving the lowest possible price on same, so as to save time and labor. , Are You Interested in Gems? We carry a full line of precious and semi-precious stones of the best quality in all sizes, both cut and in the rough, suitable for jewelry and specimens. ° Also the synthetic gems of the best quality, including rubies; pink, — yellow, blue and white sapphires. Scientific emeralds of beautiful rich color. : ; The above can be secured here at very low prices, much lower than elsewhere. § Selections sent on approval. ALBERT H. PETEREIT — 81-83 Fulton St., New York City THE \ AMERICAN JOURNAL OF SCIENCE [FOURTH SERIES.] —__¢4 9 ___—_ Art. XX XII.—Eixperimental Studies and Observations on ‘Ice Structure; by O. D. vow EncEtn. Introductory.—A detailed account of these experiments has been published under the joint authorship of the late R. S. Tarr and the present writer in the Zeitschrift fiir Gletscher- kunde (1915, vol. ix, pp. 81-139). In view of the restrictions imposed by the European situation, the fact that the Zeitschrift reaches only glacialists and that new points have come up since the writing of the longer article, it is deemed worth while to summarize the chief results obtained in an American periodical. Those specially interested in the subject are, how- ever, referred to the fuller statement which contains the records, data, and methods followed in individual experiments, references to earlier literature, and acknowledgment to those who gave assistance in various ways. The initiative in developing the work and some of the tentative conclusions are to be credited to the late Prof. R. 8. Tarr. The experiments furthermore comprise the continuation of the investigations begun by Professor Tarr in the winter of 1910-11. The earlier results were published under the title “The Properties of Ice—Experimental Studies,” by R. S. Tarr and J. L. Rich.* The experiments herein described were performed in the winter of 1911-12, when better facilities and apparatus were available. fecord of the Huaperiments.—The experiments were planned primarily to secure data bearing on the structure and flow of glacial ice. In the main they consisted of compression tests (a) on cubes of ice unsupported at the sides and (0) on cylinders of ice enclosed in metal cases of varying degrees of resistance up to complete rigidity under the pressures applied. The * Zeitschrift fiir Gletscherkunde, vi, pp. 225-249, 1912. Am. Jour, Sct.—FourtH SERizs, Vou. XL, No. 239.—NovemBER, 1915. 450 O. D. von Enyeln—Studies on Ice Structure. tests were made with screw presses designed and calibrated by Professors Kimball and Wells of the Sibley College of Mechanical Engineering, Cornell University. (See fig. 1.) Both pond and glacier ice were used, in cubes 4, 1, and 2 inches square and variously oriented with regard to the component erystal structure of the ice. It was found that the crushing Fie. 1. Fie. 1. Interior of Outdoor Laboratory. Shows large press with ice core projecting to right from lateral orifice of iron cylinder. Also shows small press (on table in background) with one of the small cubes between its plates. strength of pond ice with pressure applied in the direction of the principal axes of the component crystals (all similarly oriented) is about 1000 Ibs. per square inch. With pressure at right angles to the principal axes the crushing strength was approximately 350 Ibs. to the square inch, while that of O. D. von Engeln—Studies on Ice Structure. 451 glacier ice with variously oriented crystal grains was at some figure intermediate between 350 and 1000 lbs., and tended to the higher rather than the lower figure. These data were obtained with air temperatures between 18°F. and 20°F., but variations between 10°F. and 25°F. do not seem to have a notable effect on the strength of the ice. Great care was exer- cised to have the pressure faces as nearly parallel planes as possible and pressures were applied very slowly. In a number of instances increases of the load were made only after a number of hours interval ; in several cases a single test occupied from 16 to 20 hours time. The values of the crushing strength of ice here stated are notably higher than those obtained and approvingly cited by H. T. Barnes in recent papers.* The average values he quotes are 370 Ibs. per sq. inch, pressure parallel to the principal axes, 356 Ibs. pressure normal to the axes. The discrepancy in the first case is probably due to the following factors: (a) The ice used by Barnes was river ice and because of current action may not have had uniformly oriented crystals, especially as large size blocks, 7 inches square, were used in the tests; (0) the tests were made at air temperatures only a few degrees below freezing point and in some cases above that point ; (¢) the pressures were applied with comparative rapidity. The authors of the second paper cited use an average figure of 400 Ibs. to the square inch as the crushing strength of ice in calculating the linear thrust of ice sheets against dams. As such thrust would be exerted at right angles to the principal axes the value of the crushing strength used is probably amply high for engineering practice. It was repeatedly noted when working with cubes under pressure approaching the crushing point that cracking occurred when the pressure was relieved and that if the pressure release was at all rapid the ice actually broke apart. This indicates that ice yields elastically to pressures slowly and carefully applied and that the elastic rebound is not sufficiently rapid to enable the ice to withstand quick release of pressure, hence the differential stresses set up by the rapid release suffice to rupture the mass. On account of this elasticity it appears that there can not be any permanent deformation of ice until a certain minimum stress has been applied and the experiments showed that the yield-point is near the crushing strength of the ice. If, however, pressures just below the crushing strength of the ice are applied and continued for a sufficiently long time * Barnes, H. T.; The Crushing Strength of Ice, Trans. Roy. Soc. Canada, Third Series, vol. viii, pp. 19-22, 1914. Barnes, H. T., Hayward, J. W., and McLeod, N.: The "Expansive Force of Ice, ibid. , Pp. 29-49, 452 O. D. von Engeln—Studies on Ice Structure. the ice will yield by flow without breaking. In this way the ice can be permanently deformed. Moreover, once such flow is initiated it continues for some time despite a progressive diminution of pressure. This experiment was repeated a number of times using cubes of ice 4-inch square. Thus on January 16, 1912, air temperature 8°F., such a eube of pond ice was inserted in the press and a pressure of 260 lbs. applied in the direction of the principal axes. Nineteen hours later the cube had flattened to one and one-half its original area and pressure had fallen off to 230 Ibs. In the cases of four other similar cubes the pressures fell off to 223, 223, 230, 226 lbs., a strikingly narrow range. ‘The cubes were reduced to one half or one fourth their original thickness. Under the petrographic microscope it was apparent that the original erystal structure had been deformed, but the flattened ice was apparently as strong (coherent) as the original pieces, although the edges exhibited curved fractures parallel to the pressure direction. Glacier ice flowed and flattened similarly but only 185 lbs. of pressure could be initially applied to 4-inch eubes of such. ice and at-the end of flattening the pressure was about 159 lbs. As the air temperatures during these experi- ments were 20°F. and lower, it would seem that the yield and flow is not due to pressure melting and regelation unless this be conceived of as occurring between the particles of the erystal structure of the ice. Various kinds and forms of ice, also snow, were compressed in metal cylinders to determine (a) whether any change in erystalline structure could be brought about by pressure only, and if so of what nature, and (b) whether actual continuous flow of ice could be induced by pressure at temperatures sufliciently low to render pressure melting and regelation inoperative. Lead cylinders, open at both ends, were first used but these yielded at pressures of 720 lbs. per square inch and the ice within crushed as the cylinder sides bulged outward. In later trials copper cylinders 4 inches in diameter, 12 inches high with walls 1/16 inch thick were used. Into such a copper cylinder a rough-hewn plug of pond ice was inserted with component crystals parallel to the length of the cylinder, and water frozen in the space between the rough- hewn core and the metal walls, making a solid ice fillimg. To the ends of this cylinder a pressure of 500 Ibs. per square inch was applied, part of the load being borne by the metal edges of the casing. The air temperature when the pressure was applied was 17°F. and several hours later when the pressure was increased to 720 lbs. per square inch it had fallen to 15°F. On the next day at 9 A. M. the air temperature was 6°F. and no bulging of the cylinder or falling off in pressure was apparent. Accord- O. D. von Engeln—Studies on Ice Structure. 453 Fic. 2. Fic. 2. Ice compressed for 94 hours at air temperatures ranging between —4°F. and 20°F., under pressures up to 1400 lbs. persquare inch. Photo- graphed immediately after removal from copper cylinder. Note that there is no distortion from the cylindrical shape. Note also the etched outlines of the outcropping ends of the crystals. eae. ree 454 O. D. von Engeln—Studies on Ice Structure. ingly the pressure was raised to 750 Ibs. per square inch. The following night the temperature fell to a minimum of —4°F., and no change was apparent in the cylinder or pressure the next day. At four o’clock in the afternoon, air temperature 13°F., the pressure was raised to 1400 Ibs. per square inch, the upper limit of the calibrated scale of the press. At this pressure the cylinder was allowed to stand for 36 more hours, the air temperature meanwhile rising to 20°F. Then the cylinder was removed from the press. It had been under varying pressures for some 94 hours at air temperatures ranging between —4°F. and 20°F. On gently heating the metal cylinder it was found possible to slip the ice core out complete, indicating that there had been no bulging or distortion of the metal. (See fig. 2.) The compressed ice was found to be of crystal clearness and homogeneous, showing no line of separation to mark the junc- ture of the rough-hewn prism of pond ice and the water frozen around it. The most striking result, however, was the fact that the ice mass had been completely recrystallized. The original pond ice core was inserted with principal axes parallel to the pressure direction, the new crystals extended across the cylinder with their principal (and longer) axes at right angles to the pressure direction. On slow melting and evaporation in the laboratory the individual crystals 1/16 to 3/16 inches in diam- eter and averaging 3/4 inch to 1 inch in length were outlined by grooves along their contact planes. A further singular phe- nomenon was the extension of the crystals straight across the cylinder instead of radially inwards as might have been expected by analogy to the structure of cakes of can-frozen artificial ice. The boundaries of the crystals were irregularly polygonal, an occasional one nearly hexagonal, much like glacier grains in cross section but differing from these in that the crystals of the cylin- der were greatly elongated. (See figs. 3 and 4.) Under crossed nicols a section cut across one of the crystals exhibited parallel extinction, indicating that it comprised a single, uniform individual. Others showed wavy extinction in section, sug- gesting distortion, as is also true of glacier grains. In no ease, however, were any traces of brecciation apparent. The elon- gated form and wedge-shaped terminations in the artificial erys- tals may have owed their difference from the glacier grains to the fact that the conditions of our experiment permitted of no movement in the ice mass involved. In a later experiment this last deduction was in a sense veri- fied. An annealed, softer, copper cylinder was filled with ice fragments of pond ice, glacier ice and snow in layers, and made solid by filling up the spaces with water and freezing. Pres- O. D. von Enyeln—Studies on Ice Structure. 455 Ines, 3 Fic. 3. Same ice cylinder as shown in fig. 2 after reduction in size by slow melting and evaporation (mostly latter) in outdoor laboratory. Note outlines of crystals as shown by etching out of crystal boundaries into dis- tinct grooves. Natural size approximately 8 in. high by 4 in. wide, original cylinder 12 in. by 5 in. Fic. 4. Fie. 4. End view of ice cylinder shown in fig. 3 and photographed at the same time. Shows roughly parallel orientation of crystals, their relative lengths and wedge-shaped terminations. 456 O. D. von Engeln—Studies on Ice Structure. sure was applied to the top of ‘this composite fillmg by means of a solid iron plunger turned to a diameter that would just per- mit its easy insertion into the open end of the copper cylinder. This insured all the load being borne by the ice itself. (See fig. 5.) Pressures varying from 500 to 900 Ibs. per square inch were applied, over a period of five days. The cylinder bulged after the first few hours. As no notable change in the ice was apparent on removal from the cylinder after this treatment Fic. 5. BIB 7g DP Plunger ii ros Cal es ce a Broken Ice Pond | lee with principal axis of crystal bundles parallel to direction of pressure. i ut ) ee oe S——= i WAS tesa es Glacter Ice, granules jes variously oriented Pond Ice with principal axis of crystal bundles normal to direction of pressure. —— —————— ? ——— — Fic. 5. Diagram illustrating contents of annealed I CORDEE cylinder, using fitted iron plunger. except the absence of fracturing and apparent flow indicated by the transparency and coherence of the mass, the same ice core was utilized again. This time an unannealed harder cylin- der was probably used (our notes fail to state) and a solid ice mass assured by filling with water and freezing. Again using the iron plunger, pressures up to 1000 Ibs. per square inch were lied and allowed to remain on over night. Next day the load had fallen off to 800 Ibs. per square inch and the eylinder was very slightly deformed. The air temperatures during the time of the experiment ranged from 16° to 19° F. On removing the ice from this second cylinder it was found to be completely recrystallized. Ends of individual erystals had diameters 1/8 to 3/8 inches, some were smaller. The O. D. von Engeln—Studies on Ice Structure. 457 originally glacier ice was coarsest, the crushed top ice finest, but all parts showed the same granular structure, but not elon- gated crystals, and the grains were also variously oriented erystallographically. The appearance of this cylinder is shown in fig. 6. he the ice of this cylinder had a complicated history the experiment was repeated with an annealed cylinder filled with eracked pond ice,and snow. only, no water frozen in. This was compacted with pressures up to 1400 lbs. per square inch with much swelling of the cylinder. Then the plunger was removed and more filling of snow and cracked ice put in. This experiment was carried on for about 5 full days, air tem- peratures meanwhile ranging between 5° and 27° F. On remoyal from the cylinder the cracked ice layers were found to be clear and glassy, the snow layers glassy but clouded with innumerable air bubbles; the whole was a homogeneous ice mass. Qn melting down the characteristic, granular structure, obtained in previous experiments, was found to have developed in this also in all parts, but the grains were larger in the layers of cracked ice. Under crossed nicols extinction was found to be uniform within a given grain boundary. Similar results were obtained on compressing snow alone. It seems from these experiments that a granular ice can be developed from snow by pressure with accompanying move- ment at air temperatures eliminating the possibility of pressure melting and regelation. The pressures required to bring about flow and change of structure are also interesting, in that they correlate so nearly with the yield point pressures obtained in the unsupported cube experiments. It is to be remembered, however, that the pressure figures for most of the metal cylin- der experiments are by no means exact, inasmuch as the area to which pressure was applied must have been constantly changing as the cylinder swelled, and can hardly have been transmitted uniformly per unit of area throughout the mass during the progress of the experiment. In the second type of the compression experiments with metal cylinders it was planned to test the possibility of free and continuous flow of ice at temperatures sufficiently low to eliminate the development of pressure melting and regela- tion. A heavy iron cylinder bored out to 2 5/16 inches diam- eter, with a lateral orifice 3/4 inch in diameter, near the solid base was used. A solid steel plunger was turned to fit very accurately into the larger opening of the cylinder. It was pro- posed to force ice placed in the cylinder through the lateral orifice by applying pressure by means of the steel plunger. The arrangement of the experiment is clearly shown in fig. 1. On March 5th at an air temperature of 22° F. the cylinder 458 O. D. von Engeln—Studies on Ice Structure. Fie. 6. SNOW /0E | CRACKED /Ce ss Fic. 6. Ice cylinder of fig. 5 after second compression. Position and boundaries of different kinds of ice used in filling as marked. Note that lines of juncture of ditferent kinds of ice are nearly obliterated and that the crystal ends bridge them. Outline of the old top of ice from first compres- sion shown faintly under the snow ice, indicating transparency of latter. Granules shown to be largest in glacial ice. O. D. von Engeln—Studies on Ice Structure. 459 was filled with ice fragments and a pressure of approximately 3400 lbs. per square inch applied by 5 P. M. of that day. During the night the air temperature decreased to 4°F. Before 10:30 A. M. of the next day, when the air temperature was 14°F. 3 inches in length of solid ice core had been forced out of the lateral orifice at the base, and the pressure had fallen off to 3000 lbs. per square inch. (See fig. 7.) The projecting WIG ae Fic. 7. Flow of ice under compression. Photographed at 12.15 P. M. on March 6, 1912. Air temperatures ranged between 22°F. when pressure first applied, to 4°F. during night and 14°F. in morning when 38 inches of ice had been protruded. In this figure approximately 4 inches of the ‘‘ squeezed out” ice rod are shown. Streaks are due to iron substance from orifice and indicate the movement outward. Note downward bend of rod. Parallel rings near end are original, their cause not understood. ice rod curved downwards, its end being one-half inch below the level of the orifice. By 2:30 P. M. 4 1/2 inches were pro- jected. By 4:30 P. M. another inch had been squeezed out, the pressure had fallen off slightly and the air temperature had risen to 29°F., at approximately which point it stayed until 10 A. M. next morning when 14 inches of ice rod had been 460 O. D. von Engeln—Studies on Ice Structure. pushed forth. Two hours later the air temperature rose to the freezing point and the pressure had fallen off to 2000 lbs. per square inch. The air temperature continued to rise slowly, and by 2 P. M. water from pressure melting began to ooze out at the top of the cylinder between its bore and the steel plunger but no water came out of the lateral orifice with the ice rod, which had meanwhile increased another two inches in length, making 16 inches in all ; the load at the end being 1100 lbs. per square inch. That the water oozing out at the top was from pressure melting was indicated by the fact that it refroze in feathery crystals on the outside of the cylinder. Microscopic sections cut on March 6th from the part of the rod protruded at the lowest temperatures on examination proved the ice to be perfectly clear, glassy and compact. Under crossed nicols individual particles proved to be differently oriented but exhibited parallel extinction within a given boundary. The outer circumference was finer grained than at the core, shear lines and breccia bands could be identified but there were some apparently real crystal boundaries. ‘“Undoubted breccia bands and undoubted boundaries of breaking, but looks for all the world like a medium-grained granite rock, with a very few apparent crystal edges but not usually. The fact that this is clear compact ice shows beautiful flow.” (Tarr’s laboratory notes.) The downward bend of the rod may have been due to more rapid emergence at the top than the bottom of the orifice. If so, the differential movement may account at least in part for the breaking and shearing. However, the rod may have bent of its own weight or because of internal stresses. Clay squeezed through like orifices in clay-working machinery bends down similarly because of irregularities in the material. ~ Interpretation and Application of the Hxperiments.—Ice erystallizes in the hexagonal system. The habitus of the crystals and their orientation vary according to their origin. In the Jake ice the crystals occur in bundles of prisms with the basal planes of the mass parallel to the refrigerating surface. In glacier ice more or less nodular grains, each a erystal unit, form an interlocking aggregate with the principal axes of the individuals variously oriented, althongh some observers claim that there is a measurable tendency for the grains in the basal layers of glaciers to be disposed with their principal axes parallel to the direction of gravity. Water freezing in crevasses in glaciers has been noted to develop columnar ice crystals with principal axes extending at right angles from the opposing walls of the fissure. In artificial ice frozen in cans submerged in ammonia-cooled brine the crystals have their principal axes extending radially from the outer O. D. von Engeln—Studies on Ice Structure. 461 surface to the center of the cake. Emden™* has clearly shown that ice grains as large as hazel nuts may be grown from snow slush with originally small ice nuclei, kept in sealed vessels for several weeks at a constant temperature of 0°C. In this case growth proceeds from the center outward, the accretion and orientation of each crystal being determined by the orientation of the original nuclens. Accordingly the orientation of the grains in the mass is irregular like that of glacier grains. From these observations it seems safe to conclude that under normal atmospheric pressure the orientation of ice crystals is determined by the position of the original surface or center of refrigeration, consequently the orientation of glacier grains is determined by the position of the crystal axes in the original snow and névé grains. As originally shown by McConnel and later more fully formulated by Miigge and confirmed by the experiments of Tarr and Rich, an ice crystal is made up of molecularly thin, flexible laminze lying in a plane normal to the principal axes. Under pressure, with the ends of the principal axes supported, these lamin will glide over one another and this gliding prop-’ erty may account for such uniform orientation of ice crystals in the basal layers of glaciers as may exist. It can not be assumed that such uniformly oriented crystals develop and grow in the direction of least resistance, in ice masses under pressure, for similarly orientated mica flakes in slates and schists which are presumed to have grown under like conditions are distinctly tabular, whereas the uniformly oriented glacier erains are not described as having any greater basal diameter than adjacent ones that are irregularly oriented. On the other hand, it is quite possible that the presence of a number of similarly oriented crystals may be the result of shear along the gliding planes brought about by differential pressure in the general direction of the movement of the glacier. Re- growth of the two parts of an individual crystal completely separated by such shear would result in adjacent crystals with similar orientation. ‘This is in accord with Emden’s contention that the essentially progressive growth of glacier grains from the névé to the ends of glacier tongues results from the absorption of small granules by adjacent larger ones, and with Deeley’s assertion that the growth of glacier grains results from the transference of molecules from crystal to crystal under differential pressures in the ice. Water is at its maximum density at a temperature of 4°C. and between that temperature and the freezing point, 0°C., it *Hmden, Robt.: Uber das Gletscherkorn, Denkschriften d. schweiz. naturf, Ges., xxxiii, Ztivich, 1892. Separates published by Zircher and Furrer, Ztrich. 462 O. D. von Engeln—Studies on Ice Structure. has been suggested that the liquid phase consists of a mixture of ice molecules (H,O,) and water molecules (H,O,). Similarly it is quite possible that a residuum of water molecules exists in the solid phase, ice, at temperatures below, but near the freezing point. As the change from the liquid phase, water, to the solid phase, ice, is accompanied by expansion it follows that compression of the ice will result in the lowering of its melting point. Under uniform pressure this has been experimentally determined to amount to -00722°C. per atmos- phere of pressure. Johnston and Adams ~* contend that what is ordinarily termed the compression of solids is resolvable into uniform pressure and shearing stress. They further insist that permanent deformation of crystalline aggregates results from the effects of non-uniform pressure or shearing stress and that such deformation is occasioned by the real melting of those particles that at any instant bear the brunt of the load. On this basis they assert that non-uniform pressure lowers the melting point of ice ‘09°C. per atmosphere of pressure. Considered in connection with the possibility of a residuum of water molecules in ice at temperatures near the freezing point this concept of the nature of compression effects on solids suggests the probability of a considerable volume of water being evolved by moderate pressure on ice at or near the freez- ing point temperature and such an effect has been noted by Hess ¢ in exact experimental studies. Ice masses do not consist of single crystals but, as shown above, of aggregates of comparatively small units variously oriented. ‘The resistance of such masses to deforming stresses would appear to be due to three factors : (a) the nature of the inter- locking of the units, (2) the molecular cohesion of the ultimate particles of a unit crystal, (c) the molecular cohesion existing at the contact of adjacent crystal faces. It follows that experimental studies of ice under compression with reference to glacial phenomena should be directed toward ascertaining the behavior of ice as a erystalline aggregate rather than to the investigation of the physical properties of unit crystals. That there is a very considerable difference between the value of molecular cohesion of the particles of a unit ice erystal and that between particles of adjacent crystal faces is indicated by the differences in crushing strength exhibited by the pond ice put under compression (a) with principal axes of the component prisms parallel to the pressure direction and (b) normal to it, the crushing strength in the second case being * Johnston, J., and Adams, L. H.: On the Effect of High Pressures on in paver and Chemical Behavior of Solids, this Journal, xxxy, 200, Maren, . + Hess, H. : Uber die Plastizitat des Hises, Annalen der Physik, (4), xxxvi, pp. 449-492, 1911. O. D. von Engeln—Studies on Ice Structure. 463 only one-third that of the first. In the first case the load is borne by the columns of basal plates of the ice crystals, in the second it is transmitted through the mass along the adjacent parallel planes of the crystal prism faces. That the inter- locking, irregularly oriented, crystals of glacier ice supported loads of intermediate value lends further confirmation to this deduction. In this connection it is interesting to consider the experiment with the pond ice under compression in the unyielding cylinder which resulted in the complete recrystal- lization of the ice with principal axes normal to the pressure direction, involving a rotation of 90°. If it be assumed that films of water were produced by non-uniform pressure or shear melting, in the initial adjustment of the mass to the load, then, since ice is of less density than water, it would follow that the stable condition under pressure would be such that the load was carried by the water, and that the water be of minimum volume and maximum bearing surface. This condition seems to have demanded the reorientation of the ice crystals with principal axes normal to the pressure direction. Ice exists under natural conditions at temperatures near its melting point, hence change of phase due to increasing pressure or temperature is readily initiated and the velocity of transformation is quite high. Furthermore, as suggested in an earlier paragraph, residual water molecules may coexist with the polymerized ice molecules at temperatures near the freezing point. These relations would appear to have an important bearing on the plastic yield obtained in our experi- ments when cubes of pond ice with principal axes parallel to the pressure direction, and cubes of glacier ice, were subjected to pressures approximating the crushing strength, applied slowly. At the air temperatures at which the experiments were made these pressures were not great enough to induce melting by uniform pressure (-0072°O. per atm.) but would induce internal melting by non-uniform pressure, the effect of which is 12 times as great according to Johnston and Adams. As our bearing surfaces could not have been ideally plane-parallel, some portions of the ice must have borne a considerably higher load per unit of area than was registered by the scale of the press. Hence it seems quite probable that the yield achieved was due to internal liquefaction of the ice. The continuance of the yield, once started, with slightly diminished pressure and increased surface area would be ascribable to the fact that deformation of the cube would tend successively to exaggerate the localization of the whole load on certain small units of the bearing surface. Flow once started would also be facilitated by gliding movement between the basal plates of the ice crystals. On release of pressure the dissociated ice 464 O. D. von Engeln—Studies on Ice Structure. molecules (water) developed in the crystal structure would, owing to high reaction velocities, readily change back to the solid phase and thus preserve the ice clear and coherent throughout. The lower molecular cohesion of ice masses along the con- tact planes of adjacent crystals, as indicated by the lower crushing strength, suggests that such internal liquefaction as occurs under non-uniform pressure may be localized between the erystal units and also between the basal plates of the single erystals. Such a deduction is corroborated by other evidence, and this leads to some very interesting conclusions regarding the structure of ice aggregates. These can best be presented by first considering the premises on which they are based. Practically all natural waters carry considerable quantities of dissolved mineral salts. ‘It isto be expected, therefore, that such material will be to some extent incorporated in ice formed on ponds, lakes and rivers. In large part, these soluble min- eral salts in land waters are leached from rock substance. In the case of glacier ice, however, there is little or no opportun- ity for the incorporation of such material if it must be derived from rock leaching. But one of the commonest soluble salts, sodium chloride, in natural waters, is for the greater part not derived from rock leaching but is carried by the wind during storms as fine dust particles from the ocean and local salt deposits to points remote from its origin. The difficulty of accounting for the chlorine in inland waters and the well-known excess of chlorine in the ocean are the facts responsible for investigations that have shown, e. g., that in England at Ciren- cester an average annual deposit of 36-1 lbs. of sodium chloride has been for 26 years brought by the wind to each acre of sur- face. Hence there is no difficulty about the incorporation of a considerable quantity of this salt in glacial ice. Moreover, it is suggestive that the regions of present day notable glacia- tion are almost without exception favorably located for receiy- ing supplies of wind-borne salt, directly from the ocean, from. saline basins and from the surface of sea ice. An aqueous solution of sodium chloride will completely crystallize at its eutectic point, —22°C., forming what. has been termed ecryohydrate, a mixture of salt and ice crystals, each distinct, in the proportions 23°5 per cent NaCl and 76°5 per cent H,O. If the solution is originally more dilute than these percentages the excess of water crystallizes out progres- sively with fall of temperature until the eutectic concentration is reached at —22°C. As pointed out above, sodium chloride must be carried by the winds in appreciable quantities to the snowtields of glaciers. In the change from snow to granular névé a process akin to O. D. von Engeln—Studies on Ice Structure. 465 the freezing out of the excess of water in the dilute aqueous solution must take place and the salt content of the snow con- centrated in films between the névé granules. It the tempera- ture of the névé is at or below —22°C. these films will consist of the cryohydrate solid. But as it is unlikely that so low temper- atures as —22°C. exist far below the surface in any glacial masses it is much more probable that liquid films of salt solu- tion develop between the pure ice granules and that this salt solution is of a concentration, therefore a thickness, dependent on the temperature and the size of the pure ice granules. At comparatively high temperatures, i. e., near the freezing point of pure water, a comparatively dilute solution must remain in the liquid phase. Again, the smaller the ice grains the greater their total surface area, consequently the thinner the film of salt solution that surrounds each at a given temperature. Since glacier grains are generally observed to increase in size from the surface toward the interior and from the head to the terminus of an ice tongue and the temperature of the interior ice to approach the pressure-temperature melting point, it fol- lows that the interstitial films of salt solution will be progres- sively thicker toward the lower end of the glacier. Possibly, also, there is a tendency for the salt solution to drain away from the surface zone of the ice and to be concen- trated in the interior below the limiting depth of crevasses. Except then as the films of salt solution are held by capillarity the surface portions of the ice will be made up of ice grains that are very closely adjacent whereas there will be a corre- sponding concentration of the liquid films between the grains at greater depths. Other investigators have argued the presence of saline solu- tions between the pure ice crystals of glacier and pond ice- Buchanan* points out that rain and snow contain seven parts per million of chlorine and that when glacier ice has a temper- ature as high as —0-07°C. it will consist of 1 per cent liquid brine or water. To the presence of such films he ascribes the effect of the sun’s rays in disarticulating granular glacier ice and in developing the Forel lines that mark out the basal plates of the ice crystals. Thick pond ice similarly becomes “yotten” after a spring thaw and breaks up rapidly because it is disintegrated by melting in the planes of such films. Quincket characterizes glacier grains as ‘foam cells” filled with pure ice and separated one from another by walls of “oily” salt solution. The origin of the salt solution he ascribes *Buchanan, J. Y.: In and around the Morteratsch Glacier, Scot. Geog, Jour., XXviii, pp. 169-189, 1912 and other earlier papers. +Quincke, G.: The Formation of Ice and the Grained Structure of Gla- ciers, Nature, Ixxii, pp. 543-545, 1905. Am. Jour. Sct.—FourtH Smrizs, Vou. XL, No. 239.—Novemser, 1910. 466 O. D. von Engeln—Studies on Ice Structure. to the disintegration of rock debris incorporated in the ice. He accounts for the gliding planes of ice crystals, the development of Forel’s stripes and of capillary fissures in the ice, on the basis of the presence of such films of salt solution. Elsden* endorses the same idea, quoting Buchanan. On slowly melting and evaporating, in the experimental lab- oratory, at air temperatures near and below the freezing point, the ice crystals of the compressed ice cylinders were individu- ally marked out by well-defined grooves. A block of glacier ice (from the Illecillewaet Glacier, Canada) similarly preserved showed the same phenomenon. Like grooves commonly develop on the surface of cakes of artificial ice slowly melting in refrigerators and if such ice is split with a sharply pointed ice pick it will be found that the mass will commonly break up into prismatic crystals, the grooves marking the surface out- crop of the crystal boundaries. The fact that in ail these cases the grooves have a considerable width, from 1/32 inch to 1/16 inch, shows that the volume of the separating films is quite appreciable. The fact that these grooves develop on melting slowly in air temperatures near the freezing point indicates that the melting point of the interstitial material is lower than that of the bulk of the granules. It does not follow, however, that the width of the grooves is an index to the thickness of the interstitial films in the interior of the ice, for a very dilute salt solution would be very effective in developing a channel by melting at the surface in contact with air at temperatures above the normal melting point of pure ice. In the interior of the ice the film is probably much thinner and the solution more concentrated than in the outcropping grooves. Before the writer knew of the writings of Buchanan and Quincke his attention was called to the possibility of there being different solubilities at the centers and peripheral zones of glacier grains by phenomena attendant upon the melting of icebergs from Alaskan tidal glaciers in sea waters. On the surface of those portions of berg that had been melting below the sea-water surface a unique “ hammered silver” effect devel- oped. (See fig. 8.) The centers of each of the polygonal ice grains were hollowed out and these concave depressions met in ridges where the grains came in contact with their neigh- bors. The ice under such surfaces was solid, transparent, clear blue. When, however, such bergs melted in the air (after stranding at low tide) the surfaces became smooth, the ice lost its transparency, the grains became disarticulated, colored solu- tions could be filtered through the intergrain fissures and the surface of the grains became marked with an arboresceut sys- *Elsden, J. V.: Principles of Chemical Geology, London, 1910, p. 136. O. D. von Engeln—Studies on Ice Structure. 467 tem of grooves. Apparently the melting of the ice in air and sea water is diametrically opposite with respect to the relative rate of the process at the centers of the grains and their inter- erystal zones. The theory of the common ion seems best to fit the case of melting under sea water. In solutions so dilute that dissociation is approximately complete, the solubility of a given electrolyte will be lowered by the addition of a solution containing an ion common to it and the electrolyte, and con- Fre. 8. Fie. 8. Iceberg, stranded at low tide, with ‘‘ hammered silver” sur- face developed by melting in sea water. Note the polygonal granules hol- lowed at the center. Photographed Aug. 2, 1909, Yakutat Bay, Alaska. versely, the solubility of one salt may be increased by the presence of another salt not containing a common ion. The sea water contains dissociated Na and Cl ions, hence the solu- bility of the saline interstitial film is decreased while that of the pure, H,O, ice crystal centers is increased. The develop- ment of the “hammered silver” surface by such differential solubility probably requires the maintenance of favorable and rather delicate equilibrium of temperature in that the sea water needs to be at or near the freezing point of fresh water. 468 O. D. von Engeln—Studies on Ice Structure. Since the sea surface in the region (Yakutat Bay, Alaska) where these observations were made. is often quite completely covered with floating ice, this temperature relation is not impossible of attainment,—its existence was in fact confirmed by thermometric observations by the writer in 1909. Incident- ally, if the theory of the common ion does apply to the phe- nomenon, as here suggested, the ‘“thammered silver” berg surfaces afford a most striking and interesting natural example | of the action of a comparatively little known principle. The application of the experimental and other observations considered above to the structure and flow of glaciers has con- siderable significance. In the first place the position taken by the late Professor Tarr, that experimentation with ice to have a bearing on glacial phenomena must be done on ice aggregates and not on single crystals, seems to be borne out by the evi- dence indicating the presence of an interstitial film of saline solu- tion between ice grains. The crushing strength tests on the cubes indicate that at sufficiently low temperatures crevasses may extend approximately 2000 feet deep from the surface without closing by flow. Hence it is apparent why the upper zones of glaciers seem quite brittle. The development of crevasses to the maximum depth of 2000 feet would, however, necessitate lower temperatures than the evidence at hand shows to exist in the interior of glaciers. Hess and Bliimeke* bored holes to a maximum depth of 153 meters in the Hintereisferner glacier. They found that the temperature of the interior ice was only so much below the melting point, under atmospheric pressure, as would be brought about by the pressure of the superincumbent ice masses at an given depth. Moreover, the temperatures at different depths were very closely accordant with the caleulated depression of the melting point by a column of ice of that height. On the other hand J. Vallott observed a constant annual tempera- ture of —16°6° in the névé snow of the summit of Mont Blane at a depth of 10 to 13 meters below the surface. As the snow from the highest parts of the névé field is that which eventu- ally constitutes the bottom layers of a valley glacier, and as the deeper ice has progressively lower temperatures on account of the pressure relations, it would appear that the bottom of a glacier constitutes its coldest portion, also the thicker an ice tongue the lower its interior temperatures may be. But the factors that would need to be considered in an attempt to deduce the body temperatures of an ice tongue of large size are so various that a discussion of them would need to be long - *Hess, H.: Die Gletscher, pp. 151-153 and pp. 319-320. Later 200 meters depth was attained with same results. +Quoted by Hess, H.: Die Gletscher, p. 164. O. D. von Engeln—Studies on Ice Structure. 469 to be at all adequate. The average annual temperatures of the snow fields vary according to latitude and elevation. As snow is a poor conductor of heat it may be that quite low tempera- tures are imparted to the body of the glacier from its upper reservoir sources. Whether, during the course of its movement from the névé source to the end of the tongue, there is enough inflow of earth heat at the bottom of the ice to bring it to the pressure-temperature melting point would depend on the orig- inal degree of cold, the length of the glacier’s course and its rate of flow. In any event, earth-heat would tend to raise the temperature in some measure. Since below the névé line sufficient heat is supplied to melt some of the ice, it follows that the surface temperatures of a glacier in summer must be at the melting point. ‘The amount of surface melting is repre- sentative of the excess of heat supplied over that capable of being conducted from the surface, at the melting point, to the colder masses of the interior ice. Such conduction will con- tinue until the interior portions have the temperature of the pressure-determined melting point. But as the near-surface portions will acquire this temperature earlier than those at greater depths, the conduction of heat to the deeper interior parts will be progressively slower from the surface downward. In other words, an increasing proportion of the excess heat will be utilized in near-surface melting. Thus, while in small gla- ciers, like the Hintereisferner, pressure-temperature equilibrium may be established for a considerable thickness of the ice tongue, it does not follow that this is the case for all the mass of glaciers of great thickness. Yet it may be concluded that, while it is by no means certain that the bottom temperatures of great ice masses are at the pressure-temperature melting point, it is improbable that they exhibit any extreme degree of cold. Vallot found that the process of glacier-granule growth was distinctly in evidence in the névé of Mont Blanc at depths 10 to 13 meters below the surface (hence below the zone of sur- face water infiltration) where constant annual temperatures of —16.6°C. prevailed. Emden* has clearly demonstrated that ice granules as large as hazel nuts may be grown from origi- nally small, snow-slush nuclei, kept in sealed vessels for several weeks at a constant temperature of 0°C. This growth he contends is due to the absorption of small granules by adjacent larger ones with consequent molecular readjustment. Deeleyt also insists that the growth of glacier grains results from the transference of molecules from crystal to crystal under dif- ferential pressures. *EKmden, R.: Uber das Gletscherkorn, Denkschriften d. schweiz. naturf. eee al Zurich, 1892. Separates published by Ztircher and Furrer, uricn. +Deeley, R. M. and Fletcher, G.: Structure of Glacier Ice, Geol. Mag., Decade IV, ii, p. 155, 1895. 470 O. D. von Engeln—Studies on Tce Structure. Under these conditions of progressive granular growth it is readily perceivable how the segregation “of pure ice nuclei from the saline interstitial films takes place. Moreover, on the basis of the evidence stated above it seems unlikely that the body temperatures of glacier ice are very low. Hence, with only the development of comparatively slight pressures, through continued accumulation of snow above, it would appear that the growth of granules would be much facilitated and the flow of the ice initiated. That ice can be completely recrystal- lized under comparatively moderate pressures (1400 lbs. per sq. in. or less ; equivalent to an ice column about 3500 feet high) at air temperatures a number of degrees below the freezing point (—4°F. to 20°F.) was shown by the compression experi- ment with the unyielding copper cylinder. Moreover, the erys- tals developed were of quite large size showing that the velocity of the reaction is great; in other words, large crystals can grow in a short time. The permanent deformation by flow of the ice cubes under pressures approaching their crushing strength (650 to 1000 lbs. per sq. in., equivalent to an ice column 1600 to 2500 feet high) clearly demonstrates the ability of ice to yield plastically under differential pressure without disintegra- tion and at temperatures a number of degrees below the pres- sure-melting point. Such yield may be in part due to shear, but in accordance with the known properties of the ice crystals can be more certainly ascribed to internal liquefaction in the planes of the basal gliding plates. The cause of this readier yield or gliding in the basal planes in turn may find its explanation in the presence of a residuum of saline material between the basal plates of the crystal as argued by Buchanan and Quincke. Another possibility is that it is due to an increase, under pres- sure, in the volume of water molecules (H,O,) included with the polymerized ice molecules (H,O,) in the solid phase, ice, at temperatures near the transition point, with consequent lessened viscosity or internal friction between the crystalline particles. Whatever the explanation, it is clear that such yield would account for the continuous deformation of even quite cold ice at great depths below the surface in the bottom zones of large glaciers. That the lubricating effect of the saline interstitial films, increased in volume under pressure, and of the plastic internal yield described above, amply suffice to promote the free flow of ice without loss of coherence, was proved by the experiment in which the ice rod was caused to emerge from the lateral orifice of the rod cylinder to a total length “of 16 inches. Moreover, the fact, that, while at the beginning of the flow the temperatures were low, pressure relatively high and the emergence slow, in its later progress when pressures were O. D. von Engeln—Studies on Ice Structure. 471 much lower and temperatures higher the flow was much more rapid, indicates that with temperatures near the pressure melting point the yield of the ice is much easier and also, as shown by the experiments with the cubes, that flow once initiated continues under diminished pressure. This latter principle seems to be an important factor in that, applied to glaciers, it explains the rapid waves of advance accompanied by a swelling of the glacier tongues that have been noted in Alaska* and elsewhere. A conception of the flow of glaciers as developed from the results of the experiments and consideration of the constitution and physical properties of glacier ice does not permit of characterization as either plastic or viscous. It is a plastic flow in the sense that the ice mass as a whole is permanently deformed by the movement, but its component grains are not subject or capable of such plastic deformation except in one direction and this appears to be a minor factor. Much more important is the movement that seems to be conditioned by the presence of the interstitial film of low freezing liquid, which may be characterized as viscous movement, by analogy like that of a stiff cement-concrete mixture. From this point of view the glacier grains must be regarded as the rock-fragment units of the concrete mixture and the interstitial film as the liquid cement or lubricating substance diminishing the inter-unit, hence internal friction between them, thus facilitating movement analogous to that between the molecules of a viscous solid. It may be questioned whether the volume of the interstitial film can be sufficiently great to permit of the degree of articulation between the ice grams required for continuous flow without involving the distortion or destruction of the crystalline units, hence preventing their progressive growth in size from the upper to the lower portions of an ice tongue. In this connection a calculation by Chamberlint is interesting. He figures that in the interlocking, granular portion of a glacier six miles long, with a movement of three feet per day, an individual granule would need to move the length of its own diameter, with reference to its neighbor, only once in thirty years. On this basis it would appear that only a very slight thickness of interstitial film in the liquid state would be required to permit of all necessary readjustments between the crystal- line grains, while keeping their interlocking structure intact and permitting their growth in size. Where differential pressures exist and an accelerated local flow is demanded this will be *Tarr, R. S.: Recent Advance of Glaciersin the Yakutat Bay Region, Alaska, Bull. Geol. Soc. Amer., xviii, pp. 257-286. + Chamberlin, T. C.: A Contribution to the Theory of Glacial Motion. Decennial Publications of the Univ. of Chicago, Chicago. First series, vol. ix, p. 201, 1904. 472 O. D. von Engeln—Studies on Tee Structure. facilitated by an accompanying increase in the volume of the interstitial film in that area. Summarizing, the phenomena of a valley glacier may be conceived as follows. In the reservoir area the increasing depth of accumulating snow gradually augments the pressure on its bottom layers. In the mass, meanwhile, the growth of ice crystals proceeds by molecular transference of particles and the absorption of small grains by adjacent larger nuclei. At the same time the saline matter is being segregated in the intergranular spaces. Compression itself involves a rise in temperature and such rise is further increased by inflow of surface heat and heat from the earth. When pressure and heat attain sufficiently high points flow is initiated. During the period of snow accumulation the snow and ice mass in the reservoir is under practically uniform pressure. Once flow is started non-uniform pressures are introduced with resulting increases and decreases of the thickness of the interstitial films. The continued movement of the ice tongue so started results both from the pressure due to its own thickness and thrust from the continuation of the tongue up slope. If the snow supply in the reservoir is cut off an equilibrinm of pressures will be established and the glacier will then melt away without further flow. If the snow supply is insufficient to provide great enough accumulations for continuous flow pressures, there will be alternative waves of advance and stagnation. Appar- ently the latter condition is only attained under exceptional conditions because flow is more readily maintained than initiated. ; The surface portions of the ice tongue are relatively rigid and brittle. The viscous (?) underflow gives rise to tensional and shearing stresses in the upper, rigid ice layers with resultant crevassing. These crevasses are commonly developed on a greater scale in the upper ice tongue, partly because slopes are steeper, partly also, probably, because the ice of the upper portions of the ice tongues is colder. In the lower expanded portions of ice tongues pressure-temperature equilibrium has been established throughout the thickness of the ice, hence flow is possible over lower slopes and with thinner ice than at and above the névé line. Near-surface melting in the lower parts of the tongue is localized in the intergranular spaces, the resulting water tends to drain away and thus leaves the ice grains “loose in the socket.” On the bottom of the glacier, near the very front, the establishment of pressure-temperature equilibrium combined with the continued inflow of earth-heat and friction may result in pressure melting and thus give rise to submarginal glacial streams, whose volume is augmented by the surface-melting water percolating through the frontal O. D. von Engeln—Studies on Ice Structure. 473 portions of the crumbling ice tongue or finding its way to the bottom through crevasses and moulins. In the middle portions of the flowing ice tongue three separate zones may be conceived as existing from the surface tothe bottom. At the topis the brittle, crevassed mass, below it the viscously(?) flowing ice, and under that a colder ice layer, continuing to the bottom, which is being plastically (?) deformed by reason of the pressure of the overlying ice and the frictional pull the viscous intermediate layer is exerting on its upper surface. On this basis a decreasing rate of flow from the surface to the bottom of a glacier could be accounted for, and a greater erosive power could be ascribed to the relatively more rigid, colder bottom layers. Future Hxperimentation.—It would be very interesting to freeze a number of blocks of ice from saline solutions of progressively increasing density and slowly melt them, noting the relative development of the inter-crystal grooves. Where glacier ice is accessible the relative concentration of saline material in the granules and in the interstitial spaces might be tested by picking out a number of granules, washing them with distilled water and then analyzing them for salt content in comparison with a coherent mass of glacier ice. It wonld also be worth while to attempt the compaction of snow into ice under long-continued pressure and cause this to flow from a lateral orifice. This experiment the writer hopes to attempt in a succeeding winter. Geological Department, Physiography Laboratory, Cornell University. 474 E. 0. Case—Dimetrodon incisivus. Art. XX XIII.—A Mounted Specimen of Dimetrodon incisivus Cope, in the University of Michigan ; by E. C. Cass. Iy 1912 the author discovered the Brier Creek Bone Bed in Archer County, Texas. This bone bed in the Wichita Forma- tion of the Permo-Carboniferous deposits of northern Texas contained, by far, more bones than any known accumulation of the same age. Over 1500 specimens of separate bones were recovered, and many times that number would have been taken Hie. 1. Fic. 1. Photograph of the mount of Dimetrodon incisivus Cope. x “05d. had not a large proportion of the material been ruined by the infiltration of iron-bearing waters which rotted the bones and cemented them together in an inseparable mass. A complete description of this bed and of the new forms dis- covered has been given by the author in Publication No. 207, Carnegie Institution of Washington, 1915. Though most of the bones were isolated, due to the complete maceration of the bodies and the later movements of the material in the swamp or muck-hole in which they were imbedded, some partially complete skeletons were recovered, among them the nearly complete skeleton of an Zdaphosaurus cruciger Cope, which has been partially cleaned and will be mounted in the Uni- versity of Michigan next year. E. C0. Case—Dimetrodon incisivus. A475 From the very abundant material sufficient bones of the right size have been selected to mount a nearly complete skeleton of Dimetrodon incisivus Cope. The work of mounting this speci- men necessitated a careful study of the possible attitudes and movements of the animal which is reflected in the completed. work. The skull in the mount is restored in plaster, except the anterior portion of the lower jaws, from fragments in the collection and from careful studies upon the perfect specimen of a skull of this species in the American Museum of Natural History in New York and upon a very perfect specimen of Fie. 2. Fie. 2, Photograph of the restored skull, all plaster except the anterior portion of the lower jaw, with the teeth. x ‘13. the skull of Dimetredon gigas in the possession of the author. The notably large size of the skull led to a very critical study of the proportions to check the apparent incongruity and the author feels sure that the proportions shown in the skeleton are correct. As mounted, the skull is raised slightly from the ground but is still inclined somewhat downward in an attitude of partial repose. From the position and character of the cer- vical vertebre it is apparent that the head could not have been raised very much higher, except as it was violently forced upward in a moment of anger or attack. The body is placed in an attitude of repose; the posterior end of the interclavicle rests upon the ground and the symphy- sis of the pubes is raised only sufficiently to allow for the car- tilages which were present during life. In this attitude the 476 E. 0. Case—Dimetrodon Rogie. spines were in the most relaxed position, and they have been mounted in the most regular position and arrangement per- mitted by the condition of the spines and the vertebre to which they are attached. There can be little doubt that the © position of the spines varied in each individual and were not infrequently distorted by injuries, so the position given in the mount is probably as correct as may be. Notable is the sudden increase in height of the spines in the cervical region and the almost equally sudden decrease in height in the sacral and anterior caudal region. Fie. 4. Fie. 3. Photograph of the skull of Dimetrodon gigas Cope. eal Fic. 4. Photograph of the fore limb and shoulder girdle. x 16. The shoulder girdle is mounted nearly complete ; the inter- clavicle, the clavicles, the scapula of the left side complete, and the lower part of the scapula of the left side all being in plain view and as free as their support permits. Though a slight distortion of the bones makes the exact outline of the shoulder- girdle somewhat uncertain, the author believes that the depar- ture from the normal curvature is very slight. No cleithrum has been placed in the mounted skeleton as no bones which could be referred to this element was found in the abundant material. The humerus is placed in an almost exactly horizontal posi- tion, extending directly outward from the cotylus. It is evi- dent from the position of the scapula and its cotylus that the E.. 0. Case—Dimetrodon incisivus. ATT Fig. 5. Fic. 5. Photograph of the posterior limb and pelvic girdle. The line of the back Fic. 6. Restoration of Dimetrodon incisivus Cope. is alittle too curved. x °05 about. 478 E. C. Case—Dimetrodon incisivus. motion of the humerus was closely restricted to a plane parallel with the ground. The pelvis is complete and mounted nearly free, being attached only by the outer side of the right half: this leaves the whole sacral series of vertebrae open to inspection. The posterior limb is placed with the femur pointing slightly upward and forward; this is rather the conventional than the usual resting attitude of the creeping reptiles, but is not an impossible or even au unusual attitude. The carpus and tarsus are modeled in plaster; the first after the well-preserved carpus discovered by the author some years ago and now in the possession of the University of Chicago, and the tarsus, except for the astragalus, caleaneum and centrale 2, after the primitive type of the reptilian carpus. Elements of both carpus and tarsus are not lacking in the col- lection but were not included in the mount because of the impossibility of assembling sufficient bones of the correct size. The ribs are all restored in plaster after well-authenticated specimens. The tail has been made relatively short in consonance with _ the ideas of the author, but as no specimen has yet been dis- covered in which the tail is complete the exact number of ver- tebree is uncertain. The completion of this mount is the first attempt to assem- ble all parts of the skeleton of Dimetrodon ina natural posture. Its accomplishment is largely due to the patience and skill of Dr. E. L. Troxell, who, as preparator, codperated with the author in assembling and placing the bones. LE. L. Troxell—Fossil Ruminant from Texas. 479 Arr. XXXIV.—A Fossil Ruminant from Rock Creek, Texas, Preptoceras mayfieldi sp. nov.; by Epwarp L. Troxetr. Amone the very abundant fossil remains found at Rock Creek there was the skull of a ruminant, and possibly belong- ing to it were an atlas and some rib fragments. Associated with these bones were parts of Hlephas, Mylodon, Auchenia and Hguus, representing the fauna of the early Pleistocene. The skull resembles somewhat that of the ox, especially in the general form and position of the horns, which come out in Fre. 1. Fic. 1. Posterior view of skull of Preptoceras mayfieldi, sp. nov. Holotype, Cat. No. 10920, Yale Museuin (x 25). the plane of the face, trending upward, then downward and forward. At first it was taken to be the skull of a sheep, but it is found to be very different. It is about one-half larger than a skull of Ovis rockymontanuws and the horns, which are not so large, do not curve backward nor are they set close together. It is probably allied to a specimen from New Mexico, viz: Liops zuniensis Gidley.* But of all known skulls it resembles most that of Preptoceras sinclairi Furlong + from the eaves of California. It is, therefore, put under that genus. The new species is named in honor of Mr. Gidley Mayfield, on whose ranch the specimen was found. This is located * Proce. U. S. Nat. Mus., vol, xxx, 1906, pp. 165-167. + Univ. of Cal., Bull. Dept. Geol., vol. iv, 1905, pp. 163-169. 480 E. L. Trowvell—Fossil Ruminant from Texas. about a mile from the famous quarry where ten skeletons of Equus scotti have been unearthed. The horns of the new species arise from the frontal bones at Fic. 2. Fic. 2. Anterior view of skull of Preptoceras mayfieldi, sp. noy. Holo- type, Cat. No. 10920, Yale Museum (x °29). an angle of about 90° to each other. The cores are slightly flattened antero-posteriorly and at the bases have burrs. Fie. 3. Fie. 3. Posterior view of skull of the type specimen of Preptoceras sinclairi, after Furlong (x ‘2). The frontal bones form the crown and join the parietal well back on the head ; the suture lies in general about one centi- E. L. Troxell—Fossil Ruminant from Texas. 481 meter above the lambdoid crest. The parietal is not paired but, as in many Artiodactyla, was probably fused early in lite. The frontal and squamosal form a deep groove on the side*of the head, parallel to the suture, continuing into the orbit anteriorly and ending posteriorly at the ridge which lies along the side of the skul]. The outer extremity of the lambdoid erest continues into this sharp ridge, which marks the union of the squamosal, parietal and occipital portion of the skull later- ally, giving this region a flat appearance. The crest loops down- ward in its middle portion to join the central tubercle which formed the attachment for the nuchz ligamentum—the chord which follows along the back of the neck. Two fosse are observed on either side of this tubercle. They are shallow, however, since the lambdoid crest does not markedly over- hang the occiput. The lambdoid suture below the crest is obscured. The skull is very deep from a point between the horns to the occipital condyles and in general is very stout. The anterior portion is broken off just above the orbits. A slight rise in the frontals anteriorly suggests that the animal hada rounded forehead. The foramen magnum for the most part opens on the dorsal surface of the occiput. Dimensions. mm. tate COTM) D2 ne ee en 15 Diam. of horn cores ant.-post. .....--.----.------ 74 SOS aS te CLEANS VCLSCina po ae cine Se Se 85 Breadth of forehead, anterior to horns __---------- 144 Breadth of cranium, posterior to horns .--._---.-- 92 Depth from point between horns to extremity of Geeinmalecond ye. 222... 2 a2 o 23 - 167 Wadtmeotebasi-occipitales..—.........-.--.------- 55 Grauro-facialanole; about 0.4052... 0.6L LS. 45° Occiput-tacialtamgle, about 222-2 .2.2------!-~..-.- 683° The specimen is placed in the genus Preptoceras because of the very great general similarity to the type. It differs from Preptoceras sinclairi in that its fossee just beneath the lambdoid crest are not so deep; the lambdoid suture lies near this crest (this is like Hwceratherium); the horns are smaller, less rounded dorsally and more widely separated ; they do not come out “from the extreme posterior and lateral ends of the frontals”’ as in the type species but are situated three centi- meters from the posterior and four from the lateral borders. The present species differs very greatly from Lops zwniensis. The horns of the latter are set wide apart,-come out straight from the skull and droop decidedly ; they have no burrs on the horns, no true lambdoid crest and the skull generally is smooth in the extreme. Am. Jour. Sci.—Fourta SERIES, Vou. XL, No. 239.—NovemseEr, 1910. 32 482 E. L. Troxell—Fossil Ruminant from Texas. Ovis is frequently mentioned in the literature, but no authentic record of a true fossil sheep can be found earlier than what might well be considered Recent. The atlas found associated with the skull aud corresponding in size has some distinctly camel-like characters. The trans- verse foramen follows the wall of the lateral process and opens on the edge near the axial articulation. This articular surface is broad and flat and a wide strip extends beneath the neural canal. The atlas is over one third broader than that of a full grown camel, indicating a powerful neck. The animal must have been quite strong, for the skull also is heavy and the horns are large. The occiput-facial angle as nearly as can be determined is 684°, showing the great posterior extension of the erest or the low position of the condyles. The facio-cranial angle is, ronghly measured, 45°. Such a great deflection of the face is characteristic of the sheep and is indicative of a grazing adaptation. The angle increases with age but in the adult sheep goes beyond 50°. From the sutures, which are not closely knit, and from this angle the animal is considered as having almost reached maturity. University of Michigan, Art. XX XV.— The Separation and Estimation of Aluminium and Beryllium by the Use of Acetyl Chloride in Acetone ; by H. D. Minnie. [Contributions from the Kent Chemical Laboratory of Yale Uniy.—cclxxii. | In a former paper* from this laboratory a method was de- scribed for the quantitative separation of aluminium from iron by the use of acetyl chloride in acetone. The application of this method depends on the fact that from concentrated aque- ous solution of the two chlorides aluminium is precipitated completely as the hydrous chloride, while the iron remains in solution. This precipitation is brought about, doubtless, by the decomposition of the acetyl chloride and simultaneous formation of hydrochloric acid which, as is well known, is a precipitant for aluminium. At the same time the precipitating mixture furnishes organic liquids in which the precipitated hydrous chloride of aluminium remains insoluble and the chlo- ride of iron soluble. The main function of the acetone is simply to abate the violence of the reaction which ensues when acetyl chloride reacts with water. * This Journal [4], xxxix, 197-200, Feb. 1915. H. D. Minnig—Aluminium and Beryllium. 483 Commercial acetyl chloride can not be used in this work because of phosphorus compounds used in its manufacture, the last of which cannot be removed. This phosphorus con- ‘tamination shows itself not only in the filtrate from the alu- minium precipitate, but also in the aluminium precipitate itself, giving high results. The acetyl chloride was therefore pre- pared as described in the former article.* Acetic anhydride was distilled from a little anhydrous sodium acetate to remove contaminating phosphorus compounds. Tests for included phosphorus compounds were made by hydrolyzing the acetic anhydride with water, evaporating down with nitric acid to oxidize any trivalent phosphorus which might be present from the manufacture of the commercial product, and treating the solution with ammonium molybdate in nitric acid solution. The purified acetic anhydride was then saturated with hydro- gen chloride and the product distilled at 100° C. in a rapid current of the same gas. Redistillation gives pure acetyl] chloride. The method of procedure was the same as in the separation of iron from aluminium. Measured portions of solutions of the two chlorides in small beakers were evaporated to the smallest possible volume on the steam bath. In case the evap- oration proceeds to dryness it is well in dissolving the salts to add a drop of hydrochloric acid to prevent the formation of basic salts, particularly of basic beryllium chloride. The beaker containing the concentrated solution was then placed in a dish of cold water and the acetone-acetyl chloride mixture (4-1) was added drop by drop from a dropping funnel to the complete precipitation of the hydrous chloride of aluminium. Fifteen to twenty enbic centimeters of the*mixture usually sufficed. The settling of the crystalline precipitate is a good indication, though not an infallible one, of the complete pre- cipitation of the aluminium chloride. When the precipitation was judged to be complete, the precipitate was transferred to a weighed perforated platinum crucible and carefully washed with the precipitating mixture. The filtrate and washings were caught in a beaker under a bell jar. The precipitate was dried slowly, at first high above a low Bunsen flame and then gradually lowered until the full heat of the burner was obtained. The hydrous chloride on ignition leaves the oxide. The ace- tone-acetyl chloride solution of beryllium was copiously but cautiously diluted with water to avoid spattering, and the resulting solution in a beaker covered by a watch glass, was warmed gently on the steam bath until the volume seemed to remain constant, indicating that the easily volatile acetone had been removed. The solution was then boiled and ammonium * Loe. cit. 484 2B. D. Minnig—Aluminium and Beryllium. hydroxide added to alkaline reaction. The precipitate was allowed to settle, filtered, and treated as usual in determina- tions where beryllium is weighed as the oxide. Results of the experiments are given below. SEPARATION BY ONE PRECIPITATION. : Al,O3 INGO ey 1350) BeO BeO Al.O3 Taken as AIC]; Found Error Taken as BeCl, Found Error grm. grm, grm. erm. grim, grm. 0°0927 0°0933 +0°0006 0°0638 eee 00927 0°09384 -+0°0007 006335 =a at 0:0927 0°0943 +0°0016 00638 ° 0°0927 0°0934 +0°0007 0:0638 42s eee 0°0927 0°0931 +0°0004 0:0638) ea oe ee 0:0865 0°0878 +0°0013 00561 0°0549 —0-0012 High results in the foregoing table seemed to indicate that beryllium chloride was being included in the aluminium precipi- tate, a circumstance which does not seem surprising in view of the marked similarity between the reactions of the two elements. A double precipitation of the aluminium was there- fore tried. After the first precipitation the liquid was de- eanted through the filter, retaining as much as possible of the precipitate in the beaker. This precipitate in the beaker was washed four or five times with the precipitating mixture and finally dissolved in a small quantity of water. The excess of water was evaporated off and the process of precipitation repeated. The two filtrates were united and the beryllium determined as before. SEPARATION BY Two PRECIPITATIONS. Al.Os Al.Os Al,O; BeO BeO BeO Taken as AICI; Found Error Takenas BeCl,; Found Error erm. grm. grm. grm. grm. grm. 0°0865 00869 +0-0004 0-056) |... eee 0°0865 0°0868 + 0°0003 00561. 2S eee O°0.865) tin woe age eae 0°0561 0°0564 +0:0003 0°0865 00867 +0°0002' 0°0561 0°0575 +0°0014 0°0865 0:0867 +0°0002 0°0561 0°0561 0:0000 0:0865 0:0865 0°0000 0°0561 0°0560 —0-0001 0:0865 00870 +0°0005 0°0561 0:0554 —0:0007 0°0865 00866 +0°0001 0°0561 0°0557 —0-0004 0:0865 0'0881 +0°0016 0°1122 O'l111 —0°0011 0°0865 0°0884 -+0°0019 0°1122 0°1093 —0:0029 0°0865 00869 +0°0004 0°1122 071118 —0°0004 0:0865 00873 +0°0008 0°1122 0'1108 —0°0014 0:1720 01739 +0°0009 0°0561 0°0548 —9:0013 0'1730 01741 +0°0011 0°0561 0°0554 —0:0007 H. D. Minnig—Aluminium and Beryllium. 485 That there is danger of inclusion of beryllium chloride could readily be seen when the quantity of that substance was increased. Unless the progress of the precipitation was watched and the addition of the precipitating mixture stopped as soon as the separation of the crystalline aluminium chloride seemed to be complete, beryllium chloride could also be seen to separate out.* Its appearance is very unlike that of the aluminium chloride and can easily be distinguished. Even when the addition of the precipitating mixture was stopped before the formation of this salt became noticeable in the beaker, it appeared as soon as the supernatant liquid had been poured through the filter and an attempt was made to wash the precipitate. Increase in the quantity of aluminium chloride seems also to have a like effect because of the neces- sity for an increased volume of the precipitating mixture. The comparative insolubility of beryllium chloride in ace- tone acetyl chloride (4-1) limits this process, therefore, to sepa- ration of quantities of the two elements present as the chlorides which do not exceed the equivalent of 0°15 grm. of the oxides. Of this amount beryllium oxide should not greatly exceed one- third. The use of acetyl chloride as a substitute for hydrogen chlo- ride simplifies very much the method of Havenst and therefore the favorable criticism of Noyes, Bray and Spear{ ought to apply to this method as well as to that of Havens. These inves- tigators found the method of Havens to be very reliable even when the amount of aluminium present was exceedingly small. *Probably the same compound mentioned on page 24 of Parson’s book, “The Chemistry and Literature of Beryllium.”” That precipitate was obtained by the use of ether saturated with hydrogen chloride. + This Journal (4), iv, 111. {A System of Qualitative Analysis for the Common Hlements, III, 19. 486 C. Barus—Interferences of Crossed Spectra. Art. XX XVI.—On the Interferences of Crossed Spectra and on Trains of Beating Light Waves; by C. Barus.* 1. Lntroductory.—lf two component spectra from the same source coincide throughout their extent, the elliptic interferences will be spread over the whole surface, provided of course the respective glass and air path differences of the component rays are not too great to interfere with visibility. In the usual method of producing these interferences, where the correspond- ing reflections and transmissions of the two component rays take place at the same points of the same plane surface, the interference pattern is automatically centered, or nearly so. This is not the case when, as in the following experiments, the interfering beams are produced in some other way; and the problem of centering is one of the chief difficulties involved. Butif the four beams are to be treated independ- ently, it is difficult to obviate this annoyance. Suppose now that one of the spectra is rotated around an axis normal to both, by a small angle. Will the interferences at once vanish, or is there a limiting angle below which this is not the case? In other words, in how far ean one trench with light waves upon the case of musical beats or of interferences not quite of the same wave length? Instead of approaching the question in this form, in which it would be exceedingly ditticult experimentally, I have divided it into two component parts. Let one of the spectra be rotated 180° degrees around a longitudinal axis, parallel to the red-violet length of the spectrum and normal to the Fraunhofer lines. In such a case, interference should be possible only along the infinitely thin longitudinal axis of rotation to which both spectra are symmetrical, one being the mirror image of the other. One would not expect these interferences to be visible. It is rather surprising, however, that this phenomenon (as I have found) may actually be observed along a definite longitudinal band in the spectrum, about twice the angular width of the distance between the sodium lines and symmetrical with respect to the axis of rotation. It is independent of the width of the slit, provided this is narrow. Again let one spectrum be rotated 180° about a given Fraunhofer line (transverse axis), the mean JD line for instance. The two coplanar spectra are now mutually re- versed, showing the succession red-violet and _ violet-red respectively. Interference should take place only along the D * Abridged from a fortheonting report to the Carnegie Institution of Washington, D. C. C. Barus—Interferences of Crossed Spectra. 487 line and be again inappreciable. Experimentally, I was not at first able to find any interferences for this case in the earlier manner shown below. But this may have been due to inadequacies in the experimental means employed: for the dispersion was insuflicient and the reflecting edge of the paired mirrorstoorough. Improving the apparatus, I eventually found the phenomenon, appearing however as a single line, vividly colored above the brightness of the spectrum ; or again more jet black than the Fraunhofer lines and located in the position of the coincident wave lengths of the two superimposed spectra. It is possible, however, as will be shown below, $4, to obtain two spectra in such a way that if their longitudinal axes coincide the Fraunhofer lines intersect at a small angle, and vice versa. In such a case, for coincident Fraunhofer lines, interference occurs in a band around these lines and is absent in the rest of the spectrum; whereas, if the longitudinal axes are coincident, the interferences are arranged with reference to these axes. These results seem to bear on the question, but it is difficult to clearly resolve it. The methods used in this paper consist chiefly in bringing the two first order spectra, or the second order spectra or their equivalents, to interfere. In this respect they contain an addi- tional method of interferometry which may be useful, if for any reason it is necessary that the two component beams are not to retrace their paths. 2. Case of coincident spectra with one reversed on a gwen Fraunhofer line.—In figure 1, Z is a narrow vertical sheet (to be broadened by the diffraction of the slit) of white sunlight or are light from a collimator, G the transparent grating ruled on the side g from which the first or second order of spectra gM and gW originate. MV and WV are opaque mirrors, mounted adjustably on a firm rail, 42, each of them with three adjust- ment screws relative to horizontal and vertical axes. J/ is provided with a slide micrometer (not shown). From J and JV the beams pass to the smaller paired mirrors, m and n, which should meet in a fine vertical line at a very obtuse angle. A silyered biprism would here have been far prefer- able, but none haying the required angle was available. From n, m, the beams pass into the telescope Z. As the spectra are each divergent after issuing from g, they can be made to overlap on leaving n, m, by aid of the adjustment screws on JZ VV. Moreover as the spectra are mirror images of each other, as suggested in figure 1, any spectrum lines: (as for instance the D) may be put in cvincidence on using one of the adjustment screws specified. It is necessary that the telescope Z be sufficiently near J/ in order that the micrometer may be manipulated. 488 C. Barus—Interferences of Crossed Spectra. The D lines placed in coincidence are obviously opposites, each line being paired with the mate of the other. A fine wire must be drawn across the slit of the collimator, in order that the vertical coincidence may be tested. One should expect the interferences to appear between the D lines on gradually moving the micrometer mirror J/, parallel to itself, into the required position. As stated above, I did not at first succeed in finding the interferences, but the experiment is a delicate one. In a repetition with first order spectra, it would be advisable to replace the plane mirrors m, ” by slightly concave mirrors, about 2 meters in focal distance and to replace the telescope Z’ by a strong eyepiece. This is the method nsed in the next paragraph and it was more easily successful. Later I returned to the experiment with the same adjust- ment, except that the plane mirrors m, 2 were placed beyond the grating, with the object of using the equivalent of second order spectra to get more dispersion. ‘This plan did not fail; and having once obtained the interferences, the reproduction - seemed quite easy, as they remained visible while the micro- meter JZ was moved over about +50 cm. or more. ‘Their appearance with a small telescope was that of a single fine line, alternately flaming yellow (very bright on the yellow back- ground of the surrounding part of the spectrum) and jet black as compared with the D lines, between which the interferential line was situated. This flicker is referable to the tremor of the laboratory, which makes it impossible to keep these C. Barus—Interferences of Crossed Spectra. 489 interferences quiet. Shutting off the light from either mirror, IL or N, naturally quenches the interferences, but leaves the yellow part of the spectrum behind. Obviously coincidence of the longitudinal axes of the spectra alone is needed. Therefore upon moving the two double D lines apart, by aid of the adjustment screws on the mirror J/ and iV, symmetrically to the ends of the yellow field in the telescope, the interferences were isolated and located midway between the D doublets of each spectrum; i. e., in the center of the field of the telescope. They could now be observed to better advantage. In the small telescope there is apparently but one dark line, and if stationary, its character when centered would be surmised to be given by the intersection of a vertical diameter with a series of con-focal ellipses, successively bright and dark, as indicated in figure la. The light and dark parts alternate or flicker. On moving the micrometer, the vertical intersector A takes a more and more lateral position like B, so that the trembling interferences would soon be invisible as they rapidly become finer (not shown). On using higher magnification (larger telescope), two black lines bordering a bright, or a black line between two bright lines, seemed to be visible; but the interferences would have to be stationary to be definitely described, since the width of the pattern is not more than 1/3 to 1/2 of the distance between the sodium lines. The interferences, moreover, did not now conform to the design B, fig. 1a, anticipated, but were more of type C, with the long dark lines very slightly oblique to the vertical, and vibrating within a vividly yellow band. Sometimes these were heavier with two or three faint lines on one side. Further experiment showed that the phenomenon is not influenced by the width of the slit, except that it is clearest and sharpest with the narrowest slit possible and vanishes when the slit is made so wide that the Fraunhofer lines disap- pear. It may easily be produced by the modified method fol- lowing in any wave length, red, yellow, green, etc., with no essential difference except in size. It is present moreover in all focal planes, i. e. the ocular of the telescope may be inserted or pulled out to any distance, yet the same phenomena persist on the vague, colored background. A number of observations were made to detect the change of the pattern of the interfer- ence between its entrance into the field and its eventual evan- escence in case of the continuous displacement of the mirror M. In figure 1@ this would be equivalent to a passage of B _ into B’ through A and the fringes for a distant center should therefore rotate, as they actually do in the experiments of paragraph 4. But in the present case the type C persists, 490 C. Barus—Interferences of Crossed Spectra. the lines may become longer or all but coalesce and their incli- nation may change somewhat. They nevertheless remain fine and nearly vertical until they vanish completely and there is no rotation. Nor could the phenomenon be found again within the length of the given micrometer screw. Hence it is improbable that these interferences conform at once to the ordinary elliptic type, even if the ellipse is considered excep- tionally eccentric. The use of two slits, one following the other, does not change the pattern. The modified method of experiment was one of double diffrac- tion. In figure 1), Z is the blade of light from the collimator, which passing wnder the plane mirror m, penetrates the grat- ing G whence the diffracted first order beams reach the opaque mirrors Jf and JV. These return the beams nearly normally, but with an wpward slant, so that the color selected intersects the grating at a higher level than Z. A second diffraction takes place at the same angle, @, to the direct ray ¢ and the coincident rays now impinge on the mirror m. They are thence reflected into the telescope at 7. This method admits of an easy adjustment, as everything is controlled by the adjustment screws on JZ and J and plane mirrors JZ, JV, and m only are needed, the latter being on a horizontal axis to accommodate Z. Directly transmitted white light is sereened off. 3. The same. Further expervments.—In place of the plane mirror, m, a slightly concave mirror (two meters in focal dis- tance, say) may be used with advantage and the telescope 7’ replaced by a strong eyepiece. In this way I obtained the best results. It is to be noticed that the apparatus, fig. 1, b, may serve as a spectrometer, provided the wave length, A, of one line and the grating space, D, are known, and the mirror, J/, is measure- ably revolvable about a vertical axis. In this case any unknown wave length, X’, is obtained by rotating JZ, until 2’ is in coin- cidence with A. Supposing the ’s of the two spectra to have been originally in coincidence and that @ is the angle of which now puts )’ in coincidence with A, it is easily seen that N—A=A (28in’9 /2 + /D?/r? — 1sind). Angles must in such a case be accurately measureable, i. e. to about ‘1 minute of are per Angstrém unit, if D = 351 x 10~°, as above. Counter rotation of the mirror JV till the 2s coincide would double the accuracy. The usual grating, however, has greater dispersion and would require less preci- sion in @. Finally a still simpler and probably more efficient device consists in combining the mirror 7 and the plane grating G, C. Barus—Interferences of Crossed Spectra. 491 or of proceeding, in other words, on the plan of Rowland’s con- eave reflecting grating. In such a case the light would enter in the direction 7’G, fig. 1, }, be diffracted along GJ/, back along MG and then return along G7Z'at a slightly higher or lower level than on entering. The equation just given would still apply and many interesting modifications are suggested. _Experi- ments of this kind are in progress. Moreover, in case of the plane transmitting grating and plane mirror, as above shown, the same simplification is possible, if the lens is replaced by the telescope at Z. But in this case the spectra are intersected by strong statzonary interferences due to reflections from front and rear faces and consequently not conveniently available. A reflecting grating and telescope would not encounter this annoyance. In general, however, as in the disposition adopted fig. 1, 6, the light enters opposite the observer and the light directly transmitted can be screened off; this is a practical convenience in favor of the transparent grating. The reflected spectra used may be placed at any level by rotating the mirror m on a horizontal axis. On further repeating the work by the use of the concave mirror m, and a strong eyepiece at 7, fig. 1, 6, and using a compensator, I eventually succeeded in erecting the interfer- ence design, Q, fig. 1, a. It then took the form given at D and this seems to furnish the final clue to the subject. In other words, the design consists of a new type of extremely eccentric ellipses, with their long axes parallel to the Fraun- hofer lines, each end having the outline of a needle point, possibly even concave outward. Only one end of a closed curve is obtainable. These jet black lines dance on the highly colored background of less than half the width between the two sodium lines. The interference design, therefore, would be the same (apart from color) as that which would be obtained, if the spectrum containing ordinary elliptic interferences were to shrink longitudinally from red to violet, till it occupied less than half the space between the two DP lines. In fact I have at other times obtained just such patterns, with all the colors present, but not in the pure yellow, as in the present case. Vertically, the path difference is always due to more or less obliquity of the rays passing through the plate of the grating. Horizontally, however, the equivalent path difference results, in the present case, from the fact that one wave-length of a pair has increased, whereas the other has diminished, while both may pass through the same thickness of glass and air. 4. Case of coincident spectra with one reversed on a given longitudinal axis.—For this experiment it is necessary to reflect the first order spectra issuing at the grating G, figure 2, a, from the ruled face g (a narrow preferable horizontal blade of 492 C. Barus—Interferences of Crossed Spectra. white light is here furnished by the collimator Z with a horizontal slit and the rulings of the grating are also hori- zontal and parallel to it), twice in succession and preferably from mirrors Jf and JV and m and 2, reflecting normally to each other and inclined at an angle of roughly 45°. Each of the mirrors Jf and JV must be revolvable about a horizontal axis parallel to the slit and furnished with three adjustment screws relatively to axes normal to each other, one of which is horizontal. The mirrors m, 7 are the silvered faces of a prism right-angled at the edge. It is moreover to be placed on the slide of a Fraunhofer micrometer so that the prism may be moved, gradually, up and down for the adjustment of distances. On leaving the mirror m, n, the two spectra are carried by nearly horizontal and parallel sheets of divergent rays, which pass outward from the diagram. But it will be seen that one of the two spectra reaching the observer is reversed on the longitudinal axis relatively to the other; i. e. if one is in the op violet, tae other jog | Bottom | ¥. et Bottom » will be Top : The subsequent passage of the rays is shown in figure 3, 0, which is the side elevation and therefore at right angles to the preceding figure. The rays from m and m impinge on a distant slightly concave mirror (A (about 1°7 meters in focal distance) placed somewhat obliquely, so that when the rays come to a focus at # near the micrometer, they may just avoid it. The partially overlapping spectra at / are viewed by a strong eyepiece £. The observer at / can then control the Fraunhofer micrometer by which m, n is raised and lowered, and the three adjustment screws of JZ. The adjustment consists in first roughly placing all parts in symmetry with sunlight, until the two spectra appear at £. The lens may be removed. There should be a bright narrow spectrum band on each side of and near the edge of the prism mn; for it is clear that after passing the lens /, corresponding rays from J and VV must both enter the pupil of the eye to be seen together. To make the spectrum parallel, the mirror m n is rotated, as a whole, around a vertical axis. The three serews on the mirrors J/ and J then assist in complet- ing the adjustment; the rotation around the horizontal axis brings the sodium lines in coincidence (both must be clearly seen and sharp and at an appreciable distance apart); that around the oblique axis gives rise to more or less overlapping, as required. The need of a sharp coincidence of the sodium lines is very essential in all these experiments. After proper vertical position of m n has been found by slowly moving the micrometer screw up and down, the fringes appear. They are usually very fine lines, indicating distant position red | C. Barus—Interferences of Crossed Spectra. 493 centers of the ellipses to which they belong. The appearanee is roughly suggested in figure 3. They pass from the type a through 6 (contraction toward the violet end was not noticed), into the type ¢, when mirrors mm move in a given direction. The center of the ellipses is in the vertical through the field of the adjustment 6, in which case the lines pass from end to end of the spectrum as a narrow band near the longitudinal axis of actual coincidence of spectra, symmetrically. Find 7a The height or breadth of the longitudinal interference band, d, in fig. 3, is not greater than 1°5 to 2 times the distance apart of the sodium lines at right angles to the band. From this the angular divergence of the breadth of the band may be found, since X = D sin 8, where 2X is the wave-length of light, D the grating space and @ angle of diffraction. Hence for the two sodium lines A@—A)A/Dcosd. Since D =: 351 x 10~, cos 6 = -986, and AN=6 x 107°, A@ = 1-7 X 10~*‘ radians. Since the width of the band is about twice this, it will be 68 seconds of are, or roughly about a minute in breadth. Within the strip, when the fringes are horizontal, I counted about five of them, so that their distance apart would be about 14 seconds of are. It appears, therefore, that rays of a given color, say of the wave-lengths at MY, which leave the grating at a given point and at an angle of about one minute in the plane of the D line, are still in a condition to interfere; whereas strictly 494 CO. Barus—Interferences of Crossed Spectra. speaking only those rays which lie in the common longitudinal axis of rotation of the two coincident spectra symmetrical to this should be in this condition. Such interference should not apparently be appreciable, since the white rays are independent and come from two different points of the slit. If we con- sider the angular deviation of pencils of parallel rays crossing ths grating, to be equivalent to the divergence of their respec- tive optical axes at the collimating lens (about 45 em. in focal distance), the distance apart of two points of the slit the rays of which are still able to produce interference is v= 45 x A@ = 45 X17 X 10-* =7°6X10-* em. or nearly 1mm. Hence pe of white light in the slit about -1 mm. apart along its ength produce the band of interferences in question, extend- ing in colored light from red to violet. 5. Interference of the corresponding first order spectra of the grating, in the absence of rotation.-—This apparatus seemed to be of special interest, since the rays used do not retrace their path and are thus available for experiments in which rays trav- eling in one direction only, are needed.* I have tried both the adjustments given in figure 4a and 6. The latter, since the rays are more nearly normally reflected at the mirrors J/ and LV, has some advantages ; but the other succeeds nearly as well. The difficulty encountered is a curious one of adjustment, which was not anticipated. In other words, if the longitudinal axes of two identical spectra are in coincidence, the Frauen- hofer lines are likely to be at a small angle to each other and complete interference is therefore impossible. Again if the spectrum lines are in coincidence, the longitudinal axes usually diverge by asmall angle. Furthermore the interferences are almost always eccentric and the lines hairlike, indicating distant centers. I have not succeeded in making a perfect adjustment, systematically, but the discrepancies are themselves interesting in their bearing on the subject of this paper. In figure 4, Z is a vertical blade of white light from a collimator with fine slit and @ is the grating. The two first order spectra leaving the ruled face at the line g strike the opaque mirrors J/ and JV, the former on a micrometer moving the mirror parallel to itself. From J/ and J the rays reach the half silvered plate of glass, HS, where one is transmitted and the other reflected into the telescope Z. The coincident rays & are superfluous. After placing the parts and roughly adjusting them for sym- metry with sunlight, the finer adjustment may be undertaken. It may be noticed that the two systems J/ and J, and G as well as ZS, can be used for further adjustment, separately. All are provided with adjustment screws relatively to rectan- * Cf. this Journal, xxxiv, pp. 101, 1912, on an air column carrying elec- trical current. sl C. Barus—Interferences of Crossed Spectra. 495 gular axes. ‘T’o put the mirrors Jf and JV in parallel and in the vertical plane with the grating G, the half silver plate should be removed and replaced by a small white vertical screen of cardboard, placed at right angles to the direction of HS in figure 4, and receiving both spectra. A fine wire is drawn across the slit to locate the longitudinal axis and an extra lens may be added to the collimator and properly spaced until the doublet insures sharp focusing. Both mirrors M/ and J are £5) Oh 9 now rotated on horizontal axes, until the longitudinal black lines in their spectra cease to diverge and coincide accurately. G,M,N may now be considered in adjustment. On returning the half silvered plate, A/S, it in turn is to be carefully rotated around horizontal and vertical axes, until the horizontal black line in the spectrum and the sodium line (always incidentally present in the are lamp) both coincide. But as arnile it will be found that if the longitudinal axes ww, figure 5a, coincide, the D lines cross each other at a small angle, exaggerated in the figure. The interferences, when found by moving the micrometer at J/, are usually coarse irregular lines, indicating a center not very distant and located on the level of a band where the J lines cross. On the other hand, if the D lines are brought to coincidence by moving the adjustment screws on J/ and WV (which throws them out of parallel), the longitudinal axes ww, w'w’, figure 5b, diverge at a small angle and the interferences are found in a vertical band where the lines ww and w’w’ cross. This band is relatively wide however as compared with the cases in 496 CO. Barus—Interferences of Orossed Spectra. paragraphs 2 and 3. Nevertheless I have looked upon these results as additional proof of the possibility of interference. For in neither case ought they to oceur if the spectra are not quite coincident horizontally and vertically. If they do oceur, it would seem that a certain small latitude of wave length adjustment is permitted even with light waves. The cause of this lack of simultaneous parallelism I was at first inclined to refer to the grating itself, as it occurred with an Ames grating ruled on glass, with a Michelson reflecting grating and with a film grating, in about the same measure. But subsequently, on adopting the method of figure 4, 6, the diver- gence was largely removed and the interferences were now vis- ible throughout the whole of the spectrum. The discrepancy is probably due to insufficient normality of the plate of the grat- ing to the incident white ray, since one of the rays is twice reflected. In any case the adjustment of the coincident sodium lines must be very accurate if the fringes are to be sharp ; cer- tainly as little as half their distance apart will obscure the phe- nomenon. Though the spectra are bright the interferences are not as good as with the usual method (paragraph 1); i.e. the dark lines are not black. Neither have I found an available or systematic method for centering the fringes, so that the lines obtained are usually delicate. Again the position of the collimator, both as regards slit and lens, is here of very serious importance. Any micrometric horizontal motion of either in its own plane will throw the fringes out. Finally the whole spectrum travels ' with the motion of the micrometer mirror J/. The apparatus is thus too difficult to adjust for use, to be of practical interest when simpler methods are at hand. The effect of tremors act- ing prejudicially on so many parts is exaggerated. 6. Conclusion. The phenomena of paragraphs 2, 3 and 4, showing definite and characteristic interference in case of two coincident spectra, crossed either on a longitudinal or transverse axis, represent the chief import of the present paper. These results cannot be due to the diffraction of a slit (regarding the line of coincidence as such), owing to their relatively small magnitudes and their independence of the breadth of the slit. Since there is in each case but a single line of points or axis, the disturbance of which comes from identical sources, we might regard the image of this line in the telescope to be mod- ified by the diffraction of its objective. But if the interfer- ences originated in this way, the Fraunhofer lines of the spec- trum should show similar characteristics and the diffraction pattern should differ from those observed. Thus the conclu- sion is apparently justified that distinct and independent points of the narrow slit whose distance apart on its length is not C. Barus—Interferences of Crossed Spectra. 497 greater than ‘1 mm., are still capable of producing interference in each of the colors of the spectrum (longitudinal axes coin- eiding). This phenomenon is virtually one of homogenous light, the same type of interference occurring in each color from red to violet. They belong moreover to the elliptic cate- gory, being of the same nature as those used in displacement interferometry. With the exception of the points lying on the longitudinal axis of rotation or of coincidence, all the pairs of points of the two coincident spectra owe their light to different sources ; i. e. they are not color edimages of one and the same point in the slit. Again, in case of rotation of one of the coincident spectra around a transverse axis (Fraunhofer line), colors which differ in wave length by less than half the distance apart of the two sodium lines also admit of interference. This permissible dif- ference of wave length is thus relatively about MSs) Se olay Ne EY SSO == S< 10> or Jess than ‘1 per cent. The character of these interferences is distinctive. They are not of the regular elliptic type, but usually arise and vanish in a succession of nearly vertical (parallel to slit), regularly broken lines. Later observations however, revealed as their true form a succession of long, spindles or needle-shaped designs. The chief peculiarity is observed in their almost scintillating mobility, which in the above text has been referred to the inevitable tremors of the laboratory. It is, however, interesting to inquire into the con- ditions of the possibility of observable beating light waves. For two waves very close together of frequency m and x’ and wave lengths X and X’, if V is the velocity of light, the number of beats per second would be n'—n= V(1/XN—1/A)= VAA/Y’, nearly. Therefore in case of the two sodium lines for instance, Wi soln oe < 10/3500 x 10°" = 5X 10” i. e. about 5 X 10" beats per *1 second, the interval of flicker- ing. Naturally this seems to be out of all question: yet one is confronting a source which is an approach to a mathematical line; and though I am not apt to stretch a rather conservative imagination quite so far, I should like to see this interference produced under absolutely quiet surroundings. Its appearance is altogether singular and not like the case of paragraph 4, where there is also perceptible tremor, or with the general case of trembling interference patches, with which I am, unfor- tunately, all too familiar. Am. Jour. Sci1.—FourtH Serizs, Vou. XL, No. 239.—Novemser, 1915. 498 C. Barus—Interferences of Crossed Spectra. In this place, however, it is my sole purpose to present, at its face value, an observation which is spacial, independent of time consideration, and the laterally cramped character of the new interference, with its long hair-like lines thrust into a strip less than half the distance apart of the sodium lines, is the only evidence submitted. For, if the coincident path of two rays of slightly different wave-lengths and 2’ which interfere, is z, then there are #/d and w/X’ complete waves in the given path ; and in ease of original identity in phase, reinforeement will occur when eV | N= 17 XN) = Ose ere In other words at the nth reinforcement AX = ndi fa. Hence since X* is very small and «@ relatively very large, the exceedingly small value AX (i.e. the very narrow strip of spectrum within which the phenomenon occurs) is apparent. In the above experiments the estimates, in round numbers, were Arx =24X 10°, A° = 36x 10>". Elenceif 7) — ih ates so that one reinforcement occurs about at each 1°5 millimeter along the rays. It remains to be investigated why these nominally beating wave trains with an infinitesimal group period, can be recog- nized at all. But into this question I am not now prepared to — enter, as the answer is almost wholly dependent on experi- ments in progress. If whatever vibrates has inertia (electric displacement), the case of forced vibrations is suggested (approximate resonance), with opposite phases on the two sides of the transverse line of coincidence, essential. The characteristic feature of the new phenomenon is this, that apart from intensity, it persists without variation, through a path difference of over 10 millimeters, i.e. through 15,000 or 20,000 wave lengths. It follows, since the paths, grating mirror-grating, are alone significant, that two individual light waves of the same ray over 15,000 wave lengths apart are still appreciably identical. Beyond that the waves no longer cor- respond in orientation and cannot interfere in a way to produce alternations of accentuated brightness and darkness. Again, these 5 millimeter lengths may actually be discontinu- ous and represent successive discharges of radiation (each last- ing about 15107" seconds), separated by absence of light motion along the ray. Through the years, I seem still to hear the lament of my old friend and teacher, Ogden N. Rood, of the time he had wasted looking for the interference of differentiated light waves. If with more modern facilities I have reached a conclusion, it would be a privilege to associate it with his memory. Brown University, Providence, R. I. W. B. Clark— Brandywine Formation. 499 Art. XXXVII—The Brandywine Formation of the Mid- dle Atlantie Coastal Plain ; by Wiit1am Butiock Crarx. Name.—The name Brandywine* is proposed for this forma- tion as the deposits are extensively and typically developed in the vicinity of Brandywine, Prince George’s County, Mary- land. Synonymy.—(a) Appomattox formation.—The name A ppo- mattox was proposed by McGeet in 1888 for the older terrace accumulations and various other deposits exposed in the valley of the Appomattox River in Virginia. The same authort in 1890 discussed their southern extension and Darton§ in 1891 described the northern extension of the deposits into Maryland and adjacent areas. The term Appomattox was used by both McGee and Darton to designate two or more clearly recognizable stratigraphic units separated by well defined escarpments. In McGee’s description of the Appomattox formation he included the two higher terrace formations described by Shattuck as Lafayette and Sunderland and also extended the formation down the valley lines to embrace even later Pleistocene deposits includ- ing parts of the Wicomico and Talbot formations, as well as weathered portions of the Aquia formation of the Eocene. (6) The Lafayette Formation.—The name Lafayette was applied by MeGeel| in 1891 to the deposits of the Middle Atlantic Slope which he had previously described under the name Appomattox formation on the assumption that they belonged to the same formation as those so designated by Hilgard§ in Lafayette County, Mississippi. Hilgard’s type area was shown to be of Wilcox Eocene age by Berry** and the work of Vanghan, Stephenson, Shaw, and others has * The recognition by the U. S. Geological Survey and the various State Surveys in the Atlantic border area of the inappropriateness of the term Lafayette as employed in the Atlantic border region has led to the proposal by the author of the name Brandywine for the oldest of the terrace forma- tions of that district. This name has already been submitted to the Board of Geologic Names of the U. S. Geological Survey and adopted by it. +MeGee, W. J.: Three Formations of the Middle Atlantic Slope, this Journal, xxxv, 328-330, 1888. Tdem: Southern Extension of the Appomattox Formation, ibid,, xl, 15-41, 1890. § Darton, N. H.: The Mesozoic and Cenozoic Formations of Hastern Vir- giniaand Maryland, Bull. Geol. Soc. Amer., ii, 445-447, 1891. || McGee, W. J.: The Lafayette Formation, 12th Ann. Rept. U.S. Geol. Survey, 347-621, 1891. “| Hilgard, KE. W.: Orange sand, Lagrange and Appomattox, Amer. Geol., Vili, 139-131, 1891. ** Berry, H. W.: The age of the Type Exposures of the Lafayette Forma- tion, Jour. Geol., xix, 219-256, 1911. 500 W. B. Clark—Brandywine Formation of the demonstrated that the Lafayette as defined by Hilgard and described in detail by McGee comprised the weathered surface materials of various Cretaceous and Tertiary formations. Shattuck* in his study of the surficial formations of Maryland limited the term Lafayette to the highest of the terrace forma- tions, the formation to which the author now gives the name of Brandywine. The terms Appomattox and Lafayette as originally employed in the Middle Atlantic Coastal Plain embraced much more, therefore, than it is proposed to include under the name Brandywine formation, and likewise, the diagnosis of that for- mation is based on different physiographic conceptions from those used by McGee and Darton. Furthermore, the term Lafayette as originally employed in Mississippi was based on a misconception of the stratigraphy of the region. The use, therefore, of either the term Appomattox or Lafayette for any formational unit is impracticable. Areal outline.-—The Brandywine formation covers an exten- sive area in the southern Maryland peninsula, reaching from the eastern boundary of the District of Columbia to the northern line of St. Mary’s County with numerons outliers both to the north and to the south of these lines. It attains a maximum width, therefore, from northwest to southeast of nearly 40 miles. From this region it narrows both to the northeastward and to the southwestward, being confined largely to the landward margin of the Coastal Plain. It reaches to the northward through the central counties of Maryland into Delaware and Pennsylvania, while to the southward it has been traced in the interstream areas and along the landward border of the Coastal Plain through Virginia into the Caro- linas. Less is known of the details of its areal distribution in Virginia than in Maryland and the states which lie to the north of it. Altitude and character of landward boundary. —The alti- tude of the landward boundary reaches 400 feet in the outliers in the western part of the District of Columbia, 486 feet at Burtonville in Montgomery County, 508 feet at ‘Catonsville, Baltimore County, 480 feet at Loch Raven, Baltimore County, and 470 feet at Woodlawn, Cecil County. No one of these more western outliers covers more than a few square miles of area. Each is extensively eroded and isolated from the main body of the formation farther seaward. Ali are found in interstream positions, their coarse gravel content protecting them from destruction. *Shattuck, G. B.: Pliocene and Pleistocene, Maryland Geol. Survey, 291 pp., 75 pls., 10 figs., 1906. Middle Atlantic Coastal Plain. 501 Altitude and character of seaward boundary.—The altitude of the formation along the seaward boundary in northern St. Mary’s County is not over 200 feet while farther north- ward in Maryland, due to the more extensive later Pleistocene encroachments of the sea in those areas resulting in the re- moval of the deposits along the seaward face, elevations of 240 feet are found at Marriott Hill, Anne Arundel County, and 300 feet on Elk Neck, Cecil County. Clearly defined escarpments appear in many places along the seaward margin although at other points such escarpments are wanting, the higher terrace presenting a long featureless slope along its front which gradually coalesces with the next terrace level. A significant feature is found near Charlotte Hall in St. Mary’s County where a seaward-facing escarpment extends entirely across the divide of the peninsula of southern Maryland and separates the Brandywine from the lower terrace level of the Sunderland. The Sunderland plain also passes at this point between two of the outliers of the Brandywine and over the divide between two small streams rising within the area. fate and direction of slope of surface.—The slope of the Brandywine surface within the main area of outcrop in the peninsula between the Potomac and Patuxent rivers is some- what over five feet to the mile, although it is greater nearer the landward margin than seaward. In the former case the slope reaches 15 feet or more to the mile in the vicinity of the District of Columbia, a feature also shown to the northward in Baltimore and Cecil counties, Maryland, while nearer the sea- ward margin it declines to about three feet to the mile in southern Prince George’s and northern St. Mary’s counties. The direction of slope is from northwest to southeast at right angles to the coast lines of the period. Amount of dissection.—The Brandywine formation is exten- sively dissected, the largest continuous area of outcrop being found in the interstream area between the Potomac and Patuxent rivers. Here the margin of the formation has been extensively dissected, resulting in many outliers. In Prince George’s and Charles counties, Maryland, however, are undis- sected and undrained tracts possessing the essential features of the old terrace surface and appearing today as broad and nearly featureless plains. This is best seen in the area extend- ing from Cheltenham, Prince George’s County, southward to Hughesville and Waldorf and somewhat beyond, into Charles County, Maryland. Component materials and structure of deposits —The Brandywine formation is composed of gravel, sand, and Joam. Over considerable areas the gravel is found occurring in greater abundance at the base, while sand and loam more especially 502 W. B. Clark—Brandywine Formation of the characterize the upper portions of the formation. The beds or lenses are irregular and often the materials are mixed together in a confused manner. Cross-bedding occurs in the sands and gravels but in general the materials are imperfectly sorted and are found intermingled in varying proportions, the gravel and sand frequently containing much elay. The gravels are chiefly developed i in the landward portions of the formation and decline in frequency as well as in size of cobbles toward the seaward portions of the formation and are almost entirely absent near the seaward margin, being replaced altogether by sands and loams. The materials are of varying sizes and shapes, generally more or less angular with rounded edges and rarely affording the discoidal pebbles char- acteristic of sea beaches. The pebbles toward the landward portion of the formation are almost invariably covered with a dark-brown, ferruginous coating, but farther seaward the amount of iron decreases and the coating of iron oxide is prac- tically absent. The pebbles are largely of quartz but some are of crystalline and other rocks and a few even of Newark sand- stone, which shows that they were derived from the Piedmont and Appalachian regions. Many were doubtless redeposited from the early Cretaceous formations. Scattered among the cobbles and pebbles and also at times among the sands and loams are angular bowlders of Piedmont or Appalachian origin. The gravels especially in the landward portions of the forma- tion are often much decayed, the pebbles readily breaking down under light blows from the hammer. The sands are both coarse and fine and doubtless have largely the same origin as the gravels although more largely derived from the older Coastal Plain formations. This is doubtless especially true of the seaward portions of the formation, which were evidently derived in considerable measure from the sandy beds of the older Miocene formations, which must have extended widely over the eastern margin of the Piedmont and are today recognized in outliers some dis- tance to the west of the Coastal Plain border. At the same time the sandy members of the underlying Eocene and Ore- taceous formations must also have contributed their quota. The sands are much more heterogeneous in the landward por- tions of the formations than in the seaward and are frequently very compact when cemented by iron oxide or when they con- tain an admixture of clay, loamy sands or sandy loams being frequently found, particularly in the landward portions of the formation. Beds of pure quartz sand are infrequent and when present are not of great thickness or wide extent. They increase in prominence. however, seaward. The loams are widely extended but of very variable sorts. Middle Atlantic Coastal Plain. 503 They are more largely found at the top than lower in the for- mation but may occur at all levels. They are frequently inter- mingled with the sands and gravels, forming sandy and gravelly loams, often making a matrix for cobbles of considerable size. The loam capping of the formation, which occurs widely, varies from a few inches to 10 feet or more in thickness. In the landward portions of the formation the loam contains con- siderable iron and at times has a decided orange color. In some areas, particularly in northern Charles County, it has a pronounced mottling of drab and brick-red which is particu- larly noticeable when the material is wet. The loams in the seaward portions of the formation are generally grayish-yellow in color. Although often mixed with gravel and sand the loams are in places very argillaceous and fine in texture, espe- cially in the upper portions of the formation. They have at times irregular beds of coarse sands or even gravel interstrati- fied with them, thin seams of gravel not being an uncommon feature. The loams. are frequently very hard and compact, especially when rich in hydroxide of iron. Both the sands and the loams, particularly the gravelly or sandy loams, show a marked case-hardening on exposure to the weather. The Brandywine formation has a thickness of from 10 to 30 feet, the thickness for the most part increasing from the land- ward toward the seaward margin of the formation, although there are many exceptions to this rule where the deposits have been laid down in the inequalities of the surface of pre-Brandy- wine time. Exceptional thicknesses of over 50 feet have been found. In general the coarser materials toward the landward margin have a thickness of from 5 to 15 feet while in the sea- ward portion of the formation the finer sands and loams reach from 20 to 80 feet. No determinable fossils have been recognized in the type area althongh Darton refers to the presence of a few indeter- minable molluscan shells in Virginia which may well be reworked Miocene forms. The soils most widely distributed on the Brandywine forma- tion have been described by the United States Bureau of Soils as the Leonardtown loam, the Windsor sand, and the Norfolk sand. The Leonardtown loam is generally found where the upper layers of the Brandywine formation are loamy and con- sist of a yellow silty loam having an average depth of about 10 inches. It is underlain by a heavier yellow loam which usually grades into a mottled loam at a depth of from 28 to 32 inches. More or less sand and gravel generally appear in the subsoil. Along the borders of this soil type the sand and gravel become more prominent as the soil becomes thinner and the Leonardtown loam grades over into more stony and 504 W. B. Olark—Brandywine Formation of the gravelly types. The Windsor sands, the next most important soil type found in the Brandywine formation, consists of a medium to coarse sandy soil that generally contains about 10 per cent of fine gravel. The soil is loose and friable and reaches to a depth of abont 8 or 10 inches and is underlain by a coarse sandy subsoil. It is quite widely extended but is most common in the area midway between the landward and sea- ward borders. The Norfolk sand is less widely extended in the Brandywine formation than the two preceding soils and consists of a medium to coarse orange or yellow sand to a depth of about 10 inches and is underlain by a coarse sandy subsoil which usually becomes loamy at a depth of about 3 feet. — Vegetation and cultwre.— Extensive areas are covered with serub pine and culled hardwoods of various types, the serub pine being found largely growing on the Windsor sand. Large areas, however, are under cultivation. The Leonardtown loam, because it is capable of retaining a considerable amount of moisture during the entire growing season, is well adapted to growing grass, wheat, and corn where general farming is practised, and to cabbages, cucumbers, and late strawberries in the trucking areas. It is only producing, however, to its full capacity in the northern part of the county, where the soil has been extensively enriched and where a ready market for its products has been found in Washington. The Windsor sand is very porous but is well adapted to early truck crops, early peaches, and under certain conditions to fine grades of tobacco. Its loose, porous character makes it particularly hard to manage during a protracted drought and for this reason intensive cul- tivation is required including the incorporation of considerable amounts of organic material. The Norfolk sand, although very porous, has been utilized in general farming and truck growing and is well adapted for early strawberries, melons, sweet potatoes, and small crops of high grade tobacco. Stratigraphic relations of the deposits—The Brandywine formation overlies all of the older Coastal Plain formations of - Tertiary and Cretaceous age unconformably and at a few points rests on the crystalline rocks of the Piedmont. It is separated from the next younger or Sunderland formation throughout much of the region by a clearly marked escarpment, the Sunderland formation wrapping at a lower elevation about the lower margin of the formation unconformably and filling the pre-Sunderland valleys. One of the most striking physiographic features of the region is the Sunderland plain extending into the pre-Sunder- land valleys and abutting against their irregular slopes. Streams are often absent today in these old valley lines but where present have cut irregalar channels through the surface Middle Atlantic Coastal Plain. 505 of the Sunderland terrace. Where the streams have eroded the Sunderland, as well as at many points along the landward extension of the Brandywine formation where the interval between the terrace levels increases, the underlying formations appear in outcrop beneath the Brandywine. Local Brandywine area.—Brandywine, Prince George’s County, Maryland, is located on the slightly-eroded surface of the old Brandywine terrace not far from the center of the largest tract still preserved intact. The surface materials in this region consist of sandy loams, while the deeper trenches eut by the streams around the margin of the tract show irregu- lar beds of coarse sand and gravel, the latter generally small in size. The formation attains its maximum thickness in the general area in which Brandywine is situated. No section of the formation is exposed at Brandywine since it is situated on the uneroded surface of the formation, but the adjacent ravines both to the east and west cut through the formation, exhibit- ing the gravels, sands, and loams characteristic of the for- mation. Interpretation of lustory.—The Brandywine epoch was opened by a depression of the continent border which carried the waters of Brandywine time over the eroded Tertiary and Cretaceous deposits to well within the Piedmont area. A small ontlier of questionable origin but containing gravels similar to those of the Brandywine formation is found far within the Piedmont in the Frederick valley, at an elevation which corresponds with the possible transgression of the sea into the broader valleys of the Piedmont district. At all events, the tilting of the coastal area must have brought about renewed erosion of the adjacent land with the result that extensive and more or less heterogeneous accumulations of gravels, sands, and clays were deposited rapidly along the margin of the Brandywine sea, or more widely scattered with- out much sorting by the currents and undertow of the period. Many of the deposits, especially the coarser types, point to a fluviatile origin of the materials, and many of the beds show in their irreeular and confused stratigraphy, characters such as oc- eur from a combination of fluviatile and flood-plain deposition. At the same time commingled with these probable fluviatile elements are other materials of distinctly marine or estuarine origin, and it, therefore, seems probable that a combination of fluviatile and off-shore agencies must be predicted for the formation of the Brandywine deposits. Furthermore, the broad terrace plain, formed of materials of no great thickness, which gradually thicken seaward as so often occurs in near- shore marine beds, point to the existence of a large body of water into which the ill-sorted materials of the time were 506 W. B. Clark—Brandywine Formation. rapidly carried. Although fluviatile deposition may have actually taken place over part of the area during the period of subsidence it must have been replaced ultimately by off-shore deposition as the sea transgressed. In no other way can the form and structure of the formation be fully interpreted. Age of the formation.—There has been much discussion of the age of the deposits composing the Brandywine formation. but as no determinable fossils have hitherto been found in the beds no definite conclusions have been reached. It has been generally accepted that the post-Brandywine deposits are of Pleistocene age, while the Brandywine formation itself has been questionably referred by most authors to the Pliocene on the ground of the more extensive erosion to which the strata have been subjected and the greater decay of the constituent materials. The author raises the question, however, whether the Brandywine formation might not with entire propriety be referred to the early Pleistocene, all of the surficial terrace deposits therefore of the Middle Atlantic Coastal Plain being placed in that event within the limits of the Pleistocene. Geological Laboratory, Johns Hopkins University, Baltimore, Md. P. EF. Browning—Study of Flame Spectra. 507 Arr. XXXVIII—On Two Burners for the Demonstration and Study of Flame Spectra ; by Puitre E. Brownrne. [Contributions from the Kent Chemical Laboratory of Yale Univ.—celxxiii. | In the demonstration and study of flame spectra it is obvi- ously advantageous to impart the colors to the flames for a longer period of time than can be secured by simply inserting in the flame a platinum wire or asbestos fiber previously dipped into asolution of a compound of the element whose spectrum is desired. The purpose of this paper is to describe two forms of appar- atus which have been found convenient for securing a certain degree of permanence to the colors in the flame. Fig. 1 shows a burner constructed as follows: An ordinary salt mouthed bottle is fitted with a three-holed stopper. Through one of these holes a tube is placed which connects by means of a rubber tube with the gas supply. Through the second a tube drawn out to a jet is inserted and over this jet a porcelain tube is suspended in such a way as to make a color- less flame possible. These porcelain tubes are such as are used often as insulators in electric wiring. They are preferable to glass or metal because they impart no color to the flame and | do not corrode. Through the third opening in the stopper an ordinary thistle tube is inserted. To operate the burner a piece of marble or a carbonate min- eral of strontium or barium, for example, is placed in the bot- tle with enough water so that when the stopper is placed in position the end of the thistle tube is submerged in the water. 508 P. E. Browning—Study of Flame Spectra. The gas is then turned on and the burner so adjusted as to give a good colorless flame. Enough acid is then added through the thistle tube to start a gentle evolution of carbon dioxide, and this gas carries enough of the solution to be swept by the current of gas into the flame. Mineral carbonates have an advantage over commercial precipitated carbonates on account of their freedom from sodium. This form of apparatus may be used to obtain sodium, potassium and lithium flames, but the apparatus designated in fig. 2 is rather better for these elements. Fig. 2 is modelled after a type of burner used in the labora- tory of M. Urbain at the Sorbonne in Paris, where the author made its acquaintance. This apparatus is constructed similarly to that marked fig. 1 except that in place of the thistle tube a tube is so arranged that a glass rod attached by a rubber con- nector may slide up and down within it. Attached to the end of the glass rod by a hook made by drawing out the glass is a piece of zine bent in shape of a tube about an inch in diameter. The form here described differs from the Paris type in the sub- stitution of a porcelain tube for a glass tube, not essential if the burner is used only for sodium, and the tube-shaped piece of zinc rather than a strip of zine to secure a greater surface of metal. To operate this burner a solution of a sodium, potassium or lithium salt is placed in the bottle and enough acid is added to give a gentle evolution of hydrogen with the zine. The stop- per is placed in position and the zine lowered into the liquid by means of the glass rod. The current of hydrogen operates to bring the solution of the salt into the flame. Drushel and Knapp—Glycocoll and Diethyl Carbonate. 509 Art, XX XIX.—On the Preparation of Glycocoll and Diethyl Carbonate; by W. A. Drusxer and D. R. Knapp. [Contributions from the Kent Chemical Laboratory of Yale Univ.—celxxiv. ] Tue preparation of glycocoll by the interaction of ammonia and chloracetic acid is attended by by-reactions which make it impossible to obtain a quantitative yield of the aminoacid. In 1858 Perkin and Duppa* obtained a small yield of glycocoll by the action of ammonia upon bromacetic acid at the boiling temperature; later Cahours,+ repeating the experiment with chloracetic acid, got a yield of 10 per cent to 15 per cent. Mauthner and Suidat improved the yield by working with ammonia and chloracetie acid in the cold, and more recently Kraut was able to get a yield of 50 per cent to 55 per cent by cooling the reaction mixture to 10°. The very marked improve- ment in yield obtained by Kraut suggested the possibility of even a better yield by working at 0° and keeping the reaction mixture saturated with ammonia until all of the chlorine was fixed as ammonium chloride. In this investigation 700°" to 800° of ammonia water were cooled to 0° in an ice bath, and saturated at this tempera- ture with gaseous ammonia. One hundred grams of chlorace- tic acid were dissolved in the least possible amount of water and slowly run into the ice cold saturated solution of ammonium hydroxide, keeping the reaction mixture at 0° and saturated by passing gaseous ammonia. In our earlier experiments the mixture was kept at 0° for about fifteen hours and was then allowed to stand at room temperature for several days. In the later experiments the temperature was kept at 0° for more than four days; the reaction was found to go to completion at this temperature in about five days. The velocity of the trans- formation of the chloracetic acid was determined by pipetting off 5°’ portions of the reaction mixture from time to time, neutralizing the excess of ammonia with ice cold nitric acid and titrating the ammonium chloride present with standard silver nitrate, using potassium chromate as an indicator. A good concordance in velocity constants was obtained by mak- ing the calculation from the usual formula for second order reactions. The mean of a series of five velocity constants is 0:000223. The glycocoll was recovered from the reaction mixture and purified by the copper hydroxide method; the use of copper carbonate instead of copper hydroxide suggested * Quart. Jour. Chem. Soc., xi, 22. + Compt. rend., xlvi, 1044; evii, 147. t Monatsh., xi, 374. § Ann. Chem. Pharm., celxvi, 299. 510 Drushel and Knapp—Glycocoll and Diethyl Carbonate. by Clarke* was also tried but was found to be much less satis- factory than copper hydroxide. In no case did the yield of pure glycocoll exceed 55 per cent of the chloracetic acid used. It does not seem possible to increase the yield of glycocoll by changing the conditions of Kraut’s method in the direction of lowering the temperature or increasing the concentration of ammenia during the reaction. Diethyl Carbonate.—Since cyanformic ester is readily con- verted into aminoacetic ester by reduction, it seemed desirable to examine the methods of preparing cyanformic ester. In 1895 Nefft obtained this ester mixed with some diethyl carbonate by the action of potassium cyanide upon chlorformic ester in 25 per cent aqueous alcohol at —13°, allowing the temperature to rise slowly during the course of the reaction to —2°. Under these conditions only about 50 per cent of the chlor- formic ester was converted into cyanformic ester. In our experiments absolute alcohol was used, resulting in the abund- ant formation of diethyl carbonate without any cyanformic ester. Powdered potassium cyanide was covered in a flask with absolute alcohol and then one equivalent of chlorformie ester mixed with an equal volume of absolute alcohol was slowly added in small portions through a reflux condenser under a good draught hood. Heat was evolved, hydrocyanic acid was given off and the reaction mixture became red in color. The mixture was finally boiled for an hour on the water bath to bring the reaction to completion. The residue was then filtered off and the filtrate was fractionally distilled. The first fractionation gave almost pure diethyl carbonate, probably according to the equation C,H,CO,Cl + KON + C,H,OH = (C,H,),CO, + HON + KCl. After redistillation the product boiled at 127°, and its identity as diethyl carbonate was further established by determining its molecular weight; by determin- ing its saponification equivalent as 59, its specific gravity as 0°968 at.25°; and finally by preparing urethane from the product by the action of ammonia. No cyanformie ester could be isolated from the reaction mixture and the yield of diethyl carbonate after purification was 50 per cent of the chlorformic ester used. *H. T. Clarke, Org. Chem., p. 290. + Ann. Chem. Pharm., celxxxvii, 308. Drushel and Holden—Hydracrylic Esters. 511 Arr. XL.—On the Preparation and Properties of Hydra- crylic Esters; by W.A. Drusuet and W. H.T. Horpen. (Contributions from the Kent Chemical Laboratory of Yale Univ.—cclxxv.) Tue esters of hydracrylic acid have up to the present time received but little attention, only the methyl and ethy] esters are mentioned in the chemical literature. The direct esterification of hydracrylic acid had apparently not been attempted until one of us* prepared ethyl hydracrylate in this laboratory by the direct esterification process. As described in the literature the indirect methods of preparation are objectionable since the _ esters are obtained in impure condition, containing impurities not easily removable. Obviously mineral acids can not be used as catalytic agents in the esterification of hydracrylic acid since under the necessary conditions hydracrylic acid is con- verted into acrylic acid and the resulting esters are chiefly acrylic esters. The object of this investigation was to study the direct esterification of hydracrylic acid with the common alcohols, and the properties of the series of esters prepared in this way. Sodium hydracrylate was prepared from glycerine by the methods fully described in our previous paper.t The impor- tance of keeping the solution of sodium hydracrylate slightly acid during evaporation is to be emphasized, since in alkaline solution condensation products are formed and only very little sodium hydracrylate is obtained. The best procedure appears to be nearly to neutralize B-iodopropionic acid with sodium carbonate, taking care to leave the solution distinctly acid. The solution is then treated with an excess of silver oxide freshly prepared and free from alkali. In this way the result- ing solution contains sodium hydracrylate with a little silver hydraecrylate. After filtering off the silver iodide the solution is treated with hydrogen sulphide to decompose any silver salt present. The silver sulphide is then filtered off and the filtrate is usually sufficiently acid to permit evaporation to dryness on the steam bath without the formation of more than very small amounts of by-products. The dry sodium hydracrylate thus prepared is further purified by recrystallization from 95 per cent alcohol. The yield is almost quantitative. Weighed portions of the purified and dried sodium hydracrylate were dissolved in the least possible amount of cold water and treated with a little less than the theoretical amount of 1:1 sulphuric acid, keeping the reaction mixture cold during the process. The water in the resulting mixture was evaporated off on the steam bath attended by only a slight loss of hydraerylic acid, leaving a residue of sodium sulphate and pure hvdracrylic acid. * This Journal, xxxix, 113-121, 1915. + Loe. cit. 512 Drushel and Holden—Hydracrylic Esters. The hydracrylie acid thus prepared, in about 50 gram portions, was extracted from the residue by the absolute alcohol to be used in the subsequent esterification process. The alcoholic solution of hydraerylic acid, about 200°™ in volume, was treated with somewhat more than the theoretical amount of anhydrous copper sulphate free from sulphur trioxide. The reaction mixture was gently boiled with a reflux condenser on a sand bath until 80 per cent to 90 per cent of the hydracrylic acid had been esterified. The course of the esterification was followed by titrating from time to time 1™° of the reaction mixture with decinormal barium hydroxide. After the completion of the reaction the copper sulphate was filtered off, the unesterified hydracrylic acid almost neutralized with anhydrous sodium carbonate, the excess of alcohol distilled off in a water or oil bath and the residue of hydra- erylic ester subjected to fractional distillation under diminished pressure, usually at 12™™ to 20"". The esters which were prepared in this way and whose properties were then studied are the methyl, ethyl, propyl, isopropyl, isobutyl and isoamyl esters of hydracrylic acid. The densities of these esters were determined at 0° and at 25°; the values obtained are recorded in Table I. TABLE I. 0° 25° Methyl hydraciylates 322s se eenoe= 1140 Tae Kthyl oem eee Pos Se 1:085 1-064 Propyl Cr fe) Ses {oNGrecce gia ge aig 1°052 1:043 Tsoptojoyl nat en beat eee eee 1:071 1:058 Tsobutyl eee m tatskhes Slow RE A rte 1-013 1:003 Isoamyl . Ppa rts Se OES) 0'976 These density determinations are referred to water at 4° and were made by means of a pycnometer * of a type which was especially devised for use in the qualitative organic analysis work in this laboratory. It consists of a capillary stem about 7°™ in length, a bulb to contain approximately 1™ and a drawn out capillary about 1™ in length. The capillary stem is graduated so that the pycnometer shall contain exactly one cubic centimeter of water at 4°. This type of instrument has the advantage that by a single weighing it gives fairly accurate specific gravity data even in the hands of those not skilled in the use of the more complicated types. All of the hydracrylic esters investigated boil with decompo- sition at atmospheric pressure. Their boiling points were therefore determined directly over a range of pressures of 15™™ to about 100™™ and from the data obtained the boiling points * A cut of this type is published in the Journal of Industrial and Engineer- ing Chemistry, vii, 187. Drushel and Holden—Hydracrylic Esters. 513 for atmospheric pressure were estimated by the graphic method. The results are recorded in Table II. TaBie II. Boiling points of hydracrylic esters. Methyl Ethyl Isopropyl Propyl Tsobutyl Isoamyl ester ester ester ester ester ester (SSS (a SSS > (cmemmeiamaaan a Gamammiatammmmeny Press. Press. Press. Press. Press Press. in in in in in in JJ. ind, BLE peapanys B.P. mm. Be, sonra B.P. mm. B.P. mm 79 12 84 12 95 12 98 12 104 15 12175 «17 87 20 917% 19 gH) 20 102 19 112 20 124 20 1075 40 95°35 22 109 35 120 43 126 45 137 36 111 56 116 62 114 42 127 62 129 50 147 58 121 94 127 8§©=©.102 128°5 82 142 107 1438 88 159 95 184** 760 190** 760 196** 760 205** 760 212** 760 225** 760 ** Calculated. All of the esters of this series below the isobutyl ester are freely soluble in water in all proportions, but the higher homologues are not. The isobutyl ester is soluble in 18 parts of water and the isoamyl ester in 60 parts of water at room temperature. The refractive indices of the esters were determined by means of the Pulfrich refractometer for the D line of sodiuin light and the following results obtained : Tas_eE III. M n (n — I M Methyl hydracrylate-.-_--.-.--- 1°4306 40°10 104°0 Ethyl] 54 LE tee ee 1°4271 47°30 118-0 Propyl << Cl > soe 1°4341 54°85 132-0 Isopropyl “ Ci. 2 ee 1°4303 53°68 132°0 Isobutyl “ (CT eae ae 1°4342 63°70 146°0 Reaamipglieccn. Face ges. 174374 71°65 160-0 These indices of refraction were determined at 23° C. Conclusions :-— 1. Hydraerylic acid is obtained from glycerine in a fair yield by slight modifications of the methods described in the literature. 2. Removal of the iodine from sodium $-i0dopropionate rather than from the free acid by means of silver oxide has the advantage that the reaction goes at room temperature and requires only about half of the silver oxide theoretically required when the free acid is used instead of the sodium salt. Precautions must, however, be taken to prevent the solution from becoming alkaline. This is important since on evaporat- Am. Jour. Sc1.—Fourtu Sprizs, Vou. XL, No. 239.—Novemser, 19105, b4 514 Drushel and Holden—H. ydracrylic Esters. ing an alkaline solution of sodium hydraerylate condensation prodwets are formed and but very little sodium hydraerylate is found in the residue. 3. Hydracrylic acid in the absence of mineral acids is easily esterified directly with methyl, ethyl, propyl, isopropyl, isobutyl and isoamyl alcohols by using anhydrous copper sulphate as a dehydrating agent. The esters are obtained in 80 per cent to 90 per cent yields. 4. The hydracrylic esters are colorless liquids with faint but characteristic ethereal odor, mostly easily soluble in water ; all - below isoamyl are heavier than w ater; all are decomposed on boiling at atmospheric pressure, but ‘are distillable without decomposition under diminished pressure. SCIENTIFIC INTELLIGENCE. I. Cwemisrry anp Puysics. 1. The Volumetric Estimation of Lead.—¥. D. Mires has devised a new method for this determination, which depends upon getting the lead into the form of sulphate, decomposing this with hydrogen sulphide, filtering off the precipitate, boiling off the excess of hydrogen sulphide, and finally titrating the sul- phuric acid produced in the reaction. The complete conversion of the usual form of lead sulphate into sulphide requires a special treatment with hydrogen sulphide water in a closed flask with the aid of heat and thorough agitation in the presence of fragments of filter paper. The reaction is complete in the pres- ence of calcium sulphate, but barium sulphate prevents a com- plete conversion when an evaporation with sulphuric acid has been made, so that in its presence the method requires modifica- tion. The presence of considerable quantities of iron interferes with the method and necessitates a double separation. ‘The test analyses given by the author show satisfactory results, and he regards the method as preferable to those already in use in its application to ores, but it is evident that the atomic weight of lead is so high that great care is necessary in order to secure accurate results.—/owr. Chem. Soc., evil, 988. Teles AWE 2. Search for an Alkali-Metal of Higher Atomic Weight than Cesium.—G. P. Baxter has found that pollucite from Paris, Maine, is satisfactorily decomposed by treatment with concen- trated nitric acid. Having obtained about 3} kg. of nearly pure cesium nitrate from this source, he subjected this large quantity of the rare material to an extensive fractional crystallization. It is probable that a metal beyond cesium in the group would pos- sess a less soluble nitrate, but the crystallization gave an end Chemistry and Physics. 515 product of about 3 g. which showed no spectroscopic differences from cesium, and also showed the same atomic weight. At the soluble end the final mother liquor showed the presence of thallium, as well as lithium, sodium, potassium and rubidium. No indication was found of the presence of an unknown element with cesium in pollucite.—Zeitschr. anorgan. u. allgem. Chem., 1915. H. L. W. 3. Experimental Organic Chemistry ; by James F. Norris. 12mo, pp. 315. New York, 1915 (McGraw-Hill Book Co.). First edition.—This book is designed primarily to be used as a labora- tory guide in connection with courses in organic chemistry in which the student follows in the laboratory the subject as devel- oped in the class-room. Directions for experiments illustrating the preparation and chemical properties of all the more important classes of organic compounds are given.. These are very com- plete and cover several subjects which in many laboratory courses receive scant attention, for example, fatty amines, hydroxyacids, carbohydrates and proteins. Special attention is given to labora- tory technique and the handling of small quantities of material. Those properties of a compound which lead to its identification are particularly emphasized. As the companion book for “ The Principles of Organic Chemistry ” by the same author, this book should prove especially valuable. N. A. SHEPARD, 4. Luboratory Experiments in Organic Chemistry ; by E. P. Coox. 12mo, pp. 50. Philadelphia, 1915 (P. Blakiston’s Son & Co.).—These experiments, which are especially designed for use with Stoddard’s “Introduction to Organic Chemistry,” constitute the laboratory course in this subject given at Smith College, and are intended to require five to six hours per week for one semester for their completion. The preparations are representative and the numerous test-tube experiments well chosen to illustrate the chemical behavior of the various compounds, as well as to give considerable practice in laboratory technique. N. A. SHEPARD. 5. The Elements of Physical Chemistry ; by Harry C. Jonrs. 8vo, pp: vil, 672, 4th ed., revised and enlarged. New York, 1915 (Macmillan Co.).—In the fourth edition of this well-known text book, some new material has been added, chiefly, however, at the end of chapters, so that the plan of the book has not been changed. ‘This method at times makes statements from former ’ editions a little misleading. For instance, in a list of the “most recent measurements of thermochemistry” (p. 358) the date of the latest reference is 1904. The additions are well chosen and add very materially to the value of the book. H. W. F. 6. Alcoholometric Tables ; by Str Epwarp Tuorre. 12mo, pp- 91. New York, London, 1915 (Longmans, Green and Co.). —The numbers in these tables are printed in large, clear type ana are very convenient for reference. The tables are an extension of those given in the author’s “ Dictionary of Applied Chemistry.” Table I gives the percentages by weight and vol- ume of ethyl alcohol corresponding to specific gravities at Am, Jour. Sc1.—Fourts Series, Vou, XL, No. 239.—Novemser, 1915, 30 516 Scientific Intelligence. 15°6° ©./15°6° C. Table Il shows the indications of Sikes’ hydrometer and the percentages of British proof spirit, American proof spirit, of ethyl alcohol by weight, and also by volume at ° C. and 15°6° C. Table III compares the indications of Sikes’ hydrometer with those of Russia, Holland, Spain, and Switzer- land. H, L, W. 7. Brief Course in Metallurgical Analysis; by Henry ZinGEL. Large 8vo, pp. 72. Easton, Pa., 1915 (The Chemical Publishing Co.).—This small book has about one-half of its pages left blank for notes. A fairly extensive and well-selected list of analyses is outlined, and references are made to the best works on the subject. It appears, however, that the directions given are often not full enough for a student’s use, that the language is frequently ambiguous, and that some of the advice given is not the best. H. L. W. 8. Characteristics of Long Direct-Current Arcs.—In an earlier paper (see this Journal, vol. xxxvill, p. 362, 1914), W GROTRIAN used a direct-current dynamo capable of generating 3 amperes at 5000 volts in demonstrating experimentally that the so-called cyanogen bands owe their origin to nitrogen alone and not to the presence of carbonin the arc. In a more recent paper he has recorded the following interesting results obtained with essentially the same assemblage of apparatus. The electrical characteristics of arcs, whose lengths were varied from a few millimeters to 50 centimeters, were obtained with carbon, copper, and iron electrodes in air, nitrogen, oxygen, car- bon dioxide, hydrogen, and water vapor. It was found that the gas was chiefly responsible for the course of the characteristics while the material of the electrodes had only a very slight influence on the curves. In all cases where the axis of the are consisted in a well- defined Juminous band the Ayrton formula Viz=a+ Bl + yi + Oli reproduced the experimental data very satisfactorily. [| V= voltage across arc, 7 = current, and 7= length of are.] The following numerical results were deduced Hikdebh Vi = 32°62 + 624 + 11:4U carbon dioxide Vz = —90 + 731 + 802 + 11°71i hydrogen Vi = 180 + 2227 + 2902 water vapor Vi = 1852 + 200% + 1007 As would be expected, the temperature of the core of the ares was so high that platinum and other metals liquefied in it at once. On the contrary, the little rods of Nernst glowers with- stood the high temperature without undergoing appreciable modification and hence they could be used as the terminals of exploring circuits, that is, as electrical sounds. In the air-are it Chemistry and Physics. 51? was found that the potential gradient along the axis of the are was constant in all parts of the positive colamn and independent of the length of this region. The potential-fall per centimeter decreased hyperbolically as the current strength increased. The anode and cathode drops were determined directly. The anode- fall was independent of the Jength of the arc but decreased in the interval between 01 and 3 amperes from 150 to 30 volts, the course of the curve being partly dependent upon the material of the anode. The cathode-fall amounted to 20 volts and was inde- pendent of the arc length and of the current. It depended very slightly upon the nature of the cathode. Direct measurements of the cross-section of the arc in carbon dioxide brought out the fact that the current density increased linearly with the field strength. In the last part of the paper the author advances a theory to account for the conductivity in the are. It is based primarily on the ionization produced by the collision of electrons whose chaotic thermal velocity exceeds the inferior limit necessary for ionization. By the addition of certain auxiliary hypotheses regarding both the influence of radiation upon the conductivity and the nature of recombination, the theory is developed suffi- ciently to give the order of magnitude of the observed conduc- tivity. Such details as the limiting of the cross-section of the are, the course and position of the characteristic in different gases, and the dependence of the characteristic upon pressure are readily explained qualitatively on this theory.—Ann. d. Physik, vol. xlvul, pp. 141-196, June, 1915. it, So We 9. Prinzipien der Atomdynamik; von Dr. J. Stark. IIL Teil. Die Elektrizitat in chemischen Atom. Pp. xvi, 280, with 94 figures. Leipzig, 1915 (S. Hirzel).—As stated in the preface to Part I (see this Journal, vol. xxxii, p. 67, 1911), the atomic structure of matter would be the theme of the third and last Part of the book. By the aid of his valence hypothesis concern- ing the electrical structure of the surfaces of chemical atoms the author has succeeded in correlating and accounting for a large number of physical and chemical phenomena, A general idea of the scope and contents of the present volume may be obtained from the titles of the five chapters, which are: “ Grundlagen der chemischen Atomistik, Gleichgewicht der innermolekularen Bindung, Reaktion der innermolekularen Bindung, Zwischen- molekulare Bindung,” and “ Optik der Valenzfelder chemischer Atome.” H. S. U. 10. Zen Years’ Work of a Mountain Observatory : by GrorGE Ertery Hare. Pp. 99, 66 illustrations. Publication No. 235, Carnegie Institution of Washington, 1915.—This little book con- tains excellent photographs of the larger instrements which have been installed in the Mount Wilson Solar Observatory and in the associated physical laboratory at Pasadena, together with lucid, concise accounts of the important results obtained with each type of apparatus during the last decade. Good progress in the con- 518 Scientific Intelligence. struction of the large engine for ruling diffraction gratings and of the 100-inch reflector is indicated. It is estimated that this gigantic telescope will reveal about 100,000,000 faint stars hitherto unobservable. Among the problems under investigation may be mentioned solar meteorology, the brightness, motions, distances, and evolntion of the stars, the scattering of light in space, and the magnitude and structure of the universe. All of the figures are good, the photographs of spiral nebulz being especially attrac- tive. The author’s style and the subjects presented are of such a nature that the reviewer could not lay the book aside until he had read it through. H. 8. U. 11. The Electrical Nature of Matter and Radioactivity ; by Harry C. Jones. Thirdedition. Pp. ix, 212. New York, 1915 (D. Van Nostrand Co.).—A careful comparison, page by page, of - the first edition (see this Journal, vol. xxi, p. 465, 1906) with the third shows that the original text has been kept unaltered as far as possible. The comparatively slight changes introduced were necessitated by the conscientious and successful endeavor to bring the experimental data and theoretical considerations up to date. The number of chapters has been reduced by unity by incorporat- ing some of the material from chapter XVIII of the first edition in “chapter XVII of the third. The total number of pages has been kept constant by abbreviating the index. It is pleasing to note that this admirably written, semi-popular text has been so well appreciated as to cause it to pass through three editions in a relatively short period of time. He 8) 10 12. The Book of Wireless ;by A. FREDERICK Couuins. Pp. xv, 222, with 219 figures. New York, 1915 (D. Appleton and Co.). —In this extremely elementary book the author has endeavored to anticipate and answer all reasonable questions which may arise in the minds of boys or young men who are attempting to assem- ble and operate either bought or home-made wireless-telegraph sets. He even goes so far as to say: “But if you should have any trouble and if any questions should come up which puzzle you, if you will write to me, I shall gladly do all I can to help ou.” : The text is divided into three Parts having the following titles : I. “ A Small Wireless Outfit ” (five chapters), II. “A Long Dis- tance Wireless Set” (four chapters), and III, “Induction Coil, Transformer and Electrolytic Interruptor” (three chapters). The diagrams are large and clear, and the linear dimensions are indi- cated whenever necessary. At the end of each chapter are given itemized cost lists of parts for both commercial and home-made sets. The last chapter of the volume relates to government rules and regulations, examinations, etc. This is followed by several appendices on wire gauges, drills, types of wrials, etc , a glossary, and an index. The subject matter seems to be very well pre- sented to meet the requirements of the readers for whom it was written. H. S. U. Geology and Mineralogy. 519 13. Plane Geometry; by C. I. Parmer and D. P. Taynror. Hdited by G. W. Myers. Pp. v, 277. Chicago, 1915 (Scott, Foresman and Co.).—‘ The main aim of the authors in the prepara- tion of this text has been to approach abstract reasoning by a method that is natural and comprehensible to the youthful mind, and to vitalize the subject-matter,—making it both interesting and useful through a wide range of practical applications.” The salient points of the plans which have been adopted for the attain- ment of this end may be outlined as follows: (a) The experi- mental or inductive method is used to some extent throughout the book, but it is especially prominent in the earlier pages. After the student has acquired full comprehension of the funda- mental propositions he is required to demonstrate them in the classical, formal, deductive manner. (>) Actual work in geome- try 1s begun at once without the usual array of definitions, axioms, and principles. In general, formal definitions and axioms are placed in the text at the point where they are first needed, but not until after the student is prepared to appreciate them by hav- ing had a partial survey of the subject. (c) ‘The exercises consti- tute one of the most prominent features of the buok. They are very numerous (1239), carefully graded, both practical and theo- retical, and interesting. (d) The proofs of the more difficult as well as of the earlier theorems are given in full, but, as the work progresses, a gradual elimination of proofs is made thereby throw- ing the student more and more on his own resources. (¢) The work has been planned so that practical applications may be made as early in the course as possible. To this end, the chapter on areas precedes the one on similarity because it furnishes a great variety of applications to matters of everyday life. Special care has also been taken to make the pages as useful and attractive as possible. For example, Gothic type and italics are judiciously employed, the line diagrams and half-tone illus- trations are neat and clear, a combination linear scale and pro- tractor may be found in a pocket on the inside of the cover, formulas for reference are tabulated, and an index is appended. Unquestionably the book merits the careful consideration of all progressive teachers of elementary plane geometry. Jeb sali Il. Gerotogy anp Mineraoey. 1. Publications of the United States Geological Survey, Grorce Oris Smith, Director.—Recent publications of the U.S. Geological Survey are noted in the following list (continued from pp. 85-87, July, 1915): Monocrarnu, Volume LIII. The Pleistocene of Indiana and Michigan and the history of the Great Lakes; by Frank Leverett and Frans B. Taytor. Pp. 529; 32 pls., 15 figs. PROFESSIONAL ParErs.—No. 87 Geology and Ore Deposits of Copper Mountain and Kasaan Peninsula, ‘Alaska; by CHARLES W. Wrieut. Pp. 110; 22 pls. 11 figs. 520 Scientific Intelligence. No. 95. Shorter Contributions to General Geology. B. Kocene glacial deposits in Southwestern Colorado; by WALLACE W. Arwoop. C. Relation of the Cretaceous formations to the Rocky Mountains in Colorado and New Mexico; by Wits T. Ler. D. An ancient volcanic eruption in the Upper -Yukon Basin; by StepHEN R. Capps. Buuetins.—No. 544. Fauna of the Wewoka Formation of Oklahoma; by Greorer H. Girry. Pp. 353; 35 pls. Nos. 566, 569. Results of Spirit Leveling. R. B. Marsuatt, Chief Geographer. No. 566. Utah, 1897 tv 1914. Pp. 77. No. 569. Iowa, 1896 to 1913. Pp. 126; 1 plate. No. 587. Geology and Mineral Resources of Kenai Peninsula, Alaska; by G. C. Marrin, B. L. Jonnson and U. 8. Grant. Pp. 243; 38 pls., 43 figs. No. 591. Analyses of rocks and minerals from the Labora- tory of the U.S. Geological Survey 1880 to 1914. Tabulated by F. W. Crarxn, Chief Chemist. Pp. 376. No. 593. The Fauna of the Batesville Sandstone of Northern Arkansas; by Grorcr H. Girtry. Pp. 170; 11 plates. No. 601. Geology and Mineral Deposits of the National Min- ing District, Nevada; by WaLtpemaR LinpGren. Pp. 58; 8 pls, 8 figs. No. 602. Anticlinal Structure in parts of Cotton and Jeffer- son Counties, Oklahoma ; by Carrot, H. WEGEMANN. Pp. 108; 5 pls. Nos. 611-614. Guidebook of the Western United States. No. 611, Part A. The Northern Pacific Route, with a side trip to Yellowstone Park; by Marius R. Campsett, and others. Pp.-212; route map (in 27 sheets), 27 pls., 38 figs. No. 612, - Part B. The Overland Route, with a side trip to Yellowstone Park; by Wiiuis T. Ler, R. W. Sronz, H. 8. Gare and others. Pp. 244; 25 maps, 49 pls., 20 figs. No. 613, PartC. TheSanta Fe Route, with a side trip to the Grand Canyon of the Colorado ; by N. H. Darron and others. Pp. 194. No. 614, Part D. The Shasta ‘Route and Coast Line; by J. S. Dit_ER and others. pe 42) ive 620-B. Nitrate Deposits in Southern Idaho and Eastern Oregon; by G. R. Mansrrerp. Pp. 44; 2 pls, 2 figs. Warer-Suprty Papers. No. 340-K. Stream-gaging Sta- tions. Part XI. Pacific. Coast Basins in California ; compiled by B. D. Woop. Pp. xx, iv, 131-146. No. 342. Surface Water Supply of the Yukon-Tanana Region, Alaska; by C. E. Ertswortu and R. W. Davenrort. Pp. “B43; 13 pls., 5 figs. Nos. 356, 357. Surface Water Supply of the United States 1913, N.C. Grover, Chief Hydraulic Engineer. No. 356. Part VI. Missouri River Basin. Pp. ae 2 plates. No. 357. Lower Mississippi River Basin. Pp. 86; 2 pls. No. 358. Water Resources of the Rio Grande Basin 1888- 1913; by R. Fottanssee and H. J. Dean. Also Surface Water Geology and Mineralogy. 521 Supply of the United States, Part VIII, Western Gulf of Mexico Basins. Pp. 725; 3 pls. Nos. 375, B, C, D, E. Contributions to the Hydrology of the United States, 1915. Pp. 51-130. 2. Relation of the Cretaceous Formations to the Rocky Moun- tains in Colorado and New Mexico ; by Wiis T. Lez. Prof. Paper 95-C, U.S. Geol. Surv., 1915, pp. 27-58, pl. V, text figs. 12-22.—A carefully wrought out paper showing that the present Rocky Mountains in Colorado, New Mexico, and Wyoming did not exist as such in late Mesozoic time, and that these areas were covered with many thousands of feet of Cretaceous strata. There were no granitic island masses in the Cretaceous sea of Colorado, as held by some geologists. These results have a direct bearing on the problem of the Cretaceous-Tertiary boundary in the Rocky Mountain region. The author says: “Certain con- glomerates that rest unconformably on Cretaceous beds are regarded as basal Tertiary by some geologists and as Cretaceous by others. These conglomerates contain great numbers of peb- bles of crystalline and metamorphic rocks such as are now found in the mountains, and they are so distributed as to prove that they were derived from the present mountainous areas. Inasmuch as the Cretaceous formations were originally continuous over the site of these mountains, it follows that there must have been uplift and erosion sufficient to remove them and to reach the pre-Creta- ceous rocks before the materials for the conglomerates could be obtained. In the Rocky Mountain region of Colorado and New Mexico all deposits above these conglomerates are of the nonma- rine type that characterizes the undisputed Tertiary formations of the same regions” (p. 57). He then concludes: “ It naturally follows that the conglomerates and other sediments derived by erosion from the newly uplifted mountains—such as those of the Denver, Arapahoe, Dawson, Raton, and related formations— belong to the Tertiary system” (p. 58), Cc. 8. 3. Conceptions regarding the American Devonic; by Joun M. Crarxe. N. Y. State Mus., Bull. 177, 1915, pp. 115-133.— An excellent paper, written in Clarke’s characteristic style and prepared to commemorate the seventieth birthday of the Euro- pean authority on the Devonian, Professor Emanuel Kayser. it brings within a small compass much of our knowledge of the American Devonian sequence and distribution, and its correlation with the occurrences of the Devonian in the rest of the world. C. 6: 4, Fauna of the San Pablo Group of Middle California ; by Bruce L. Crarx. Univ. Calif. Pub., Bull. Dept. Geology, vol. 8, No. 22, 1915, pp. 385-572, pls. 42-71.—This work brings together all that is known of the U pper Miccene strata and fauna of Middle California. There are 165 species, and of these 135 (72 new) are determinable molluscs ; 32 are living (21 per cent). C..S: 5. The Oretaceous Sea in Alberta; by D. B. Dowtine. Trans. Royal Soc. Canada, ser. III, vol. ix, 1915, pp. 27-42, 522 Scientific Intelligence. | 11 pls.—This good paper summarizes what is known of the sue- cession, physical character, and distribution of the various forma- tions of the Cretaceous period in Alberta. Of great value are the ten paleogeographic maps showing the extent of the western land, the marine, brackish-water, and continental deposits. It is to be hoped that the author will continue his studies northward and present his conclusions as to the extent of the various forma- tions into the Arctic Ocean. eos 6. Wabuna Iron Ore of Newfoundland ; by AtBERT Orion Hayes. Canada, Geol. Surv., Memoir 78, 1915, pp. 163, 28 pls., 4 text figs.—This excellent report shows that the bedded iron ores of eastern Newfoundland (Bell island) are of early Ordovician age (Arenig), though they have long been considered of Silurian (Clinton) time. The greater part of the work is devoted to the genesis of the iron ores, which were deposited in a shallow agi- tated sea devoid of limestones and igneous rocks. The little lime present in the ore (2°5 per cent) occurs in the form of fossils. ‘The phosphorus of the ore is also derived from the remains of organic life preserved in it. No evidence of diagenetic trans- formation from an original oolitic limestone to an oolitic iron ore has been found and no concentration of iron has occurred since the deposition of these ferruginous sediments. They are primary bedded iron ore deposits, mined to-day in essentially the same condition except for induration, faulting, and the addition of small amounts of secondary calcite and quartz in fault cracks, as when they were laid down” (page 93). Cree 7. The Yukon-Alaska International Boundary, between Por- cupine and Yukon Rivers; by D. D. Catrnes. Canada, Geol. Surv., Memoir 67, 1914, pp. 161, 16 pls., 2 text figs, 2 maps.— The author describes the living flora and fauna, the topography, geology, and sedimentary sequence of a narrow belt between the Yukon and Porcupine rivers along the 141st meridian on the International Boundary between Canada and the United States. The greater portion of the Paleozoic sequence is well developed here, and as a rule the deposits are limestones and dolomites. Doctor Cairnes is to be congratulated on having brought out con- siderable of the detailed sedimentary sequence. Cc. 8. 8. The Ordovician Rocks of Lake Timiskaming ; by M. Y. Wixuiams. Canada, Geol. Surv., Mus. Bull. No. 17, 1915, pp. 8, 1 pl. 1 text fig.—The author records the first finding of Black River limestones with fossils beneath this Silurian remnant of once widely distributed formations. Cc. §. 9. Structural Relations of the Pre-Cambrian and Paleozoic Rocks North of the Ottawa and St. Lawrence Valleys ; by EK. M. Kinpiz and L. D. Burtine. Canada, Geological Survey, Mus. Bull. No. 18, 1915, pp. 23, 2 pls., 6 text figs.—The authors point out that the abrupt elevation of the Laurentian plateau (“ Lau- rentian Plateau escarpment”) above the horizontal lowlands of the Ottawa and St. Lawrence valleys represents a fault line scarp. The early Paleozoic strata are as a rule here faulted down, and Miscellaneous Intelligence. 523 in the lowlands have been preserved from denudation; accord- ingly “the Palzeozoic seas extended very widely, if not completely, over the Laurentian upland southeast and east of Hudson bay.” The paper is a valuable contribution toward a better understand- ing of the paleogeography of Ordovician, Silurian, and Devonian times. 6.8: 10. Geology of Franklin County; by A. M. Mitumr. Ky. Geol. Surv., 4th series, vol. 2, 1914, pp. 7-87.—This excellent report maps and describes in great detail the various Ordovician formations of Franklin County, central Kentucky. The Mohawk- jan series is divided into nine members, and the Cincinnatian into five. Commonly the Eden is regarded as at the base of the Cin- cinnatian series, but in this report the Cynthiana is the founda- tion, a formation unknown in the standard section of New York. The soils of the county are described in an additional paper (pages 89-144) by 8. C. Jones. C. Ss. 11. Revision of the Tertiary Mollusca of New Zealand ; by Hewry Suter. New Zealand Geol. Surv., Palzontological Bull. No. 2 (Part: I), 1914, pp. 64, 17 pls.; No. 3 (Part II), 1915, pp. 69, 9 pls.—In these publications all of the new species defined many years ago by Captain F. W. Hutton and others are revised in the light of modern knowledge of the Recent Mollusca. They comprise a first step toward a revision of the Tertiary Mollusca of New Zealand. CAS: 12. Third Appendix to the Sixth Edition of Danas System of Mineralogy ; by Wuitam E. Forp. Completing the work to 1915. Pp. xii, 87. New York, 1915 (John Wiley & Sons).— This third appendix to Dana’s System has been prepared by Prof. W. E. Ford, to whom the science owes also the completion of Appendix II, seven years ago. Notwithstanding the fact that mineralogy is sometimes regarded as the most nearly complete of all the sciences, descriptions of new occurrences still go on actively and no fewer than one hundred and eighty new names have been added between 1908 and 1915; of these about one-third are regarded by Professor Ford as probably well established species. This Appendix follows the lines laid down in its predecessors, but special attention is given to the literature of X-rays and crystal structure. III. Miscerztanrovs Scientiric INTELLIGENCE. 1. Publications of the Carnegie Institution of Washington.— Recent publications of the Carnegie Institution are noted in the following list (continued from p. 94, July, 1915): No. 85. Index of Economic Material in Documents of the States of the United States; New Jersey (1789-1904). Prepared for the Department of Economies and Sociology by ADELAIDE R. Hasse. Pp. 705, 4to. No. 175. Researches of the Department of Terrestrial Magne- tism. Volume II. Land magnetic observations, 1911-1913, and reports on special researches ; by L. A. Bauer and J. A, FLemine. Pp. v, 278; 13 pls., 9 figs. 524. Scientific Intelligence. No. 207. The Permo-Carboniferous Red Beds of North America and their Vertebrate Fauna; by E. C. Case. Pp. iii, 176; 24 pls., 50 figs. To be noticed later. No. 209. Acidity and gas interchange in Cacti; by Hurpert M. Ricuarps. Pp. 107. ; Nos. 211, 212. Papers from the Department of Marine Biology. No. 211, vol. VII. Homing and related activities of birds; by J. B. Warson and K. S. Lasutzey. The acquisition of skill in archery ; by K. 8. Lasutry. Pp. 128; 9 pls., 19 figs. No. 212, vol. VIII. Contains nine papers by F. A. Ports, H. L. Crarx, Grace Mepzs and others. Pp. 256; 23 pls., 73 figs. Nos. 221, 222. Contributions to Embryology. No. 221, vol. I, No. 1. On the fate of the human embryo in tubal pregnancy ; by F. P. Maru. Pp. 103, 4to; 11 pls. 24 figs. No. 222, Nos. 2-6. Pp. 108, 4to; 10 pls., 25 figs. Contains the following : Descrip- tion of two young twin human embryos with 17-19 paired somites ; by James C. Warr. An anomaly of the thoracic duct with a bearing on the embryology of the lymphatic system; by Etior R. Crarx. Fields, graphs, and other data on fetal growth ; by A. W. Meyer. The corpus luteum of pregnancy, as it is in swine; by GrorcE W. Corner. ‘Transitory cavities in the corpus striatum of the human embryo; by Cuarues R. Essicx. OBITUARY. M. Jean Henni Fasre, the eminent entomologist, known the world over for his delightful descriptions of the habits and lives of insects, died on October 11 at the advanced age of ninety-two. He has been called a poet as well as a naturalist and the interest- ing story of his life has been well told in a recent volume by C. V. Legros (see vol. xxxvil, 284). Dr. Turovor Boveri, the eminent German biologist, since 1893 professor of comparative anatomy and zodlogy at Wiirzburg, died recently at the age of fifty-three years. Henry Gwyn Jerrreys Mosevey was killed in action at the Dardanelles on August 10 at the age of twenty-seven years. He was the son of Professor H. N. Moseley of Oxford, and though so young had already given proof of his rare intellectual gifts ; he was particularly interested in physical investigations and had done important work on the X-ray spectra of rare earths and the study of erystal structure by X-rays. Proressor D. T. Gwynne-VauGHan, the distinguished plant anatomist of Reading, England, died on September 4 at the age of forty-four years. Witiiam Watson, secretary of the American Academy of Arts and Sciences since 1884 and an authority on technical education, died in Boston on September 30 in his 82d year. Aveustus Jay Du Bots, professor of civil engineering in the Sheftield Scientific School of Yale University, died on October 19 in his sixty-seventh year. ys NATURAL SCIENCE ESTABLISHMENT A Supply-House for Scientific Material. Founded 1862. Incorporated 1890. - A few of our recent circulars in the various departments: : Geology: J-3. Genetic Collection of Rocks and Rock- _ forming Minerals. Mineralogy: J-109. Blowpipe Collections. J-74. Meteor- ites. Paleontology: J-134. Complete Trilobites. J-115. Collec- tions.. J-140. Restorations of Extinct Arthropods. Entomology: J-30. Supplies. J-125. Life Histories J-128. Live Pupae. ~ Zoology: J-116. Mate .l for Dissection. J-26. Compara tive Osteology. u-v4. Casts of Reptiles, ete. Microscope Slides: J-135. Bacteria Slides. Taxidermy: J-138. Bird Skins. J-139. Mammal Skins. Human Anatomy: J-16. Skeletons and Models. General: J-100. List of Catalogues and Circulars. Wards Natural Science Establishment 84-102 College Ave., Rochester, N. Y., U.S. A. IMER & AMEND Complete Laboratory Furnishers Chemical Apparatus, Balances, etc. C. P. and T. P. Chemicals and Reagents Best Hammered Platinum Ware, Blowpipe Outfits and Assay Goods WE CARRY A LARCE STOCK QF MINERALS FOR BLOWPIPE WORK, ETC. EST’B - 1851 _203-211- THIRD -AVE NEW-YORK- CITY. CONTENTS Art. XXXII.—Experimental Studies and Observatiaas on Ice Structure; by O. D. von EneELy XXXIII.—A Mounted Specimen of Dimetrodon incisiyus Cope, in the University of Michigan; by E. C. Casu -- XXXIV.—A Fossil Ruminant from Rock Creek, Texas, Preptoceras mayfieldi sp. nov.; by E. L. Trox Ext XXXV.—The Separation and Bistasenien of Aluminium and Beryllium by the Use of Acetyl Chloride m Acetone ; by H. D. Minnie XXXVI.—On the Interferences of Crossed Spectra and on Trains of Beating Light Waves; by C. Barus XXXVII.—The Brandywine Formation of the Middle Atlantic Coastal Plain ; by W.-B. Clark XXXVIII.—On Two Burners for the Demonstration and ~ Study of Flame Spectra; by P. E. Brownine _-.. _--- XXXIX.—On the Preparation of Glycocoll and Diethyl Car- bonate ; by W. A. Drusuet and D. R. Knape XL.—On the Preparation and Properties of Hydracrylic Esters ; by W. A. Drusnex and W. H. T. HoLbEneeeaane SCIENTIFIC INTELLIGENCE. Chemistry and Physics—Volumetric Estimation of Lead, F. D. Mires; Search for an Alkali-Metal of Higher Atomic Weight than Cesium, G. P. Baxter, 514.—HExperimental Organic Chemistry, J. F. Norris: Labora- tory Experiments in Organic Chemistry, E. P. Cook: Elements of Physical Chemistry, H. C. Jonrs: Alcoholometric Tables, E. THorpn, 515.—Brief Course in Metallurgical Analysis, H. Zrmcmu: Characteristics of Long Direct-Current Ares, W. Grotrian, 516.—Prinzipien der Atomdynamik, J. STARK: Ten Years’ Work of a Mountain Observatory, G. E. HALE, 517. Electrical Nature of Matter and Radioactivity, H. C. Jonms: Book of Wire- — less, A. F. Cox~zins, 518.—Plane Geometry, C. I. Patmmr and D. P. TAYLOR, 519. Geology and Mineralogy—Publications of the United States Geological Sur- vey, G. O. Smits, 519.—Relation of the Cretaceous Formations to the Rocky Mountains in Colorado and New Mexico, W. T.LeE: Conceptions regarding the American Devonic, J. M. CLarxe: Fauna of the San Pablo Group of Middle California, B. L. CLarK: Cretaceous Sea in Alberta, D. B. Downe, 521.—Wabana Iron Ore of Newfoundland, A. O. Hays: Yukon- Alaska International Boundary, between Porcupine and Yukon Rivers, D. D. Carryzes: Ordovician Rocks of Lake Timiskaming, M. Y. WiLitams: Structural Relations of Pre-Cambrian and Palzeozoic Rocks North of the Ottawa and St. Lawrence Valleys, E. M. KinpLE and L. D. Buriine, 522.— Geology of Franklin County, A. M. Miuuer: Revision of the Tertiary Mol- lusca of New Zealand, H. Sutpr: Third Appendix to the Sixth Edition of Dana’s System of Mineralogy, 028. Miscellaneous Scientific Intelligence—Publications of the Carneete Institu- tion of Washington, 523. Obituary—J. H. Fasre: T. Bovert: H. G. J. Moseztmy: D. T. GWYNNE- Vaucuan: W. WArson: A. J. DuBots, 524. Library, U.S. Nat. Museum. VOL. XI. Established by BENJAMIN SILLIMAN in 1818. 4 THE ee. - Ss | ; | b Eprror: EDWARD S. DANA. | x eee Be a ASSOCIATE EDITORS _ || Proressons GEORGE L. GOODALE, JOHN TROWBRIDGE, | a W. G. FARLOW anv WM. M. DAVIS, or Camsricz, | || Prorrssors ADDISON E. VERRILL, HORACE L. WELLS, s LOUIS V. PIRSSON, HERBERT E. GREGORY 3 ; and HORACE S. UHLER, or New Haven, | || __ Prorsssorn HENRY S. WILLIAMS, or Irnaca, . 1 Prorrsson JOSEPH S. AMES, or Batrimore, f ia Mr. J. S. DILLER, or Wasuinaton. a FOURTH SERIES a: [ee . i Ee _ VOL. XL_[WHOLE NUMBER, CXC]. “a AN | ag No. 240—DECEMBER, 1946)" | NEW HAVEN, CONNECTICUT—21! | f9o.Eo* | THE TUTTLE, MOREHOUSE & TAYLOR CO., PRINTERS, 123 ippMPLE STREET, | shed monthly. Six dollars per year, in advance. $6.40 to countries in the | m ; $6.25 to Canada. Remittances should be made either by money orders, | letters, or bank checks (preferably on New York banks). | New and Rare Minerals EUXENITE, near Tritriva, Antsirabe, Madagascar. Sharply developed crystals; small and large. ‘de to $5.00. : AMPANGABHITE, nr, Tritriva, Antsirabe, Madagascar. Small and large crystals; new mineral. $1.00 to $8.00. BETAFITE, Betafe, ESD AOTOSY, and Antsirabe, Madagascar. ‘Tac to $5.00. MANGANOSITE, Franklin Furnace, New Jersey. New discovery. $1.00 to $2.00. HODGKINSONITE) Rania Furnace, New Jersey. $1.00 to $5.00. Just received a large collection of excellent specimens suitable for work- ing material. This is a fine opportunity,to secure rare minerals at low. figures. Kindly send for list. Prices range from 10c to $2.00. Important to Collectors If you want to sell your collection or some of your specimens, you will now have an excellent opportunity. We are making a specialty of selling collections on a fair commission basis. We have the best facilities and are always in contact with the greatest collectors, who are willing to pay a good price for choice minerals and gems, but they must be first class. The min- erals, your property, are just as safe with us as at your place. We are lo- cated in the financial district in a fire-proof building and have vaults for the valuable articles. Desire to state that the specimens and collections we have sold in the past always gave perfect satisfaction. % We are always in the market for fine showy crystallized minerals and pay the highest prices for them, The minerals must be sent on approval — and when not accepted will be returned promptly. Letters must accom- pany shipments, giving the lowest possible price on same, so as to save time and labor. Are You Interested in Gems? We carry a full line of precious and semi-precious stones of the best quality in all sizes, both cut and in the rough, suitable for jewelry and specimens. Also the synthetic gems of the best quality, including rubies; pink, yellow, blue and ae SED ne Scientific emeralds of beautiful rich color. The above can be secured here at very Nh prices, much lower than elsewhere. . Selections sent on approval. ALBERT H. PETEREIT 81-83 Fulton St, New York City THE AMERICAN JOURNAL OF SCIENCE [FOURTH SERIES.] —_~++>—__- Arr. XLI—A Metallographic Description of Some Ancient Peruvian Bronzes from Machu Picchu; by C. H. Maruzw- SON. [Contribution from the Hammond Laboratory of the Sheffield Scientific School, Yale University.] IntTRODUCTION. I am indebted to Professor Hiram Bingham for an oppor- tunity to examine the ancient Peruvian bronzes collected at Machu Picchu by the National Geographic Society-Yale Uni- versity Peruvian Expedition of 1912 (1).* The structures exhibited by objects of this character are of interest not only to metallographists but also in archeological circles, inasmuch as certain definite facts concerning the methods used in shaping the objects can be learned from the structural examination. Such work, however, necessitates some mutilation of the speci- mens, which doubtless accounts for the scarcity of metallo- graphic work devoted to rare objects. Collectors frequently submit their specimens for chemical analysis, since this merely involves the preparation of a small sample by drilling, while the scope of the forthcoming information is known in advance and the undertaking can be intelligently provided for. Metal- lographic examination, on the other hand, may yield much or little information and generally necessitates subsequent restora- tion of the object. As to the character of the information likely to be obtained by the use of metallographic methods, the following general * For references here and on later pages, see the Bibliography at the end of this article. Am. Jour. Sct.—Fovurta Srerizs, Vou, XL, No, 240—DzcumsBerr, 1915. 526 C. H. Mathewson—Metallographic Description statements may be of service to those who are unfamiliar with this class of work :* 1. An unaltered cast structure can es recognized at once and a qualitative distinction between rapidly cooled metal, e. g., metal cast in a small unit, and slowly cooled metal, e. g., metal from a casting of considerable size, can be made. 2. An annealed casting can be distinguished from an unannealed casting, but such annealing would serve no useful purpose in small bronzes of low tin-content and, in any ease, would hardly be undertaken except as a preliminary to other operations resulting in more or less complete obliteration of the casting structure. 3. Mechanical alteration of a cast piece, or of a piece with more extended but not necessarily known past history, by rolling, hammering, etc., can be detected with certainty pro- vided the piece has been worked with considerable intensity. A few gentle blows with the hammer, a bend, or a twist, would not alter the visible structure of the metal, but mechanical treatment resulting in pronounced change of form gives rise to characteristic structural changes as well. ‘These effects do not, however, serve to differentiate at all sharply between dif- ferent forms of mechanical treatment, whether by rolling, hammering, etc. It seems necessary to rely mainly upon sur- face observations for evidence of this character. Under the influence of mechanical treatment, the metal stiffens and gradu- ally loses its ductility until, eventually, further change of form cannot be brought about without fracture. The precise nature of these changes cannot be detected by observations under the microscope. Certain coincident effects, such as elongation and etching peculiarities of the crystalline grains, serve, together with physical tests, to establish the worked condition of the metal. 4, Annealing restores the ductility by producing an entirely new growth of grain in which certain unfailing characteristics may be recognized. Thus, a recrystallized bronze can be recog- nized at once and the evidence of previous mechanical treatment, with following anneal, is thereupon complete. The nature of the recrystallization, whether fine- or coarse-grained, localized or evenly distributed, incipient or sufficiently developed to entirely obliterate the pre-existing structure, depends upon the compo- sition of the metal, the nature, intensity and distribution of the mechanical treatment, as well as the temperature and dura- * Ample discussion and illustration of these statements will be given in ensuing portions of this paper.. of Bronzes from Machu Picchu, Peru. 527 tion of the anneal. Fortunately, careful study of the recrys- tallization phenomena in a particular group of alloys, such as the present group of alpha-bronzes (see discussion of constitu- tion, p. 540), leads to certain associations and generalizations so that the treatment resulting in a given structure may frequently be specified without the too free admission of bewil- dering assumptions and conditional premises. Certain short- comings must, however, be admitted. For example, a ductile object may be lengthened at will by intensive rolling, drawing, or hammering, if an adequate number of intermediate soften- ings, or annealing, operations are conducted. The final struc- ture and physical properties are determined by the last co-ordinated draft and anneal unless unusually light mechanical or annealing treatment has prevailed. No knowledge of pre- ceding operations can ordinarily be derived from the structural characteristics at this point. If the duration of the last anneal is approximately known, the temperature region of annealing may be estimated from the mean grain size. Some distinctions can be made without regarding the time factor, e. g., an anneal at bright red heat can be distinguished from one at nascent red, since the coarse grain produced in a very few minutes at the higher temperatures is not much affected by longer exposure and cannot be duplicated by even a very prolonged anneal at the lower temperature. It has thus been possible to fairly esti- mate the temperature which must have been reached in the finishing anneal of some of these old bronzes. While, in completely recrystallized structures, the grain characteristics give no indication of the total extension of the piece or the number of stages in the process, partially recrys- tallized structures, in which traces of the cast condition are apparent, indicate that the finished piece corresponds closely in form to the original castings. Thus, a number of the bronzes examined were obviously cast in the rough and wrought into shape, the process requiring a moderate amount of work and a limited number of annealing operations; perhaps only one. 5. Drastic treatment whereby the object, by successive drafts and anneals, suffers manifold extension in one direction or another may sometimes be recognized by characteristic migrations of insoluble impurities which have remained sensibly unaffected during the heat treatment. Such evidence is purely qualitative, since the extension of individual units within the moving mass is not proportional to the extension of the piece — ——-—_— —— 528 O. H. Mathewson—Metallographic Description as a whole. Most of the bronzes contain sulphur in small amount which, from the chemical relationships involved, must occur in the form of cuprous sulphide. In the thin, flat bronze knives of the collection, the particles of cuprous sulphide are elongated to a degree attainable only through several successive drawing and annealing operations. A copper knife of similar form contains in the neighborhood of one per cent of cuprous oxide which, in the broader and thicker parts of the specimen, occurs in normal eutectic form as small globules characteristi- cally grouped throughout the copper matrix. In the intensively worked parts of the specimen, notably the blade, each group of oxide particles has been dragged out into a continuous train which gives the appearance of a dotted line in the photo- micrograph (fig. 71). 6. Positive identification of hot-worked metal is not always possible, since, in hot-working at effective temperatures, the deformation is followed by recrystallization as in the anneal of cold-worked metal. Variables of the same character affect the result. In hot-working, however, recrystallization always takes place at temperatures* which would produce relatively coarse. grain if time enough were allowed. ‘Thus, the grain size is determined mainly by the time available for recrystallization between the blows, or after working has ceased. With small specimens, this is almost invariably brief, since the small mass of metal cools rapidly. Upon re-heating, a coarser grain will develop a relatively coarse grain while other portions are work the whole mass of metal simultaneously, attention will first be directed to some particular portion which will later develop. Since, with primitive facilities, it is impossible to being worked. This results in a highly non-uniform conglom- erate. Hot working is, therefore, indicated where coarse and fine grains are found closely associated and where neighboring parts of the specimen exhibit wide variation in grain size. The foregoing statements will indicate in a general way what manner of conclusions may be drawn from metallographic examination as applied to metal of sensibly uniform composi- tion throughout, viz., in the case of pure metals or solid solu- tions. Some of these statements will be elaborated further on * Where both hot-working and cold-working are permissible, as in the case of these alpha bronzes, the former is undertaken in order to avoid the necessity of annealing between stages of intensive working, i. e., the metal stays soft, owing to rapid recrystallization and relief of strain, if the temperature is high “enough. A full ‘red-heat is required for continued softness under a rapid succession of blows. of Bronzes from Machu Picchu, Peru. 529 and the specific application of the principles involved will be given in individual descriptions of the several objects. Additional characteristics are encountered when constitu- tional changes occur during thermal treatment of the metal, i. e., when the alloy is made up of two or more distinct struc- ture elements differing in composition and subject to influences _ affecting their composition, distribution, or even identity. These characteristics may be used, in conjunction with the others already mentioned, to supply some knowledge of the past history of the object. A brief discussion of constitutional relationships between tin and copper will be introduced shortly, along with a table of analyses pertaining to the present set of bronzes. To further the immediate appreciation of pre- vailing conditions, it may be observed at this point that, while all of the bronzes examined are normally composed of a single structure element, the alpha solution of tin in copper, certain secondary constituents of a transitory nature occur in the cast metal. Aside from the purely micrographic aspects of the work, some facts of interest may be learned from first hand observation. Thus, certain of the objects, notably the long cloak pins, show surfaces which may be closely simulated in wood by whittling; the separate surface elements are approximately flat and long enough in some cases to indicate the use of a very broad-faced hammer or even a rolling face in shaping the metal. One large pin has a flattened head bearing a single perforation, which, from its eccentricity and decreased diameter on one face, must have existed before the head was flattened. (A cylindrical hole starts to close on the hammered face.) Duplex objects, in which different parts are joined mechanically, by welding, or by duplex casting, may sometimes be recognized by superficial examination. A case of the last named variety was found in the present collection. Where surface indications are not adequate, examination of a properly selected section will serve to settle the question. The archzologist will doubtless be inclined to look through these pages for some expression as to the probable age of the bronzes. Unfortunately, no inferences of this sort can be drawn from examinations of this character, or from any other form of examination, as far as 1 am aware. The rate of oxida- tion, or patina formation, depends upon the purity and struc- tural condition of the metal, but no laws governing the process have been formulated. Cast structures are more porous and 530 C. H. Mathewson—Metallographic Description structurally less homogeneous than worked structures and oxidize more rapidly (cf. descriptions of Objects No. 2 and No. 5). Certain structure elements are attacked selectively on long exposure to atmospheric influences. Garland (2) concludes from examination of some ancient Egyptian bronzes that some recrystallization of worked por- tions has very likely been effected by ageing. Rose (3) points to possible recrystallization of an old trial plate of gold used as an assay standard in the Royal (British) Mint. Such conclusions are interesting but they are based wholly upon circumstantial evidence. I have observed cases of incipient recrystallization of vigorously worked metal in the present collection of bronzes, but precisely the same effect may be pro- duced by a form of annealing which is clearly indicated from other considerations. Cf. description of Object No. 3; in particular, photo-micrograph No. 48. In other cases, the very fine recrystallized grains show deformational characteristics which proves that they were produced by the original craftsmen, e. g., Object No. 16. Numerous mechanically hardened and subsequently unaltered (by perceptible recrystallization) struc- tures were observed. In no case does it seem necessary to resort to ageing for an explanation of the structures observed in these bronzes. Whatever the actual facts involved, any eventual reversion of unstable, mechanically hardened, struc- tures to more stable (recrystallized) form cannot be defined as a determinate function of the time involved in the transfor- mation. It must likewise be conceded that metallographic examination of the finished object cannot be expected to furnish any clue to the smelting process used in preparing the metal or, indeed, to indicate whether the alloy was produced directly by smelting a mixed ore, or by alloying tin and copper. 1 Summary or ANALYSES AND STRUCTURAL EXAMINATION. The present collection embraces about one hundred articles which may be classed as tools, including axes, hatchets, knives, chisels, bars and pointed instruments; domestic implements, including’ mirrors, tweezers, small knives, pins or needles, spatulas or spoons and various small articles ornamentally cast; articles of adornment, such as. rings, bracelets, spangles, bells, ete.; and crude, or irregular pieces. A number of specimens (thirty-three in all) intended to approximately represent the diversity of the collection were analyzed. ‘Twenty-one of these of Bronzes from Machu Picchu, Peru. 531 were cut into appropriate sections and examined structurally. The latter are illustrated from drawings in figs. 9 to 29. A general summary of analytical data is given in Table I, and metallographic data, with references to photo-micrographs, in Table II. TaBLe I. No. Character and Weight of Object. Cu Sn |, Ag Fe Ss) 1 | Knife, T-form. Weight, 26g.--._-.-_- 94.26) 4.82] ____- O32); 082) eee 2 |Spatula-shaped object. Ornam. head, ing shonmieee ete eet ee ee be 86.03 | 18.45 | _.-.. | --_ | _-- | Zn-0.82 3 | Cloak-pin, star head. 37g.__-.----.------ GEO OTE SEGO): [haute | She targa ene GL |) TBH, “QRS ese ee 96.90 | 2.11 CO eM een ets a fe ae 5 | Ornam. knife. Fisher-boy. 41g.__-_---. Sfe-(s}|| G68) || 25 - se |) see | AoeOaly/ 6 || Chnggly 2s eee eee COBO | Boil | ascee Pee nee ere e 7 | Knife, T-form. Llama head, 20g.------- 96.79 | 3.00} tr. SS bP | ree ee 8 | Axe, broken short below head. 242g. ___- GBs7O4, Oily) 2-25. 0.87 | 0.44) ------ 9 | Thin, flat copper knife, T-form. 22g.---| 99.73} -... | ----- LO tree (Peace aas 10 | Axe, broken short 1" above edge. 62g.._-| 94.41] 5.12] _____ a O29) eae ene 11 | Cloak pin, pierced flat head. 387g._._---- S831 |] G82 |) sess. Eh pe oe | ebb Bees NMipandemurnons WeOr 2. <2 256 sce = 2 G483 5 Oc oay tae ae se [hi aye llh eral ae 2 Soe 13 | Tweezers, in process of manufacture. 9g.) 90.05 | 9.72} ___-- Eo inl easy Slee eh 14 | Thin, flat knife, T-form. 21g._.__-----.| 96.26] 3.67) tr. aye SORLSt eee see 15 | Axe with double-branched head. 414g.__| 95.63] 3.99 ORS ie een LOFLON eam as 16 | Chisel, twisted and broken short. 100g.-| 98.90] 5.53 | ~___- 0.06 | 0.15 | ------ 17 | Irregular mass. 286g.___.---.---------- 95.68 | 4.20] _.--- ASE NAS wate 28 pac 18 | Large, crude needle. 61g.__.----------- 94.69} 5.16) ___-- Pak Pe ere rere MOMpEweezerse SS) 2-2 55-.2-2-5--e5 sees ook. 94.69 | 5.53 | ___-- eS spe ss e Aes QOM plancelbarmelO0g 2. e2- S2 see] ea 94,52 | 5.45 | ____- pee eat | ee eae 21 | Light rectangular piece. 20g.+-.-------- 94.42 | 5.96) ..--- Bee all, meee sate 22 | Thin, flat knife, T-form. 9g._...-------- 95.35 | 4.22) ____- aa OS 20) hes 2 pe Oe a8 op So ore aes 94,70 | 6.60 aT | eee | coa eeepae Aa Bee mesh ue IQ pease soe Seer COCO) CEB OG) Sa) O18 oaesa- 25 a Ol ema oo HEA Yo e e e rS 92.55) 7.14] tr. oe eel OreO) eae se Gm imi yes <6 “s AG cree eens Dee 94,52 | 5.12) tr. (He, |. OU cose 27 OU a ioe a Ii a ee eee 95.08 | 5.12] tr. SER SM eee 23} |) Simnalll lomie,. Cees Soe eee eee ee ae C2 Oa) Westy |S Ossi) S28] 2s5 |i sesae- 20) || Sionelllsxocls - leh Se ee ee eee CBP || G8 | Ss Less Wee oleae seats 30 | Irregular wrinkled sheet. 160g._-_..---- S999 ose = oe 5 | eee Sissies Bile ithum) kmitemblade:;- 6g4-2 2-2) --=--.252--- 91.24) 8.89) tr. He |] OBE |) secese 32 | Small silver dise (spangle). 0.77g.------- eae tee en OOS eC ee |e 33 ee te et we 9Gpe eee ee ee heer |i OOLO Fe |keae ol eee —br The analyses shown in Table I were made in the Sheffield Chemical Laboratory by Mr. Frank G. Mooney under the super- vision of Professor H. W. Foote. After a preliminary quali- tative examination of each object, the elements indicated were determined quantitatively according to standard methods. Some of the bronzes are remarkably pure, aside from small quantities of sulphur in the form of cuprous sulphide. This C. H. Mathewson—Metallographic Description 532 *x @J, JO UOlyogIUseU B 4¥ OU] ,¥E Suole pozttNod sUlBIs Jo IequINU OdBIOAY y u0}y -BINSYUOO [V1oeds Fo uoTser 10401990 @ UL SUIM} (%) [ezluesu00 gc ysnory4 WpIM erngonays yseo [worddy, | “Te ‘(Q¢) “Sty WoT}eg ol “Sta ued v prey Pus 4s poyst -ulq ‘(01g oUt Aem jaed 4u10d 4sn1yj) peey spieMoy yutod | pezitjeysf1oe1 Ajo}0[duI00 : ee—jutod | woiy AyisueqUL SuIsveroep | ey} 0} prey 4svo [emIou Sixe [edrourrd WO} SeyOUr | YIM polvouuy “SulyOM | 9} ULOIZ SuIssed soinjons4s 67 ‘RP suoje suory peoy 124g yeteaes -pjoo Aq pepue;xe yusys | sursuvyqo jo uorsseoons | ‘1% “(QF) “SI | -008 pozefosy| TTL “Si ‘aid yRorp | woryezty [24 -sf1001 juerdiour pue uo1y -euojep Aq pesezje ATqe -1lepIsuod ynq pue eBpnyeds UL “148 4880 o[QeysInsulysiq [eouue [Vo0] O7¥Ip | [Iq poyenuesye ul pout10; CP ‘PP pue 07 pus -euie}ur 4ysi, A1eA RIM | -op Ajotoaes ‘pitq Jo c1nsy | ‘ep ‘ep ‘LP WOIZ WOT}OES moods Sol}IWe1}Xo 4e PoYIOM-pjoyH | yNoysno1yy emnyonsys 4se_ | “TP (Op) “SLA | [eurpnqrcuoT OL Sia Io enyeds | g §o—AUeYs ar ‘004 98 apelg [eouue wpe | pue Yurys ul Fys poyst JojJV ‘7848S | -UlY ‘“[eouUe oJeIpeUTIEZUL SONSIIOJORIVYO [VUOT} pue 0} pue [BUISIIO 4U SI, UAIM AT[eoo, pax AA | -eUIOJep [BOOT WIM oan 62 ‘se WLO.AF WOTJOES UL ‘Iojopuy | ‘wa0j [euy ‘xordde 0} 4suQ | -ony4s suysvo yueulmoperg | “Le ‘(9g) “stq | [eULpnyisuoT 6 ‘SI asIU yy T 4 SULBID ydeaisoioryy poutmexy "i : jo sequin guoulyeery, erqeqorg SoNsiejyowlegy jeanjon14s -00g 101}104 wojwysniy| yoolqo ONE ‘THOIYeULUIeXG oLyderdo][e,07/, ‘TI Idv jo Arewmuns jog 533 of Bronzes from Machu Picchu, Peru. *X BJ, JO UolywoyluseM ve 4v oul], ,fg Suole peyuNod sure«d Jo 1oquinu odeAOAY y ope] q UI WoIste}xe jeouue yuenbesqns SUIMOYS woTyez[peyshs0 -o ‘“woIsueyxe Aq poeieqye diy rw38Ua puw 103190 ye e[pury oO} Terresed opriq yon yim eoeid podeys | AT[eI1eye0n sepoaed oprxo rar eug Ysno144 aylUy -— Ysnot wlory poeyxtoM | JO UoyNqiaysip o40eynq | ‘Ty “(OY) ‘Stq | suooesssorg | ZT “Siq ieddop 6 0[—peurme qeoy pot [INF eplydins jo yunows pues peoiq -x9 UOljIod | ye pafwsuuy ‘pues proiq | ese, e@ pue urlels os1e0o WOIZ WOLy g[suis uy | oy} ye AlyYSI,T peyxtoA, | Sq poziatoqovseyo ATie1oedsq | gg ‘(g9) “Sty | -008 19UT0g Ql ‘siq oxy 8 spoyjeu jensn Aq odeys OjUL poxy1oM sv ‘opelq pure yurys 03 yueys YIM ‘e100 vy, ‘yuewjzvery yuenbesqns Aq perez jo uorjiod v -je ATqisues you ‘uo141sodu100 JueIeFIp Jo [BjJeUL UL 4sKO pur peoy ony SI Peo VULRI[ OY} JO coURZsqNs UIBVUI ey} YOIYA Puno ysno1y4 proy e100 B SUMOF 4 CLO M AZLUIOI}XS TaYyJO oY} 07 SeqoReT 19 013008 -eBUe|T 8I—e10 | epRtq 949 Jo [eToyVUL eyy YOIYM ur eoaid xejdnp VW | ‘99 ‘(GQ) “Stq | yeurpnysuoy | eT ‘Sig | osmy | 4 ATOYMOS[9 ULLIS IVT elT— esUBI TOVeZIT[e} | -nserat ATYSIy 4nq ‘esrv0H #9 ‘(e9) espe [estyo edpe [estyo | -sAtoea oy} MOTeq poysiuryg | ‘espe oyg avou peyeto | ‘eg ‘(T9) ‘99 | pue aynq ou9 ey} WoIy | ‘odpo oy} Avou 4sve, ye | -osse Ajosojo ureads pozty | (gg) ‘gq ‘1G qonoiyy Your euQ | vey por 4YSIAq 4e payxToM | -[eysA10e1 ouy pue osx¥v0g | ‘9g ‘(cc) “BI BUOTIOES tL Sta e469) 9 uwoeutloeds pezip “xo ATpeg ‘Tvouue 4y9I] epriq Sutpeovord §qyIM Sury.toM jody, ‘om quem4¥e1} SUI[vouUR | -ploo Jo yooTa oY} sMoOYsS -3y 09}e1480.1d qysI] o}yeIpetIeyUL ouOs | YoIyM ‘opriq O49 ut 4doo ey} [onory4 foq-194stq UIIM Opvlq OY} WL peyIoM | -xe eAnjon.sys ysvo yeolddy, | pe ‘(eq) “s1q | Uor}0eg el Sta eyuy | G I Rares quomyeely, o1qeqo1g SOLSlIOjOwIeYO jeanjons49 eens ane worerqsnfit| y0efqQ | “ON ‘WOVULMeXY o1gdersoyejoyy Jo L1vumunsg jog “‘panuyjuog—jj WIavy, C. H. Mathewson—Metallographic Description 534 *x Zl JO WoMvogluseur v 48 oUl]:, fe Suoys pojunoo sureis Jo tequinu osvIeAY IeAO0 Te prey poysiury “yeou eu QZ—peulme | pox [Np 48 |vouue 4se7 S01} peay -x0@ UOTJO0S | 4UOUI}BeI} SuljeouUe oZeIp | -SIIOJORIBYO [eUOTFeUIIOJOG UIIe 9uO pournte-oM4 o[SuIs UT | -OUILEZUL YALA poyLoM-pToo | ‘peztresdxoer ATUMoFruy | 6g “(9g) “Bt | Jo AjwMeTyx| | es Sry | WM ex | CT e[puey, suol{eIedo SuLvouue | SUIT[OL JO UoT}JOeIIp ey Ut jo Aq1u10814 exBIpeutojur Alesseoou oy] | sejoyzed oepiydjns ey, jo -xe pedvys eg—oe[puvy | JIM Sulyjor Jo waox ve Aq | woIsuejxe peounouor1d Aq cg -yooy qsnoigy oFLUy I0j oovIOAW | sdeyied yuowuyve14 o1jseaq | pozimeqovaeyo Ajarepnowyarg | ‘Fg ‘(eg) “S17 m0ly0eg 08 “Sta qupy ‘Ulu, | FT ezIs ULeisd qUOUL | UL WOMBIIeA o[qviepisuod pue 0} pue -4vo1] SUL[BVOUUe 4YYSI] ITM | WITM exngon.4s poztijeysAr1o WOIy WOTjOes sseo0id UL CZ} ST | eos1d poytom Ajo}eIepopy | -e1 poztuesomoy Ayjenszeg | gg (Tg) “sta | TeUlpnyisuoT | ey “sq S19Z00M J, | EL Suljeouues yy SI, ee MOSTO e[puey Jo sixe Sg—AoIItor | yIIM AjeAIsuezxe poeHIOM | uoryeztpeyshioe1 ojo]du09 tofeur ysno1q4 jo Apoq | yoefqo jo Apog ‘edvys | ‘ejpuey ureanjona4s3ulysvo 08 suissed 101 1OF OSeIAAW | 4880 8} SoAtosord oe[pueyy | peziuesomoy ATL eV141eq | ‘Gy (QL) “Sly | Worj0es sso1p 1¢@ Sta | -al puey | ST TO BULMe prey peyst -x9 oF WMOpP C[—peul | -ul_ “yRey per ][nz ye queu SOMSILoJORIeYO [UOT] peqqna prot -mlexe gied | -y¥01, SuljeouuR IM SUI | -eUlIojep pexIeM SuIMOYS jo ooryins peoy yep JO oovIEAY | -Iomurvy Aq pouszyrp poy | urers pozipjeysAr0er osavop | 2) (92) Stq | 7ep [eingen 6L “SLA ‘ard yeolp | IT e[— peur jeoq eSpe Ieou UOT} espe -xX@ 01}0e8 | pod [[NJ ye Suljeouue oelp | -vuZojop oleAes SulMoys cy 0} .tepno1pued Q[SUIS UT | -eWIoJUL YIM poyYIOM-pTOL | uler3 pozi[eysAtoea osaeoH | ‘Fy, (ey) “Stq | -aed uomoeg SI Sq oxy OL dvis Sate queUIyvery, e[qeqorg. SOI4SIIejowIeyO jeimjonsi4sS us eee ans uorjeaysniiy| yoelqQ | ‘ON ‘UOleUulMexG oLgdeisojejeyy Jo Areumung jog “panuyuog—jJ #1aV 1, "x gl JO UOLBoyIUSeU v 4e OUT, , fe SuoTe poyuNod sule1s Jo Iequinu eseIOAYV y ve) espe [e104] eoerd rey oD jueulyee14 Sul | uoljIpuoo pajveuue Ut pur re9su W0140eSs -nSue,0e1 us) ee -[BeTUe YJIM peYIOM-pjopH | peztjeysXhi1oer Aje,e1du109 jeulpnyismoy | 6g ‘Sty Teas Tz WOIjVplxO JO souepLAP Sulleeq seljiavo snoreunu pary AjeyeLopour peystury | Suturezuoo ‘aatjoojap AOA 5 ‘qUOULyVeI} SUITROUUR YONUL | [ee] + “Bet0D Jo e0vay ON pue suo 1e98u = BE YjIA peytoM AJOAISMO}XY | “poztipeyshatoer Ajeqye;du10p WOTJOSS SSOID) 9g ST eq esieTyT | OZ SY) ay 3 . T6 “OfH Ul UMOYS 4Uq} Gp JO ULLOF oF 3 04 IV]IULIs o1njon14s A1epuo rojo doceMe pro) Gents) S jUeUyve1} Stuijeouue | -des pozt[eyshioe1 peureis Taqyye Wo140es = JUSI] YjIM pexOM-plop |-eulg ‘se1oo yuourmorg [eulpnyisuoy | 1g “Sty S1eZ00M J, | 6T AY = eke Jo = sdep [ereqe[ | AyluTolA UI soljsl1ojovareyo 3 YJIA SUISO[O pUB POL Fo pud | [vuUOTJeUIIOJOp yuSULMOIg aXe ysno14y e[peeu = 8 ey} Sutdooy Xq peuog aA | *poztpeysX1oer AJoyordurop | 16 ‘(96) Sia Wo1}00g 9 “SLA esivy | ST S S pue euo teeu am WOMYVUILOF WIM} [LUOISv000 C6 eoerd ysno1ryy SSvUL % pe}veyed 10 peyooo A[Mo]G | yyIM eanjonsjs SutyseQ | ‘FE (eG) “Sta moTj08g CZ “SIT qe[nse11y | LT xX S peulmexe uouttoeds 480 S -piey, ‘jeouue yysi; Area RQ euo 4Ssve] 4B YIM ATsno10 SONSIIOJOVIVyO [eUOLy = “SIA peyxIOM SeM 4T “AjO11y | -eUAojeop Suoyg ‘sjavd espe [e1eqey “Ud OINJONIIS 4SVO S}I 4SOT | [[e UL WoTZeZI[[B}SA100r QUOT 26 Tweu Worlqoes ce Ieaeu sty uewmtoeds styy, | -diouy ‘setoo queututorg | ‘16 ‘(06) “Sty | [eulpnysu07 $6 “SLT JeslayQ 91 SUTeID deaiso.101yy eulmexy é yo 1equinyy queuIyvoery, e[qeqoig SOSoyoRIVyO jeinjon1y9 q -0400d Dene twoneysniyT| yoelqo On ‘MOlMeUulMeXxY o1gderSoy[vjyeyy] Jo AreuUMNG jog "“panuyuojg—]]I W1UV CC SAS*SCSS:~=—S—_—CC 536 C. H. Mathewson—Metallographic Description constituent was identified metallographically (cf. observations on cuprous sulphide in copper by Heyn and Bauer (4)) in every bronze examined, although sulphur was not reported in a number of the analyses.* One specimen, No. 8, contains nearly a per cent of iron and another, No. 4, contains about the same quantity of silver. Others contain smaller percent- ages of these elemnts, while small amounts of zine were found in two specimens. ‘The rest of the bronzes contain practically no metallic impurities. Two silver dises are also very pure, one containing a trace of lead. Perhaps the most noteworthy specimen is No. 30, which consists of very nearly pure tin. Other specimens were found in which strips of the metal were rolled into the form of a ball, presumably intended for convenience in cutting small pieces to alloy with copper. This metal was not analyzed but it is unquestionably tin of good quality. Probably the dis- covery of these specimens constitutes the first direct proof that Inca metallurgists were acquainted with tin in elementary - form. It is fair to infer that they used it in preparing their bronzes. On the other hand, it is certainly true that the proportion of tin present in any given case has not been chosen with par- ticular regard to the use for which the object must have been intended. The largest percentage of tin was found in a small spatula- or spoon-shaped object (No. 2) which is particularly distinguished by a perfectly executed cast figure of a humming bird at the extremity of the handle. This specimen contains 13.45 per cent tin. Next in order, we have another artistic casting representing a prostrate fisher-boy with line and fish which, although worked (mildly) into a blade below the figure (cf. fig. 13), could never have been intended for severe use. This object contains 9.39 per cent tin. The ordinary axes and knives, which would seem to require high percentages of tin to give them maximum hardness and strength, carry from 3 to 9 per cent of this element, most of them in the neighborhood of 5 per cent. One knife, No. 9, is composed of very nearly pure copper. Boman (5) points out similar anomalous rela- tionships in discussing a table of some 65 analyses representing the compositions of various objects found in Argentina, Bolivia, Peru, and Ecuador. * When we consider that one part by weight of sulphur yields about five parts of cuprous sulphide, the constituent which appears under the micro- scope as a distinct structure element, it is apparent that the micrographic method may well be superior to methods of chemical analysis in detecting small quantities of sulphur in bronze. of Bronzes from Machu Picchu, Peru. 537 At first sight, these facts seem to indicate that the Inca bronzes were produced by smelting mixtures of tin and copper ores, since, by such a process, it would be difficult to control the composition of the resulting alloy. According to Joyce (6), “it seems almost certain that the presence of tin is acci- dental, since it is found in greatest quantity in those implements which require it least.” I have already remarked that metallographic testing methods cannot be expected to furnish any clue to the manner in which tin and copper were brought into association. Without addi- tional facts of more direct application, it does not appear possible to prove the genesis of these bronzes. I wish, how- ever, to draw attention to a few general considerations which have strengthened my belief that the present objects were produced by alloying the metals, tin and copper, after obtaining them in comparatively pure form. In the first place, both tin and copper were known in elemen- tary form and the former metal must have led a transitory existence (as raw material for use in bronze making), since no finished objects of tin have been found in Inea ruins or burial places. From a metallurgical standpoint, while it is true that tin and copper occur closely associated in some Peruvian ores, I find it difficult to believe that these primitive people could have smelted mixed ores which would almost certainly contain other metals (lead, zinc, silver) and, very likely arsenic and antimony, as well as sulphur, so as to pro- duce the remarkably pure copper-tin alloys in question. A private communication from Mr. D. C. Babbitt, of the Cerro de Paseo Mining Co., contributes analyses of table concentrates from Peruvian sources showing copper and tin in association together with silver, gold, tungsten, zinc, antimony, arsenic, sulphur and earthy material. These complex concentrates are handled with reluctance by the highly equipped smelters of Swansea (Wales), on account of smelting difficulties. Mr. Edmond A. Guggenheim, who has taken considerable pains to gather metallurgical opinions from his associates in South America, calls attention in a recent letter to the frequent occur- rence of tin and copper in chemical or mineralogical combina- tion; but on the other hand, to the frequent occurrence of lodes carrying copper ores in association with tin ores. This is true of the Bolivian tin-fields in which tin predominates over the copper. He also refers to the difficulty of smelting the associated ores, stating that in Cornwall, even in very remote times, the two classes of ores were separated before smelting. 538 C. H. Mathewson—Metallographic Description It is far easier to believe that the Incas learned to recognize certain characteristic copper and tin minerals which were care- fully hand-picked and smelted independently, or perhaps in association. Native copper, which occurs in small quantities throughout the Cordilleras, was probably known to them, while stream tin (cassiterite), also of likely occurrence in a limited sense, would furnish a very pure source of tin. The idea that native copper may, itself, contain percentages of tin equivalent to those encountered in these bronzes cannot be entertained. The fact that the percentage of tin contained in the Inca bronzes is not governed by the use for which they were intended raises the question as to what has governed the percentage of tin in case this was subject to some controlling influence. Joyce’s observation that tin is found in greatest quantity in those objects which require it least is particularly interesting, since it suggests that the objects which require the least tin for proper service may require the most tin for some other reason. While I have not found time to search carefully through South American archeological literature for analyses pertaining especially to objects of known character, the present collection offers certain evidence bearing upon this question. This evidence deals with the casting properties of bronze. As far as the present collection goes, those objects which would require the least tin in service are the more delicate, or orna- mental pieces. As already pointed out, the two finest castings, Nos. 2 and 5, contain maximum percentages of tin (13.45 and 9.39, resp.). A few rough experiments indicated that these high percentages of tin yield the best impressions in casting. Bronzes of this character expand in solidifying, whereby the finer details of the mould are registered in the metal, even though subsequent contraction, on cooling to ordinary tempera- ture, determines a total shrinkage effect. Wust (7) determined the value of this expansion (also the total shrinkage) in alloys containing approximately 5, 10, and 20 per cent tin. The percentage expansion was found to be 0.085, 0.122, and 0.01, respectively. Later, Haughton and Turner (8) investigated the same property, using a somewhat different method, and located the maximum expansion at 10 per cent tin. Thus, we see the advantage of choosing alloys containing in the neighbor- hood of 10 per cent tin for casting purposes. It is also worthy of note that the 10 per cent alloy begins to freeze at a tempera- ture some 50° lower than the 5 per cent alloy, or some 80° lower than pure copper. This means that, as the tin-content of Bronzes from Machu Picchu, Peru. 539 increases, alloys from the same initial heat will remain longer in the fluid condition, whereby the casting operation is facili- tated, particularly in the case of small objects which tend to chill rapidly. Modern bronze compositions used in casting objects of art (statuary, etc.) usually contain fair percentages of zinc. This lowers the freezing point, increases fluidity, soundness, and adaptability to hammering, chipping, ete. (lead is also an important addition agent where the latter property is particu- larly important), decreases cost and produces a pleasing color effect. Thurston (9) specifies several suitable compositions, as follows: (1) Copper—92, Tin—2, Zine—6; (2) Copper— 85, Tin—5, Zine—11; (3) Copper—65, Tin—3, Zinc—382. The expansion during solidification and total shrinkage are generally favorable in well chosen ternary alloys of this type. Cf. Miller (10). Zinc occurs only as an incidental impurity in these bronzes and the Inca metallurgists were unable to avail themselves of its useful properties in this connection. Nor were they acquainted with other addition agents (e. g., phosphorus) which, by reducing the heavy, mechanically entangled tin oxide, itself formed during alloying by the reducing action of tin on the copper oxide commonly present in molten copper, slags out the oxide and cleans up the metal. Thus, their castings leave much to be desired by way of soundness and strength. Having accounted for the high percentage of tin in orna- mentally cast objects, some attention will now be devoted to the objects which require strength and hardness for general industrial uses. No unusual characteristics were observed in any of these bronzes. They are simple alloys of tin and copper in which the physical properties were modified by the ordinary operations of forging, annealing, and cold-working. ‘There is no evidence that any special heat treatment was adopted in order to facilitate the working of alloys high in tin. The metallographic tests indicate that cast pieces were usually hammered in their original condition and annealed as occasion demanded. Only moderate percentages of tin, preferably below 7 or 8, are safe when free working of the cast metal is a primary consideration. Foote and Buell (11), in a recent examination of three Peruvian bronze axes of uncertain origin, found 12.03, 5.58, and 3.36 per cent tin, respectively. They were unable to cold-work cast metal similar in composition to the first of these axes without first applying a special form of heat- 540 C. H. Mathewson—Metallographic Description treatment, the significance of which will be pointed out later on (p. 548). The four axes and chisels of the present collection (Nos. 6, 8, 15 and 16) contain from 3.71 to 5.53 per cent tin. Shepherd and Upton (12) have shown that the ductility of cast specimens decreases rapidly beyond 5 per cent tin. The latter alloy gave an ultimate elongation of about 20 per cent, scarcely inferior to that obtained from pure copper, while, at 10 per cent tin, the elongation dropped to half this value. In the words of the author, ‘“‘by suitable heat treatment, it is possible to vary the ultimate elongation of a bronze containing 90 per cent copper from 10 per cent to 37 per cent without affecting the tensile strength materially.” As will be explained later on, in dealing with the constitution of these alloys, the heat treatment indicated is adapted to bronzes containing the 6 constituent and permits the use of higher percentages of tin than any found in the present collection. It is probable that the early Incas, at least, were unfamiliar with refined methods of heat treatment and were compelled to sacrifice the extra hardness and strength obtainable by increas- ing the tin-content in favor of very free working properties. We would hardly expect the many different objects to show the same composition within narrow limits, since, aside from certain marked variations quite likely to arise in foundry prac- tise which is not subject to analytical control, a greater or lesser degree of softness (measured by the tin-content) would be desired to favor particular requirements in shaping the object. We may add, at this point, that cold-working was invariably depended upon to produce the final stiffness and hardness of an object. ConstTITUTIONAL RELATIONSHIPS. From a metallographic standpoint, the most noteworthy conclusion to be drawn from the table of analyses given on p- 531 may be expressed as follows: The full range of con- centration covered by this set of alloys corresponds almost exactly with the natural range of stability in the form of so-called alpha solutions of tin in copper. As shown in the equilibrium diagram, fig. 1, the alpha solutions reach from 0 to 18 per cent tin at moderately low temperatures. At higher temperatures, the tin concentration of the saturated alpha solution is lower and amounts to some 9 per cent at 800°. Of the bronzes analyzed, No. 9 (Table I) contains no tin and No. 2 contains the maximum percentage of tin, 13.45. All of Bronzes from Machu Picchu, Peru. 541 others lie between these limits. In the lower left hand corner of the diagram, fig. 1, the tin-concentration of each analyzed bronze is indicated by a short vertical line at the corresponding abscissa value. Where the neighboring concentrations differ very little from one another, the lines run together and a general survey of the distribution of these alloys according to tin- 5 to 15 20 25°30 Wt. % Sn. content can be gathered from the widths of the bands thus produced. The mean average concentration of all specimens analyzed lies in the neighborhood of 5.75 per cent tin. The abridged diagram, fig. 1, is drawn from data considered most reliable by Guertler (13), who, in his Handbook of Metallography, has critically reviewed the constitutional work of various authors in this field (notably Heycock and Neville and Shepherd and Blough). The more important concentra- Am. JouR. ae ye SERIES, VoL. XL, No. 240.—DrcremsBer, 1915. 542 C. H. Mathewson—Metallographic Description tion values, as well as the temperatures of the two horizontals, are written directly into the diagram. Bronzes which, under equilibrium conditions, should be composed exclusively of the alpha constituent immediately after solidification imvariably show a zonal structure after ordinary casting and, even when the tin-content is less than 2 per cent, still contain liquid metal when the lowest temperature of alpha crystallization is reached, whereupon solidification is completed in the form of a secondary, or beta constituent. The minute detail of the transformations which occur in this beta constituent on further cooling cannot be fully described at present. The principal features are, however, well under- stood and corresponding effects upon the working properties of the metal are not difficult to determine. Probably most of the unstable beta constituent, whose presence in these alloys is due to rapid solidification, is formed in the normal manner by reaction between the residual liquid (after crystallization of alpha) and the neighboring alpha constituent. This would yield a beta of minimum tin-content (22), at least along the line of contact with alpha. The interior of this constituent may be protected from reaction with alpha (by envelopment), in which ease, its tin-content will be higher. In view of the facts that the entire beta is small in amount, that it must form within the enveloping alpha, high in copper, and that its corresponding solidus and liquidus concentrations are not far removed from one another, it does not appear that the tin-con- tent at any point, in spite of rapid cooling, can be high enough to permit the separation of anything but alpha (along the line 22-28) on subsequent cooling. During such cooling, the prevailing tendencies are (1) for the enveloping alpha to absorb the beta and to homogenize generally, (2) for the copper-rich beta at the boundaries to separate alpha according to the curve 22-28, and (3) for the peripheral and central portions of ‘the beta to effect a concentration adjustment. However these individual tendencies may operate in conjunction, when the temperature has fallen to 500°, or thereabouts, (4) the residual beta of eutectoid composition will undergo transforma- tion into alpha and delta of distinct eutectoid appearance. Such a structure is represented by fig. 43. The white delta is considerably more abundant than the dotted enclosures (dark) of alpha. The present experiments have shown that this trans- formation always occurs so as to produce a distinet eutectoid structure in small chill-cast specimens which must have cooled of Bronzes from Machu Picchu, Peru. 543 through the transformation temperature at the rate of several degrees per second. It can be suppressed only by active quenching. Before referring to the effect of this transformation on the malleability of these bronzes, it may be well to add that the recent discovery by Hoyt (14) of an additional transformation point in this portion of the diagram has not been disregarded in seeking to interpret the solidificational phenomena. Accord- ing to Hoyt’s provisional diagram, we should add to the ten- dencies enumerated above, (3a) for the beta to transform into an eutectoid conglomerate of gamma with little alpha when the temperature has fallen to 590°, (3b) for the gamma to separate alpha as the temperature falls toward 525°, and (3c) for the residual gamma, instead of the residual beta of (4), to develop the final eutectoid of alpha and delta at 525°, instead of at 500°, as previously accepted. All of the num- bered statements are representative of tendencies towards change in the direction of equilibrium. The only one which is dis- tinctly realized under the imposed conditions is that (3c or 4) which results in the complete replacement of an original beta by an eutectoid conglomerate of alpha and delta. Thus, the intermediate changes are not important in this connection. Hoyt’s principal findings are (a) that there are two thermal effects below 600° in certain of the copper-rich bronzes and (b) that annealing, or slow cooling, between these two tem- peratures (590° and 525°) develops a new phase, thus far unidentified, between neighboring alpha and beta grains. He considers his results to be of a preliminary nature and only attempts a provisional reconstruction of the diagram. Bronzes containing the alpha-phase associated with moderate quantities of the beta phase are malleable and may be worked either hot or cold. This is the condition recognized by Foote and Buell (1. ¢., p. 539) in their treatment of a cast bronze containing 12 per cent tin. The metal was either forged above the transformation temperature or quenched from such a tem- | perature to preserve the beta structure for cold working. By long annealing at a suitable temperature (high enough to pro- mote diffusion and low enough to permit the formation of an alpha solution of high tin concentration—probably 550° is most suitable) the malleability is still further increased, owing to complete absorption of the secondary constituent with forma- tion of a tin-rich alpha bronze whose properties are independent of the mode of cooling, except that overheating will cause 544 C. H. Mathewson—Metallographic Description beta to separate again and a subsequent slow cooling will again develop the brittle delta. Alloys containing up to 12 or 13 per cent tin may be made very malleable by such treatment. Shepherd and Upton’s high values of elongation (1. ¢., p. 540) are readily explainable upon this basis. Malleability in the cast condition is highly dependent upon the quantity and distribution of the alpha-+delta eutectoid. The principal constituent of this eutectoid, delta, is specifically brittle. A cast bronze with 10 per cent tin will ordinarily contain several per cent of this constituent, but it will not be sufficiently abundant to form a destructive network around the alpha grains and the material may still be worked effectively. With 13 per cent of tin, the metal may usually be worked if handled with care and annealed frequently at low red heat. Such annealing causes gradual absorption of the brittle constit- uent with continual improvement in malleability. The small object, No. 2, Table I, illustrated in fig. 10, contains a little over 13 per cent tin and was hammered into the shape of a spatula or spoon at one end. In order to test the cold mal- leability of the metal in its present condition, the end was cut off and a new spatula-tip formed by hammering. This was accomplished without fracture. The grain characteristics of the metal show that light annealing treatment was used to facilitate working. Neither this nor any other specimen examined was quenched after annealing, since no beta struc- ture was observed, while the eutectoid structure was recognized in a number of cases. The unaltered cast structure of this specimen is shown in fig. 41. Aside from the black dendritic cores, a lighter sharply-defined and angular structure-element may be seen within the white matrix of tin-rich alpha. This is the eutectoid constituent which is represented under higher power in fig. 43 on the same page. In quantity, it falls a little short of that required to form a continuous membrane around the casting grains. The structure in this vicinity does not indicate very favorable working properties. As previously stated, the highly worked bronzes of the col- lection contain moderate percentages of tin, i. e., less than 10, and, in most cases, only 4 to 6. While such alloys contain a little of the eutectoid constituent in the original cast condition, this is rapidly absorbed on annealing, e. g., in a few minutes at red heat, and the metal is easy to work without any amneal- ing treatment. Finished objects were, as a rule, almost entirely homogeneous and hammer-hardened. It seems fairly obvious —— ie ee of Bronzes from Machu Picchu, Peru. 545 that, in the fashioning of these objects, ease of working was placed above other considerations, and the composition of the metal adapted to this requirement. RECRYSTALLIZATION AND GROWTH OF GRAIN. Two prominent metallographists, E. Heyn and G. Tammann, have lately dealt with the conditions of recrystallization and growth of grain in overstrained metal. E. Heyn (15) gives curves representing the mean size of grain developed by anneal- ing cold drawn copper wire as a function of the time at four different temperatures, 500°, 700°, 900°, and 1000-1050° C. Experiments were also made with soft iron wire. Owing to the transformations which occur in iron, y to B at 900° and B to a at 780° C. and their influence upon grain development, particularly that of the former, attention need not be directed to the latter experiments in considering the grain growth of alpha bronzes, which are not subject to transformation. On the basis of the experimental curves obtained and certain fun- damental considerations, mainly of surface tension, Heyn develops generalizations relative to the crystallization of cold worked metal. It is quite evident that, in any particular case, an estimate of the previous treatment to which a given alloy may have been subjected by observation of its grain characteristics requires systematic study of the grain structure developed in a similar alloy under a variety of carefully adjusted conditions. This feature of the present investigation was served by hot working, cold working and annealing under varying conditions a large number (150) of small bronze specimens ranging in composi- tion in three stages from 92 to 96 per cent copper (92, 94, and 96 per cent). It is not the intention in this paper to tabulate the results of these experiments nor to describe them fully. Some pertinent conclusions derived from them will be made use of in the forthcoming description of the individual structures of the ancient bronzes. A number of curves, figs. 2, 3, and 4, representing the development of grain upon ‘annealing (a) chill cast specimens, (b) chill cast and homo- genized specimens, and (c) chill cast, homogenized, cold worked and recrystallized (850°) specimens, respectively, all of which had been reduced 25 per cent in area of section by cold ham- mering, will, however, be shown, together with homogenization curves to be described later. These curves are particularly important in that they show the characteristic annealing proper- Number of grains counted along 34” at magnification of 72x. 546 C. H. Mathewson—M etallographic Description Fig. 2. 96% Cu See OCR. — TIME IN MINUTES. Growth of grain on annealing after 25% reduction by cold-working. of Bronzes from Machu Picchu, Peru. 547 ties of the bronzes under present consideration. They may be used to illustrate the generalizations formulated by Heyn and by Tammann. Following the discussion of these curves from such a standpoint, other significant facts will be pointed out. The curves, a, b, and ¢, of fig. 2, for example, represent the number of grains counted at a magnification of 72 along a 314” line superimposed upon the images of polished and etched sections of cold worked, chill cast bronze containing 96 per cent copper, after different annealing periods at the respec- tive temperatures, 700°, 775° and 850° ©. Heyn, from an experimental standpoint, emphasizes the principle of grain size equilibrium as opposed to ordinary equi- librium in terms of the phase rule. According to this principle, the grains, when brought to a given temperature, coalesce at first rapidly and then more slowly until the mean grain size for this temperature approaches a maximum equilibrium value. This characteristic is distinctly seen in the curves mentioned above. Here, the mean number of grains per linear unit, instead of the mean grain size, is plotted against the time of anneal. Obviously either of these graphical relationships may be used equally well for the present purpose. Coalescence represents a natural tendency resulting in decreased surface area and a higher order of stability in accord- ance with the principles of energetics. Heyn remarks that a certain mobility of the ultimate particles which constitute the grains is necessary in order that coalescence may occur. This mobility increases with the temperature. Hence the grains can grow more effectively as the temperature is raised. Ata given temperature, the grains, after sufficiently long exposure, will reach a size corresponding to the mobility for this temperature. The earliest stages of rearrangement and growth in over- strained metal, which are of great interest from a theoretical standpoint, are not dealt with in these generalizations. As will be seen later, the curves in this region are indeterminate as far as ordinary counting methods are concerned. Phenomena of this order are subject only to qualitative interpretation. The later stages, i. e., those actually observed and recorded in graphical form, are clearly defined and may be characterized at any point. Thus, we see that, within certain limits, a given grain size may be produced by anneal at different temperatures when the time factor is properly adjusted. For example, a count of about 12 grains is obtained in a 5-minute anneal of 548 C. H. Mathewson—Metallographic Description the cast material containing 92 per cent copper at 850° (fig. 4, curve c) or in a 60-minute anneal of the same material at 775° (fig. 4, curve b). On the other hand, if we assume that the annealing period is neither extremely brief nor excessively prolonged, e. g., occupies from 15 to 60 minutes, the grain size is fairly independent of the time, and will serve to indicate the approximate temperature of anneal. In the present case, the values for 700° range from 24 to 20, those for 775°, from 15 to 12, and those for 850°, from 10 to 744. It would thus be possible, by measuring the grain size, to estimate whether a cold worked cast bronze containing 92 per cent copper had been annealed at dull red (700°), cherry (900°), or at an intermediate temperature, provided no other complications affecting the growth of grain were present. Similarly, the data plotted in figs. 2 and 8 may be used to interpret the annealing treatment of bronzes containing 96 and 94 per cent copper, respectively. Tammann (16), in discussing the relationship between tem- perature and grain size, defines an end condition for each temperature in terms of equilibrium between the tension at the inner surfaces (inter-granular boundaries and intra-granular surfaces of slip, see Theories of Deformation, p. 559) and the rigidity of grain. Thus, the temperature—grain size curves for increasing annealing periods at constant temperature (or for decreasing rates of heating towards a convergent tempera- ture) will lie superposed, the uppermost position (maximum grain size) corresponding to the final condition at which the above equilibrium obtains. Curves of this character, plotted from the data obtained in this investigation, are shown in fig. 5. Annealing periods of 5, 10, 30, and 60: minutes, respectively were used. Since the rate of adjustment between the opposing forces increases with the temperature, the highest value on the uppermost curve more nearly represents a true equilibrium condition for the corresponding temperature than the lower values on this curve. The position of the true equilibrium curve with respect to that of the uppermost curve is accordingly similar to that indicated by the adjacent dotted curve. The grains counted at 775° and 850° C., even after very brief periods of anneal, in all probability represent completely recrystallized metal in which none of the internal changes due to the original deformation persist. At 625°, however, after a comparatively long anneal (30 m.) there is evidence that of Bronzes from Machu Picchu, Peru. 549 some of the grains are fragments of the original distorted grains which would further recrystallize at higher temperatures. The higher temperature ranges of all curves described therefore represent normal coalescence of the natural grains. Experience Fig. 5. Grain size (area) in square microns (/’). 625° 700° 775° 850° TEMPERATURE Growth of grain on annealing after 25% reduction by cold-rolling. Bronze containing 96% Cu. has shown that the grain size produced by anneal under uniform conditions is very nearly independent of the magnitude of previous deformation, provided this has been considerable. While it is difficult to clearly specify the amount of reduction 550 C. H. Mathewson—Metallographic Description by cold work which is necessary in order to bring about this result, it may be safely asserted that, with this class of material, any reduction above 15 per cent (in area of section) by cold rolling or hammering will suffice to bring about uniform grain characteristics on anneal for moderate periods of time above 700° C. As we pass to lower temperatures of anneal, the visible grain is not in a natural condition throughout, but is constrained, in part, owing to the presence of strained patches or grain fragments which are potentially subject to recrystal- lization. It is thus evident, that the count cannot be extended to the low temperatures of incipient recrystallization and still retain its original significance. The prolongation of all annealing curves in the direction of their origin must, therefore, be based largely upon theoretical considerations. A discussion of the lower, as well as the upper ranges of the temperature-grain size curves has been undertaken by Tammann (lc., p- 548). The principal arguments are as follows: Reerys- ail tier in a conglomerate containing grains artificially reduced by deformation will start at a temperature where the tension at some of the inner surfaces overcomes the rigidity at these points. At other points where the rigidity is greater, recrystallization will not occur until a higher temperature is reached. At any of these equilibrium temperatures, the mean grain size is smaller than would correspond to a natural grain size for the same temperature, owing to increased rigidity. A temperature is ultimately reached where the tensions at the inner surfaces become equal to the rigidity of the mean size of grain for the natural condition. At this point, all evidences of deformation will have disappeared and the curve merges into the natural curve. I have attempted in the following remarks to give a general outline of recrystallization phenomena with the aid of a pro- visional diagram, fig. 6, which seems to account for the main facts developed by experiment, This diagram is based upon the assumption that the mean size of the new grains which form at any given temperature by coalescence of existing grains or by local break-down of internal surfaces within the original strain-hardened grains is determined by the temperature of anneal. Thus, the size of the recrystallized grain will be some function of the anneal- ing temperature and the conglomerate will be composed of recrystallized grains along with unrecrystallized fragments of the mechanically altered grains. Such structures are com- of Bronzes from Machu Picchu, Peru. aml monly observed in strained metal which has not been annealed sufficiently to cause complete recrystallization throughout the mass, e. g., fig. 48. The entire field is traversed by a set of lines, ab, ed, etc., which cut any vertical line into a number of segments intended to represent the mean number of recrys- tallized grains which would normally be counted along a unit length upon the prepared specimen after anneal at the corre- sponding temperature. A linear relationship is assumed in this construction, but, obviously, any experimentally proven rela- »§ N TE TLL LLIN Mi /T) ees} ll (Fragmental resolution of grain.) s 8 N N N Ml /] Ni LMM A LMT a. N Ss Ny N Lene ' i Percentage recrystallized (cumulative). \ \ \ 1 la T% Te T7 ai Temperature, (Number and size of grains, resp., fragments.) o tionship between grain size and annealing temperature may be depicted in similar manner. By cold-working metal of definite grain characteristics, e. g., uniformly recrystallized metal of the mean grain size which would be produced by complete anneal at T,, the grain struc- ture is broken down, inner surfaces develop and new, latent, grains of a lower order of size and stability are formed. As already stated, these secondary particles cannot be distinguished under the microscope. Etching peculiarities, diminishing contrast, lines of deformation, ete., are, indeed, visible when deformation has been severe, but no direct indication of the size or shape of the ultimate secondary units can be obtained. As far as direct observation goes, these new structure elements are latent only. Upon annealing, they assert a modified individuality wherein mutual readjustment and coalescence cause them to become visible units with definite orientation and the dimensions characteristic of the annealing temperature adopted. 552 C. H. Mathewson—Metallographic Description In order that the annealing effect may be felt at a given temperature, it is clear that the mechanical destruction of the original grain must have been sufficiently pronounced to pro- duce fragments inferior in size to the recrystallized grain characteristic of this temperature. Thus, if the T, grain previously cited is to recrystallize at T,; it must have been artificially reduced by deformation to a point where the grain size characteristic of T; can develop. Im general, the grain fragments produced by deformational processes will vary widely in size, so that certain of them will be able to coalesce below T,;, while others will remain unaffected at this tempera- ture. The particular factors which apply in any given deformational process will combine to determine a curve of Fragmental Resolution of Grain* for this particular process in which the cumulative percentage of fragments below a given size appears as a function of the size of the fragments. The curve T,T, is a hypothetical curve of this sort and signifies, in the present case, that, by deformation of the original T, grain, approximately 65 per cent of the material has been sufficiently reduced (by internal slip) to recrystallize by the time the temperature T,; is reached. At T, then, 65 per cent of the material will have reached the mean grain size charac- teristic of this temperature, while 35 per cent will remain unannealed. Little can be said relative to the distribution of the unan- nealed fragments and the recrystallized grains. In general, the former occur in patches of irregular outline and, when present in comparatively small amount, as in the present case, cannot usually be distinguished from the latter. When present in large amount, however, the patches are large, compared with the size of the recrystallized grains, and often bear evidence of strain. Such a condition is shown in fig. 48. Large patches of metal bearing distinct lines of deformation can be seen along with small recrystallized grains. The first formation of recrystallized grains usually occurs at the boundaries of the parent grains, since, at such localities, maximum inhomogeneity of the lines of force may be expected and, consequently, maximum fragmental resolution of grain. Initial recrystallization around the primary grains is well *T have introduced this phraseology to aid in defining the disintegrating effect of cold-working without regard to the specific nature of the frag- ments, or particles, formed. Most theories freely admit that sub-granular units possessing some degree of individuality are produced during permanent deformation of the metal. of Bronzes from Machu Picchu, Peru. 553 illustrated by fig. 64, which represents spontaneous recrystal- lization during hot-working. As the degree of deformation increases, whether by rolling, drawing, hammering, etc., the curve of fragmental resolution recedes in the direction of maximum resolution, a specific condition for each metal or alloy beyond which further destruc- tion of grain will bring about fracture. Recession in this direction is indicated by the set of curves, T,T;, T,T4, and T,T;. As maximum resolution is approached, the abscissa range covered by the curve becomes narrower, since large fragments give place to smaller ones and the size of the latter eannot be reduced without limit. Ultimately we would reach a condition of uniform resolution into fragments of minimum size. Jt is improbable that such a condition can be realized without overworking the metal to the point of manifold frac- ture. Thus, the limiting curve, T,T., is drawn to represent a 50 per cent resolution into fragments of minimum size. There is, for each metal or alloy, a minimum temperature at which recrystallization will start from a condition of maximum resolution into fragments of minimum size. This temperature is represented by T, in the present diagram. One half of the metal deformed according to the curve T,T, would recrystallize at T,. At the higher temperature, T;, all of the metal deformed according to the curve, T,T;, about 85 per cent of the metal deformed according to T,T,, about 35 per cent of the metal deformed according to T,T;, but no part of the metal deformed according to T,T, would recrystallize. At T,, the grain of all four would be identical and would grow uniformly beyond this temperature. These curves all represent severe deformation, in that the percentage of coarse fragments is low compared with the percentage of fine fragments. In the alpha bronzes, such a condition occurs when the reduction by rolling or hammering is carried beyond some 15 or 20 per cent. Above 700°, provided the annealing period is not con- fined to a short term of minutes, a bronze of definite composi- tion will give uniform grain characteristics whatever the extent of previous reduction beyond the minimum value given. Hence, we may draw certain conclusions relative to heat treat- ment from these grain characteristics without intimate knowl- edge of the previous mechanical treatment. Turning now to the effect of anneal upon metal which has received very light deformational treatment, it appears that the curves of fragmental resolution will here assume a some- 554 C. H. Mathewson—Metallographic Description what different form, in that the percentage of coarse fragments will very likely exceed the percentage of fine fragments. This condition is represented by the curves, T,T,, T,T,, and T,To. Moreover, none of the fragments are likely to be very small and, hence, recrystallization will not start until moderately high temperatures haye been attained. Recrystallization accord- ing to the curve T,T, starts at T,, but is confined to small portions of the mass until the temperature rises well towards Ts. At the latter temperature, the entire alloy is recrystallized. According to the curve T,T, the extremely light deformation indicated has left 50 per cent of the original T, grain unaltered, while the remainder has been broken down sufficiently to cause recrystallization at T,, a temperature not greatly below the original annealing temperature. At T,, the metal has assumed its original grain characteristics and growth will continue normally as the temperature is further elevated. The curve T,T, represents an intermediate condition which may easily be interpreted. , In cases of very light deformation, recrystallization cannot be detected under the microscope, since the recrystallized grains are not greatly inferior in size to the original grains. In this connection, it should be noted that a section through the con- glomerate shows large and small grain sections whatever the true size of grain, since a grain may be cut at any point according to its position with regard to the cutting plane. This renders it difficult to detect recrystallization except where the new grains are considerably smaller than the original grains and are sufficiently numerous to form groups of characteristic appearance. The curves of fragmental resolution are purely hypothetical and only represent in a general way what seem to be predomi- nant characteristics of the ordinary deformational processes. The precise form of any curve will be determined by, the - nature and intensity of the deformation sustained by the metal. _ It is clear from the discussion thus far that the form of the curve of fragmental resolution determines the form of curves representing the cumulative percentage of metal recrystallized as a function of the temperature. Without seeking to estab- lish the exact relationship between these curves, it may be assumed that they are substantially identical in form, the former representing an obscure condition and the latter a visible effect. By careful counting under suitably chosen magnification, the approximate form of some of these curves of Bronzes from Machu Picchu, Peru. 555 may be determined. The present discussion is intended to convey a rational conception of the connection between defor- mational and recrystallization phenomena from a qualitative standpoint. As far as the present bronzes go, it is possible to detect incipient recrystallization after severe deformation with- out inquiring into the details of the treatment effected. Cases of this sort will be pointed out later. Semi-quantitative conclusions bearing upon subsequent growth of grain seem justified from every point of view. It may be remarked that a diagram of this sort may be drawn to represent the condition developed by anneal during a fixed period of time, or by anneal for the period necessary in order to bring about an equilibrium effect. The latter (ideal) condition was kept in mind throughout the foregoing discussion. ‘The former condition would entail displacement of all recrystallization curves without affecting the general principles involved. Owing to the difficulty of distinguishing between recrystal- lized and unrecrystallized units, counting data obtained from partially recrystallized conglomerates are not especially useful or significant. It is possible to obtain at first an increase and, later on, a decrease in the number of visible grains as the period of anneal at constant temperature, or as the temperature for constant period of anneal, increases. This does not signify that the individual grains first disintegrate (contrary to thermodynamical requirements) and later coalesce, but that the original strain-hardened grain fragments (each of which is. counted as a single unit) are rapidly developing secondary grains in the early stages of the anneal, while, later, coalescence itself constitutes the predominant factor. Such a condition is represented by the curve lettered (a) in fig. 7. This curve represents the number of distinguishable grains after annealing an alloy containing 96 per cent copper at 625° for periods ranging from one half hour to five hours. Beyond one hour, growth is normal, in that the distinguishable grains increase in size and decrease in number. Between one half hour and one hour, however, the reverse effect is encountered. This reversal evidently takes place earlier in the alloy containing 94 per cent copper, as indicated by the trend of the curve (b) in this vicinity, while, in the alloy containing 92 per cent copper, it must take place at a still earlier stage of recrystal- lization, since the corresponding curve (c) shows normal growth of grain at the very start. Viewed under the lowest powers, 556 CO. H. Mathewson—Metallographic Description these incipient effects are completely masked by the primary casting structure, e. g., fig. 47, magnified 12. At ordinary magnifications, both effects may usually be seen, e. g., fig. 91, magnified 72>, while, at the higher magnifications, the finer detail of recrystallization is very apparent, e. g., fig. 48, and fig. 92, both magnified 220. A number of the structural changes resulting from thermal and mechanical treatment of a chill-cast bronze containing 94 per cent copper are illustrated by the six photo-micrographs Number of grains counted along 34" at 72x. TiMe IN HOURS ! 2 3 + Growth of grain on annealing after 25% reduction by cold-rolling. Annealing temperature 625°. assembled on p. 603. Fig. 30 shows the cast structure of this bronze magnified 10 after a contrasty etching with ferric chloride. A hole was drilled in the lower part of the specimen as shown. The photo-micrograph represents the upper portion of a transverse section through the small bar and, accordingly, the extension of the grains at right angles to the upper cooling surface can be plainly seen along with the grouping of small grains in the vicinity of this surface. The internal structure of the grains can also be seen. We may call particular atten- tion to the small grain section with striations inclined at an angle of about 45° to the horizontal which is enclosed within a larger grain with vertical striations near the upper right-hand corner of the photo-micrograph. The next figure, numbered of Bronzes from Machu Picchu, Peru. 55) 31, represents the same surface, re-polished and etched, after the specimen had received a homogenizing anneal of one and one-half hours at 775°. Here, the contrasts are greater owing to the homogeneous character of the grains. Practically no change in the individuality of the grains has occurred, i. e., there has been no growth of grain. The large individuals of fig. 30 can be recognized in fig. 31 and the small grain previ- ously mentioned has retained its orientation during the heat treatment in spite of the directive influence exerted by the large grain which totally surrounds it. In repolishing the specimen, about 0.003 in. was removed from the surface. Consequently, the small grains, some of which are only a few thousandths of an inch in diameter, do not wholly correspond in the two photo- micrographs. ‘The general effect of re-polishing and etching may be seen by comparing fig. 31 with fig. 32, the latter repre- senting a re-treated surface, which, although subsequently strained, shows how the grain-contours of fig. 31 are altered by a second preparation of the surface. Any ordinary annealing treatment of cast bronze of this character does not bring about growth of grain. Even the small grains of fig. 30 have not coalesced upon annealing. This may be due to the fact that the grains, while small, are still larger than the normal grains produced by annealing after strain at this temperature. Moreover, the surface conditions in the case of cast grains are less favorable to coalescence than in the case of deformed grains, since, here, a film of impurity may separate the grains, while, after deformation, recrystal- lization starts within the original grains, at the internal surfaces of slip, where there are no impurities to prevent coales- cence or destruction of free surfaces under the influence of surface tension. Drilling of the specimen has produced a condition of strain which has caused recrystallization upon anneal, as shown in fig. 31. The metal surrounding the aperture now shows the usual fine-grained recrystallized structure in which twins are abundant. A few small twinned grains are also seen at other points upon the surface. This shows that not enough of the surface was removed by polishing after anneal to entirely remove the layer which was strain-hardened in the preceding polishing operation and which recrystallized during the anneal. After the second preparation of the surface (fig. 32), no twinned grains of this sort can be seen. The small section from which these photo-micrographs were obtained was rectangular with deeply rounded edges at the Am. Jour. Sein ei SERIES, VoL. XL, No. 240.—DrcremBeEr, 1915. 558 C. H. Mathewson—M etallographic Description top. When placed in the vise and strained, the first effects were felt in the lower part of the specimen, while, later, the metal near the top came under the influence of the load. Sufficient pressure was applied to cause permanent deformation of the order shown by the changed dimensions of the aperture in fig. 32. At the lower part of the specimen, deformation has been sufficiently intense to plainly ruffle the surface. Above the aperture, the degree of deformation decreases to a very small value at the top. Upon annealing this specimen, we should, therefore, obtain the characteristics of recrystallization after heavy and light deformation, respectively, upon different portions of the surface. The structure after an anneal of 15 minutes at 775°, is shown in fig. 33. Here, we observe a fine- grained, but rather uniform structure, in the lower left-hand corner of the specimen where deformation was most severe. This structure corresponds to the ordinary effect obtained in annealing metal which has been cold-worked to a point of con- siderable reduction in area of section. It may be characterized by application of the usual counting methods (ef. figs. 2, 3, 4, with discussion). In other parts of the specimen, only partial refining of the grain has occurred. Near the top, frag- ments of the original grains may be identified. The small light grain, to which early reference was made, may still be seen in the upper right-hand corner. Below this, a little to the right, a sharp-edged fragment of a large white grain, which reaches to the top in preceding figures, can be distinguished. As already stated,* it is difficult to characterize recrystallization of this sort (after light deformation), since the recrystallized grains cannot be distinguished from the residual fragments. As related to an original cast structure, the fact of recrystalliza- tion is clearly apparent, but, as related to a structure which itself shows twins and the general characteristics of reerystal- lization, even the fact of a subsequent recrystallization is likely _ to remain obscure. In ‘every case, the grain size and general appearance is highly dependent upon the previous structure. A partially refined grain of this character may be seen in fig. 60. This represents a spot near the broad end of a chisel belonging to the present collection. Large fragments of the casting grains have remained after working and subsequent annealing. The alteration of shape at this point must have been slight and the temperature of anneal high enough to remove cores, viz. bright red heat unless the period was considerably prolonged. *See p. 554. of Bronzes from Machu Picchu, Peru. 559 The two ensuing photo-micrographs, figs. 34 and 35, repre- sent further stages in the treatment of this specimen, viz., an anneal of one hour at 775° and one of 15 minutes at 850°, respectively. In both cases, there is very little change in the appearance of the regions which were originally subjected to moderate strain. In the severely strained regions, however, the grains have grown in accordance with the temperature-time requirements previously set forth (see earlier discussion under the present heading). THEORIES OF DEFORMATION. The finest grain which can be produced within a metal results from reorganization of the material either by spontaneous transformation according to the principles of heterogeneous equilibrium whereupon new and initially minute crystalline individuals are formed, or by disperse readjustment of the component crystalline particles without change of phase. ‘Since no transformations (phase changes) are involved in the heat treatment of the alpha bronzes, only the latter condition need be considered here. An extremely fine grain is produced when highly worked metal recrystallizes at the lowest effective temperatures. The growth of grain by coalescence has been described. We have still to account for the condition which causes the initial development of extremely fine grain from a primitive coarse grain. Deformation produces a condition in the primitive grain which causes numerous new grains to develop on anneal- ing. The starting point for each new grain is obviously a fragment of the primitive grain, dislocated, or forced out of alignment with neighboring particles teats the process of deformation. The first changes which occur in an individual grain when it is strained to the point of permanent deformation are well understood, at least from a proximate standpoint; sliding movements occur along planes which correspond to the cleavage planes of minerals so that a set of cleavage elements is formed. These elements do not fall apart, however, but are held in place by adhesive forces of some description. As deformation proceeds, the planes of slip become more numerous and the cleavage elements become smaller. The process may be fol- lowed to a certain extent under the microscope by observing the polished surface while under strain, but the precise nature of the movements under severe deformation, leading finally to rupture, cannot be clearly recognized. 560 C. H. Mathewson—Metallographic Description Three theories of deformation have lately come into promi- nence. A brief characterization of each follows: The Translation Theory, proposed by Tammann, (1. ec. (16), p- 56-74), rests upon a purely crystallographic basis. All deformational properties are explained by movements of the above character, in which the original molecular space lattices are maintained; every movement proceeding strictly in accord- ance with the crystallographic relationships within the primitive grain. The Displacement Theory (cf. Mollendorff and Czochralski (17)) is based mainly upon ideas of molecular aggregation developed by Lehmann (18) and associates a gradual destruc- tion of the space lattices with the progress of deformation whereby a final condition of ‘forced homotropy” results, in which the molecules of a primitive grain are not thrown utterly into disorder but are forced into a modified configuration governed by their own mutual attractions and the play of imposed forces. A Modificational Theory elaborated by Rosenhain (19) from ideas originally advanced by Beilby (20) builds upon the ordinary conception of translation by assuming that thin layers of an amorphous cementing material are formed wherever intercrystalline sliding movements occur. Similar layers are supposed to exist wherever two different grains meet, whether these are recrystallized grains or primary grains of solidifica- tion. This theory is especially serviceable in explaining the remarkable adhesive properties of the grains whereby, at ordinary temperature, a fracture normally crosses rather than follows the grain boundaries; while, at elevated temperature, the tendency to pull apart along the boundaries begins to assert itself. These facts are easily explained by attributing suitable properties to the intercrystalline cement, i. e., it is assumed to be rigid and strong at ordinary temperatures, but weakens as the temperature rises. There is no difficulty in proving that a thin, seemingly isotropic, layer of metal is formed at the surface during the operation of fine polishing. This constitutes the experimental basis upon which this theory was developed. Direct proof that such layers are actually amorphous, or indeed, that metal can be transformed into an amorphous modification has not yet been forthcoming. Lehmann believes that these layers are semi-isotropic after the manner of certain liquid crystals. It is clear that modificational and displacement theories can- not be sharply differentiated. without the aid of molecular- of Bronzes from Machu Picchu, Peru. 561 theoretical hypotheses which are, themselves, in an imperfect state of development with respect to the ultimate constitution of solid material. Without considering the nature and arrangement of the molecules or molecular aggregates within the individual grains, we may freely observe, apart from any particular theory of deformation, that severe deformation pro- duces numerous subdivisions or dislocations within the original grain which are unstable in form and re-orientate, forming new individuals, on annealing. Relations between surface tension and grain size seem most adaptable in explaining the characteristic alterations of grain on annealing, although Rosenhain (19) dissents from this view and presents his own ideas of grain growth. MicrocrarHic OHaractEristics or DrorMaATIon. Micrographie methods are not particularly serviceable in detecting the strained condition within a body of metal. The moderate stresses which serve to develop slip bands upon a pre- _ viously polished surface leave the metal with sensibly unaltered internal structural characteristics. The metal may be strained sufficiently to show a distinct increase in hardness by direct test and to recrystallize on anneal, without visible microscopic - alteration of structure. As deformation increases, the normally rectilinear junctions between some of the twinned grains become visibly curved and, later, the etching contrast between neigh- boring grains decreases along with the appearance of lines of deformation. The latter are straight or curved lines which have etched selectively at the expense of the surrounding material and indicate a profound alteration of the metal. They are typical in appearance and always occur when the deformation has been severe. Their arrangement is not such as to indicate the order of fragmental resolution or destruction of grain by deformation. They merely constitute the principal etching characteristics of highly deformed bronze, brass, and other alloys, and can only be interpreted in a qualitative sense. Ordi- narily, the sectional area of a specimen must be reduced at least 15 or 20 per cent by rolling in order to bring about a distinct development of these lines all over the (small) section. The nature of the deformational process cannot be disregarded in seeking to characterize the appearance of the lines or the point at which they first appear. Thus, the same reduction by roll- ing, drawing, or hammering would not develop precisely the same etching characteristics and, moreover, the mechanical 562 C. H. Mathewson—Metallographic Description details of a given operation, i. e., the number and intensity of blows, size of hammer face, etc., in hammering; the speed, number of passes, etc., in rolling, would also influence the final result. There are so many of these variables that it is doubtful if the structural effects can ever be sufficiently distinctive to permit a clear interpretation of the contributing factors. The appearance of these lines of deformation may be seen in fig. 62, which represents the effect of continuous forging from red-heat down to ordinary temperature on an initially coarse recrystallized grain. Partial recrystallization in the form of fine grain has occurred during the early stages of the work at high temperature, but a large number of the original grains have remained and these are traversed by dark markings, mostly curved or wavy. These are positive evidence of the strained condition. In fig. 48, the deformational alteration of a cast structure may be seen under a magnification of 220. Part of the metal has recrystallized owing to light annealing treatment, but large patches of strain-hardened material may be seen; one in the upper right hand corner and another near the center of the left edge, in which the dark lines of deforma- tion give a crevassed appearance. Other structures which show lines of deformation and, therefore, indicate drastic mechanical treatment subsequent to the last anneal are as follows: figs. DO, Ol > Mes TO.) (i. tgs Oo, O2ssamd ne noe When the Character of the deformational treatment has been such as to produce marked elongation of the specimen, the grains themselves show elongation. ‘This is also an effect which cannot be used in formulating quantitative conclusions, since the elongation of the grains is not strictly proportional to the elongation of the mass but varies with their size and distribu- tion. In an ordinary rolling process, such elongation can usually be distinguished when the reduction in area of section reaches some 25 per cent. ‘The structures shown in fig. 62 and fig. 75 illustrate this condition. It is hardly necessary to remark that annealing removes all evidence of this character. DirFrusioNaL CHARACTERISTICS. All cast bronzes differ internally in composition from point to point. Homogenization is effected by annealing treatment. Portevin (21) was able to effect complete homogenization of a bronze containing 95 per cent copper by annealing for a period of three hours at 750°, while a six-hour period at 400° of Bronzes from Machu Picchu, Peru. 563 produced no change in the original casting structure. In a later communication (22), the same author shows a set of photo-micrographs which represent stages in the homogenization of an alloy of the same copper-content at 800°. Homogeniza- tion was likewise complete in this case after an annealing period of three hours. The condition of inhomogeneity in cast alloys of the solid solution type is due to incomplete diffusion during the solidification interval and throughout the subsequent period of cooling. It is, therefore, relieved, in accordance with the laws of diffusion, by exposure for an adequate period of time at an effective temperature. Fick, in 1855 (23), gave a mathematical analysis of diffu- sional phenomena based upon Fourier’s theory of thermal conductivity and fundamentally related to the principle that the rate at which a dissolved substance diffuses into the solvent is proportional to the difference in concentration from point to point. According to Nernst’s theory of diffusion (24), osmotic pressure is the driving force which causes the move- ment of diffusion from regions of high concentration towards those of low concentration. Roberts-Austen (25) showed that fluid metals diffuse in one another according to Fick’s law. He also measured the diffusivity of several solid metals in one another at a number of temperatures. Recent progress in the study of constitutional relationships has enabled us to clearly specify those metallic readjustments which are essentially diffusional in nature and not complicated by other molecular changes due to solution, precipitation, chemical reaction, ete. The homogenization of a solid solution is distinctly a process of this sort which, in all probability, proceeds according to Fick’s law and is very likely due to osmotic pressure between regions of unlike concentration. It is thus clear that we must consider primarily the variables of time, temperature and concentration in dealing with this subject. Any experimental result is not clearly defined unless it includes proper specification of all three factors. Accord- ingly, we cannot make broad use of data similar to that mentioned above (Portevin). The rapid increase in the rate of diffusion with rise of tem- perature is a general condition of particular significance in the annealing of such alloys as german silver and cupro-nickel which require very high temperatures for rapid effects. It is well known that ordinary bronzes, brasses, ete., when abnor- mally inhomogeneous and, therefore, unstable do not spontane- 564 C. H. Mathewson—Metallographic Description ously homogenize at atmospheric temperatures, even after long periods of time. Garland, 1. c. (2, second paper), p. 331, describes conditions of this sort affecting Egyptian objects some 3000 years old. Many of the present Peruvian bronzes have likewise retained their initial inhomogeneity through the centuries which must have elapsed since their preparation. Thorough investigation of the diffusional properties of a group of alloys is, in itself, a task of some magnitude. Since time could not be spared for the complete experimental correla- tion of time, temperature and concentration in these alloys, it was necessary to make certain approximations and rely upon a moderate amount of suitably selected experimental work. A cast bronze of the present type is composed of grains, each of which possesses an internal dendritic structure representing its development from the original nucleus by thickening, extend- ing and branching from the primary stem. The average thickness of a stem or branch, as well as the difference in con- centration from center to edge, is determined mainly by the rate at which the whole mass solidifies and this is, in turn, determined mainly by the relations between the mass of metal, its temperature at the moment of pouring and the temperature, size, configuration, and heat conductivity of the mold. When the alloy solidifies rapidly, the numerous nuclei, or centers of crystallization, mutually interfere before they can develop to great size; the completed grains will, therefore, be small and their internal structure fine. On the other hand, when cooling is more gradual, fewer nuclei are formed and these grow to greater size and show coarser internal structure. It is true that other factors; supercooling, convection currents, ete., may influence the final structure of a cast alloy. More- over, we know very little about the actual variation in internal concentration from point to point in specimens produced under different conditions. Gulliver (26) (27), in two papers pre- sented before the Institute of Metals, has dealt quantitatively with the concentration changes which occur during solidification with incomplete adjustment of equilibrium in alloys of the solid solution type, but conditions are highly complicated and it does not seem possible to make any practical use of his deductions in the present case. As a matter of fact, the temperature-time data of homogenization may be fairly well defined on the basis of the average distance between centers of adjacent dendritic branches,* without knowledge as to the average difference in *This gives an approximate measure of the distance which must be covered in the diffusional migration of the molecules. of Bronzes from Machu Picchu, Peru. 565 concentration between such points or the actual rate of cooling which gave rise to the structure in question. It is probable that very similar distribution of material is encountered when- ever the average distance between branches (in different specimens) is the same. The Peruvian bronze objects are, for the most part, small and they were cooled rapidly enough during casting to produce grain characteristics quite comparable with those obtained in the laboratory by chill-casting 60 gram ingots in an open iron mold from a pouring temperature just high enough to hold the metal fluid during the filling of the mold. The laboratory specimens measured 0.015 to 0.04 mm. between branches and a large number of them was used in homogenization experi- ments with wholly consistent results. In these homogenization experiments, the effect of time and temperature was investigated with respect to this single type of casting structure, which, as already explained, is approximately comparable to some of the structures observed among the Peruvian bronzes. One bronze only (92 per cent copper) of distinctly coarser structure was homogenized at one temperature only (775°) in order to obtain a rough idea of the effect of the third variable on the rate of. homogenization at a given temperature. There is no standard test which may be applied to determine the degree of homogenization after a given treatment. We must rely upon the appearance of the specimen after etching and seek to characterize this in some way. Different etching agents produce different effects and the same etching agent may also produce different effects on the same specimen unless used in wholly similar manner. In all of the experiments, a mixture of ammonia and hydrogen peroxide was used under uniform conditions as far as possible. Six stages in the appear- ance of a specimen were distinguished and designated as fol- lows: (1) VD, very distinct zones, or cores; (2) D, distinct zones; (3) F, faint zones, clearly visible when the principal detail of the specimen is in focus; (4) VF, very faint zones, visible only as a shadowy effect by manipulating the fine focussing adjustment; (5) VVF, less distinct than the preced- ing; and (6) H, homogenized, no traces of zones. The final condition does not imply complete homogenization, since these specimens commonly tarnish on long standing in such manner as to indicate some residual inhomogeneity. It does, however, represent a distinct stage approximating the end-point and clearly comparable in different specimens. of Bronzes from Machu Picchu, Peru. 567 Three compositions were investigated, viz., bronzes contain- ing 96, 94, and 92 per cent copper, respectively. Instead of tabulating the large number of experimental results secured, a single typical diagram is shown (fig. 8) and a general summary introduced, as follows: (1) The time required to homogenize a given alloy at any given temperature is very nearly independent of the composition within this narrow range (92-96 per cent Cu). Alloys containing 96 per cent copper seem to lag a little behind those containing 92 and 94 per cent copper. (2) Homogenization is complete in very few minutes (5 to 8— with preheating period of 9) at 850°, but requires a period of hours (4-5) at 625°. (3) The curve (ab, fig. 8) showing the time required for com- plete homogenization as a function of the temperature indicates a rapidly decreasing temperature-rate of homogenization as lower temperatures are approached. It shows that an infinitely long time would be required for complete homogenization at atmospheric temperatures. (4) The time required for homogenization at any temperature is substantially the same whether the metal is annealed in the unaltered cast state, or after strain hardening by cold work, 1. e., it is unaffected by simultaneous recrystallization of the metal. (5) As opposed to (4), the grain size produced by recrystalliza- tion of strain-hardened metal is highly dependent upon the state of homogenization. By referring to figs. 2, 3, and 4, it may be seen that the two curves showing the number of eraing as a func- tion of the time of anneal at given temperature and composition in metal which, in one case, has been cold-worked after casting only, and, in the other case, has been homogenized after casting and then cold-worked (e. g., the curves lettered 700°, cast, and 700°, homogenized, respectively, in fig. 2) lie widely separated from one another. ‘They lie nearer together at high annealing tem- peratures than at low annealing temperatures, obviously because the cast metal homogenizes in the first few minutes of anneal at the higher temperatures and thus becomes identical (in internal composition) with the metal which had been homogenized previous to cold-working. It does not appear, however, that two respective curves at any temperature would ever meet, whatever the annealing period. Thus, the nature of growth in the early stages of anneal- ing cast metal (of zonal structure) exerts some effect on the final grain size, no matter how prolonged the annealing treatment may be. It has already been shown that the size of the recrystallized grains in metal of given composition is a function of both the temperature and the time of anneal. Within certain limits, 568 C. H. Mathewson—Metallographic Description the time factor may be disregarded and an approximate idea of the temperature of anneal deduced from the grain charac- teristics. Since, the degree of homogenization (as revealed by etching) is also a function of the temperature and time of anneal, it should be possible to obtain a graphical solution for both variables through the medium of the homogenization curve and the curve of grain size. Unfortunately, however, every adjustment of the two variables (temperature and time) to a given degree of homogenization corresponds to about the same grain size, 1. e., the two curves are parallel, as far as could be determined by these experiments. This points to some relationship between the phenomena of diffusion (homogeniza- tion) and of recrystallization. Pure metals recrystallize after strain-hardening and, since they are of uniform composition throughout, the ordinary process of diffusion cannot occur in them. Molecular movements from one sphere of attraction into another do occur during recrystallization and these are not incomparable with the diffusional migration of molecules, particularly if we argue on the basis of a modification hypo- thesis, such as that advanced by Rosenhain (1. ¢c., p. 560). Apart from theory, the present experiments indicate that, in the structural reorganization of the metal at elevated tempera- ture, the degree of concentration adjustment is not unrelated to the growth of grain. With respect to the past history of the Peruvian bronzes, the present experiments show that a structure of the sort represented by fig. 49, which is characterized by a grain count of 33 and a degree of homogenization between D and F, cannot be accurately placed as regards time and temperature of anneal, e. g., it might have been produced by a 5-10 minute period of anneal at 700°, or by a period in the neighborhood of 2 hours at 625°. In any event, it is practically certain that the temperature did not reach bright red in the annealing of this specimen, since, under these conditions, homogenization would have resulted almost immediately and the grain would have grown to much greater size. Broadly speaking, small grains and inhomogeneous structures are characteristic of anneal at the lower temperatures. For comparison with the chill-cast bronzes of fine texture, a single ingot (92 per cent copper) was cooled at the much slower rate of 150° per minute through its freezing range. This sample measured about 0.1 mm. between branches; approximately three times as much as in the ease of the chill- of Bronzes from Machu Picchu, Peru. 569 castings. A period of 60 minutes at 775° was required for homogenization. This is about three times as long as the corresponding period for the chill-cast specimens. On the basis of this single experiment, the period of homogenization is roughly proportional to the distance between branches. The general diffusional characteristics of these alloys are shown in diagrammatic form in fig. 8. Here, the region of homogeneity is shaded and the boundary curve, ab, represents the time-temperature values for complete homogenization. The wedge-shaped outlines represent the gradual disappearance of cores from broad end to point; the width at any marked position corresponding to a given degree of homogenization, D, V, VVF, ete. Along with this lettering, the corresponding grain count is given in figures taken from the original data and otherwise plotted in figs. 2, 3, 4, and 7. Curves, which may be con- veniently called isozonal, are drawn through the temperature- time points corresponding to the same degree of homogenization (same zonal characteristics). It will be observed that the isozonal grain count is practically constant. For example, the counts along the VVF curve are 20 at 850°, 21 at 775°, 22 at 700°, and 21 at 625°. These experiments do not prove that a definite grain count invariably corresponds to a given degree of homogenization however this may be effected. They merely indicate approximate parallelism of the isozonal curves and the curves of grain size in bronzes of this particular character and treatment. The general question is a broad one, requiring for its solution a wealth of experimental data in which difficult counting problems as well as difficulties in the way of characterizing the degree of homogenization will be encountered. MerTattocraruHic Description oF THE INDIVIDUAL OBJECTS. The objects are numbered according to Table I.* Drawings, diagrams and photo-micrographs are assembled in figs. 9 to 97, pp- 559, et seq. The weights of the objects are given in Table I. A brief summary of the structural characteristics and probable treatment of the objects, along with a key to the illustrations and photo-micrographs and designation of the parts examined, is given in Table II.+ The following list of prin- cipal dimensions will serve to fix the scale to which illustrations and drawings have finally been adjusted: *See p. 531. 7 See pp. 532 to 535. 570 C. H. Mathewson—Metallographic Description Object No. 1, fig. 9. Extreme length, 2 3/4”; extreme width (blade), 2 1/2”, Object No. 2, fig. 10. Extreme length, 2 3/4”; thickness (stem), 3/32”. Object No. 8, fig. 11. Extreme length, 12’; diameter of star, 1 1/8”. Object No. 4, fig. 12. Diameter, 3/4”. Width of aperture, 3/16”, Object No. 5, fig. 13. Rear of figure to tip of blade, 5 1/8”. Object No. 6, fig. 14. Extreme length, 5 1/4”; average width, 1 3/16”. Object No. 7, fig. 15. Extreme length, 3’’; rear of head to nose, 9/16”. Object No. 8, fig. 16. Extreme length, 3”; extreme width (at edge), 2 7/8”. Object No. 9, fig. 17. Top to bottom, 2 5/8”; length of blade, 3 7/16”. Object No. 10, fig. 18. Extreme height, 1”; diagonal, 2 3/4”. Object No. 11, fig. 19. Extreme length, 10 3/4”; Extreme width (at head), 15/16”. Object No. 12, fig. 21. Diameter of disc, 3 1/8”. Object No. 13, fig. 22. Extreme length, 2 5/8”; extreme width, 9/16”. Object No. 14, fig. 20. Top to bottom, 3 1/4”; length of blade, 3 3/8”. Object No. 15, fig. 23. Top to bottom, 4 1/16”; width at top, 4”. Object No. 16, fig. 24. Extreme length, 2 3/4”; average width, 1 7/16”. Object No. 17, fig. 25. Irregular, approximately 3 1/4” x 2 1/4”. Object No. 18, fig. 26. Extreme length, 18”; diameter of loop, 1/4”. Object No. 19, fig. 27. Height, 1 1/4”. Object No. 20, fig. 28. Length, 17 1/8”; maximum width, 1”. Object No. 21, fig. 29. Length, 1 7/16”; width, 1/2”. The photo-micrographs have been reduced one-eighth in the engravings. Stated magnifications refer in all cases to the original photos. Object No. 1 (cf. Table I). The knife shown in fig. 9, while similar in general outline to the numerous thin knives of the collection which are repre- sented by figs. 17 and 20, is more substantial and far better executed. From its appearance, the tapered shank might have been hafted or forged into the wedge-shaped blade which is of Bronzes from Machu Picchu, Peru. 571 considerably thicker at this point than the shank itself. With this possibility in view, a longitudinal section through the center of the entire blade, shank, and handle was cut for metal- lographic examination. The outline of this section is shown in fig. 86. The entire surface of the section was explored under the microscope after the usual preparation. The general structure, characteristic of cast metal, is illus- trated by fig. 37, taken from the base of the blade, b, as indi- cated in the sketch. In this micrograph, the greatest possible hetheogeneity was developed by etching rather deeply with ammonia-hydrogen peroxide and then lightly with ferric ehloride-hydrochloric acid. The former reagent does not reveal the black cores but attacks the metal rather uniformly as far as the tin-rich boundaries of the crystallites, showing a raised network of rounded units, each possessing a black border which is not a dark etching but is due to inequality of focus at the edges. The large black cores developed by the ferric chloride represent the copper-rich centers of the grains. Under high power, a small quantity of bluish-gray, or slate-colored, constituent can be seen. This is probably cuprous sulphide, which was invariably found in these bronzes and could be brought into harmony with the analytical figures where sulphur was reported in the analyses. (Cf. description of Object No. 2 for distinction between this constituent and the a + 6 com- plex.) It is always located within the tin-rich network, described above, as are other comparatively insoluble impuri- ties, mechanically mixed foreign material such as tin oxide, of which indications were obtained in some cases, contraction cavities, ete. Rapidly cooled bronzes, which should normally consist of the a constituent alone, usually show traces of the a + 65 complex even where the tin-content is lower than the present value, 4.82 per cent. This is true in the case of our own chill-cast specimens ranging from 4 to 8 per cent tin and is due to the comparatively slow rate of diffusion as explained earlier. In the present case, however, no traces of any eutec- toid structure element could be detected by thorough search with the highest power at our disposal; a one twelfth inch fluorite immersion objective of 1.32 mm. aperture. Such a condition could be produced by slow cooling, but numerous experiments in casting metal at widely different rates of cool- ing have shown beyond question, that the size attained by the primary dendrites when homogenization in this respect is secured by rather uniformly retarding the rate of cooling 572 C. H. Mathewson—Metallographic Description throughout the whole casting operation is very much greater than in the present case. The absorption of the a + 6 complex was, therefore, brought about by reheating. Since reheating at bright red heat causes practically complete homo- genization in a very few minutes,” it is, on this basis alone, almost certain that the reheating was moderate, i. e., at a low red heat, in order that the present heterogeneity of the specimen may be accounted for. Other reasons for the same conclusion will be stated presently. The normal site of the «a + 6 complex, viz., the centers of the rounded tin-rich excrescences already described, is fre- - quently occupied by very small patches which are faintly dis- tinguishable from the surrounding somewhat lighter material. The particles of cuprous sulphide sometimes appear as kernels within these faint patches. These patches were not observed in any other specimen, either Peruvian or synthetic, and their identity is undetermined. ‘They may represent a stage in the absorption of the a + 6 ‘complex under the influence of the 0.38 per cent of iron which is present in this specimen. The only other specimen containing an appreciable amount of iron, viz., the axe numbered 8, is entirely homogeneous aside from the considerable quantity of sulphide which it contains and cannot be used for comparison in this respect. The whole knife was undoubtedly cast in one piece. The structure shown in fig. 37 is continuous across the specimen in the region where welding was suspected. Not the least trace of a weld or other type of joint can be found. This structure is, however, modified in the regions, a and ec, which are differentiated from b and d in the diagrammatic sketch. This modification has been effected mainly by cold-working. In these pages, we shall use the term, cold working, with reference to the temperature range within which work upon the metal produces rather permanent distortion visible as lines of deformation within the crystalline grains. The upper limit of this range approximates incipient red heat. That the piece was heated to a temperature of recrystallization either during or after working is shown by the occurrence of small polyhedral grains, largely twinned, as a secondary structure.t It cannot * Compare Diffusional Characteristics, pp. 562 to 569. + The possible occurrence of congenital twins does not invalidate the conclusion drawn above. In common with others (cf. Desch, Metallog- raphy, London, 1913, p. 183) we have, in this laboratory, observed twin formation in annealed castings (bronze, german silver, cupro-nickel). In no case, however, was it certain that the metal had been rigorously guarded from strain incidental to handling, ete. Thus far, we have been of Bronzes from Machu Picchu, Peru. 573 be determined by metallographic examination whether this reheating followed cold working of the metal or marked the beginning of the mechanical treatment. It appears most probable that reheating was solely for the purpose of softening the metal to facilitate cold-working. Im any event, the tem- perature employed was relatively low, or the period at red heat was very short as shown by the small size of the grains and the general subversion of this secondary structure to the primary east structure. Furthermore, even if the working was started hot, the metal was vigorously worked as the temperature fell below red heat and finally left in the cold-worked condition. It is inconceivable that this piece was repeatedly hammered and reheated to an annealing temperature, as was almost cer- tainly the case with certain other specimens to be described later. Such treatment results in complete refining of the grain with little or no mdication of the pre-existing cast structure. As already suggested, the cast structure was modified mechanically only in the parts a and ec. The amount of reduc- tion desired was small and, according to my interpretation of the structures encountered, the shank was constricted towards the center and the blade tapered to a blunt edge through a process of hammering somewhat facilitated by mild reheating. The complete development of the structure of this specimen was a matter of some difficulty. As already indicated, three distinct, but superimposed, structures were observed; (1) the cast structure, somewhat modified (partially homogenized) by heat treatment, (2) the recrystallized structure, appearing as an incipient refining of the cast structure, and (3) the inter- granular deformational detail of (2) due to overstrain during cold working. Since each of these structures possesses its own etching characteristics, it is clear that all of them will not be clearly indicated in a single etching. Thus, etching of the character shown in fig. 37, in which maximum heterogeneity was developed by using ammonia and hydrogen peroxide, fol- lowed by ferric chloride and hydrochloric acid, rather effectively masks the (2) type of structure. This is due to the fact that unable to produce twin erystals by annealing castings which have been handled with extra care. Vigorous agitation, or stirring, during crystal- lization does not lead to their formation in bronze. Whatever the facts relative to the genesis of twin erystals in cast metal which has not been intentionally strained, they cannot be mistaken for those developed after vigorous working of the metal. In the former case, the twinning is seen only in isolated regions, while in the latter case, it is everywhere abundant in the case of copper and its alpha solutions. Am. Jour. Scr.—Fourtx Series, Vou. XL, No. 240.—DrcemsBer, 1915. 39 574 C. H. Mathewson—Metallographic Description the recrystallized units are visible only in the cores of the primary dendrites, which are themselves blackened by this treatment. Since the original etchings were made in this manner, it was not at first supposed that the specimen had been reheated. The evidence of deformation was broadly apparent in the threading out of the tin-rich network in a direction perpendicular to the forces applied, particularly where the greatest elongation occurred, viz., towards the apex of the blade. By etching very lightly with an extremely dilute solution of ferric chloride in hydrochloric acid (showing no color) the detail of (2) became apparent within the central portions of the original crystals in the regions a and e¢, according to the diagrammatic sketch. The reason why the (2) structure can- not be seen in the tin-rich network is plain when we consider that the network is left comparatively bright (unattacked) after etching, i. e., while it is true that a still more inert material within this network, such as cuprous sulphide, can be shown by selective etching of the tin-rich material, the finer detail of the latter is difficult to develop and cannot be developed in a manner continuous with that of the surrounding zones. That the recrystallized units (2) extend through the tin-rich network constituting the boundaries of the original crystals is, however, apparent by their incompleted character as seen up to this point. The contrast usually shown between differently orientated crystals and different parts of twin crystals cannot be seen with distinctness in this case, since they are located within the cores of the primary crystals which themselves etch and darken selectively as opposed to the tin-rich boundaries. In other words, the secondary (2) structure is more or less masked by the primary cast structure and this fact alone renders very close observation necessary in order to detect the former. The lines of deformation within the secondary crystals which | we have designated as the third superimposed structure can be seen only with the higher powers. Their uniform direction in any given grain, or symmetrical part of a twinned grain, aids in distinguishing the outlines of the grains making up this (2) structure. Obviously, these lines can be seen only in the cold worked portions, a and ec, of the specimen. It may be urged that a description of structures seen under the microscope is unjustifiable without the presentation of photo-micrographs for comparison. Every metallographist is aware that the successful preparation of photo-micrographs at of Bronzes from Machu Picchu, Peru. 575 magnifications which are too high to permit the use of pris- matic illumination through the objective requires the develop- ment of considerable contrast upon the specimen by etching. The difficulties in this direction have already been set forth. In view of the fact that similar but more distinct secondary structures obtained in the study of other objects will be described and illustrated shortly, it has seemed appropriate to forego the introduction of illustrative photo-micrographs here and apply an annealing test to confirm the results of the original examination relative to the distribution of the shaping forces (hammer blows) over the surfaces of the specimen. A temperature of 700° and a time of one-half hour was adopted for the anneal. The entire longitudinal section was annealed and then explored under the microscope. The general effects obtained in annealing different classes of material have already been discussed. As may be anticipated from the original examination, the regions b and d of the annealed metal are now nearly homo- geneous, showing a grouping of irregularly bounded polygonal grains without twin formation. This proves that the metal in these regions had never been worked. On the other hand, the regions a and ec possess a fine-grained structure in which twins are abundant, i. e., the type of structure which would be developed after anneal of cold-worked metal. The two conditions are represented by figs. 39 and 38, respectively. _ The points at which the photo-micrographs were taken are shown on the diagrammatic sketch. The structures are quite uniform in the areas indicated with rather sharp transition zones. The degree of homogeneity attained in the anneal was strictly com- parable with that attained in the anneal of our own material under similar conditions, i. e., only faint shadows accentuated by changing the focus or moving the illuminating prism are seen after anneal for one half hour at 700°. These are due to slight changes in elevation developed by selective etching determined by the varying copper-content at different points. Some of these shadows are faintly seen in fig. 38, particularly across the upper left-hand corner of the photo-micrograph. The fact that this specimen has homogenized to about the same extent as our own cast and hammered specimen proves that any previous heat treatment was not drastic, as previously deduced from other considerations. The size of the recrystal- lized grains after anneal is also comparable with that obtained under similar conditions with our own specimens. The num- 576 C. H. Mathewson—Metallographic Description ber of grains counted along 314” at 72 for a half hour anneal at 700° was 25. At 775° it was 16, while, in the case of the Peruvian specimen, the value 23 was obtained at a temperature which ran a little above 700°. In the region d the metal has been slightly overstrained as shown by the somewhat indiscriminate occurrence of a few twin crystals. This may have been caused by our own handling of the piece in the vise, etc. In any event, the effects in ques- tion are of little magnitude and are the result of light local deformation. It is inconceivable that the hole at d, which is trapezoidal in section, could have attained its present form by any kind of mechanical treatment after castmg. In such case, no spot similar to that shown in fig. 39 at which no recrystallization has occurred, could have been found at the edge of the hole. (See p. 557 for description of the effeet pro- duced by annealing around a drilled hole.) We can only con- clude that the hole was not bored, or otherwise pierced, but was cast intentionally in the metal. In this connection, it may be well to state that this was the general practise with the artisans of these objects. Other cases in point will be described later. In every case examined, where no other complications, i: e., indiscriminate mechanical and thermal treatment after casting, have served to mask the original structure, it is evident that these perforations were attained in casting practise rather than by mechanical means. Obviously, the perforation of so tough a metal as copper or bronze in the absence of steel tools was a task incommensurate with the mechanical ingenuity of these Inca craftsmen; require- ments in this direction were effectively reduced through their skill in foundry practise. Judged from the standpoint of soundness, this specimen is the best casting examined. Most of the imperfections occur in the central portion of the thickest part of the metal, at b, but the number of blow-holes, contraction flaws, ete., is small and the piece is of good quality according to present standards. The large number of holes shown in the two photo-micrographs numbered 38 and 39 are in no way typical of the original metal as these structures were obtained after a laboratory anneal and it was thought unwise to sacrifice the specimen by removing enough metal to get entirely below the partially oxidized surface layer. of Bronzes from Machu Picchu, Peru. 5 Object No. 2 (ef. Table I). The small object shown in fig. 10 was cast in substantially its present form. As in the case of the preceding piece, the casting has been somewhat altered by cold-working. The altered portions are the delicate elongated bird’s bill and the lower part of the stem, including the flattened end. As this object was too small to yield satisfactory drillings for analysis, a sec- tion was cut from the stem for this purpose. Of the two pieces remaining, the larger, including the bird’s figure, was embedded in plaster, ground, and polished for examination. A diagram of this portion is given in fig. 40, along with other figures representing the microstructure. The other piece, flattened at the lower end, was thereupon flattened without fracture at the other end by hammering on an anvil, thus demonstrating the cold malleability of the metal. It was thought desirable to make this test in view of the considerable tin-content of the metal (13.45 per cent) and the comparatively large amount of embrittling a + 6 complex. Of., in this con- nection, discussion on pages 543-545. This is shown in fig. 41 as a delicate light-colored constituent within the comparatively broad, white, tin-rich zones which envelop the black cores. Maximum heterogeneity was developed in etching by using a light wash of acidified ferric chloride after the usual treatment with ammonia and hydrogen peroxide. The casting structure is distinctly predominant all over the surface of the section. The flattened end of the object had, however, received a light annealing treatment between the stages of cold working. This is shown by the occurrence of very fine recrystallized grains, which are abundant in this region, but gradually disappear as we proceed upwards along the stem. Grain characteristics of this sort were developed in the laboratory by annealing at 650°. It is probable that the present object was annealed by thrusting the point into the fire and then withdrawing it, when a dull red color began to creep up the stem. Tests along the stem showed a scleroscopic hardness of 32-35, while, within the contours of the figure at the head of the object, a hardness value of 14-15 was obtained. Corresponding to these results, the characteristics of deformational treatment were observed only in the stem and at both ends of the object. The effect of working a small casting locally in this manner can be seen on the etched surface without the aid of the microscope. There is a visible difference in the quality of the etching in 578 C. H. Mathewson—Metallographic Description worked and unaltered regions, respectively. This may be traced to a change in contrast values due to irregular move- ments of the structural elements during cold working and the appearance of dark-etching lines of deformation within the grains. The photo-micrographs, figs. 44 and 45, afford a com- parison between the altered and unaltered casting structure, respectively, after etching with acidified ferric chloride. In the latter figure there is abundant contrast between the small number of grains which are collectively visible under this moderate magnification (72>). An unsound part of the object was purposely chosen for representation in order that charac- teristic oxidation fissures might be shown without recourse to another micrograph. One of these fissures (black) extends from the left hand edge towards the center near the bottom of the figure. In fig. 44, one grain cannot be distinguished from another and a general effect of elongation in the direction top- bottom is seen in the arrangement of the light constituent (a + 6 complex) lengthwise in this direction. This condition was brought about by cold working. ‘The lines of deformation are not clearly resolved at this magnification. It was not thought desirable to provide another photo-micrograph devoted to this feature, since the same effect is shown in fig. 48, leaving out the small recrystallized units which are not distinguishable in the present object except near the spatula-tip.+ The a + 6 complex is the transformation product of a B constituent which was the last material to solidify, and conse- quently, filled in the interstices between dendritic grains of the primary, a, material. Every change of state in an interstitial material is likely to disturb its alignment with the surrounding material and thus develop cavities. Aside from this, evolution of gas (SOx, ete.), as a result of diminished solubility or of reaction between dissolved impurities, is likely to occur most freely during the latter stages of solidification when the corre- sponding concentration values are high. As a matter of fact, cavities could be seen in almost every patch of complex when examined under high power. Very small ones would be indis- tinguishable from the minute, dark-etching particles of a. The rounded black spot in the uppermost branch of the large patch shown in fig. 43 (300) is one of these cavities. A number of similar black spots may be seen among the numerous patches of complex shown under low power in fig. 41. Oxidation of the finely divided « of the a + 8 complex starts at the surface of the casting and its inward creep is facilitated *See p. 561. ; 7 See p. 562. of Bronzes from Machu Picchu, Peru. 579 by the prevailing porosity. The massive grains of a are attacked only superficially, resulting in the formation of a thin patina. Consequently, the very considerable oxidation which has affected only those objects which contain a comparatively large amount of the complex (this specimen and the one num- bered 5) is confined to this structure element and its path may be plainly followed under the microscope. Reference has been made above to the oxidation cavities of fig. 45. In this photo- micrograph, neighboring patches of unaltered complex are visible and the general similarity in appearance leaves no doubt as to the nature of the material which originally filled the dark cavities. In fig. 42, an altered patch is shown under higher power (300). This may be compared with the light patch shown in fig. 43. Small bright particles of residual 6 are seen. A number of stages in the process of alteration were observed. The a constituent is attacked first, but oxidation . ultimately extends throughout the complex. The sulphide constituent, undoubtedly Cu,S, is unattacked by any of the etching agents used, but, while it is left bright, it may be distinguished from the silvery 6 by its bluish-gray or slaty color. The distinction may be made without difficulty in photo-micrograph, fig. 43. Here, the sulphide constituent is a little darker than the neighboring 6. Three rather large rounded units are seen closely grouped at the extremity of the thin projecting branch of the main patch of complex and one unit half-way down this branch. In fig. 42, where the complex has been altered so as to appear dark, the sulphide constituent, in the shape of three units somewhat larger than those of fig. 43 (two, very close to one another above the center at the left, and one, below the center at the right) is particularly prominent by reason of enhanced contrast. The amount of sulphide in this object is inconsiderable and sulphur was not reported in the analysis. It always occurs in association with the com- plex, usually as a neighboring constituent, but, sometimes, completely enveloped by the latter. Accordingly, cuprous sulphide is appreciably soluble in the 8 liquid solution and most of it separates when this solution begins to crystallize. _ Object No. 8 (cf. Table I). This is a symmetrical and well-shaped pointed object about 12” long, of the type commonly used to fasten the outer gar- ment. As may be seen in the illustration, fig. 11, it has a six-pointed star-shaped head and tapers gradually to the rather 580 CO. H. Mathewson—Metallographic Description blunt point. In order to study the structure of the metal in the head, a cut was carried from the knob at the top along the principal axis of the pin well into its shank. The section was then detached by a transverse cut. The location and sur- face configuration of this section is shown in the diagram, fig. 46. A short section was also taken out of the lower part of the shank, as indicated. A fair idea of the structural characteristics of the entire object should be obtainable by examination of these two pieces. The lower part of the object bears evidence of far more effective annealing treatment than was encountered in the case of the two preceding specimens. Here, the recrystallized struc- ture is predominant and only a trace of the original casting structure remains in the form of shadowy cores, as shown in fig. 49. As pointed out in the section on Diffusional Charac- teristics, the degree of homogenization and the grain size correspond so that a particular combination of the two may be produced by annealing for a short time at a high temperature or for a longer time at a lower temperature. In this particular case, the grain count of 33 and the degree of homogenization, D-F might be produced, according to the diagram, fig. 8 which applies here, by a short annealing period of minutes at 700° or a period of several hours at 625°. It is unhkely that the Inca metallurgists would have resorted to a period of hours in which to soften the metal when a brief period at higher temperature would have been equally effective. They desired to soften the lower part of the object so that further extension could be effected. Very likely they annealed it more than once (after corresponding working stages), but in such manner that the temperature attained in the upper part, near the head, fell considerably short of that attained in the lower part. This is beyond question, smce the upper part has just started to recrystallize in the locality marked by a black dot in fig. 46, and structurally represented by fig. 48, while, higher up, beyond the dotted region of fig. 46, no recrystallization has occurred and the original casting grains are still preserved in a severely deformed condition. All the evidence indicates that this object was annealed with decreasing intensity from the point towards the head, i. e., by thrusting it part way into the fire. While we cannot deduce the exact temperature-range and time of anneal from the struc- ture at any point, it is practically certain that the temperature of the lower portion did not greatly exceed 700°, since a very * See section on Diffusional Characteristics. of Bronzes from Machu Picchu, Peru. 581 short exposure at a materially higher temperature would have produced a distinctly coarser grain. It is also probable that the temperature of this portion was not greatly inferior to 700°, since a drop of 50 or 75° would necessitate a very lengthy period of annealing in order to produce the present character- istics. It is the writer’s opinion that, in the annealing of this pin, a moderate red heat was developed near the point with gradual fall to black heat at the head. ' Deformational characteristics abound all over the object. Fig. 48 has already been cited in this connection (p. 578). This is an interesting photo-micrograph as it shows (1) a primary zonal structure, e. g., dark-etching centers as in the lower left hand corner, (2) portions of the casting grains altered by severe deformation, and (3) small but clearly distinguish- able recrystallized grains. In fig. 47, magnification 12, the field of view is carried from the lower part of the head-section (vicinity of fig. 48) as far as the massive part of the head. It reveals a casting structure of normal appearance. Altera* tion of the order shown in fig. 48 is not indicated under this low power. Exploration of the upper part of the head-section under high power has shown that the star-branches and the knob at the top preserve the general shape of the original cast- ing but were brought to a smooth finish by hammering or some deeper-seated process than that of simple abrasion. The sur- face characteristics of the shank were reproduced in the laboratory by hammering on the anvil with a broad-faced hammer. Object No. 4 (cf. Table I). This is one of the several very nearly spherical objects included in the collection, all of them provided with a point of attachment in the shape of a pin sunk into a hemispherical cavity. They may have been used as plumb-bobs, sounding, or fishing weights, ete. This specimen is illustrated in fig. 12, and a diagram of the section examined is given in fig. 50. The cut along this plane traversed the pin centrally from end to end and divided the object symmetrically into two parts. The finer structural detail of the object is of secondary importance. It is particularly desirable to illustrate the two types of crystallization encountered, one in the body of the object and the other in the pin, as well as the form of transi- tion from one to the other. ‘This is best effected through the medium of photo-micrographs at low power prepared after a 582 C. H. Mathewson—Metallographic Description contrasty etching with acidified ferric chloride. One of these, fig. 51, shows the structure of the pin and the change of struc- ture which occurs on passing through the pin into the adjacent metal, while the other, fig. 52, shows the structure at the base of the hemispherical depression just below the pin. The latter is typical of the crystallization at large. The present alloy contains a low percentage of tin (2.11) and a high enough percentage of silver (0.81) to give it a ternary aspect. Without special knowledge of the ternary relationships involved, it is not possible to fix the identity of a bright secondary constituent which was observed in the loca- tion commonly held by the a X 6 complex. This constituent is not of eutectoid appearance, nor may it be positively identi- fied as the sulphide constituent usually found in these bronzes, although similar in many respects. It is barely distinguishable between the lobes of some of the grains in fig. 52 (in the form of bright points). The somewhat deeper etching in fig. 51 has obscured this structural feature at the present magnification. It may be observed, in this connection, that sulphur was not reported in the analysis of the present object (cf. Table I). For present purposes it is not essential that the nature of this structure-element be ascertained and no further experiments were devoted to this question. Far more important is the fact that it occurs with well-marked characteristics both in the pin and in the body of the object, whereby we may affirm that both parts are composed of the same metal. While the structure of the pin is distinctly different from the structure of the adjacent metal, in that twinned elements abound in the former but are entirely lacking in the latter, the transition from one to the other is gradual and continuous; no mechanical union, welded or soldered joint, or duplex casting could furnish the degree of continuity revealed in fig. 51. There is no doubt that the pin was cast along with the rest of the metal, i. e., the object was completely formed in one casting operation. In the tabulated summary of metal- lographic results, it is suggested that congenital twinning may have occurred in this object. The comparatively small volume of metal comprising the pin probably solidified first and realized an appreciable fraction of its total shrinkage by the time it was gripped by the adjacent metal. Subsequent contraction of the latter might then have overstrained it at a temperature favorable to recrystallization. This is purely a provisional explanation. The facts and conditions pertaining to congenital of Bronzes from Machu Picchu, Peru. 583 twinning are not well enough understood at present to permit any clear identification of such phenomena. It is possible that _the rough edges of the pin as cast were dressed in such manner as to overstrain the metal. In such case, recrystallization with twin formation would have occurred on annealing. ‘There is, however, no evidence that the object has ever been annealed. Moreover, a dressing of this sort would almost certainly pro- duce corresponding alteration of the adjacent metal and prob- ably of the metal below the pin. There is, however, no recrystallization in these parts. In conclusion, it must be admitted that no clear explanation of the segregated twin- formation which has occurred in this object can be offered. Object No. 5 (cf. Table I). This object is the finest example of casting practise fur- nished by the entire collection. As may be seen in fig. 13, it represents a knife of some sort surmounted by an ornamental group comprising a prostrate fisher-boy with line and fish. The metallographic description of this specimen may be made very brief. A section through the body and head of the figure shows a characteristic casting structure. The outline of this section is given in fig. 53, and the corresponding photo-micro- graph in fig. 54. Corresponding to the high tin-content of the object some patches of a + 6 complex may be seen within the tin-rich matrix even under the moderate magnification chosen for this micrograph, e. g., a sinuous patch just beyond the right-hand margin a little above the center. Most of this constituent, however, has been destroyed by oxidation. The black patches represent oxidation cavities. They are very numerous and the highly oxidized condition of the specimen was apparent from the crumbly character of the drillings, the friability of the metal, and the nature of the analytical results (total metallics 97.64 per cent, cf. Table I). No direct determination of oxygen was mace It seemed unwise to mutilate this unusual object except on one side only. Consequently, a section was not taken through the blade. However, some parts of it were polished as well as possible by manipulating the whole object on the polishing discs. The subsequent examination showed that the blade had been shaped after casting but without sufficiently drastic annealing treatment to remove the casting cores or even the patches of complex. Very fine-grained recrystallization had occurred and prominent deformational characteristics were 584 C. H. Mathewson—Metallographic Description observed. The structure in this part of the object is similar to that illustrated in fig. 92 (ef. description of Object No. 16). This specimen, in common with many others, seems to have been cast with a light body of outlying metal corresponding to the blade and from which the latter was wrought without very pronounced change of shape and without the necessity of drastic annealing. Probably this part was momentarily brought to red heat after partial working. Object No. 6 (cf. Table I). The chisel shown in fig. 14 is one of the most interesting pieces examined, since it is the only one in which distinct evidence of hot working was obtained. This evidence is con- fined to the blade of the chisel and leads to the conclusion that this part was formed by bringing the object to bright red heat (neighborhood of 850°) and hammering both in this condition and during the fall of temperature to a point well below that of redness. A large number of forging experiments were undertaken in order to acquire some familiarity with the structures produced under different conditions. In these experiments, the following factors were varied within reason- able limits; temperature, previous condition with regard to grain size, degree of homogenization, ete., number and intensity of blows, temperature range of hammering, conditions of intermittent working with reheating, and period of rest at the forging temperature. Presentation of these results is beyond the scope of this paper. Only the structures found in this object and duplicate structures obtained in the laboratory by applying the appropriate treatment will be shown. The structure shown in fig. 58 corresponds to a scleroscopic hardness of 14. The spot at which this photo-micrograph was taken is indicated on the diagram at the left, fig. 55, and corre- sponds to a distance of about one inch from the extreme edge. A little further on, the hardness drops a few points and averages 9-10 along the main body of the object, showing that the metal is in a soft condition. It will be noted that the grains shown in fig. 58 are abundantly twinned, of rather uniform appearance, and equiaxed. No lines of deformation are visible and it is evident that the metal has reerystallized after permanent distortion at some stage of its existence. Whether the deformation was effected by cold-working or by forging at a high temperature cannot be foretold from this structure alone, since, in either case, the metal must have been of Bronzes from Machu Picchu, Peru. 585 left at rest at high temperature after working (ordinary annealing after cold-working or a sojourn at the forging tem- perature while work was confined to another portion) and the same ultimate grain characteristics are developed in either ease. The observed grain count of 13 indicates that this temperature was in the neighborhood of 800.° The condition of the metal near the edge, however, throws much light upon this question. Passing towards the edge from the point at which photo-micrograph, fig. 58, was taken, the grains very soon begin to show lines of deformation and elongation in the direction of the principal axis. Moreover, smaller recrystallized units begin to appear among the distorted parent grains. This structure is shown im fig. 57, after ordi- nary etching with ammonia and hydrogen peroxide. (This method of etching permits a clearer photographic reproduction of the finer detail than is ordinarily the case with ferric chloride, which gives pleasing contrasty effects with coarser structures, as in fig. 58.) As the edge of the chisel is approached, fig. 56, the grains become finer, the lines of defor- mation more abundant, and the elongation more pronounced. In other words, the proportion of strain-hardened metal is greater. This was tested by the scleroscope which gave average values of 15, one half inch from the edge; 24, five sixteenths of an inch from the edge; and 30, three sixteenths of an inch from the edge. The structures shown in figs. 56 and 57 are characteristic of hot-working at a temperature in the vicinity of 800° with continuation of working as the temperature falls to a low value. As a grain is deformed under the blow of the hammer, it spontaneously recrystallizes at those points where the strain was greatest. In tapering an object, as in the present case, deformation is greatest in the thinner portions, hence there will be more points of recrystallization in these regions and more abundant lines of deformation if the cold-work is confined largely to these regions, as is likely to be the case. The same succession of structures was obtained in the laboratory by hot- and cold-working a piece of metal from the shank of the chisel to a tapered edge.* The structure at one point is shown in fig. 62, with diagram of position at the left, fig. 61. It will be observed that this structure corresponds closely with that shown in fig. 57. Cold-working has been somewhat more * Metal from this chisel was used in order to obtain the best possible comparison. Precisely the same effects were obtained with alloys prepared in the laboratory. 586 C. H. Mathewson—Metallographic Description severe, as shown by the more prominent lines of deformation — (wavy lines giving a crevassed appearance to the grain-frag- - ments). If the metal is worked only while the temperature is at red heat and thereupon allowed to cool undisturbed (rapidly), the same grain characteristics are observed except for the absence of lines of deformation. The effect of this treatment is shown in fig. 64, with diagram of position at the left, fig. 68. In fig. 14, it is observed that the edge of the chisel has been turned in use. This condition must have influenced the structure somewhat at the extreme edge. Such alteration does not, however, extend far from the edge. For example, the position represented by fig. 56, is near enough the edge to show a fissure where the metal has split in use. Yet there seems to be no interference with the normal elonga- tion of the grains as prescribed by the original process of hammering from the sides. It seems probable that more than one heat was used in forging the blade of this chisel. This would explain the uniform coarse grain of fig. 58, representing the condition of the metal where the taper begins, and the fact that the grain is more highly refined in this vicinity than along the body of the object. According to this interpretation, the whole blade was partly forged to shape in the first heat. On reheating to the forging temperature, coarse grain developed and was subsequently unaltered in this portion of the piece, but was altered, as described above, further along towards the edge. The structure at the head of the chisel is shown in fig. 60, with diagram of position at the left, fig. 59. This represents the general structure in the body. It clearly indicates that the object preserves approximately the shape of the original cast- ing, since large unrecrystallized fragments of the original grains may still be seen. Owing to the use of high tempera- tures in forging or annealing no traces of the internal grain structure remain. The small dark spots represent particles of sulphide. Normally a large part of this material separates out between the casting grains, and in this phcoto-micrograph some of it may still be seen in the grain boundaries. In completely recrystallized structures the sulphide is distributed without relation to the grain boundaries, e. g., fig. 69. The present structure may have been developed by reheating after hot- working or by annealing after cold-working. Since the blade was almost certainly hot-worked, the former interpretation seems more reasonable. of Bronzes from Machu Picchu, Peru. 587 Object No. 7 (cf. Table I). That the Inca craftsmen were able to show considerable variety in their methods of shaping these small implements became evident on examination of the small knife with orna- mental llama head at the extremity of the handle shown in fig. 15. A number of cast structures, some of them unaltered, some altered by cold-working with annealing, and one altered by hot-working, have been described. The present specimen is of a duplex character, in that two distinct varieties of metal were used in its construction. A vertical section sawed through the center of the ornamental head and part of the shank is shown diagrammatically in fig. 65. This diagram was made from the polished, but unetched section, in which the boundary line between the two kinds of metal was plainly apparent owing to a difference of color and sharp discontinuity at the junction. These facts are very well represented in the photo-micrograph, fie. 66, taken at a magnification of 12 from the unetched section. The color difference is shown by the well developed contrast between the two alloys, further accentuated by a den- dritic structural resolution of the other alloy, which is mainly due to a difference of hardness, and, consequently, of elevation between the central cores and boundary zones of the branched erystallites. The presence of numerous pits, or blow-holes, located in the boundary zones where final solidification occurred has also contributed to this effect. It is not necessary to resort to etching, or to a higher magnification, in order to demonstrate that the metal composing the main portion of the ornamental head is now in the condition originally produced by casting. This is evident from the casting structure revealed in fig. 66. When drillings were taken from this specimen for analysis, its duplex character was not recognized and only the lower part of the handle and the upper part of the blade were sampled. This is the metal which projects up into the head and is of a reddish tint, corresponding to its low tin-content of 3 per cent. It is evident from the white color of the surrounding metal that this contains a much higher percentage of tin. Fig. 67 represents the structure of the metal composing the shank and blade of the knife. It is entirely homogeneous and has been annealed after deformation. The grain count of 18 indicates that a temperature of 700-800° was reached in annealing. Since the outside metal possesses a casting structure, it is natural to conclude that it was cast in place around a core of 588 C. H. Mathewson—Metallographic Description the metal which forms the shank and blade. It is seen from the two sections, a and b, of fig. 65, which are parallel and about one sixteenth of an inch apart, that the central core bends over at the top and develops an appendage which is also embedded in the second metal. Such a complicated fitting could only be obtained by flowing the outside metal around the imner piece. The general shape and appearance of the aperture indicates that it was left in casting. Naturally the details of this interesting casting practise cannot be described. Object.No. 8 (cf. Table I). Only a single section was cut from the axe illustrated in fig. 16. The position of this is shown in fig. 68, and the structure well under the skin in fig. 69 of the same plate. The most striking feature of this photo-micrograph is furnished by the large amount of cuprous sulphide which is normally bluish-gray in appearance but has fallen away in places, as a result of grinding and polishing, leaving black pits. According to the analytical results given in Table I, there is 0.44 per cent of sulphur in this specimen, a value which corresponds to about 2.2 per cent of cuprous sulphide. By comparing the present photo-micrograph with the preceding one, fig. 67, which illustrates the structure of the copper-rich metal in the knife just deseribed, a good idea of the value of the microscope in estimating the sulphur-content of a bronze may be obtained. A trace of sulphur was indicated in the qualitative examination of this knife. It is plainly evident in fig. 67; less abundant and structurally finer than in fig. 69, but of the same general characteristics. In spite of the large amount of sulphide contained in this object, its cold-working properties are retained to a surprising degree. A piece measuring 9 X 7 X 16 mm. was eut from the head of the axe and cold-rolled in several passes to a round rod 4 mm. in diameter and 35 mm. long. Further extension could not be obtained without fracture. As a result of this treatment, the particles of sulphide have adapted their own shape to the changing form of the bronze matrix and now appear elongated without evidence of fracture. (They do not show cracks or fissures on the polished surface.) The photo- micrograph, fig. 87, which represents this condition, was taken after momentary annealing at a full red heat. The coarse grain of fig. 69 indicates that the object has been thoroughly annealed at a temperature not lower than of Bronzes from Machu Picchu, Peru. 589 800°. The observed grain count of 10 can be duplicated in ordinary bronze of approximately the same tin-content by annealing for a period of some fifteen minutes at 850°, while, at 775°, more than an hour is required. Separated impuri- ties, such as cuprous sulphide in the present case, as a rule interfere with coalescence and retard growth of grain. The large quantity of sulphide shown in fig. 69 would probably raise the normal annealing temperature corresponding to the grain represented. The object has not been worked severely enough to sensibly alter the rounded character of the sulphide particles. The original casting was, therefore, similar in size and shape to the finished object. Corresponding to the thorough annealing treatment adopted, the metal is completely homogenized. Object No. 9 (cf. Table I). The knife illustrated in fig. 17 is the only one of the analyzed objects which contains no metal alloyed to the copper base. Cuprous oxide ordinarily occurs in copper and this is the case with the present object. The amount of oxide is considerably greater than that encountered in ordinary com- mercial copper of the present day, in which reduction of the oxide has been effected by the operation of poling, leaving some 0.3-0.6 per cent of oxide in the tough pitch metal. Photo- micrograph, fig. 72, represents the structure of the metal in the shank just where it joins the blade. Cf. diagram of position, fig. 70, a. Here, the casting structure may be recognized in the dark areas of dotted appearance which represent the eutectic of copper and cuprous oxide somewhat distorted out of the original casting position by mechanical treatment. According to a rough estimate, the eutectic constituent occupies about one third of the total area of the photo-micrograph. Since this constituent itself contains all of the cuprous oxide in the pro- portion of three and one half per cent of oxide to ninety six and one half per cent of copper, there are about one and two tenths per cent of oxide in the alloy (0.33%, 0.0314 X 100). This is twice the quantity previously mentioned. That the material of the shank has been moderately worked and then annealed is evident from the moderate extension of the eutectic areas and the recrystallization with twin lamellae shown in fig. 72. The material of the blade has been worked much more extensively, as would be expected from its thinness (minimum, 0.040”). This is clearly shown by the fact that Am. Jour. Sci.—FourtH SrErtes, Vou. XL, No. 240.—Drcemper, 1915. 590 C. H. Mathewson—Metallographic Description each group of oxide particles, originally scattered through the - copper matrix so as to constitute an equiaxed patch of eutectic, has been dragged out into a continuous train giving the appear- ance of a dotted line in the photo-micrograph, fig. 71; dia- gram of position, fig. 70, b. The annealing characteristics of copper containing oxygen in this form have not been sufficiently defined to justify a prediction as to the probable annealing temperature. The direction of elongation is shown in the diagram, fig. 70, a and b. It is the same in both shank and blade, viz., from handle to blade and from base of blade to edge. More- over, in the flare of the blade, the metal decreases in thickness from the center both towards the edge and the base of the blade. These facts suggest that the knife was made from a flat T-shaped casting in which the stem was nearly as long as the finished shank and handle and the cross-piece practically equal in length to the finished blade from tip to tip. Object No. 10 (ef. Table I). As may be seen from the illustration, fig. 18, this object represents a broken axe blade. ‘The axe previously described (Object No. 8) is also broken. Baessler (28) states that some of the objects in his collection were mutilated in a manner which could not have occurred through use. He believes that they were intentionally mutilated before burial with the owner. The pres- ent fragment shows a comparatively coarse recrystallized grain with a small amount of tin-rich material which has not passed into solution during the annealing treatment. This structure is shown in fig. 74, diagram of position, fig. 73. There are no indications of hot-working on the section examined. If hot-worked at all, the piece received a subsequent anneal at bright red heat whereby the characteristics of forging were replaced by comparatively uniform coarse grain. The structure very near the edge is notable in that it shows severe deformation of a character which could only result from an upsetting operation; not from the ordinary method of shaping the edge by lateral application of force (hammering or rolling along the sides). Fig. 75 represents this feature. This photo-micrograph is mounted to conform with the direc- tional requirements of the diagram, fig. 73. Lines of defor- mation and elongation of the grains are plainly evident. In its present condition the blade is nicely rounded to a blunt edge bearing scratches which indicate that it was finished on of Bronzes from Machu Picchu, Peru. 591 a stone or in a similar manner. It appears reasonable to con- clude that this implement was used in heavy work, perhaps upon stone, whereby the edge suffered severely and was appropriately dressed from time to time. Object No. 11 (cf. Table I). Only the flattened head of the cloak pin illustrated in fig. 19 was examined. A number of hardness tests were made on different parts of the surface after removing the outer skin. Scleroscopic numbers varying from 14 to 18 were obtained. This indicates that the metal was finished hard and the recrystallized structure shown in the photo-micrograph, fig. 77, supplies additional information to the effect that inter- mediate annealing treatment was applied in flattening the head. Some undissolved tin-rich material (light centers with broad black boundaries due to inequality of focus) is present. Abundant lines of deformation, corresponding to the hard finish, may be seen. The head of this object bears an aperture just above the union with the shank, as shown in the diagram, fig. 76. This aperture almost certainly existed before the head was flattened, since it is now asymmetrically conical in shape, a condition which may easily be duplicated by hammering over one face of an ordinary cylindrical aperture. In an experiment of this sort, the face held on the anvil remains open, while the ham- mered face tends to close. Thus, the head of the present object appears to have been hammered on one side only, viz., the side opposite to that represented in fig. 76. In all prob- ability, the aperture was originally provided in casting the metal. Object No. 12 (ef. Table I). This object, illustrated in fig. 21, is a thin circular disc shghtly dished from rim to center and bearing a slotted handle. Many objects of this sort have been found in Peru (the present collection embraces several) and they are generally supposed to have been used as mirrors. Baessler (1. c. 28) describes a similar object containing about 9 per cent tin. The present piece contains about 5.34 per cent tin and is distinctly inferior in whiteness to alloys containing 9 or 10 per cent tin. It is worthy of note that bronzes of this character, while capable of high polish, would require frequent polishing to preserve a highly reflecting surface as they tarnish readily. They are in 592 C. H. Mathewson—Metallographic Description no sense comparable to the copper-tin alloys known as speculum metal which contain in the neighborhood of 30 per cent tin and retain their lustre for a long time. The order of curvature possessed by this dise may be sensed by referring to the diagram, fig. 78, which was prepared by running the pencil around the edges of a centrally located trans- verse section of the object placed face down on the paper. This serves as a diagram of position and two representative photo-micrographs are shown on the same plate. Fig. 79 represents the structure of the handle. This is an unworked casting structure, nearly homogenized by heat treatment, as shown by the faint, but readily distinguishable cores (shadowy markings all over the surface). Passing from the handle into the disc, the large irregularly bounded casting grains give place to a much finer recrystallized structure of the usual characteristics. This is quite uniform all over the section of the disc and is fairly represented by fig. 80, which, at the stated magnification (72), practically spans the section from one side to the other at the point indicated. The grain count of 22 indicates a moderate annealing temperature; neighbor- hood of 750°. As in the case of most of these specimens, it is not possible to decide whether hot-working did or did not occur during some stage of the shaping process. The section examined bears evidence of undisturbed annealing after the hot- or cold-working had been effected and subsequent alteration in some places by cold-working, notably in the vicinity of the handle. Thus, whatever may have preceded, the final changes were produced by cold-working and the last severe work was probably done in the region of the handle. Object No. 18 (cf. Table I). This piece is chiefly of interest as an example of a partially completed object. By comparing fig. 22, illustrating the present object, with fig. 27, illustrating a form of tweezers commonly used for removing hair (cf. Joyce (6), p. 130), it is readily seen that the finished tweezers, except for the per- foration, will result when the present shape is bent around into proper form. A section was prepared by sawing the piece along its middle line from end to end. A diagram of this section is given in fig. 81, and a photo-micrograph of the spot indicated is shown in fig. 82. The grain count varies along the section from 13 to 25. This indicates that some parts were worked considerably more than of Bronzes from Machu Picchu, Peru. 593 others before the last annealing treatment, which, in view of the residual inhomogeneity (shadowy cores and distinct traces of the original tin-rich network), was of moderate intensity (e. g., brief period, neighborhood of 750°). Object No. 14 (ef. Table I). There are a large number of thin, flat knives in the collec- tion of the form illustrated in fig. 20. Eight of these, in addition to the copper knife of similar form which has already been described, were analyzed (cf. Table I) and one was examined in the blade, shank and handle. As would be expected, a structure indicative of severe working with asso- ciated annealing was observed. ‘The general conclusions drawn from the examination of the present object should apply fairly well in all cases. Two structures at the upper end of the object have been chosen for reproduction. One of these, fig. 85, as may be seen from the adjacent diagram of position, fig. 83, represents the shank just before it bends around into the hook-shaped handle. Here, as well as all along the shank and handle, the sulphide constituent is threaded out to an extent indicative of at least 100 per cent elongation in the working of these parts. This condition was very closely duplicated by cold-rolling a piece of metal taken from Object No. 8, which contains a large quantity of normally equiaxed sulphide, so as to obtain an extension of 13714 per cent (from 16 mm. to 38 mm.). After extension to 35 mm. the piece was annealed by bringing it momentarily to red heat. The final condition is shown in fig. 87. This surface was prepared so as to show the continuity of the elongated sulphide particles as well as possible without regard to the detail of the matrix. In preparing fig. 85, this feature was not primarily considered and the smudgy appear- ance of the lines is due to imperfect preparation in this respect. The sulphide particles are, in reality, enormously extended without apparent fracture. In the blade of the knife, the particles of sulphide are small, numerous and often grouped, but not distinctly elongated in any one direction. It thus appears that the blade has been worked so as to extend the metal more or less uniformly in several directions. Photo-micro- graph, fig. 84, represents the structure passing from the inner edge at the point of the bend towards the center. Very fine recrystallized grain is observed at the edge where the original grains were most severely overstrained in bending. Towards 594 C. H. Mathewson—Metallographic Description the center, where the minimum strain was felt, the original coarser grain is relatively little altered. These conditions would be produced by suitably annealing a bend formed in soft metal of the structural characteristics shown in fig. 85 and they indicate that the piece received annealing treatment after the handle was bent into shape. Scleroscopic tests on this object show hardness numbers of about 14 on the shank and 27 on the thin parts of the blade, which was set im plaster for these tests. Thus, there is considerable temper in the blade and some in the shank (soft, annealed alloys of this composition give hardness numbers below 10). It may be said, in conclusion, that nothing can be learned from this examination relative to the original size and shape of the casting from which the knife was made. Several knives may have been made from one cast piece although there is no particular reason for believing this to be true. Im any event, the casting structure is totally obliterated and the piece in its present condition cannot bear much resemblance to original stock used in its construction. ae Object No. 15 (ef. Table I). Only a small piece from the axe illustrated in fig. 23 was examined. ‘This was cut from the extremity of one of the arms or branches forming the head, as shown in the diagram, fig. 88. The corresponding photo-micrograph is shown in fig. 89. Its comparatively fine grain (counting 28) indicates final annealing treatment hardly above dull red heat unless extremely brief, and the presence of faint cores indicates that the object has never been held at a temperature corresponding to bright redness for any considerable period of time. Deformational characteristics are present and the axe has been left in the hammer-hardened condition. Object No. 16 (ef. Table I). The object illustrated in fig. 24 appears to be a chisel which has been broken off about 234” above the edge and twisted so that vertical planes passing through the edge and the opposite end, respectively, would intersect at an angle of about 12°. One of the vertical edges of the chisel was ground down flat, polished, etched, and examined along its entire length. A diagram of this surface is given in fig. 90. At 72, the finer detail of the structure is imperfect owing to the small size of the recrystallized grains and the presence of prominent primary of Bronzes from Machu Picchu, Peru. 595 zonal structure. The general effect at this magnification is shown in fig. 91. All parts of the section look very nearly alike, but, under the higher powers, it is seen that the deformational characteristics are more pronounced in the vicinity of the edge; the cores, or zones, themselves show elongation near the edge and lines of deformation are abundant, both in the frag- ments of the primary grains and in the recrystallized grains. The structure of this portion is shown in fig. 92 at a magni- fication of 220. Owing to the selective etching in copper- rich zones, the finer detail cannot be brought out with uniform clearness all over the surface; if the etching is carried far enough to bring out the detail in the light zones, it will be obscured in the dark zones (fig. 91); if it is only carried far enough to bring out the detail in the dark zones, there will be very little detail in the light zones, as in the present instance, fig. 92. Scleroseopic tests confirm the micrographic conclusions, in that they show a maximum hardness number of 27 near the edge (14” from the extreme edge) and a lower number, 15-16, in the body of the chisel. The grain count of 55, which was originally obtained at 220 and recalculated for the standard magnification, 72, is only approximate but indicates, in con- nection with marked residual inhomogeneity, that the chisel has received only very light annealing treatment. From the character of the zones, in particular their equiaxed appearance, except near the very edge, it is clear that the present form of the object is not widely different from that of the casting used in making it. Object No. 17 (cf. Table I). This object is described in Table I as an irregular mass. The surface illustrated in fig. 25 is covered with large and small warty excrescences, while the opposite side is smooth and partially rounded at the edges. It is not improbable that the object is, in reality, a button which was left in the pot, or crucible, after pouring a heat of metal. In conformity with this assumption, the grain, as shown in fig. 94, at 12 and in fig. 95, at 72, is coarse enough to indicate comparatively slow cooling. The tin-rich zones appear dark in both photo- micrographs owing to their rounded character after etching and the corresponding inequalities of focus. Some patches of the a + 6 complex may be seen when the specimen is examined with the higher powers. Experiments with small castings of 596 C. H. Mathewson—Metallographic Description approximately this composition have shown that the metal can- not be brought to red heat without absorption of the complex. It is highly improbable that the present object has ever been annealed. A number of twinning bands were observed on the surface of the section examined, fig. 93, which was cut, at random, from one side of the object. Several of these bands are shown near the center of fig. 95. They may represent congenital twinning in the sense that the object was forcibly handled so as to overstrain it locally just after it had set, or at a tempera- ture high enough to cause recrystallization. Thus, the red-hot button may have been pulled or dumped from the pot. Object No. 18 (cf. Table I). The long, heavy needle illustrated in fig. 26 was examined only along a short longitudinal section passing down through the eye. A diagram of this is shown in fig. 96. The struc- ture, at the point indicated, is shown in fig. 97. This point represents the mechanical union between the looped end of the metal and one of lateral flaps which is hammered down to hold it in place. The nature of this construction is well enough evident from the illustrations previously cited without further explanation. As may be seen in fig. 97, abundant lines of deformation remain in evidence of the severe hammering effected at this vicinity, i. e., the metal has not since been annealed. Scleroscopic tests have shown hardness numbers of 21, 18, and 15, at the points a, b, and c, respectively, of fig. 96. The grain count of 22 indicates that the last annealing treatment was of the type most frequently observed in these objects, viz., probably a brief period at moderate red heat. Object No. 19 (cf. Table I). Tweezers ‘of the form illustrated in fig. 27 are well repre- sented in the present collection. A partially completed speci- men has already been described (No. 13). One of the finished specimens was flattened into the form which it obviously pos- sessed before bending and polished on one side in those spots which reached the abrasive without resorting to deep cutting. After the examination, it was readily bent back into shape without any sign of fracture. This indicates that the metal was comparatively soft when originally bent into final shape. After being worked flat into proper form, it was probably annealed, pierced by driving a pointed object through the thin of Bronzes from Machu Picchu, Peru. Bom central portion, and then bent around into shape. The set of the aperture indicates that it was punched rather than drilled. The metal shows a finer recrystallized structure and more prominent cores than were encountered in the case of the partially completed specimen. It, therefore, received less drastic annealing treatment. Its structural characteristics are intermediate between those shown in fig. 49 and fig. 91, more closely approximating the latter. Such an effect would be produced by annealing for a few minutes (5-10) at 700°, or by just allowing the object to reach a temperature some 50° higher. Object No. 20 (ef. Table I). This object, which weighs over two pounds, is the most mas- sive piece of the entire collection. As may be seen in fig. 28, it represents the lower portion of what may be conveniently called a crowbar. Scleroscopic tests on smooth-filed portions of the surface show that the metal was finished hard, i. e., the hardness numbers vary between 18 and 25. The bar was used in a tensile test which gave an elongation of 6 per cent in 2 inches and an ultimate strength of 27,800 pounds per Square inch. Worked bronze of this composition, when hardened so as to possess very little ductility, should show far greater strength than was obtained in this case. The present combination of ductility and strength indicates metal of very poor quality. A polished and etched section near the fracture revealed a multitude of large and small holes with blackened, or oxidized surfaces. These predominate near the center of the cross section. Nearer the surface the holes have been closed in working the metal but without removing the source of weakness, since no true welding has occurred. The metal appears to have oxidized as a result of some initial porosity and much annealing treatment. Object No. 21 (cf. Table I). The small piece of metal illustrated in fig. 29 is one of several small shapes, bars, rods, etc., which probably constituted material intended for subsequent shaping into objects of some particular significance. This piece has been worked and annealed to a condition of complete homogeneity. ‘The absence of deformational characteristics and the observed hardness number of 9-10 show that it is now soft and well adapted to further mechanical treatment. Sheffield Scientific School of Yale University, New Haven, Conn., June 20, 1915. 598 C. H. Mathewson—Metallographic Description BiIsLioGRAPHY. (1) For a description of the work of this expedition, see papers by Hiram Bingham; The Wonderland of Peru, Nat’l. Geogr. Mag., April, 1913, and The Story of Machu Picchu, Nat’] Geogr. Mag., Feb., 1915. (2) Garland, Communication on a paper by Hudson, The Microstructure of German Silver, J. Inst. Metals, ix, 118-119, 1913. Later paper by the same author, Metallographical Researches on Egyptian Metal Antiquities, J. Inst. Metals, x, 329-343, 1913. (8) Rose, On the Annealing of Coinage Alloys, J. Inst. Metals, viii, 86-125, 1912. (4) Heyn, and Bauer, Kupfer und Schwefel, Metallurgie, iii, 73-86, 1906. (5) Boman, Antiquités de la Region Andine, de la Republique Argentine, et du Désert D’Atacama, Vol. IJ, Paris, 1908. (6) Joyce, South American Archeology, N. Y., 1912, p. 210. (7) Wust, Metallurgie, vi, 769-792, 1909. (8) Haughton and Turner, J. Inst. Metals, vi, 192-212, 1911. (9) sue A Treatise on Brasses, Bronzes and Other Alloys, N. Y., 1907, p. 214. (10) v. Miller, Studien itiber die Einwirkung der wichtigeren metal- lischen und nicht-metallischen Zusiitze auf normale Kupfer—Zinn—Bronze, Metallurgie, ix, 63-71, 1912. (11) Foote and Buell, this Journal, xxxiv, 128-132, 1912. (12) Shepherd and Upton, J. Phys. Chem., ix, 441-476, 1905. (13) Guertler, Metallographie, Berlin, 1911, Vol. 1, p. 660-690. (14) Hoyt, On the Copper-rich Kalchoids, J. Inst. Metals, x, 235-274, 1913. (15) Heyn, Materielenkunde fur den Machinenbau, Berlin, 1912, ILA, p. 212-218; 225-241. (16) Tammann, Lehrbuch der Metallographie, Leipzig, 1914, p. 74- 34. (17) v. Moellendorf and Czochralski, Technologische Schlusse aus der Kristallographie der Metalle, Zeitschr. Deutsch. ‘Ing., Juni, 1913; also Czochralski, Gegen die Translations-Hypothese als Ursache der Bildsam- keit von Metallkristallen, Intern. Zeitschr. fiir Metallographie, vi, 289-296, 1914. (18) Lehmann, Spontane und Erzwungene Homéotropie, Intern. Zeitschr. fiir Metallographie, vi, 217-237, 1914. (19) Rosenhain, Der kristallisierte und der amorphe Zustand der Metalle, Intern. Zeitsch. fiir Metallographie, v, 65-106, 1913. (20) Beilby, The Hard and Soft States in Metals, J. Inst. Metals, vi, 5-43, 1911. A bibliography of Beilby’s earlier papers is given by Rosenhain (19), p. 103. (21) Portevin, Revue de Metallurgie, vi, 814-818, 1909. (22) Portevin, Revue de Metallurgie, x, 677-721, 1913. (23) Fick, Pogg. Ann., xciv, 59, 1855. (24) Nernst, Zeitschr. Phys. Chem., li, 613, 1888. (25) Roberts-Austen, Phil. Trans. Roy. Soc., London, IS7A, 383-415, 1896. (26) Gulliver, J. Inst. Metals, ix, 120, 1913. (27) Gulliver, J. Inst. Metals, xi, 252, 1914. (28) Baessler, Altperuanische Metallgeriite, Berlin, 1906. of Bronzes from Machu Picchu, Peru. 599 600 C. H. Mathewson—Metallographic Description of Bronzes from Machu Picchu, Peru. 601 602 CER Mathewson—M, achu Picchu Bronzes. Magnification, 10x. EHtched with acidified ferric chloride.—These photo-micrographs, to- gether with those on the following pages, have been reduced one-eighth in the engravings. Fig. 38. Magnification, 72 x (see remarks on preceding page). Fig. 37, etched with ammonia and hydrogen peroxide, followed by acidified ferric chloride. Figs. 38, 39, etched with ammonia and hydrogen peroxide. Fig. 41. - Fig. 45. _ Figs. 41, 44, and 45, magnification, 72x. Figs. 42 and 43, magnification, 300x. Fig. 41, etched with ammonia and hydrogen peroxide, followed by acidified ferric chloride. Figs. 42, 43, 44, and 45, etched with acidified ferric chloride. Am. Jour. Scit.—Fourts Srries, Vou. XL, No. 240.—DecremBer, 1915. Al Fig. 47, magnification, 12x. Fig. 48, magnification, 220x. Fig. 49, magnification, 72x. Figs. 47 and 48, etched with ammonia and hydrogen peroxide, followed by acidified ferric chloride. Fig. 49, etched with ammonia and hydrogen peroxide. Fie. 51. Figs. 51 and 52, magnification 18x. Fig. 54, magnification 72x. Figs. 51 and 82, etched with acidified ferric chloride. Fig. 54, etched with ammonia and hydrogen peroxide, followed by acidified ferric chloride. Vig. 55. Magnification, 72x. Figs. 56 and 57, etched with ammonia and hydrogen peroxide. Fig. 58, etched with ammonia and hydrogen peroxide, followed by acidified ferric chloride. Trig. 59. Magnification, 72x. Fig. 60, etched with ammonia and hydrogen peroxide, followed by acidified ferric chloride. Figs. 62 and 64, etched with ammonia and hydrogen peroxide. Fig. 66. Tig. 68. Fig. 66, magnification, 12x. Figs. 67 and 69, magnification, 72x. Fig. 66, unetched. Figs. 67 and 69, etched with ammonia and hydrogen peroxide. Fig. 71. Iie, (0: == Wie, (2. aes Magnification, 72x. Etched with ammonia and hydrogen peroxide. Fig. 74. Fig. 75. K IG. 77. TG. 76. Tr Etched with ammonia and hydrogen peroxide. 2x. 7 n, Magnificatio ae oo rT eee goog ha Sie) Tic. 81. ne eae bee PO Magnification, 72x. Etched with ammonia and hydrogen peroxide. ce Fic. 84. Fig. 86. Magnification, 72x. Etched with ammonia and hydrogen peroxide. Fig. 88. Fic. 89. iS ‘ 2 we Nt: eae 14 be RE 4 LES FO TL Vic. 90. eg eee RTS Bs Figs. 89 and 91, magnification, 72x. Fig. 92, magnification, 220x. Etched with am- monia and hydrogen peroxide. Fig. 94, magnification, 12x. Figs. 95 and 97, magnification, 72x. Etched with ammonia and hydrogen peroxide. Ruth R. Mook—A New Cephalopod. 617 Art. XLIL—A Mew Cephalopod from the Silurian of Pennsylvania ; by Rura Razprr Moox. Introduction. Dvrine the summer of 1913 while working on the Blooms- burg Red Shale and the overlying limestones, in a limestone quarry about 7 miles southwest of Bloomsburg, Pennsylvania, the writer’s attention was called to a rare cephalopod col- lected and owned by Mr. Guy A. Mowry of Grovania. The specimen was first discovered by quarrymen while excavating rock to be sent to the lime kilns. Mr. Mowry, who has care- fully collected fossils for some years, was called in to make a eareful note of the exact position and horizon of the specimen. The fossil was found in a bed of limestone, generally known in eastern Pennsylvania as the Bossardville, of Upper Silurian age. : 2S far as can at present be determined, a similar form has not yet been described for the United States, and because of its apparent rarity it has seemed worth while to call attention to it. Mr. Mowry has therefore kindly loaned the specimen for the purpose of description. _ The name Trochoceras grovaniense is here proposed for this new species. The genus Trochoceras. The genus Zrochoceras was proposed without any concert of action. by Barrande in 1848 and Hall in 1852 for fossil species generically similar but not generically identical. At a meet- iug of the ‘Freunde der Naturwissenschaften in Wier,” on September 10th, 1847, Dr. Franz vy. Hauer called the attention of the members to some new cephalopods from the Silurian of Bohemia, which had been sent in by Barrande. The proceed- ings of this meeting were published in 1848.* The report included Barrande’s original description of the genus Z’rocho- ceras, which reads as follows: “ Trochoceras (Barrande). The shell is characterized by the peculiar nature of its enrollment. The revolutions are laid upon one another in a spiral manner, so that the shell itself is not symmetrical. Zrochoceras parallels therefore the genns Turrilites in the family of the Ammonitidae. All the species which Barrande discovered belong to the lower division of the * Berichte tiber die Mittheilungen von Freunden der Naturwissenschaften in Wien; gesammelt und herausgegeben von Wilhelm Haidinger, iii, Nr. 1-6, p. 264, July-December, 1847. Vienna, 1848. 618 Ruth R. Mook—A New Cephalopod. Upper Silurian System.* No genotype was named. In a later volume of the same publication a list of species was Fig. 1. Fie. 1. Pencil drawing of type specimen of Trochoceras grovaniense, sp. noy. 7/16 nat. size. given. In this list Zrochoceras davidsoni was mentioned first and this is therefore Barrande’s genotype.t+ * Loe. cit., p. 266: ‘‘ Trochoceras (Barrande). Durch die Art der Kinrol- lung der Schale characterisirt. Die Umgiinge sind naimlich in einer Schrau- benlinie aneinandergelegt, so dass die Schale selbst nicht symmetrisch ist. Trochoceras entspricht demnach dem Geschlechte Turrilites aus der Familie der Ammonitidae. Alle Arten, die Barrande auffand, gehoren der unteren Abteilung des oberen silurischen Systems an.” + Haidinger’s Berichte, IV, p. 208, 1848. Ruth Rk. Mook—A New Cephalopod. 619 The name proposed by Hall was printed in the second vol- ume of the Paleontology of New York in 1850, although the volume was not issued until 1852.* Hall’s genotype was Trochoceras gebhardi. His original description of the genus follows: “ Turbinate or trochiform ; spire elevated, more or less ven- tricose, umbilicated ; aperture rounded or round oval; volu- tions above the outer one septate; siphuncle submarginal or dorsal. “Tn the specimen from which the generic description is principally made, the septa are strongly arched from the inner basal angle of the volution to the outer one, advancing on the outer angle towards the aperture.” In the above description owing to a confusion of terms the sipliuncle is recorded as sub- marginal or dorsal, whereas in present day usage it would be recorded as submarginal or ventral. In 1894 Hyatt pointed out that Hall’s species are quite dis- tinct from Barrande’s, and that they do not belong to any genus yet described from Bohemia. Zvrochoceras Barrande is not generically identical with Zrochoceras Hall. Hyatt sug- -gested that the name Zvochoceras Barrande be retained and for Hall’s forms Zvochoceras gebhardi and turbinatum he pro- posed the new name Mttroceras with Mitroceras (Troch.) gebhardi as the genotype.t Jitroceras Hyatt differs from Trochoceras Barrande in the high turbinate spire, and in the deep and sharply angulate umbilicus. Reasons for assigning the new species under discussion to the genus Trochoceras. The formation in which the specimen under discussion was found has undergone little diastrophic movement. The beds dip at an angle of about 40 degrees to the south, but outside of uplifting and tilting there is no evidence of any great dis- turbance such as twisting, mashing, ete. In the overlying formation, the so-called Bastard limestone, fossils are abundantly preserved and show no evidences of distortion. The speci- men under discussion is considerably warped, apparently indi- eating that it was a low spired trochoceracone rather than a flat nautilicone ; its warped condition then can be reasonably explained by the normal crushing of a cone-shaped shell. Trochoceras grovaniense sp. nov. Shell large, probably trochiform ; spire very slightly ele- vated ; maximum diameter 158™™; diameter of outer whorl a Palaeontology of New York by James Hall; vol. ii, p. 335. Albany, 1852 + Phylogeny of an Acquired Characteristic. Alpheus Hyatt. Proc. Amer. Phil. Soc., vol. xxxii, No, 148, p. 502, 1894. 620 Ruth R. Mook—A. New Cephalopod. about 25™" (varying according to state of compression); num- ber of whorls 4, the outer one incomplete ; increase in diameter of whorls very gradual; diameter of inner whorl about 10™™; septa simple, very slightly curved, about 4™™ apart on outer whorl and 3"™ apart on inner side of inner whorl; annulations on outer whorl about 2°5™™ apart; siphuncle ventral. Owing to the erushed condition and poor preservation of the shell the extent of the elevation of the spire is unknown. From the irregular warped surface of the specimen it seems reasonable to suppose that the spire was very broad and low. The exact nature of the umbilicus is indeterminate ; presum- ably it was broad, shallow, and not angular. ‘The precise character of the siphunele is also unknown. There is only one place on the shell where the siphuncle appears; here it is only very poorly preserved. Its ventral position, however, can be determined. Trochoceras grovaniense seems most nearly related to Z. priscum Barrande, Etage, of Bohemia. It differs from the latter in its greater size, more gradual increase in the size of the whorls, and n having the septa straighter and slightly farther apart. Paleontological Laboratory, Columbia University. Jaggar—Activity of Mauna Loa. 621. Art. XLIII.—Actwity of Mauna Loa, Hawaii, December- January, 1914-15 ; by T. A. Jaaear, JR. Tue activity of Mauna Loa, inaugurated November 25, 1914,* continued through December and part of January at ~ the summit crater and then ceased, this movement constituting the first phase of a new eruptive period, quite in accord with the habit of this voleano. No lava flow from the flank of the volcano has occurred up to the time of writing (July 22, 1915), and none need be expected for some years. The notes which follow are taken from the records of the Hawaiian Voleano Observatory. The following corrections should be noted of statements made by the writer in his previous notes.* The interval of repose preceding this outbreak was the maximum recorded since 1868, for it appears that within the eight years from 1888 to 1895, a summit outbreak occurred, namely November 30—December 2, 1892, previously overlooked when that inter- val was cited as the maximum. This outbreak of 1892 is considered a first preliminary of the eruptive period which culminated in the lava flow of 1899, the second preliminary occurring in 1896. If we treat a complete eruption as consist- ing of one or more preliminary summit fountainings leading to a concluding lava flow, then the intervals of repose have been as follows: eos A phil, to L670, Jamiaty. 2 2213242 20 months PSii webruary., to, 1880; Mays = <2 ee bey th 1881, November, to 1887, January -_------ 2a 1887, February, to 1892, December._-.-... 70 “ 1899, August, to 1903, September ____-_--- DOE 1907, February, to 1914, November -.--... 93 “ There is clearly a suggestion here of gradually increasing inter- vals. A second correction concerns the hour of outbreak Novem- ber 25,1914. It was at first supposed to be about 3.45 p. m. according to observers at Pahala (fig. 1), but later advice made it clear that cattle herders of Kapapala Ranch saw the fumes rising from the summit crater about 12.25 p. m. just when the seismographs at the Observatory were registering prolonged motion. Two -other corrections are to the effect that part of the north Innate platform in Mokuaweoweo erater still persists, although not shown in Mr. Palmer’s map,t+ and that while no *The Outbreak of Mauna Loa, Hawaii, 1914, by T. A. Jaggar, Jr., this Journal, vol. xxxix, pp. 167-172, Feb, 1915. {+ This Journal, Feb. 1915, page 171. Am. Jour. Sci.—FourtTH Series, Vou. XL, No. 240.— December, 1915. 42 622 JSaggar— Activity of Mauna Loa. Fic. 1. HAWAII Compiled from Govt Survey Maps by Baldwin & Alexander Civil Engineers MILES (a, 15 20 1907 Fic. 1. Map of Hawaii (from Baldwin’s Geography of the Hawaiian Islands). Jaggar—Activity of Mauna Loa. 623 instantaneous sympathy with the eruption of Mauna Loa was shown by Kilauea, yet there was a very pronounced rise of the Kilauea lava column which immediately followed the date of the Mauna Loa outbreak November 25. (Fig. 2.) The seismic prelude to the eruption has been described by ~H. O. Wood.* The number of local shocks per month was as follows: April... 17 shocks. | August-September, incompletely recorded. Meiers lk 2“ October ==---—- 14 shocks. Jumey. ls. November ----_-_-- Uo Willige 34 December ._-.-.. 13 “ Fig, 2. RIM isi 3700 Fr JANIFER|MAR|APRIMAY|JUN|JULJAUG|SEP|OCT|INOV|DEC FEET 100 3600 DOWN 200 3500 200 3400 400 ; 3300 500 3200 Fic. 2. Diagram showing rise of lava column of Kilauea volcano, in Halemaumau Pit during year 1914, and rapid acceleration after November 25, the date of Mauna Loa outbreak. By T, A. Jaggar. During the year preceding October 1914 there were at least two shar p shocks on Hawaii, one of them in the Mauna Loa axis; and in July and September there were seismic spasms of numerous earthquakes lasting two or three days. The com- puted distance of origin of many of the local earthquakes of the year, notably from September to December, corresponded to the distance from the Observatory to the Mauna Loa rift line, namely eighteen to thirty miles. Immediately after the ontbreak, from November to December 8 inclusive, eleven days, no earthquakes were registered at the Observatory, show- *The Seismic Prelude to the 1914 Eruption of Mauna Loa, by Harry O. Wood, Bull. Seis. Soc. Amer., vol. v, pp. 39-50, March 1915. 624 JSaggar— Activity of Mauna Loa. ing a sudden cessation in seismic activity. In the first half of 1915 the seismicity has been normal. All the evidence accumulated concerning this summit out- break of Manna Loa shows that the main eruption expended itself during the first twelve hours in a line of tremendous lava fountains rising through a rift in the crater floor and probably spouting to heights over five hundred feet in some instances. No one saw the fountains from the rim of the crater during Fic. 3. Fic. 3. Mauna Loa and Kilauea from east rim of Kilauea crater Nov. 26. 1914, 12.30 a.m. The width of the fume column in Mauna Loa calculated fron: measurement of photograph was 7300 feet, or approximately 1°4 miles ; the distance 116,250 feet or 22 miles. South is on the left. T. A. Jaggar, phot. the first two days, and it is probabiy no exaggeration to assert that no one has ever seen an outbreak of Mauna Loa from a viewpoint close at hand. ‘These first twelve hours presented a spectacle quite different from what was seen by Palmer and Forrest on the third day, for those observers reported spatter heaps of large size beside dwindled fountains which had built them, and night photographs taken by the writer November 26* and November 27 under like conditions of distance, light * This Journal, Feb. 1915, page 165, Jaggar—Activity of Mauna Loa. 625 and exposure show striking dwindling in the luminosity of the fume column. On November 26 the brilliantly-lighted jets of vapor formed a wide band corresponding to more than one mile of summit rift (fig. 3), while on November 27 the band had narrowed, was brightly Juminous only in one southern streak with a faint secondary streak north of it. (Fig. 4.) The sudden outbreak and intense activity of the first few hours correspond to the explosive phase of the first few hours of Fic. 4. Fic. 4. Camp and glow over Mokuaweoweo, Puu Lehua, Kona, Nov. 27, 1914, 10 P.M. Photographed by moonlight; 25 min. exposure; Wratten panchromatic plate, Tessar lens; conditions similar to fig. 3 as to clearness and distance but on opposite side of mountain. South is on the right. T, A. Jaggar, phot. eruption in other volcanoes such as Vesuvius, and undoubtedly marked a sudden release of gas pressure in a foaming liquid of relatively low viscosity. The seismic prelude implies that for some months the lava was pushing its way upward, probably by a combination of blowpiping, stoping and wedging. On November 26, from the observatory, a clear view of the profile of Mauna Loa at about 9 a. m. showed a cluster of fume columns, merging into a fluted curtain of thin bluish-white 626 Juggar— Activity of Mauna Loa. fumes rising quietly to a position subtending an angle above the summit about two-thirds as great as that subtended by the summit above the level of Kilauea, hence attaining an altitude of 6000 to 7000 feet above the summit, where a scanty cumu- lus crown of condensed vapor hung. (Fig. 5.) Presumably this was condensed water vapor from the atmos- phere. All about it was a very thin, much diffused bluish haze of uncondensed fumes. By day the effect was disap- Fie. 5. MAUNA LOA Nov 26 1914 pointingly slight. Drifting clouds during forenoon and after- noon made seeing uncertain. Toward sundown, however, clear seeing was again afforded. Seen in sunset lighting, partly by reflected, partly by transmitted light, the thin fume curtain— now chiefly a north and a south column—displayed a succession of color tints of distinctly fluorescent character, as follows: (@) just before sunset, with the sun on the mountain profile well south of the summit, the fumes showed a dirty saffron tint, seen largely by reflected light; (6) at about ten minutes after sunset this had gradually turned to brown, in which a distinetly greenish cast was seen; it was still a muddy color, though seen largely by transmittéd light; (¢) about twenty minutes after sunset the tone had become a deep, translucent brown, still of somewhat muddy consistency. Seeing in the evening was much interrupted by mist. Meteorological conditions at the summit were unquestionably less favorable for a spectacular display than on the evening of JSaggar— Activity of Mauna Loa. 627 outbreak. Nevertheless it appeared certain that the active area had decreased. There were two well-marked columns of rising fumes, the southern larger and better illuminated. Usually these were separated by a blank space of clear sky, but at times illuminated drifting fumes intervened. On November 27, by day only brief glimpses of the top of the fume column were obtained. At night the scene was essentially unchanged, though somewhat more spectacular than on the previous evening. On November 28, forenoon views were like those of the pre- ceding days. Heavy rain set in about noon and there was no further seeing. On November 29, no view was obtained till late afternoon and evening. The summit was seen to be covered with snow down to about 12,000 feet. There was no significant change in the action from that seen on Fr iday evening, November 27. On November 30, at about 10 a. m. the north fume column was seen to be very much lessened. At night in bright, hazy moonlight the display was brilliant and quite as spectacular as at any time except the first evening and night. Nevertheless the output of fumes was less than on the preceding evenings. The south column of fumes was much the larger, and better illuminated. The dark space intervening between it and the north column was occasionally filled with drifting fumes well lighted. On December 1, an overcast day with very high clouds, the fumes showed faint brown tints against the cloud background in early forenoon. A stratum of very thin, cerulean blue fumes, with a faint but distinctly defined limiting plane at the bottom, rested upon the mountain above, about 10,000 feet. In the afternoon only the south fume column was any longer visible. This rose, straight and slender, from 9,000 to 10,000 feet above the summit, and there a ragged cumulus crown hung stationary in a streaked, wavy network of blue haze. Im the evening the slender fume column was faintly illuminated and soon obscured by drifting clouds. On December 2, no view was obtained until evening. Then, at first the single south fume column appeared dimly lighted, almost invisible in the bright light of the full moon, but soon fumes appeared to be rising more copiously and more rapidly and the column became very brilliantly lighted, as much so as at any time except on the night of outbreak. Seen from the Observatory, on December 3 in the evening, Manna Loa cleared under high clouds, showing a single smoke column more voluminous than on previous days and more spectacularly illuminated. December 4 and 5 there was rain, but a faint diffused illumination of the cloud cap was seen at 628 Jaggar—Activity of Mauna Loa. night. December 6, at half-past nine p.m., Mauna Loa cleared off, showing a single slender column of fume with orange light of about the same color as Halemauman, rising rapidly in puffs discernible even at this great distance (twenty-two miles) to a height of about 6000 feet, and there spreading into a diffuse mushroom or balloon of thin vapor, illumined ruddy. (Fig. 6.) The topmost detectable glow reached to at least 10,000 feet above the summit. The next morning, Monday, December 7, Fic. 6. MAUNA LOA DEC 6 at 8 a.m., Mauna Loa was revealed with a wide snow eap, and from its middle rose a very slender column of white vapor, diffused above into a bluish haze. On the next morning, December 8, the mountain was brilliantly clear and showed not the faintest trace of a fume column, and there was no night glow seen, though the mountain was clear at 4a. m., De- cember 10. At 8.30} p- m., however, a faint glow showed and this continued on the 11th. Dec. 12 there were faint fumes in the morning and fumes with glow in the evening. Dec. 13 a compact cloud appeared over the crater in the morning, there were light puffs rising at 5 p. m. and the glow was distinct at night. We succeeded on December 15 in reaching Mokuaweoweo from the east side and taking photographs and notes of the fountaining activity which continued there. There had now been five parties which reached the summit area, and only one of these, the first, Messrs. Forrest and JSaggar— Activity of Mauna Loa. 629 Palmer, on November 27, succeeded in spending the night there. Owing to high winds and snow the December parties had to content themselves with daylight views. The parties named were Forrest and Palmer, November 27; Jaggar, No- vember 28; Charles Ka, December 3; Baker and Bowdish, December 11; Voleano Observatory expedition, December 15. Three of the expeditions attempted to camp at the summit and failed, owing to weather conditions. The expense in the ease of the Observatory’s trips has been enormously out of pro- portion to the small results achieved, owing wholly to the lack of any shelter in the summit area. The lessons learned in this respect, however, have been of great value. The Mauna Loasum- mit region cannot be a place of good scientific observation and survey until a shelter hut and stable have been erected there. Dr. A. 8. Baker and Mr. A. C. Bowdish ascended Mauna Loa from Kona December 10, and Mr. Bowdish described what he saw in the Honolulu Star Bulletin of December 15, 1914. He reached the western edge of Mokuaweoweo on the morning of December 11, and “saw one cone still active, throwing lava up 150 feet or more, while nearby was a bowl that was boiling, splashing lava to the height of 50 feet or more. Just beyond these to the south was a narrow line of fire where a stream of lava had not fully cooled on the surface. There was smoke issuing at a dozen or more places in compar- atively small volumes, but no fire was visible or other cones in sight. The whole floor was a vast black surface showing the chilled walls of lava streams apparently no longer active. The active cone is from 50 to 150 feet high, and the lava was thrown up. three times the distance of the height of the cone above its crest. The lava lost its first color before it reached the highest point and became black.” (Fig. 7.) The Observatory expedition left Volcano House Monday, December 14, at 9 a. m., by motor car for the upper cattle pen of Kapapala ranch, below the Halfway House. At 10:30 the pack train was loaded and it reached the water tank and camp ground in the forest reservation, at an elevation of about 8000 feet, at 4 p. m. and camp was made for the night. There were eight men with riding animals and five pack animals. The party was H. O. Wood, D. Lycurgus, Mr. Withers, Mr. Hannon, Alex Lancaster, Joe de Mello and H. Kaukine, besides the writer. December 15 start was made for the summit with the packs at 7.30 a. m., and before noon, in the snow-covered summit region, a gale of wind from the southwest sprang up. This was bit- terly cold and the animals could barely make progress against it. The weather was cloudy but without snowfall or rain. The snow of the summit plateau was deep and crusted over, \ 630 JSaggar—Activity of Mauna Loa. so that it generally supported horses without their plunging through the crust, but one or two of the animals went through. There was perhaps the equivalent of a foot of snow on the level, drifted deep into the hollows, and revealing points of rocks. sire! As there was no diminution of wind on the summit platean, which was reached at 12.30 p. m., I sent the pack animals back to the lower camp and all but the two packers proceeded to Fig. 7. Fie. 7. Southwest edge of Mokuaweoweo crater, Feb. 1912. J. F. Rock, phot. the crater. It was quite impracticable to make camp in such a gale, and in deep snow, with every prospect of a possible storm. There are no ridges to offer protection, only a waste of pahoehoe and aa blanketed with snow and occasional concealed crevasses. This plateau extends about four miles from the edge of the crater on the east side, but the west or Kona side is the actual summit of Mauna Loa and slopes off rapidly westward. We reached the east margin of Mokuaweoweo at 1.15 p. m. — Jaggar— Activity of Mauna Loa. 631 above the east end of the south lunate platform as mapped by Alexander in 1885. We looked across the south half of the main crater circle, which in general is much like the greater erater of Kilauea, and saw a large red fountain playing continu- ously in the southwest part of the crater. The fountain rose from the northwest side of an oval pool of crusted pumiceous lava, and back of the fountain was a huge half-cone of its own building. All of this was a mile away, as though one looked Fie. 8. Fic. 8. Lava pool and fountains of Mokuaweoweo from east margin of crater, Dec. 15, 1914, 1.30 Pp. wm. The main fountain shows white with a dark half-cone behind it and fume cloud above. T. A, Jaggar, phot. from the Voleano House of Kilauea at a fountain playing near the foot of Uwekahuna bluff, the great west cliff of Kilauea erater. The fountain played steadily to a height of about a hundred feet, and its horizontal diameter was about the same. Above this it sent up jets fifty feet higher, which parted into many fragments, cooling through shades of cherry-red to claret-color and black, and these black ejecta, instead of falling heavily, floated away and fell slowly like burnt paper, showing that the lava was of a very light, pumice-like quality. (Fig. 8.) The heights may have been greater than above stated. 632 Jaggar— Activity of Mauna Loa. The falling spatter from the fountain was to the west and north, and here on the edge of the pool was a black mound, probably crescent-shaped in plan, and steep or overhanging on the side of the fountain like the oven ramparts that build over grottoes around the borders of the Kilauea lake. This spatter heap was at least seventy-five feet high and made part of the background of the great fountain. Lower ramparts engirdled the “oval lake, which stood relatively high above a region of black pools and flows south of it and from it. The west side of the lake exhibited other lower fountains, one of them building a small mound, and mostly in the line of the big fountain and along the shore of the pool. Other fountains broke through the crust of the pool from place to place and time to time, seemingly indicating that the black crust was foamy or light and easily punctured. The fountains were indescribably different from the relatively heavy domes like Old Faithful in Halemaumau. They were strikingly like the flamy protuberances in the pictures of the solar corona. They appeared to me much redder in daylight and more like flames than a heavy liquid. The suggestion was rather as of an exceedingly light and gas-charged liquid, which cooled and changed color even more quickly than the fluid of Kilauea, and which boiled to much greater heights because of its aeriform consistency. In the high wind which was blowing very little cuibke showed. Above the fountains, however, - a blue fume devel- oped in yolutes and rose. Between the large fountain and the fume above it, appeared a semi-transparent space with strong uprushing heat-movement lines, which gave me the impression of being a bluish flame which in darkness would have shown as such. To Mr. Arthur Hannon, an artist in the party, this _ appeared of violet color and a distinct flame. The only smok ing area was in the vicinity of the heated pool and the flows below it to the south. The rest of the crater appeared much as in 1912 and 1913, except that new detail of small mounds appeared along a fissure line northward from the large fountain and beyond the great mound or cone of 1903. Mr. Palmer’s map will serve to illustrate the line of the rift, and no doubt there are new flows from this fissure over the middle region of the erater. The larger central mounds, however, shown on his map are not new, but are the old cones of 1903- 1907, some- what moditied by the new eruption. (This Journal, Feb. 1915, 171.) No changes were seen in the walls of Mokuaweoweo nor in the north or south pits, though we could not see into the latter. The southern lunate bench and parts of the crater Jaggar— Activity of Mauna Loa. 633 floor were covered with snow, but not the central, southern and north-central parts of the main crater, implying, perhaps, that all of that region was warm. Remnants of the north -Junate bench still persist but show change since last year. The north gateway to the crater formed a frame for a striking view ot Mauna Kea, with its snow-covered upper cones. (Fig. The ‘party left the summit at 2 p. m. and reached the Fie. 9. Fic. 9. Northeast wall of Mokuaweoweo, Dec. 15, 1914, showing floor devoid of snow and distant Mauna Kea. T. A. Jaggar, phot, lower camp at 5. One of the animals was quite exhausted, owing to feet cut by the rough trail in aa lava, and was unable to hold out through the return trip next day. On December 16 we left the camp at 8.45 a. m. and reached the highroad soon after noon. The summit glow and fume column on Mauna Loa during the fortnight ending December 24 were visible in clear weather over Mokuaweoweo without marked change. Decem- ber 22 there was a deep mantle of snow on the mountain. (Fig. 10.) The fume column was very clearly visible at the end of December owing to exceptionally transparent air. It made a visible gauge of wind direction on the summit, often 634 JSaggar—Activity of Mauna Loa. quite different from the wind below, and gradually developed one habit of puffing up in distinct volutes, as the supply became ess. December 26, the slender single fume column was glowing in the evening. December 27, from 7 to 9 a. m. (fig. 11), — with light northeast wind at Voleano House, the fume column on Mauna Loa rose high, bent eastward and thinned out over Puna to a bluish iridescent fume cirrus, quite unlike any rain cloud in color and luster. It showed a slight transverse ripple- marking in the middle part of its course. There was possibly Fic. 10. a slight increase in volume of fumes, which continued through the week, but owing to varying conditions of wind and of clear atmosphere, with frequently no seeing at all, it is impos- sible to judge slight changes accurately. An increase of fumes was to be looked for, by analogy with the habit of Kilauea, if the fountains in Mokuaweoweo were still diminishing. The glow at night continued through the week. December 28, in the morning, small volutes of fume were rising; December 29, a straight funnel column in still air with traces of vortex motion in a spiral (fig. 12); Decem- ber 30, the same, bending in a counter current eastward in upper atmosphere, while the wind was northeast below. (Fig. 13.) At 1 p. m. the high column was bending south- ward, the puffs showing brownish below and blue above. The JSaggar— Activity of Mauna Loa. 635 morning of December 31, with northeast wind continuing below, the fume puffs of Mokuaweoweo were blowing away westward, implying that the wind below and above was now one and the same, and that the trade wind stratum had in the last few days thickened to above 14,000 feet. The first week of the new year 1915 found the Hawaiian vol- eanoes Mauna Loa and Kilauea both in a state of activity, the former with one or more fountains of brilliant lava foam con- tinuously playing in the southwestern part of the summit erater Mokuaweoweo, the latter with a lively lake and three Fig. 11. MAUNA LOA DEC 27 [a1 smaller lava ponds 368 feet below the rim of the erater pit Halemaumau. . The fume and glow over Mokuaweoweo continued constant. December 31, the fume was blowing to the northwest. Jan- uary 1, the fume blew westward, making the column a low dome as seen from the Observatory, with glow at night. January 2, there was a northeast drizzle with tremendous squalls of wind at the Observatory; the Mauna Loa fume appeared to be blowing southeast. In the afternoon there was a crown of rain-cloud above the snow-line on Mauna Loa, with the blue fume funnel rising vertically from its midst and bend- ing east above. The sunset was brown-yellow, with the fume- streak crossing it. The night was clear, showing not only the 636 . Jsaggar—Activity of Mauna Loa. illumined fume-column but a marked radiant glow spreading upward in the atmosphere above the summit fountains. January 3, very small puffs of fume showed, apparently blowing west, and this continued January 4, the nights showing the radiant glow and occasionally the low illumined fume. January 5 was similar, but the fume dome rose higher and at night there was a high luminous zone and rare glimpses of the fume top. There was fresh snow on the mountain. January 6, the fume was blowing northwestward. By the middle of January every expectation based on the hypothesis that Hawaiian volcanoes respond to the solstice and Fie. 12. subside thereafter, was fulfilled, for the lava of Halemau- mau was rapidly sinking and on January 11 the fume column and glow of the Mokuaweoweo fountain ceased. The time of the solstice, or maximum declination of the sun south, was from December 20 to 24 inclusive, the turning point being reached December 22; calling the eruptive period of Mokuaweoweo forty-eight days, from November 25 to Jan- uary 11 inclusive, the middle day of the period, twenty-four days after the outbreak, was December 18-19. J anuary 4, 1915, appears to have been the date of highest tide in Hale- maumau, when the last overflows of the lake took place at an elevation 363 feet below the southeast rim of the pit; on Jan- uary 13 the lake had fallen about thirty-five feet below this, or 398 feet below the rim. JSaggar—Activity of Mauna Loa. 637 Kilauea was intensely active and in general rising all through December, and in accordance with the measurements of preceding years it was fully expected at the Observatory that it would sink in January. The movements of the lava column, though as a whole lower, have been remarkably paral- lel to those of 1911-12, when also there was a maximum on January 4. There was glow over Mokuaweoweo January 6, 1915, and a tuft of fume like white cotton appeared on the morning of January 7, apparently blowing away to the west. For the previous fortnight there had been no marked changes, except Hie. 13. MAUNA LOA _ DEC $0 191% that a diffuse glow had appeared at night distinct from the light reflected from the obvious fume cloud, and frequently in the air around the fume column. This was attributed to atmospheric conditions. At and around 2 p. m. Jannary 6, streamers of white cloud rose in tufts from the higher cone region on the northeast slope of Mauna Loa, quickly disappear- ing and having no resemblance in color to the fume. On Jan- uary 7 between 1 and 2 p. m. the same thing happened again, mostly above the snow-line, along the northeast cone line, and at one time over Mokuaweoweo itself, the color of the white tufts being quite distinct from blue fume against which they showed. At 3:30 p. m. these white jets were still at work near the two upper cones where similar jets have been seen Am. Jour. Sci1.—FourtH SERIES, Vou. XL, No. 240.—Drcrmpnr, 1915. 43 638 Jaggar— Activity of Mauna Loa. from time to time during the past three years. January 9 there was a strong column of fume rising high and bending southward, and in the evening bright glow with some visible fume, blowing to the southwest. January 10 at the Observatory the weather was brilliant with strong northeast wind. The fume was seen in Kona, but day and night there was no trace of fume or glow seen from the Observatory at Kilauea. No fume or glow was seen there- after, nor was any reported from other points. On January 14 at 8 a. m. the weather was calm with the Kilauea fume rising high and spreading into a mushroom. There was no evidence of a high wind current which might blow the fume on Mauna Loa away. With clear seeing and powerful field glasses not the slightest trace of fume was detected over Mokuaweoweo. From January to June 1915 inclusive, Mauna Loa has been under incessant inspection from a distance and no trace of fume or glow recurred until after the summer solstice. With the migration of the sunset position back of Mauna Loa thin but definite fume was detected from the Observatory by Mr. H. O. Wood at sundown July 10, 11 and 13, 1915, rising, or standing, above Mokuaweoweo. Shortly after noon July 14 several persons saw fumes above the summit. No glow has been certainly seen at night, though one inexperienced observer reported glow of short duration 11.30 p. m. July 10. Whether this episode is a slight revival of gas activity or merely ex- ceptional seeing conditions is not clear. Kilauea lava had risen slightly in June. It may be concluded that the outbreak of 1914-15 was a preliminary summit ebullition of the same type as the prelim- inary outbreaks of previous eruptive periods; those of 1870, 1872, 1873, 1874, and 1875, and 1876 anticipated the sub- marine flow of 1877 at Kealakekua; that of May 1880 heralded the great Hilo flow of 1880-81; those of 1892 and 1896 preceded the flow from the Dewey Orater July 4, 1899; the summit action of October-November 1903 preceded the Kahuku flow of 1907. The phenomenon in this eruption which most impressed the writer, and which seemed contrary to some of the early accounts of the Mauna Loa lava fountains, notably those of William Lowthian Green, was the extraordinarily light and foamy character of the spurts seen above the large fountain on De- cember 15 in Mokuaweoweo. The quick chilling and change of color in the upper part of the fountain through cherry-red and purple to black, and the blowing away of the fragments like pieces of burnt paper, was utterly unlike the heavy splashes of orange-red melt seen in the fountains of Kilauea. The Jaggar—Activity of Mauna Loa. 639 material of consolidation of this eruption of 1914 will probably prove to be in part a light limu or what Dana ealled “thread-lace scoria.”’ Such material, often sherry-colored and with iridescent surfaces, the writer collected about hot cracks near Mokuaweoweo in 1912, this being presumably the summit product of the 1903— 1907 activities. There is much limu occurring as smail lapilli about the summit area, showing that this material was blown about during recent eruptions, and the central cones in the crater, built in 1903, are largely of pumiceous material. The occurrence of gas-impelled and gas-heated fountains of lava foam as the initial manifestation of a rising lava column - at the summit of Mauna Loa, 10,000 feet higher than the vent of Kilauea, is more compatible with the possibility of a subter- ranean connection between the conduits of the two voleanoes than would be the case if the Mauna Loa fountains were jets of heavy melt without any sign of gas or flame. It may well be that with the expansion, escape and oxidation of the gases rising as bubbles through the lava, a subterranean cooling and congealing effect determines for the time being the duration of the preliminary phase. There follows a term of accumulation accompanied by some change of state which is less frothy and the final lava flow relieves the accumulated stress and ushers in a repose period. The question of sympathy with Kilauea must still be considered an open one. In the long and large, the activities of Kilauea have shown tendency to alternate with the eruptive periods of Mauna Loa, this being especially striking since 1887. The repose period 1908-1913 of Mauna Loa was a time of rise, culmination aud fall in Kilauea. In 1913-1914 Kilauea lava subsided to smoky depths for a year and then came the Mauna Loa outbreak. With this outbreak, however, the Kilauea column gradually rose to a relatively low culmination coincident with the month of Mauna Loa activity and there- after it subsided, but has slightly revived with the solstice of June, 1915. The writer inclines to the belief that future research will demonstrate a complicated sympathy of alter- nation between the two volcanoes dependent upon gas release at one vent impoverishing the other. But the response is not instantaneous and phenomena of what may be called back-kick are to be looked for. Something of this kind may explain the rise of Kilauea in 1914-15, if that rise proves to be temporary. Volcano Observatory, Hawaii. 640 North and Conover—Mineral Sulphides. Arr. XLIV.— Decomposition of Mineral Sulphides and Sul- pho-Salts by Thionyl Chloride; by H. B. Norrs and C. B. Conover. iy another journal* the authors have recently described an investigation of the action of thionyl chloride on sulphides, the study being limited entirely to compounds prepared in the laboratory by precipitation or by fusion. Reaction was found to proceed according to the following general equation in which M represents a divalent element: MS + 2SOCl, = MCI, + SO, + S,Cl,. Deviation from this was found only in two or three cases in which oxidation as well as chlorination took place. Inasmuch as minerals are frequently less readily attacked by reagents than are artificially prepared compounds, the study has been continued to the reaction of thionyl chloride on a number of mineral sulphides and sulpho-salts. The minerals used were carefully selected specimens. The fragments chosen were powdered in an agate mortar, the pow- der dried at 110° and preserved in tightly stoppered tubes. The reactions were carried ont in sealed glass tubes, about a gram of the powdered mineral and an excess of thionyl chloride being heated together in each experiment. The temperature employed was 150-175° C. When reaction appeared to be complete, the tubes were opened, the solid contents well washed with carbon disulphide, dried and analyzed. Sulphur dioxide and sulphur monochlo- ride were found to be present in every experiment in which reaction took place. In nearly all cases a small amount of white insoluble matter remained when the sample taken for analysis was dissolved. This proved to be silica. In the three quantitative experiments this silica was weighed and the weight deducted from that of the sample taken. Galena.—Thiony] chloride did not react with galena when the two were brought together at the ordinary temperature. When heated at 150° the solid contents of the tube gradually whitened. Reaction appeared to be complete after 40 hours. Analysis of the white solid proved it to consist almost entirely of lead chloride with a trace of ferric chloride. Reaction takes place as follows: PbS + 2SOCI, = PbCl, + SO, + §,Cl,. Pyrites.—Pyrites and thionyl chloride did not react in the cold, but at 150° reaction took place readily with the formation * Journal of the American Chemical Society, Nov. 1910. a ae ee SS eee Le ee ee a eee ~——S . North and Conover—Mineral Sulphides. 641 of beautiful hexagonal green crystals of ferric chloride. Reac- tion undoubtedly follows the equation: 6FeS, + 20SOCI, = 6FeCl, + 1080, + 11S,Cl,. Cinnabar.—The reagent appeared to be without effect on cinnabar at the ordinary temperature, but reaction was com- plete after a few hours heating at 150-175°. The solid product formed consisted of a mass of white needle crystals which upon analysis were found to be mercuric chloride. Reaction takes place according to the equation : HeS + 2S0Cl, = HeCl, + SO, + S,Cl,. In addition to the above mentioned experiments in which the product was analyzed quantitatively, the following tests were made merely to ascertain whether decomposition of the mineral results from heating with thionyl chloride. Pyrargyrite.— Decomposition took place slowly, but a very appreciable change was noticed after 48 hours. After heating for several days reaction appeared to be complete and the solid matter in the tube was found to consist entirely of chlorides. Proustite.—Reaction proceeded much more slowly than with pyrargyrite. After 48 hours only a slight change had taken place, as was indicated by the appearance as well as by qualita- tive tests. Thionyl chloride decomposes proustite with great diticulty. Argentite.—The tube containing thionyl chloride and argen- tite was heated for several days without the slightest change in the appearance of the contents. Qualitative tests failed to show even the slightest traces of silver chloride. The results of this experiment are interesting inasmuch as the authors have previously shown* that the artificially prepared silver sulphide is decomposed readily by thionyl chloride giving silver chloride. Covellite.—This mineral was completely decomposed by thionyl chloride within 48 hours. Boulanger ite.—Boulangerite reacted readily with the re- agent, the color of the powder becoming white within a few hours. Qualitative tests showed the mineral to be completely decomposed. Lnargite.—Decomposition of enargite took place rapidly, as was shown by the production of brown anhydrous cupric . chloride. Marcasite.—Specimens of marcasite from two widely sepa- rated localities were employed. In both cases decomposition took place completely within a few hours, each yielding the green crystals of ferric chloride. * Loc. cit. 642 North and Conover—Mineral Sulphides. Molybdenite and Cobaltite—These two minerals were not affected by the reagent even after heating for several days. Orpiment and fealgar.—These minerals were completely dissolved by thionyl chloride within a few hours. No attempt was made to identify the arsenic trichloride formed inasmuch as it is a liquid and ditficult to separate from thionyl] chloride. Reaction had certainly taken place, as was shown by the sul- phur dioxide and sulphur monochloride which were easily identified in the tubes. Stibnite.—This mineral likewise was completely dissolved by the reagent, undoubtedly with the formation of antimony pentachloride. The tube contained sulphur dioxide and sulphur monochloride, leaving no doubt that reaction had taken place. Sphalerite.—No visible change took place in the sphalerite, but after heating for several days the solid contents of the tube were shown by analysis to be zine chloride. Reaction there- fore had taken place. Tetrahedrite.—Tetrahedrite was found to react slowly with the formation of anhydrous cupric chloride. The antimony and sulphur went entirely into solution. Arsenopyrite.— Arsenopyrite was completely decomposed after heating for a few hours, the arsenic going entirely into solution while the iron remained in the form of crystallized ferric chloride. feésumé.—Of the minerals studied, the only simple sulphides which do not react with thionyl chloride are argentite and molybdenite. In general the more complex mineral sulphides and sulph-arsenides are likewise decomposed by the reagent though the time required for complete reaction varies greatly. Rutgers College, New Brunswick, N. J. W. A. Drushel—Simple and Mixed Alkyl Phosphates. 643 Arr. XLV.—On Simple and Mixed Alkyl Phosphates; by W. A. DrusHeEt. [Contributions from the Kent Chemical Laboratory of Yale Univ.—celxxvi. ] Wuen a sodium alcoholate free from the alcohol is treated with the theoretical amount of phosphorus oxychloride at room temperature in the presence of ether a good yield of the corresponding trialkyl phosphate is obtained.* By means of this reaction Lossen and Kohlert in 1891 prepared trimethyl phosphate, triethyl phosphate, and from these esters they pre- pared by a method to be described the mixed esters, dimethyl- ethyl phosphate and methyl-diethyl phosphate. They studied these simple and mixed alkyl phosphates with special reference to their behavior toward barium hydroxide at room tempera- ture. From their observations they drew the following con- clusions: (1) That the esters of phosphoric acid are far more stable than the esters of organic acids toward aqueous barium hydroxide. (2) That the salts of dialkyl phosphoric acids are much more stable than the trialkyl phosphates. (8) That more than one alkyl group of a trialkyl phosphate may be hydrolyzed off only with the greatest difficulty, if at all. (4) That the decomposition of mixed alkyl phosphates by aqueous barium hydroxide proceeds in an unusual manner, in that in mixed alkyl phosphates one alkyl group is hydrolyzed off to the exclusion of the other. Apparently the same methyl-ethyl-barium phosphate was obtained by these inves- tigators by the action of barium hydroxide upon both methyl- diethyl phosphate and dimethyl-ethyl phosphate. The purpose of the present investigation was to prepare a larger number of the simple and mixed esters of orthophos- phorie acid, to study their properties and rates of hydrolysis in decinormal hydrochlorie acid, and to examine the validity of Lossen and Kohler’s fourth conclusion since this conclusion is contrary to the rule observed in the case of esters of dibasic and polybasie organic acids. In general the methods described by Lossen and Kohler were used in the preparation of the esters prepared and studied in this investigation. For the preparation of.the simple esters sodium alcoholate in each case was prepared by dissolving a weighed amount of pure metallic sodium in a sutiicient excess of the desired alcohol. After the sodium was completely used up the excess of alcohol was dis- tilled off in an oil bath, heating the bath finally at 180° as *Limpricht, Ann, Chem. Pharm., exxxiv, 347. f¢ Ann. Chem. Pharm., cclxii, 209-14, 644. W. A. Drushel—Simple and Mined Alkyl Phosphates. long as any alcohol distilled over. The residue of sodium aleoholate was cooled, covered with absolute ether and treated with the theoretical amount of phosphorus oxychloride, diluted with an equal volume of absolute ether, by adding the diluted phosphorus oxychloride in small portions through a reflux condenser, taking care to keep the reaction mixture cool by immersing the flask in cold water. At the conclusion of the reaction the residue of sodium chloride was filtered off, the filtrate distilled on a warm water bath to remove the ether, and the residue subjected to fractional distillation. The frae- tionation was made at diminished pressure for all of the esters except trimethyl phosphate, triethyl phosphate and dimethyl- ethyl phosphate. These three esters were distilled at ordinary pressure without decomposition. The mixed alkyl phosphates were prepared from the simple esters by making successively the barium salts by the action of aqueous barium hydroxide upon the simple esters at room tem- perature, the free dialkyl phosphoric acids by the action of sulphuric acid upon the purified barium salts, the silver salts of the dialkyl phosphoric acids by the neutralization of these acids with silver oxide, and finally the mixed alkyl phosphates from the silver salts ‘by the action of the appropriate alkyl iodides in ethereal solution. The silver iodide was then filtered off, the ether distilled off from the filtrate and the residue fractionally distilled, generally under diminished pressure. The boiling points of these simple and mixed esters at dimin- ished or ordinary pressure, their densities at 0° and at room temperature and their solubilities in water at room temperature were determined, and are recorded in Table I. According to Winssinger’s* observation the tripropyl phosphate decomposes when boiled even in vacuo. In order to test the correctness of this observation this ester was redistilled four times at a pres- sure of 15™" without any change in the boiling point observed in the first distillation and without any other evidence of the decomposition of the ester. The isobutyl ester was likewise redistilled at 15"™ without any evidence of decomposition. In order to determine whether one alkyl group in mixed alkyl phosphates is hydrolyzed to the exclusion of the other alky] group contrary to the rule in the case of esters of organic acid, the barium salts of a number of simple and mixed alkyl phosphates were prepared by the action of a slight excess of barium hydroxide upon the esters in aqueous solution either at room temperature or upon the steam bath, were purified and analyzed for barium, obtaining the results recorded in Table II. It is to be observed that when barium hydroxide reacts with simple alkyl phosphates the barium content of the result- * Bull. soe. chim., xlviii, 111. _ —— ae W. A. Drushel—Simple and Mixed Alkyl Phosphates. 645 TaBLe I. Boiling Presswre ——Density Solubility in Esters point in mm. Or 22° water at 25° Mero ** 197 760 1-218 1°200 ilggall Kt, PO, 215 760 1:071 1°056 Mog dh er ‘PO, 1351 15 1°025 1:007 1:155 Isobu,PO, 152 15 0978 0°965 1 : over 1000 Me, EtPO, 203 760 1:176 1-161 Tesi Me, PrPO. 116 15 1195 1-180 gil MePr. PO, 129 20 1077 1°059 22 Et, PrPO, 130 20 1098 1°077 1g 1 EtPr PO, 145 20 1-046 1:025 1:45 ** Me denotes methyl ; Et, ethyl, etc. ing salts agrees well with the calculated amount, and that when barium hydroxide under the same conditions acts upon mixed alkyl phosphates in no case does the amount of barium found in the resulting salts agree with the amount calculated for the theory that only one of the two different alkyl groups is hydrolyzed off to the exclusion of the other. It is safe to conclude, contrary to Lossen and Kohler’s theory, that in mixed alkyl phosphates barium hydroxide does not remove one of the different alkyl groups to the exclusion of the other and that, although one alkyl group may be more readily attacked than the other of two different alkyl groups, the reac- tion does not by any means go wholly in that direction. The reactions involved in the preparation and hydrolysis of alkyl phosphates may be briefly summed up in the following equations, where R denotes an alkyl group and R’ a different alkyl group: 2ROH + 2Na > 2RONa + H,, 3RONa + POC], > R,PO, + 3NaCl, 2R,PO, + BaO iH, > BaR, (PO,), + 2ROH, (also to a very ‘slight extent, R -PO,+ BaO,H, > BaRPO,* + 2ROH), BaR,(PO,), + H,SO, > 2HR,PO, + BaSO,, 2HR,PO, =e Ag,O = 2AeR,PO, ee EO; AgR, PO, ets R,R’PO, ae ‘Ag, 2h, R ‘PO, + BaO,0W, > BaR, R’ (PO,), ), + 2H,0, also 2h, R'PO, + BaO, H, > BaR,(PO, i ee 2H Oy R,PO, + HOH + HCl > HR, PO, + ROH + HCl (First stage), R,PO, + HOH + HC] > H,RPO, + ROH + HCl (Second ‘stage), H,RPO, + HOH + HCl > H,PO, + ROH + HCl (Final stage). * This salt being insoluble is readily removed by filtration. 646 W. A. Drushel—Simple and Mixed Alkyl Phosphates. R,R/PO, + HOH + HCl > HRR’PO, + ROH + HCl (also R,R'PO + HOH + HCl > HR,PO, + R/OH + HC), (First stage); similarly for the second and final stages. In each case the first alkyl group is hydrolyzed off more easily than the second . and third groups, which is in agreement with Lossen and Kohler’s observation for the saponification of these esters by barium hydroxide at room temperature. TABLE II. Barium salts of dialkyl phosphoric acids prepared and analyzed for barium. Esters Per cent. of barium and Salts found theory Ba(Me,PO,), * 35:2 35°4 35°38 Ba(Et,PO,), xr 31°2 30°7 31°3 Ba(MeEtPO,), rl 34:0 33°0 Salt probably contained some from r 34:1 BaMe,(P9Q,),. Me, EtPO,,. rII 34:1 35°4 calculated for BaMe,(PO,).,. Ba(MePrPO,), rI 32:0 306 Salt probably contained some from Ot S29 BaMe,(PO,),. Me,PrPO,, 8 34:1 from sI 28°8 27°5 calculated for BaPr,(PO,),. MePr,PO,. slI 28:9 Salt probably contained some Bakr (RO) Ba(EtPrPO,), s 31°6 29°1 Salt probably contained some from 31°5 Bakt,(PQ,),. Ht, PrPO,, from i 27°7 Probably chiefly BaPr,(PO,), EtPr,PO,. 27°7 formed. * y, Saponification was made at room temperature; s, on the steam bath. The acid hydrolysis of the esters recorded in Table III was made in flasks contained in a boiling water bath especially con- structed for the purpose and fitted with a reflux condenser. Decinormal hydrochloric acid was used as the hydrolytic agent and the course of the reaction was followed in the usual way by making titrations with decinormal barium hydroxide, using phenolphthalein as an indicator. In each case weighed amounts of ester were used, making the concentration of ester in the reaction mixture initially deeinormal. Since the end-point of the primary hydrolysis stage could not be determined by titra- tion on account of the interference of the secondary and ter- tiary stages, it was calculated in each case from the weight of ester taken. However, since the secondary and tertiary hy- drolysis products form insoluble barium salts, it was possible to | | — W. A, Drushel—Simple and Mixed Alkyl Phosphates. 647 observe directly how far the titrations furnished reliable data for calculating the velocity constants for the primary hydrolysis stage. The constants recorded in Table III represent only the primary hydrolysis stage, the latter constants in some cases being perhaps slightly influenced by the secondary reaction. In calculating the constants the reaction was considered of the first order and no account was taken of the possible hydrolytic effect of the very weak dialkyl phosphoric acids set free in the primary hydrolysis stage. The acid hydrolysis was first attempted at 25° but was found to proceed so slowly as to_ make it impracticable at this temperature. o¢ ey th M ou ty, NS Lae “3, Ch, CONN SELON A ot) cnc os i : ae! W 3 Ml e n> Si a, ao, pS < " SMITHSONIAN INSTITUTION LIBRARIES “TAOTMENTT 3 9088 01298 5917